Nothing Special   »   [go: up one dir, main page]

WO2018119902A1 - 一种地面环境的检测方法和装置 - Google Patents

一种地面环境的检测方法和装置 Download PDF

Info

Publication number
WO2018119902A1
WO2018119902A1 PCT/CN2016/113089 CN2016113089W WO2018119902A1 WO 2018119902 A1 WO2018119902 A1 WO 2018119902A1 CN 2016113089 W CN2016113089 W CN 2016113089W WO 2018119902 A1 WO2018119902 A1 WO 2018119902A1
Authority
WO
WIPO (PCT)
Prior art keywords
ground environment
sub
laser
scanning
different
Prior art date
Application number
PCT/CN2016/113089
Other languages
English (en)
French (fr)
Inventor
曹彤彤
邵云峰
姚骏
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Priority to KR1020197021867A priority Critical patent/KR102243118B1/ko
Priority to PCT/CN2016/113089 priority patent/WO2018119902A1/zh
Priority to EP16925037.0A priority patent/EP3553566B1/en
Priority to JP2019535919A priority patent/JP6798032B2/ja
Priority to CN201680091952.6A priority patent/CN110114692B/zh
Publication of WO2018119902A1 publication Critical patent/WO2018119902A1/zh
Priority to US16/456,057 priority patent/US11455511B2/en

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/48—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S17/00
    • G01S7/4802—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S17/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/48—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S17/00
    • G01S7/491—Details of non-pulse systems
    • G01S7/4912—Receivers
    • G01S7/4913—Circuits for detection, sampling, integration or read-out
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
    • G01S17/02—Systems using the reflection of electromagnetic waves other than radio waves
    • G01S17/06—Systems determining position data of a target
    • G01S17/42—Simultaneous measurement of distance and other co-ordinates
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
    • G01S17/87—Combinations of systems using electromagnetic waves other than radio waves
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
    • G01S17/88—Lidar systems specially adapted for specific applications
    • G01S17/93—Lidar systems specially adapted for specific applications for anti-collision purposes
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
    • G01S17/88—Lidar systems specially adapted for specific applications
    • G01S17/93—Lidar systems specially adapted for specific applications for anti-collision purposes
    • G01S17/931—Lidar systems specially adapted for specific applications for anti-collision purposes of land vehicles
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/48—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S17/00
    • G01S7/4808—Evaluating distance, position or velocity data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/25—Fusion techniques
    • G06F18/251—Fusion techniques of input or preprocessed data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/25—Fusion techniques
    • G06F18/253—Fusion techniques of extracted features
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/60—Extraction of image or video features relating to illumination properties, e.g. using a reflectance or lighting model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/762—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using clustering, e.g. of similar faces in social networks
    • G06V10/763—Non-hierarchical techniques, e.g. based on statistics of modelling distributions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
    • G06V10/803—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level of input or preprocessed data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
    • G06V10/806—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level of extracted features
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Definitions

  • Embodiments of the present invention relate to the field of automatic driving, and in particular, to a method and apparatus for detecting ground environment.
  • Autonomous vehicles also called driverless cars, are smart cars that are unmanned by computer systems or terminal devices.
  • One of the important prerequisites for realizing automatic driving is the detection of the ground environment, the detection of the ground environment, the determination of the road surface condition, the trafficable area of the vehicle, etc., thus serving the planning, decision-making and control of subsequent automatic driving.
  • embodiments of the present invention provide a method and apparatus for detecting a ground environment based on multi-wavelength lidar, which uses a laser radar based on different working wavelengths to scan the ground according to the reflection intensity of the ground environment under different wavelengths of laser light.
  • a laser radar based on different working wavelengths to scan the ground according to the reflection intensity of the ground environment under different wavelengths of laser light.
  • the present application provides a method for detecting a ground environment, comprising: scanning a ground environment with laser pulses of different working wavelengths, receiving a reflected signal reflected by a ground environment for a detection signal; and determining each ground environment according to the reflected signal.
  • Scanning point information of the scanning point the scanning point information includes a direction angle, a distance, and a laser reflection intensity of the scanning point relative to the laser radar; determining spatial coordinate information and laser reflection characteristics of each scanning point according to each scanning point information, and the ground environment Minute
  • the sub-areas are cut into different laser reflection characteristics, the laser reflection characteristics include reflectance to lasers of different wavelengths; and the ground environment type of each sub-area is determined.
  • a plurality of laser radars with different working wavelengths can respectively emit laser detection signals of their own working wavelengths to scan the surrounding ground environment, or a laser radar with multiple working wavelengths can respectively emit laser detection signals of different wavelengths.
  • the laser detection signals of different wavelengths are used to scan the surrounding ground environment, the reflectivity of each scanning point to different wavelengths of laser light is determined, the ground environment is divided into sub-regions having different laser reflection characteristics, and the ground of each sub-area is determined.
  • Type of environment According to the reflection intensity of the ground environment at different wavelengths, the ground environment type is judged, the perception effect on the complex ground environment is improved, and the passable road surface is better determined.
  • the type of ground environment for each sub-area is determined based on the laser reflection characteristics of each type of ground environment.
  • the laser reflection features of each sub-region are input into the neural network to obtain the ground environment type of each sub-region output by the neural network.
  • the neural network used here uses the reflectivity data of lasers of different wavelengths as the input by different ground environment types, and the ground environment type as a set of data of the output is trained as a sample data set, and a data model for distinguishing different ground environment types is obtained. According to the data model obtained by the training based on the neural network, the ground reflection type of each sub-area is determined for the laser reflection characteristics of each sub-area, and the accuracy of the recognition of the ground environment type is improved.
  • the position information of each scanning point is transformed into the same coordinate system, and the scanning points of each scanning point acquired by each laser radar are scanned.
  • the point information is fused, and the spatial coordinate information and the laser reflection characteristic of each scanning point are determined; the spatial division information and the laser reflection characteristic of each scanning point are used for region division, and the ground environment is divided into sub-regions having different laser reflection characteristics. That is, firstly, the scanning data of the plurality of laser radars are fused according to the scanning point position information, and then the fused data is clustered according to the reflectance information of different wavelengths, and the ground environment is divided into sub-regions having different laser reflection characteristics.
  • the scanning points acquired by each laser radar are separately segmented according to the scanning point information to generate a clustered sub-region with different laser reflection intensities for each laser radar; clustering of each lidar
  • the position information of the post subregion is transformed into the same coordinate system for each laser thunder
  • the scan point information of each scan point in the sub-area after the clustering is fused according to the position information of the transformed sub-area, and the ground environment is divided into sub-areas having different laser reflection characteristics. That is to say, the scanning data of each laser radar is first clustered and segmented, and then the data of the different laser radar regions is merged according to the position information, and the ground environment is divided into sub-regions with different laser reflection characteristics.
  • the region segmentation is performed by using a region growing mode or a K-means method.
  • the image data of the ground environment is collected by a plurality of cameras, and the image data collected by the camera is processed to identify the type of the ground environment, and the ground environment type determined by the multi-wavelength laser radar is integrated to increase the ground. Robustness of environmental detection systems.
  • an embodiment of the present invention provides a ground environment detecting device, including: a laser scanning unit, configured to scan a ground environment by using laser detection signals of different working wavelengths, and receive a reflected signal reflected by a ground environment for a detection signal; data collection a unit, configured to determine scan point information of each scan point of the ground environment according to the reflected signal, where the scan point information includes a direction angle, a distance, and a laser reflection intensity of the scan point relative to the laser radar; and a scan data processing unit, configured to The scanning point information determines spatial coordinate information and laser reflection characteristics of each scanning point, and divides the ground environment into sub-regions having different laser reflection characteristics, the laser reflection characteristics including reflectance to different wavelength lasers; ground environment determining unit, Used to determine the type of ground environment for each sub-area.
  • a laser scanning unit configured to scan a ground environment by using laser detection signals of different working wavelengths, and receive a reflected signal reflected by a ground environment for a detection signal
  • data collection a unit configured to determine scan point information
  • the laser scanning unit is a mechanical rotary laser radar or a solid state laser radar.
  • the above ground environment detecting device scans the surrounding ground environment with laser detection signals of different wavelengths, and determines the reflectance of each scanning point to different wavelength lasers according to the reflected signal reflected by the ground environment, and divides the ground environment into different laser reflection characteristics. Sub-regions that determine the type of terrestrial environment for each sub-region. According to the reflection intensity of the ground environment at different laser wavelengths, the ground environment type is judged, the perception effect on the complex ground environment is improved, and the passable road surface is better determined.
  • the ground environment determining unit is configured to determine the ground environment type of each sub-area according to the laser reflection characteristics of each type of ground environment.
  • the ground environment determining unit is configured to input laser reflection features of each sub-region into the neural network, and obtain a ground environment type of each sub-region output by the neural network;
  • the neural network is used to distinguish the data models of different ground environment types by using different ground environment types as the input of the reflectivity data of different wavelength lasers, and the ground environment type as the output set of data is trained as the sample data set.
  • the scan data processing unit includes: a fusion subunit, configured to convert position information of each scan point to the same coordinate system according to scan point information of each scan point and installation positions of each laser radar And merging the scanning point information of each scanning point acquired by each laser radar to determine spatial coordinate information and laser reflection characteristics of each scanning point; the area dividing subunit is used for spatial coordinate information and laser according to each scanning point
  • the reflection feature performs region segmentation to segment the ground environment into sub-regions having different laser reflection characteristics.
  • the fusion subunit is configured to first fuse the scan data of the plurality of lidars according to the scan point position information, and then the region division subunit is configured to cluster the merged data according to the reflectivity information of different wavelengths, and the ground environment is Segmented into sub-regions with different laser reflection characteristics.
  • the scan data processing unit includes: a region segmentation subunit configured to separately segment the scan points acquired by each lidar according to the scan point information, and generate different laser reflection intensities for each lidar.
  • the fusion sub-unit is used to transform the position information of the sub-regions of each lidar into the same coordinate system, and scan points of each scanning point in the sub-region after each lidar clustering The information is fused according to the position information of the transformed sub-region, and the ground environment is divided into sub-regions having different laser reflection characteristics.
  • the region dividing sub-unit is used to first cluster the scanning data of each lidar to perform regional segmentation, and then the fusion sub-unit is used to fuse the data of the different laser radar regions according to the position information, and divide the ground environment into Sub-regions with different laser reflection characteristics.
  • the region segmentation sub-unit performs region segmentation by using a region growth mode or a K-means method.
  • an embodiment of the present invention provides a ground environment detecting device, including: a laser scanning unit, configured to scan a ground environment by using laser detection signals of different working wavelengths, and receive a reflected signal reflected by a ground environment for a detection signal; a unit, comprising a processor and a memory, wherein the memory is used to store computer execution instructions, and the processor executes computer execution instructions for determining scan point information of each scan point of the ground environment according to the reflected signal, and determining each according to each scan point information Scanning point space coordinate information and laser reflection characteristics, dividing the ground environment into different lasers A sub-region of the reflection feature determines a ground environment type of each sub-region; wherein the scan point information includes a direction angle, a distance, and a laser reflection intensity of the scan point relative to the laser radar, and the laser reflection characteristic includes a reflectance to the laser of different wavelengths.
  • the laser scanning unit is a mechanical rotary laser radar or a solid state laser radar.
  • the above ground environment detecting device scans the surrounding ground environment with laser detection signals of different wavelengths, and determines the reflectance of each scanning point to different wavelength lasers according to the reflected signal reflected by the ground environment, and divides the ground environment into different laser reflection characteristics. Sub-regions that determine the type of terrestrial environment for each sub-region. According to the reflection intensity of the ground environment at different wavelengths, the ground environment type is judged, the perception effect on the complex ground environment is improved, and the passable road surface is better determined.
  • the processor is configured to determine a ground environment type of each sub-area based on laser reflection characteristics of each type of ground environment.
  • the processor is configured to input laser reflection features of each sub-region into a neural network to obtain a ground environment type of each sub-region output by the neural network; wherein the neural network is used to distinguish different ground environments.
  • the type of data model is obtained by using the reflectivity data of different wavelengths of the laser for different ground environment types as input, and the ground environment type as a set of output data for training as a sample data set.
  • the processor is configured to convert the position information of each scanning point to the same coordinate system according to the scanning point information of each scanning point and the installation position of each laser radar, and obtain the information obtained by each laser radar.
  • the scanning point information of each scanning point is fused, the spatial coordinate information and the laser reflection characteristic of each scanning point are determined, and the spatial coordinate information and the laser reflection characteristic of each scanning point are used for area division, and the ground environment is divided into different laser reflections.
  • Sub-region of the feature That is, the processor is configured to first fuse the scan data of the plurality of lidars according to the scan point position information, and then cluster the merged data according to the reflectivity information of different wavelengths, and divide the ground environment into different laser reflection characteristics. Sub-area.
  • the processor is configured to separately segment the scan points acquired by each laser radar according to the scan point information, and generate a clustered sub-region with different laser reflection intensities for each lidar;
  • the position information of the sub-regions after the clustering of the lidars is transformed into the same coordinate system, and the scanning point information of each scanning point in the sub-region after each lidar clustering is converted according to the transformation.
  • the positional information of the sub-areas is fused to divide the ground environment into sub-regions having different laser reflection characteristics. That is, the processor is used to first cluster the scan data of each lidar to perform regional segmentation, and then combine the data of the different laser radar regions according to the position information, and divide the ground environment into sub-regions with different laser reflection characteristics. .
  • the processor performs region segmentation by using a region growth mode or a K-means method.
  • the ground environment detecting method and device scans the surrounding ground environment by using laser detecting signals of different wavelengths, and divides the ground environment into having the reflectance of laser light of different wavelengths according to each scanning point.
  • Sub-areas of different laser reflection features determine the type of ground environment for each sub-area. Since the lasers of different wavelengths are used to scan the ground, and the ground environment type is judged according to the reflection intensity characteristics of the ground environment under different wavelength lasers, the sensing effect on the complex ground environment is improved, and the passable road surface is better determined.
  • FIG. 1 is a schematic diagram of a ground environment detection scenario applied according to an embodiment of the present invention
  • FIG. 2 is a flow chart of a method for detecting a ground environment
  • Figure 3a is a schematic diagram of a method of dividing a sub-area
  • Figure 3b is a schematic diagram of another method of dividing a sub-area
  • FIG. 4 is a schematic structural view of a ground environment detecting device
  • FIG. 5 is a schematic structural view of another ground environment detecting device.
  • FIG. 1 is a schematic diagram of a ground environment detection scenario applied in an embodiment of the present invention.
  • the entire ground environment detection system consists of a laser radar mounted on an autonomous vehicle and a scanning data processing unit.
  • the laser radar can be a mechanical rotary laser radar or a solid-state laser radar for transmitting laser signals to the road surface for scanning the surrounding environment. Each laser radar will receive the laser signal reflected back from the ground and return to a series of scanning points. information.
  • the existing laser radar mainly uses a working wavelength of 950 nm, and the laser radar using the single working wavelength is often difficult to identify a complex road surface environment.
  • a multi-wavelength laser radar is used, that is, a laser radar with different working wavelengths is used. ground.
  • FIG. 1 shows that three laser radars are installed on the self-driving vehicle. In practical applications, the number of installed laser radars can be flexibly selected according to factors such as demand and cost, and the present invention does not limit this.
  • the scanning data processing unit extracts the information of all the scanning points received by each laser radar on the ground, and determines the distance and the reflection intensity of each scanning point returning to the ground through the time difference of the laser signal transmission-reception and the signal information, thereby extracting the road surface.
  • the three-dimensional structure and the intensity of the reflection determine whether it is a passable road surface.
  • the scanning data processing unit may further divide the ground into sub-regions having different laser reflection characteristics according to original data of different wavelengths of the laser radar, that is, The sub-areas with different reflectances of lasers of different wavelengths finally determine the type of ground environment of each sub-area.
  • the embodiment of the present invention provides a method for detecting a ground environment, as shown in FIG. 2, and the specific process includes:
  • Step 201 Scanning the ground environment with laser detection signals of different working wavelengths.
  • the laser detection signal emitted by the laser radar with different working wavelengths scans the surrounding ground environment.
  • multiple laser radars can be used to separately emit laser detection signals of different wavelengths for scanning, and multiple laser radars with different working wavelengths can be used respectively.
  • a laser detection signal that emits its own working wavelength can also be used to emit laser detection signals of different working wavelengths by using a laser radar having multiple working wavelengths.
  • Step 202 Receive a reflected signal reflected by the ground environment for the detection signal.
  • the laser detection signal emitted by the laser radar encounters a target (also called a scanning point) in the surrounding ground environment, and will reflect, and the laser radar will receive a reflected signal reflected from the target.
  • a target also called a scanning point
  • Step 203 determining scanning point information of the scanning point according to the received reflection signal, where the scanning
  • the trace information includes the direction angle, distance, and laser reflection intensity of the scan point relative to the lidar.
  • the reflected signal information received by each radar can be transmitted to the scan data processing unit, and the transmission mode can be wireless transmission mode (such as Bluetooth) or cable connection transmission mode (such as direct connection of signal lines).
  • the scanning data processing unit determines the direction angle, the distance and the laser reflection intensity of each scanning point according to the transmission signals returned by the respective scanning points received by the respective laser radars, such as the time difference of the laser detecting signal transmitted to the receiving and the signal strength information of the transmitting and receiving signals. And other information.
  • the method for calculating the scanning point information of the scanning point by receiving the reflected signal can be a common method in the existing laser radar detection, which is not described in detail in the present invention and does not affect the applicable range of the present invention.
  • Step 204 Determine spatial coordinate information and laser reflection characteristics of each scanning point according to each scanning point information, and divide the ground environment into sub-regions having different laser reflection characteristics.
  • the laser reflection characteristics here include the reflectivity of lasers of different wavelengths.
  • the scan data processing unit combines the scan point information of all the scan points, that is, according to the direction angle, the distance and the laser reflection intensity of all the scan points, the scan point data returned by the laser radars of different working wavelengths are merged to generate relative to the same coordinate system. Scanning point information, clustering points with similar spatial coordinates and similar laser reflection features in all scanning points, and clustering the ground environment into sub-regions with different laser reflection characteristics.
  • Step 205 determining a ground environment type of each sub-area.
  • the ground environment type of each sub-area is determined according to the laser reflection characteristics of each type of ground environment, that is, the reflectivity of lasers of different wavelengths.
  • three laser radars having different operating wavelengths are taken as an example, and the operating wavelengths are respectively ⁇ 1, ⁇ 2, and ⁇ 3. It is assumed that there are three types of ground environment to be judged, A, B, and C (for example, road, water, and vegetation, respectively), and q is the reflectivity.
  • the laser reflection characteristics of each type of ground for lasers of different wavelengths satisfy the following relationship. :
  • Type A q( ⁇ 1) ⁇ q( ⁇ 2) ⁇ q( ⁇ 3), the reflectivity of the laser for three wavelengths is substantially the same;
  • Type B q( ⁇ 1) ⁇ 0, q( ⁇ 2) ⁇ 0, q( ⁇ 3) ⁇ 0, the reflectance of the laser light of wavelength ⁇ 1 is not 0, and the reflectance of the laser light of wavelength ⁇ 2, ⁇ 3 is close to 0 ;
  • Type C q( ⁇ 1)>q( ⁇ 2)>q( ⁇ 3), and the reflectance of the laser light having wavelengths ⁇ 1, ⁇ 2, and ⁇ 3 is sequentially decreased.
  • a machine learning based method may also be used to extract each The laser reflection characteristics of the sub-areas are judged.
  • a different ground environment type to reflect the reflectivity data of lasers of different wavelengths as input, corresponding ground environment type as a set of data as a sample data set, training the neural network to obtain the neural network for distinguishing different ground environment types Data model.
  • the ground environment type of each sub-region output by the neural network according to the data model can be obtained.
  • T ⁇ T1, T2,..., TM ⁇ represents the type of ground environment (including M classifications) as an output of the neural network.
  • a series of Q, T-to-one data is used as a sample data set to train the neural network, and a neural network is used to distinguish the data model of the ground environment type.
  • the laser reflection characteristic data of a group of sub-regions that is, the reflectance data of the laser beams of different wavelengths into the neural network are input into the neural network, and the neural network can output the ground environment type corresponding to each sub-region according to the data model obtained by the training.
  • the ground environment detecting method provided by the embodiment of the present invention shown in FIG. 2 uses laser detection signals of different wavelengths to scan the surrounding ground environment, and determines the reflectance of each scanning point to different wavelength lasers according to the reflected signals reflected by the ground environment.
  • the ground environment is divided into sub-regions with different laser reflection characteristics, and the ground environment type of each sub-region is determined. Since the lasers of different wavelengths are used to scan the ground, and the ground environment type is judged according to the reflection intensity of the ground environment under different wavelength lasers, the sensing effect on the complex ground environment is improved, and the passable road surface is better determined.
  • step 204 in the foregoing embodiment spatial coordinate information and laser reflection characteristics of each scanning point are determined according to each scanning point information, and the ground environment is divided into sub-regions having different laser reflection characteristics, which may be as follows Method: according to the scanning point information of each scanning point and the installation position of each laser radar, the position information of each scanning point is transformed into the same coordinate system, and the scanning point information of each scanning point acquired by each laser radar is fused to determine The spatial coordinate information and the laser reflection feature of each scanning point are segmented according to the spatial coordinate information and the laser reflection feature of each scanning point, and the ground environment is divided into sub-regions having different laser reflection characteristics. That is After receiving the scanning point information from multiple laser radars, the scanning data of the plurality of laser radars are first fused according to the scanning point position information, and then the fused data is clustered according to the reflectivity information of different wavelengths.
  • the scanning data of the plurality of laser radars are merged according to the position information of the scanning points (including the direction angle and the distance), that is, the scanning of the scanning points scanned by the respective lidars.
  • the point information is merged according to the scanning point. If the installation positions of multiple laser radars are different, it is necessary to first change the position of the scanning points acquired by each laser radar into the same coordinate system, such as a vehicle coordinate system. After the coordinate transformation is completed, if the position information of the scanning points acquired by each laser radar is not one-to-one correspondence but is misaligned, it is necessary to newly construct a new scanning point set to store scanning information of different laser radars, each of the scanning point sets.
  • the position of the scanning point can be obtained according to the position information of the scanning point of each laser radar, or can be a previously defined value.
  • the reflection intensity information of different lidars of each point in the set of scanning points ie, the reflectivity information for lasers of different wavelengths
  • the position of the scanning points and the intensity of the reflection for different wavelengths of the laser radar that is, the positional information of each scanning point and the reflectivity information of the scanning point for the laser of different wavelengths are finally obtained. For example, suppose that all three laser radars scan N scanning points. Of course, different laser radars may scan different scanning points.
  • the number of scanning points scanned by each laser radar is the same, and It has an impact on the scope of application and the scope of protection of the embodiments of the present invention.
  • the first laser radar has a scanning wavelength of ⁇ 1
  • the scanned scanning point set A is ⁇ P1, P2, ..., PN ⁇
  • the second laser radar has a scanning wavelength of ⁇ 2
  • the scanned scanning point set B is ⁇ Q1.
  • the scanning wavelength of the third laser radar is ⁇ 3
  • the scanning scanning point set C is ⁇ L1, L2, ..., LN ⁇
  • the newly constructed scanning point set D is ⁇ Z1 for fusion.
  • the reflection intensity q( ⁇ 1) of the first laser radar corresponding to the Z1 point can take the reflection intensity of the scanning point of the nearest neighbor point in ⁇ P1, P2, ..., PN ⁇ , corresponding to the reflection of the second lidar
  • the intensity q( ⁇ 2) can take the reflection intensity of the scanning point of the nearest neighbor point in ⁇ Q1, Q2, ..., QN ⁇
  • the reflection intensity q( ⁇ 3) corresponding to the third laser radar can take ⁇ L1, L2, ..., the reflection intensity of the scanning point of the nearest neighbor point in LN ⁇ .
  • Other scan points are similar, complete the newly constructed scan point set D There is a fusion of scan points.
  • the fusion is completed, according to the characteristics of the reflection intensity information of the laser beams of different wavelengths for each scanning point in the merged scanning point set D (such as the laser reflection characteristics of the lasers of different wavelengths of the three types of ground environments mentioned in the previous embodiments)
  • the area division is performed, that is, the reflectance of lasers of different wavelengths is clustered according to the scanning points.
  • the manner of the growth of the connected domain region, or the K-means method may be used for the region segmentation.
  • the manner of the growth of the connected domain region is taken as an example, and the scope of application of the embodiment of the present invention is not applicable. The scope of protection is limited.
  • the whole process is as shown in FIG. 3a.
  • the distance information of each position point is often used for segmentation.
  • the area division may be performed together with other information in the traditional method, for example, combining distance information.
  • the position of each scanning point is represented by a distance and a scanning angle.
  • a scanning point is selected as a starting point, and the difference between the scanning point of the adjacent scanning angle and the scanning point and the scanning of the adjacent scanning angle are respectively calculated.
  • the similarity of the characteristics of the reflection intensity information of the different wavelengths of the point to the scanning point is classified into one class if the difference and the similarity are both smaller than the threshold.
  • the scanning points acquired by each laser radar are respectively determined according to the scanning point information.
  • Perform regional segmentation to generate clustered sub-regions with different laser reflection intensities for each lidar.
  • cluster the laser radars After the clustered sub-regions of each lidar are transformed into the same coordinate system, cluster the laser radars.
  • the scanning point information of each scanning point in the area is fused according to the position information of the transformed sub-area, and the ground environment is divided into sub-areas having different laser reflection characteristics. That is, after receiving the scanning point information from a plurality of laser radars, the scanning data of each laser radar is first clustered separately for region division, and then the regions of different lidars are divided. The data is fused based on location information.
  • the scan data of the plurality of lidars are respectively clustered according to the reflection intensity information, that is, the reflectances of the laser beams of different wavelengths are clustered according to the scan points, and a plurality of sub-regions having similar reflection intensity characteristics are generated.
  • clustering region segmentation
  • the point is used as a starting point, and then the similarity between the reflection intensity information of the adjacent scanning point and the reflection intensity information of the scanning point is calculated.
  • the scanning wavelength of the first laser radar is ⁇ 1
  • the scanning scanning point set A is ⁇ P1, P2, ..., PN ⁇
  • the scanning wavelength of the two laser radars is ⁇ 2
  • the scanning scanning point set B is ⁇ Q1, Q2, ..., QN ⁇
  • the scanning wavelength of the third laser radar is ⁇ 3
  • the scanning scanning point set C is ⁇ L1, L2, ..., LN ⁇ .
  • the first lidar of the first lidar is clustered to generate two sub-areas A1 and A2
  • the second lidar of the lidar is clustered to generate three sub-areas B1, B2 and B3.
  • the position information of each sub-area clustered by each lidar is first transformed into the same coordinate system, and based on the boundaries of all sub-regions.
  • the method for detecting the ground environment involved in the embodiments of the present invention is combined with the conventional method.
  • the image data of the surrounding ground environment can be collected by multiple cameras, and the data fusion of the laser radar scans of multiple wavelengths is added, and the processing and fusion of the image data collected by the camera are also added, for example, to the camera.
  • the image data is processed to identify the type of ground environment, and compared with the type of ground environment determined by the multi-wavelength lidar to further confirm the ground environment type and increase the robustness of the ground environment detection system.
  • FIG. 4 shows a possible structural diagram of a ground environment detecting device according to the present application.
  • the detecting device can implement the functions of the ground environment detecting device in the method embodiment in FIG. 2 above.
  • the ground environment detecting device 40 includes a laser scanning unit 41, a data collecting unit 42, a scan data processing unit 43, and a ground environment determining unit 44.
  • the laser scanning unit 41 is configured to scan the surrounding ground environment with different working wavelength laser detection signals, and receive the reflected signals reflected by the ground environment for the detection signals.
  • the data collection unit 42 is configured to determine the ground environment according to the reflection signals received by the laser scanning unit 41. Scanning point information of each scanning point, where the scanning point information includes a direction angle, a distance, and a laser reflection intensity of the scanning point with respect to the laser radar; the scanning data processing unit 43 is configured to determine each scanning point according to each scanning point information. Spatial coordinate information and laser reflection characteristics, the ground environment is divided into sub-regions having different laser reflection characteristics, where the laser reflection characteristics include reflectance to lasers of different wavelengths; the ground environment determination unit 44 is used to determine the ground of each sub-region Type of environment.
  • the laser scanning unit 41 may be a mechanical rotary laser radar or a solid state laser radar. It can be multiple laser radars with different operating wavelengths, or a laser radar with multiple operating wavelengths.
  • the ground environment detecting device scans the surrounding ground environment by using laser detection signals of different wavelengths, and determines the reflectance of each scanning point to different wavelength lasers according to the reflected signal reflected by the ground environment, and divides the ground environment into having Sub-areas of different laser reflection features, determined The type of ground environment for each sub-area. Since the lasers of different wavelengths are used to scan the ground, and the ground environment type is judged according to the laser reflection intensity of the ground environment under different wavelength lasers, the sensing effect on the complex ground environment is improved, and the passable road surface is better determined.
  • the ground environment determining unit 44 is specifically configured to determine a ground environment type of each sub-area according to laser reflection characteristics of each type of ground environment.
  • the laser reflection characteristics of each type of ground environment here can be represented by a predefined formula.
  • the ground environment determining unit 44 is specifically configured to input laser reflection features of each sub-region into the neural network, and obtain a ground environment type of each sub-region output by the neural network.
  • the neural network used here uses the different surface environment types as the input to the reflectivity data of different wavelength lasers, and the ground environment type as the output set of data as the sample data set to train, and obtain the data model for distinguishing different ground environment types.
  • the neural network outputs the ground environment type of each sub-area for the laser reflection characteristics of each sub-area input by the ground environment determining unit 44 through the data model.
  • the scan data processing unit 43 includes a fusion subunit 431 and a region division subunit 432.
  • the fusion sub-single 431 is configured to convert the position information of each scanning point to the same coordinate system according to the scanning point information of each scanning point and the installation position of each laser radar, and scan the scanning points acquired by each laser radar.
  • the point information is fused to determine spatial coordinate information and laser reflection characteristics of each scanning point;
  • the area dividing sub-unit 432 is configured to perform area segmentation according to spatial coordinate information and laser reflection characteristics of each scanning point, and divide the ground environment into different A sub-region of the laser reflection feature.
  • the fusion subunit 431 first fuses the scan points obtained by each lidar scan according to the position information, and the region segmentation subunit further performs region segmentation according to the position information of each merged scan point and the reflection intensity information of different wavelength lasers. Sub-regions with different laser reflection characteristics are obtained. Correlation of the method and detailed description of the scanning points obtained by scanning the respective lidars according to the position information and the area division according to the scanning points after the fusion, and the description of the embodiment shown in FIG. 3a for the step 204 in the previous method embodiment Basically the same, no longer repeat them here.
  • the scan data processing unit 43 includes a fusion subunit 431 and a region division subunit 432.
  • the area dividing sub-unit 432 is configured to separately segment the scanning points acquired by each laser radar according to the scanning point information, and generate clusters with different laser reflection intensities for each laser radar.
  • Sub-region; the fusion sub-unit 431 is configured to transform the position information of the sub-regions of each lidar into the same coordinate system, and scan the scan point information of each scan point in each sub-region after the lidar clustering according to the transformation
  • the positional information of the sub-areas is fused to divide the ground environment into sub-regions having different laser reflection characteristics.
  • the region dividing sub-unit 432 first clusters the scanning data of each laser radar separately, and then fuses the divided sub-regions of the different lidars by the sub-unit 431.
  • the data is fused according to the location information to segment the ground environment into sub-regions with different laser reflection characteristics.
  • the region dividing sub-unit 432 performs region segmentation by using a region growing manner or a K-means method.
  • Fig. 5 schematically shows another ground environment detecting device 50 of an embodiment of the present invention.
  • the ground environment detecting device 50 includes a laser scanning unit 51 and a data processing unit 52.
  • the laser scanning unit 51 is configured to scan the surrounding ground environment with different working wavelength laser detection signals, and receive the reflected signals reflected by the ground environment for the detection signals.
  • the data processing unit 52 includes a processor 521 and a memory 522.
  • the memory 522 is configured to store computer execution instructions, and the processor 521 executes computer execution instructions stored in the memory 522 for determining scan point information of each scan point of the ground environment according to the reflected signal, and determining each scan according to each scan point information.
  • the spatial coordinate information of the point and the laser reflection feature divide the ground environment into sub-regions with different laser reflection characteristics, and determine the ground environment type of each sub-region.
  • the scanning point information here includes the direction angle of the scanning point relative to the laser radar, the distance and the laser reflection intensity, and the laser reflection characteristic includes the reflectance of the laser light of different wavelengths.
  • the processor 521 can be a general-purpose central processing unit (CPU), a microprocessor, an application specific integrated circuit (ASIC), or one or more integrated circuits for executing related programs.
  • CPU central processing unit
  • ASIC application specific integrated circuit
  • the memory 522 may be a read only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM).
  • ROM read only memory
  • RAM random access memory
  • the program code for implementing the technical solution provided by the embodiment of the present invention is stored in the memory 522 and executed by the processor 521.
  • memory 522 can be used to store computer-executed instructions, as well as to store various information, such as laser reflection signature formulas for each type of terrestrial environment.
  • the processor 521 can read the information stored by the memory 522 or store the collected information to the memory 522.
  • the processor 521 is configured to determine a ground environment type of each sub-area according to a laser reflection characteristic of each type of ground environment.
  • the processor 521 is configured to input laser reflection features of each sub-region into the neural network, and obtain a ground environment type of each of the sub-regions output by the neural network, where the neural network is used to distinguish data models of different ground environment types.
  • the ground environment type is trained as a set of data of the output as a sample data set.
  • the processor 521 is configured to convert position information of each scanning point to the same coordinate system according to the scanning point information of each scanning point and the installation position of each laser radar, and to scan points acquired by each laser radar. Scanning point information is fused, determining spatial coordinate information and laser reflection characteristics of each scanning point, and performing regional segmentation according to spatial coordinate information and laser reflection characteristics of each scanning point, and dividing the ground environment into sub-regions having different laser reflection characteristics. .
  • the processor 521 is configured to separately segment the scan points acquired by each laser radar according to the scan point information, and generate a clustered sub-region with different laser reflection intensities for each lidar, and gather each lidar.
  • the position information of the sub-sub-region is transformed into the same coordinate system, and the scanning point information of each scanning point in the sub-region after each lidar clustering is fused according to the position information of the transformed sub-region, and the ground environment is divided into different lasers.
  • a sub-region of the reflection feature is configured to separately segment the scan points acquired by each laser radar according to the scan point information, and generate a clustered sub-region with different laser reflection intensities for each lidar, and gather each lidar.
  • the position information of the sub-sub-region is transformed into the same coordinate system, and the scanning point information of each scanning point in the sub-region after each lidar clustering is fused according to the position information of the transformed sub-region, and the ground environment is divided into different lasers.
  • a sub-region of the reflection feature is configured to separately
  • the processor 521 can also perform region segmentation by using a region growth mode or a K-means method.
  • a communication interface and a bus wherein the communication interface can employ a transceiver such as, but not limited to, a transceiver for implementing between the data processing unit 52 and the laser scanning unit 51.
  • the bus can include a path for transferring information between processor 521 and memory 522.
  • the bus may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus.
  • PCI peripheral component interconnect
  • EISA extended industry standard architecture
  • the bus can be divided into an address bus, a data bus, a control bus, and the like.
  • the ground environment detecting device 50 shown in FIG. 5 may also include hardware devices that implement other additional functions, depending on the particular needs.
  • the disclosed systems, devices, and methods may be implemented in other manners.
  • the device embodiments described above are merely illustrative.
  • the division of the unit/module is only a logical function division.
  • there may be another division manner for example, multiple units or components may be used. Combinations can be integrated into another system, or some features can be ignored or not executed.
  • the mutual coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or unit, or an electrical, mechanical or other form of connection.
  • the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the objectives of the embodiments of the present invention.
  • each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
  • the above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.
  • the integrated unit if implemented in the form of a software functional unit and sold or used as a standalone product, may be stored in one computer computer readable storage medium or as one or more instructions or code embodied on a computer readable medium.
  • Computer readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one location to another.
  • a storage medium may be any available media that can be accessed by a computer.
  • computer readable media may comprise RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage media or other magnetic storage device, or can be used for carrying or storing in the form of an instruction or data structure.
  • the desired program code and any other medium that can be accessed by the computer may be stored in one computer computer readable storage medium or as one or more instructions or code embodied on a computer readable medium.
  • transmission Computer readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one location to another.
  • a storage medium may be any available media that can
  • connection may suitably be a computer readable medium.
  • the software is transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave
  • coaxial cable , fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, wireless, and microwave are included in the definition of the medium to which they belong.
  • a disk and a disc include a compact disc (CD), a laser disc, a compact disc, a digital versatile disc (DVD), a floppy disk, and a Blu-ray disc, wherein the disc is usually magnetically copied, and the disc is The laser is used to optically replicate the data. Combinations of the above should also be included within the scope of the computer readable media. Based on such understanding, the technical solution of the present invention is essential or part of the prior art, or all or part of the technical solution may be stored in a storage medium, including a plurality of instructions for causing a computer device (may be a personal computer, server, or network device, etc.) performing all or part of the steps of the methods described in various embodiments of the present invention.
  • a computer device may be a personal computer, server, or network device, etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Software Systems (AREA)
  • Multimedia (AREA)
  • Artificial Intelligence (AREA)
  • General Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Databases & Information Systems (AREA)
  • Remote Sensing (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Data Mining & Analysis (AREA)
  • Electromagnetism (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Biomedical Technology (AREA)
  • Mathematical Physics (AREA)
  • Molecular Biology (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Optical Radar Systems And Details Thereof (AREA)
  • Optics & Photonics (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Automation & Control Theory (AREA)

Abstract

一种地面环境检测的方法和装置，方法包括：采用不同工作波长激光探测信号扫描地面环境，接收地面环境反射回来的反射信号，根据反射信号确定地面环境的每个扫描点的扫描点信息，根据每个扫描点信息确定每个扫描点的空间坐标信息和激光反射特征，将地面环境分割成具有不同激光反射特征的子区域，确定各子区域的地面环境类型。由于利用了不同工作波长的激光雷达扫描地面，并依据地面环境在不同波长激光下的反射强度判断地面环境类型，提高了对复杂地面环境的感知效果，更好地确定可通行路面。

Description

一种地面环境的检测方法和装置 技术领域
本发明实施例涉及自动驾驶领域，尤其涉及地面环境检测的方法和装置。
背景技术
自动驾驶汽车也称无人驾驶汽车，是一种通过计算机系统或者终端设备实现无人驾驶的智能汽车。实现自动驾驶的重要前提之一就是对地面环境的检测，通过对地面环境的检测，确定路面状况、车辆可通行的区域等，从而服务于后续自动驾驶的规划、决策与控制。
现有的技术往往采用基于单一工作波长的激光雷达扫描或者图像识别的方式，依据路面的空间信息或者图像特征确定可通行的路面,然而这些方式仅对简单地面环境(如平整的高速公路的路面)或图像特征明显的地面环境(如有清晰的车道线的路面)有较好的效果，难以正确检测复杂的道路环境，如无法检测路面不平整，或者路面有草丛、积水的道路，或在道路标线不全、路面有阴影或者夜间时检测效果非常差。
发明内容
有鉴于此，本发明实施例提供了一种基于多波长激光雷达检测地面环境的方法和装置,该方法通过采用基于不同工作波长的激光雷达扫描地面，依据地面环境在不同波长激光下的反射强度特征判断地面环境类型，提高对复杂地面环境的感知效果，更好地确定可通行路面。
第一方面，本申请提供了一种地面环境检测的方法,包括：采用不同工作波长激光探测信号扫描地面环境,接收地面环境针对探测信号反射回来的反射信号；根据反射信号确定地面环境的每个扫描点的扫描点信息，扫描点信息包含扫描点相对于激光雷达的方向角、距离以及激光反射强度；根据每个扫描点信息确定每个扫描点的空间坐标信息和激光反射特征，将地面环境分 割成具有不同激光反射特征的子区域，激光反射特征包括对不同波长激光的反射率；确定各子区域的地面环境类型。
可以理解的是，可以采用多个具有不同工作波长的激光雷达分别发射自身工作波长的激光探测信号扫描周围地面环境，也可以采用具有多个工作波长的激光雷达分别发射不同波长的激光探测信号。
采用上述的方法，用不同波长的激光探测信号扫描周围地面环境，确定每个扫描点对不同波长激光的反射率，将地面环境分割成具有不同激光反射特征的子区域，确定各子区域的地面环境类型。依据地面环境在不同波长下的反射强度判断地面环境类型，提高了对复杂地面环境的感知效果，更好地确定可通行路面。
在一种可能的实现中，根据每种类型地面环境的激光反射特征，确定各子区域的地面环境类型。
在一种可能的实现中，将各子区域的激光反射特征输入神经网络，获取神经网络输出的各子区域的地面环境类型。这里采用的神经网络通过采用不同地面环境类型对不同波长激光的反射率数据作为输入，地面环境类型作为输出的一组数据作为样本数据集进行训练，得到用于区分不同地面环境类型的数据模型。通过神经网络根据训练得到的数据模型来针对各子区域的激光反射特征确定各子区域的地面环境类型，提升了地面环境类型识别的准确率。
在一种可能的实现中，依据每个扫描点的扫描点信息和各激光雷达的安装位置，将各扫描点的位置信息变换到同一坐标系，对通过各激光雷达获取的各扫描点的扫描点信息进行融合，确定每个扫描点的空间坐标信息和激光反射特征；根据每个扫描点的空间坐标信息和激光反射特征进行区域分割，将地面环境分割成具有不同激光反射特征的子区域。即首先依据扫描点位置信息将多个激光雷达的扫描数据进行融合，然后再将融合后的数据依据不同波长的反射率信息进行聚类，将地面环境分割成具有不同激光反射特征的子区域。
在一种可能的实现中，将各激光雷达获取的扫描点根据扫描点信息分别进行区域分割，生成每个激光雷达具有不同激光反射强度的聚类后子区域；将每个激光雷达的聚类后子区域的位置信息变换到同一坐标系，对各激光雷 达聚类后子区域内的各扫描点的扫描点信息按照变换后子区域的位置信息进行融合，将地面环境分割成具有不同激光反射特征的子区域。即首先将各激光雷达的扫描数据分别聚类进行区域分割，然后将不同激光雷达的区域分割后的数据依据位置信息进行融合，将地面环境分割成具有不同激光反射特征的子区域。
在一种可能的实现中，采用区域生长方式或者K-means方式进行区域分割。
进一步的，通过多个摄像头采集地面环境的图像数据，对摄像头采集到的图像数据进行处理，识别地面环境类型，与通过多波长激光雷达确定的各子区域的地面环境类型进行融合，增加了地面环境检测系统的鲁棒性。
第二方面，本发明实施例提供了一种地面环境检测设备,包括：激光扫描单元，用于采用不同工作波长激光探测信号扫描地面环境，接收地面环境针对探测信号反射回来的反射信号；数据采集单元，用于根据反射信号确定地面环境的每个扫描点的扫描点信息，该扫描点信息包含扫描点相对于激光雷达的方向角、距离以及激光反射强度；扫描数据处理单元，用于根据每个扫描点信息确定每个扫描点的空间坐标信息和激光反射特征，将地面环境分割成具有不同激光反射特征的子区域，该激光反射特征包括对不同波长激光的反射率；地面环境确定单元，用于确定各子区域的地面环境类型。
在一种可能的实现方式中，激光扫描单元为机械旋转式激光雷达或固态激光雷达。
上述的地面环境检测设备采用不同波长的激光探测信号扫描周围地面环境，并根据地面环境反射的反射信号，确定每个扫描点对不同波长激光的反射率，将地面环境分割成具有不同激光反射特征的子区域，确定各子区域的地面环境类型。依据地面环境在不同激光波长下的反射强度判断地面环境类型，提高了对复杂地面环境的感知效果，更好地确定可通行路面。
在一种可能的实现方式中，地面环境确定单元用于根据每种类型地面环境的激光反射特征，确定各子区域的地面环境类型。
在另一种可能的实现方式中，地面环境确定单元用于将各子区域的激光反射特征输入神经网络，获取该神经网络输出的各子区域的地面环境类型； 其中，该神经网络用于区分不同地面环境类型的数据模型通过采用不同地面环境类型对不同波长激光的反射率数据作为输入，地面环境类型作为输出的一组数据作为样本数据集进行训练得到。
在一种可能的实现方式中，扫描数据处理单元包括：融合子单元，用于依据每个扫描点的扫描点信息和各激光雷达的安装位置，将各扫描点的位置信息变换到同一坐标系，对通过各激光雷达获取的各扫描点的扫描点信息进行融合，确定每个扫描点的空间坐标信息和激光反射特征；区域分割子单元，用于根据每个扫描点的空间坐标信息和激光反射特征进行区域分割，将地面环境分割成具有不同激光反射特征的子区域。即，融合子单元用于首先依据扫描点位置信息将多个激光雷达的扫描数据进行融合，然后区域分割子单元用于将融合后的数据依据不同波长的反射率信息进行聚类，将地面环境分割成具有不同激光反射特征的子区域。
在另一种可能的实现方式中，扫描数据处理单元，包括：区域分割子单元用于将各激光雷达获取的扫描点根据扫描点信息分别进行区域分割，生成每个激光雷达具有不同激光反射强度的聚类后子区域；融合子单元，用于将每个激光雷达的聚类后子区域的位置信息变换到同一坐标系，对各激光雷达聚类后子区域内的各扫描点的扫描点信息按照变换后子区域的位置信息进行融合，将地面环境分割成具有不同激光反射特征的子区域。即，区域分割子单元用于首先将各激光雷达的扫描数据分别聚类进行区域分割，然后融合子单元用于将不同激光雷达的区域分割后的数据依据位置信息进行融合，将地面环境分割成具有不同激光反射特征的子区域。
在一种可能的实现方式中，区域分割子单元采用区域生长方式或者K-means方式进行区域分割。
第三方面，本发明实施例提供了一种地面环境检测设备,包括：激光扫描单元，用于采用不同工作波长激光探测信号扫描地面环境，接收地面环境针对探测信号反射回来的反射信号；数据处理单元，包括处理器和存储器，存储器用于存储计算机执行指令，处理器执行计算机执行指令，用于根据反射信号确定地面环境的每个扫描点的扫描点信息，根据每个扫描点信息确定每个扫描点的空间坐标信息和激光反射特征，将地面环境分割成具有不同激光 反射特征的子区域，确定各子区域的地面环境类型；其中,扫描点信息包含扫描点相对于激光雷达的方向角、距离以及激光反射强度，激光反射特征包括对不同波长激光的反射率。
在一种可能的实现方式中，激光扫描单元为机械旋转式激光雷达或固态激光雷达。
上述的地面环境检测设备采用不同波长的激光探测信号扫描周围地面环境，并根据地面环境反射的反射信号，确定每个扫描点对不同波长激光的反射率，将地面环境分割成具有不同激光反射特征的子区域，确定各子区域的地面环境类型。依据地面环境在不同波长下的反射强度判断地面环境类型，提高了对复杂地面环境的感知效果，更好地确定可通行路面。
在一种可能的实现方式中，该处理器用于根据每种类型地面环境的激光反射特征，确定各子区域的地面环境类型。
在另一种可能的实现方式中，该处理器用于将各子区域的激光反射特征输入神经网络，获取神经网络输出的各子区域的地面环境类型；其中，该神经网络用于区分不同地面环境类型的数据模型通过采用不同地面环境类型对不同波长激光的反射率数据作为输入，地面环境类型作为输出的一组数据作为样本数据集进行训练得到。
在一种可能的实现方式中，该处理器用于依据每个扫描点的扫描点信息和各激光雷达的安装位置，将各扫描点的位置信息变换到同一坐标系，对通过各激光雷达获取的各扫描点的扫描点信息进行融合，确定每个扫描点的空间坐标信息和激光反射特征，根据每个扫描点的空间坐标信息和激光反射特征进行区域分割，将地面环境分割成具有不同激光反射特征的子区域。即，处理器用于首先依据扫描点位置信息将多个激光雷达的扫描数据进行融合，然后将融合后的数据依据不同波长的反射率信息进行聚类，将地面环境分割成具有不同激光反射特征的子区域。
在另一种可能的实现方式中，该处理器用于将各激光雷达获取的扫描点根据扫描点信息分别进行区域分割，生成每个激光雷达具有不同激光反射强度的聚类后子区域；将每个激光雷达的聚类后子区域的位置信息变换到同一坐标系，对各激光雷达聚类后子区域内的各扫描点的扫描点信息按照变换后 子区域的位置信息进行融合，将地面环境分割成具有不同激光反射特征的子区域。即，处理器用于首先将各激光雷达的扫描数据分别聚类进行区域分割，然后将不同激光雷达的区域分割后的数据依据位置信息进行融合，将地面环境分割成具有不同激光反射特征的子区域。
在一种可能的实现方式中，该处理器采用区域生长方式或者K-means方式进行区域分割。
通过上述技术方案，本发明实施例提供的地面环境检测方法和装置，采用不同波长的激光探测信号扫描周围地面环境，并根据每个扫描点对不同波长激光的反射率，将地面环境分割成具有不同激光反射特征的子区域，确定各子区域的地面环境类型。由于利用了不同波长的激光扫描地面，并依据地面环境在不同波长激光下的反射强度特征判断地面环境类型，提高了对复杂地面环境的感知效果，更好地确定可通行路面。
附图说明
为了更清楚地说明本发明实施例的技术方案，下面将对实施例描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1为本发明实施例应用的地面环境检测场景示意图；
图2为一种地面环境检测方法流程图；
图3a为一种分割子区域的方法示意图；
图3b为另一种分割子区域的方法示意图；
图4为一种地面环境检测设备的结构示意图；
图5为另一种地面环境检测设备的结构示意图。
具体实施方式
下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行描述。
图1是一个本发明实施例所应用的地面环境检测场景示意图。整个地面环境检测系统包括安装在自动驾驶汽车上的激光雷达以及扫描数据处理单元组成。
其中的激光雷达可以采用机械旋转式激光雷达，也可以采用固态激光雷达，用于向路面发射激光信号进行周边环境扫描，每个激光雷达将接收地面反射回的激光信号，返回一系列的扫描点信息。现有激光雷达主要采用950nm工作波长，而采用该单一工作波长的激光雷达往往难以识别复杂路面环境，本发明实施例的技术方案中采用了多波长激光雷达，即采用不同工作波长的激光雷达扫描地面。图1中示出了自动驾驶车辆上安装了3个激光雷达，实际应用时可根据需求和成本等因素考虑，灵活选择所安装激光雷达的数量，本发明对此不做限制。
扫描数据处理单元通过汇总地面各激光雷达接收的所有扫描点的信息，通过激光信号发射--接收的时间差及信号信息，确定返回地面的各扫描点的距离及反射强度等，从而提取出路面的三维空间结构和反射强度，判断是否为可通行的路面。而本发明实施例中，因为采用了不同波长的激光进行地面扫描，扫描数据数据处理单元进一步的可以根据不同波长激光雷达的原始数据，将地面划分成具有不同激光反射特征的子区域，即对于不同波长激光的反射率不同的子区域，最终确定各子区域的地面环境类型。
结合图1所示的地面环境检测应用场景示意图，本发明实施例提供了一种地面环境的检测方法,如图2所示，具体的过程包括：
步骤201，采用不同工作波长激光探测信号扫描地面环境。由具有不同工作波长的激光雷达发射激光探测信号扫描周围地面环境，可选的，可采用多个激光雷达分别发射不同波长的激光探测信号进行扫描，可以采用多个具有不同工作波长的激光雷达分别发射自身工作波长的激光探测信号，也可以采用具有多个工作波长的激光雷达分别发射不同工作波长的激光探测信号。
步骤202，接收所述地面环境针对所述探测信号反射回来的反射信号。激光雷达发射出的激光探测信号遇到周围地面环境中的目标(也称作扫描点)，会进行反射，激光雷达会接收到的从目标反射回来的反射信号。
步骤203，根据接收到的反射信号确定扫描点的扫描点信息，这里的扫 描点信息包含扫描点相对于激光雷达的方向角、距离以及激光反射强度。各雷达接收到的反射信号信息可被传送给扫描数据处理单元，这里的传送方式，可以采用无线传输方式(例如蓝牙)、线缆连接传输方式(如信号线直接相连)等。扫描数据处理单元根据各激光雷达接收的各扫描点返回的发射信号，如激光探测信号发射到接收的时间差以及发射和接收的信号强度信息等，确定各扫描点的方向角、距离及激光反射强度等信息。这里通过接收反射信号计算出扫描点的扫描点信息的方法，可以采用现有激光雷达探测中的常用方法，本发明对此不做详细说明，不对本发明的适用范围造成影响。
步骤204，根据每个扫描点信息确定每个扫描点的空间坐标信息和激光反射特征，将地面环境分割成具有不同激光反射特征的子区域。这里的激光反射特征包括对不同波长激光的反射率。
扫描数据处理单元结合所有扫描点的扫描点信息，即根据所有扫描点的方向角、距离以及激光反射强度，将不同工作波长的激光雷达返回的扫描点数据进行融合，生成相对于同一坐标系的扫描点信息，将所有扫描点中具有相近空间坐标和相似激光反射特征的点进行聚类，聚类后将地面环境分割成具有不同激光反射特征的子区域。
步骤205，确定各子区域的地面环境类型。
可选的，根据每种类型地面环境的激光反射特征，即对不同波长激光的反射率，确定各子区域的地面环境类型。在本实施例中，以3个具有不同工作波长的激光雷达为例，设工作波长分别为λ1、λ2、λ3。假设待判断的地面环境有3种类型，A、B、C，(例如分别表示道路、积水、植被)，q表示反射率，每种类型地面对不同波长激光的激光反射特征满足如下关系：
类型A：q(λ1)≈q(λ2)≈q(λ3)，对于三个波长的激光的反射率基本相同；
类型B：q(λ1)≠0，q(λ2)≈0，q(λ3)≈0，对于波长为λ1的激光的反射率不为0，对于波长为λ2、λ3的激光的反射率接近0；
类型C：q(λ1)>q(λ2)>q(λ3)，对于波长为λ1、λ2、λ3的激光的反射率依次降低。
根据上述在不同波长下激光的反射率关系判断各子区域的类型，确定可 通行路面。即通过如上所示的预定义的每种类型地面环境的激光反射特征的公式，确定子区域为哪种地面。
可选的，在实际应用中，对于具有复杂不同波长的激光的反射率关系的情形，除了可以采取如上所示的通过预定义的公式进行判断外，还可以采用基于机器学习的方法，提取每个子区域的激光反射特征进行判断。采用不同地面环境类型对不同波长的激光的反射率数据作为输入、对应的地面环境类型作为输出的一组数据作为样本数据集，对神经网络进行训练得到该神经网络用于区分不同地面环境类型的数据模型。在具体应用时，将各子区域的激光反射特征输入神经网络，就可以获取神经网络根据数据模型输出的各子区域的地面环境类型。例如，用向量Q＝[q1，q2，…，qN]代表一组某种地面环境类型对N个不同波长的激光的反射率数据，作为神经网络的输入，T∈{T1，T2，…，TM}代表地面环境类型(包含M个分类)，作为神经网络的输出。利用一系列Q、T一一对应的数据作为样本数据集训练神经网络，得到神经网络用于区分地面环境类型的数据模型。这样，将一组各子区域的激光反射特征数据，即个子区域对不同波长激光的反射率数据输入该神经网络，神经网络就可以根据训练得到的数据模型输出各子区域对应的地面环境类型。
图2所示的本发明实施例提供的地面环境检测方法，采用不同波长的激光探测信号扫描周围地面环境，并根据地面环境反射的反射信号，确定每个扫描点对不同波长激光的反射率，将地面环境分割成具有不同激光反射特征的子区域，确定各子区域的地面环境类型。由于利用了不同波长的激光扫描地面，并依据地面环境在不同波长激光下的反射强度判断地面环境类型，提高了对复杂地面环境的感知效果，更好地确定可通行路面。
进一步的，对于前面实施例中的步骤204中，根据每个扫描点信息确定每个扫描点的空间坐标信息和激光反射特征，将地面环境分割成具有不同激光反射特征的子区域，可以通过如下方式：依据每个扫描点的扫描点信息和各激光雷达的安装位置，将各扫描点的位置信息变换到同一坐标系，对通过各激光雷达获取的各扫描点的扫描点信息进行融合，确定每个扫描点的空间坐标信息和激光反射特征，根据每个扫描点的空间坐标信息和激光反射特征进行区域分割，将地面环境分割成具有不同激光反射特征的子区域。即在接 收到来自多个激光雷达的扫描点信息后，首先依据扫描点位置信息将多个激光雷达的扫描数据进行融合，然后再将融合后的数据依据不同波长的反射率信息进行聚类。
具体的，在接收多个激光雷达的扫描点信息后，首先依据扫描点的位置信息(包括方向角和距离)将多个激光雷达的扫描数据融合，即将各激光雷达扫描到的扫描点的扫描点信息按照扫描点进行合并。如果多个激光雷达的安装位置不同，需要首先将各激光雷达获取的扫描点的位置变换到同一坐标系中，如车辆坐标系。完成坐标变换后，如果各个激光雷达获取的扫描点位置信息并非一一对应而是有错位的情况，需要新构造一个新的扫描点集合存放不同激光雷达的扫描信息，该扫描点集合中每个扫描点的位置可以根据各个激光雷达的扫描点位置信息取值，也可以是事先定义好的取值。该扫描点集合中每个点的不同激光雷达的反射强度信息(即针对不同波长激光的反射率信息)根据各个激光雷达的扫描点反射强度确定，可以采用就近匹配或者插值的方式，最终输出每个扫描点的位置和针对不同波长激光雷达的反射强度，即，最终得到每个扫描点的位置信息和该扫描点针对不同波长激光的反射率信息。例如，假设三个激光雷达均扫描了N个扫描点，当然也可能不同的激光雷达扫描出不同的扫描点数量，为描述方便，此处假设各激光雷达扫描到的扫描点数量相同，并不对本发明实施例的适用范围和保护范围造成影响。假设，第一个激光雷达的扫描波长为λ1，扫描的扫描点集A为{P1，P2，…，PN}，第二个激光雷达的扫描波长为λ2，扫描的扫描点集B为{Q1，Q2，…，QN}，第三个激光雷达的扫描波长为λ3，扫描的扫描点集C为{L1，L2，…，LN}，为进行融合而新构造的扫描点集合D为{Z1，Z2，…，ZN}，且位置为事先定义好的{(x1，y1，z1)，(x2，y2，z2)，…，(xN，yN，zN)}，其中的1，2，…，N表示各扫描点集中的扫描点，x，y，z表示扫描点的三维坐标。则Z1点的对应第一个激光雷达的反射强度q(λ1)可以取{P1，P2,…,PN}中最邻近位置点的扫描点的反射强度取值，对应第二个激光雷达的反射强度q(λ2)可以取{Q1，Q2,…,QN}中最邻近位置点的扫描点的反射强度取值，对应第三个激光雷达的反射强度q(λ3)可以取{L1，L2，…，LN}中最邻近位置点的扫描点的反射强度取值。其他扫描点类似，完成新构造的扫描点集合D中所 有扫描点的融合。融合完成后，根据融合后的扫描点集D中每个扫描点对不同波长激光的反射强度信息的特征(如前面实施例中提到的3种地面环境类型对不同波长激光的激光反射特征)进行区域分割，即按照扫描点对不同波长激光的反射率进行聚类。可选的，可以采用连通域区域生长的方式，或者K均值(K-means)方式等进行区域分割，这里以连通域区域生长的方式为例进行说明，并不对本发明实施例的适用范围和保护范围造成限制。首先取一个扫描点作为起始点，然后计算临近扫描点的不同波长的反射强度信息的特征与该扫描点的相似度，如果小于设定的阈值则视为连通，归为一类，否则视为不连通，再继续计算其他相邻扫描点，以此类推，直到完成所有扫描点的区域分割，生成具有不同激光反射强度特征的子区域。整个处理过程如图3a所示，先将各激光雷达扫描得到的扫描点按照位置信息进行融合，再根据每个融合后的扫描点的位置信息和对不同波长激光的反射强度信息进行区域分割，获得具有不同激光反射特征的子区域。
在传统的激光雷达数据处理过程中，往往利用各个位置点的距离信息等进行分割，在本发明实施例中，除了可以根据每个位置点的不同波长的反射强度信息的特征进行区域分割外，可选的，还可以结合传统方法中的其他信息一起进行区域分割，例如结合距离信息。例如，将每个扫描点的位置以距离和扫描角度表示，首先选取一个扫描点作为起始点，分别计算相邻扫描角度的扫描点的距离与该扫描点的差值以及该临近扫描角度的扫描点的不同波长的反射强度信息的特征与该扫描点的相似度，如果差值和相似度都小于阈值则归为一类。
可选的，对于前面实施例中的步骤204的实施方式，除了采用前面如图3a所示的实施方式外，还可以采用如下的实施方式：将各激光雷达获取的扫描点根据扫描点信息分别进行区域分割，生成每个激光雷达具有不同激光反射强度的聚类后子区域，在将每个激光雷达的聚类后子区域的位置信息变换到同一坐标系，对各激光雷达聚类后子区域内的各扫描点的扫描点信息按照变换后子区域的位置信息进行融合，将地面环境分割成具有不同激光反射特征的子区域。即在接收到来自多个激光雷达的扫描点信息后，首先将各激光雷达的扫描数据分别聚类进行区域分割，然后将不同激光雷达的区域分割后 的数据依据位置信息进行融合。
具体的，首先将多个激光雷达的扫描数据依据反射强度信息分别进行聚类，即按照扫描点对不同波长激光的反射率进行聚类，生成多个具有相似反射强度特征的子区域。与前面实施例中所述的方法一样，可以采用连通域区域生长的方式，或者K-means方式等进行聚类(区域分割)，同样以采用连通域区域生长的方式为例，首先取一扫描点作为起始点，然后计算临近扫描点的反射强度信息与该扫描点的反射强度信息的相似度，如果小于设定的阈值则视为连通，归为一类，否则视为不连通，再继续计算其他相邻扫描点，以此类推，得到每个激光雷达针对自身波长具有不同反射率的子区域。同样的，与前面实施例中所述方法一样，除了可以根据每个位置点的不同波长的反射强度信息的特征进行区域分割外，还可以结合传统方法中的其他信息(如距离信息)一起进行区域分割，分割的方法和细节描述与前面实施例基本相同，这里不再赘述。接下来，将不同激光雷达聚类后得到的分割的子区域按照子区域的位置信息进行融合。与前面实施方法类似，为保证融合的准确性，首先需要将不同激光雷达聚类后生成的子区域的位置信息进行融合，将子区域的位置信息变换到同一坐标系中。对各激光雷达聚类后子区域内的各扫描点的扫描点信息按照变换后子区域的位置信息进行融合，将地面环境分割成具有不同激光反射特征的子区域。首先在新的坐标系中根据所有激光雷达聚类后的子区域边界划分出一系列新的子区域，然后根据每个激光雷达聚类后子区域的反射强度信息确定新生成的各子区域对不同波长激光的反射率信息。与前面实施例一样，这里也假设三个激光雷达均扫描了N个扫描点，第一个激光雷达的扫描波长为λ1，扫描的扫描点集A为{P1，P2，…，PN}，第二个激光雷达的扫描波长为λ2，扫描的扫描点集B为{Q1，Q2，…，QN}，第三个激光雷达的扫描波长为λ3，扫描的扫描点集C为{L1，L2，…，LN}。如图3b所示，第一个激光雷达的扫描点集A聚类后生成了A1、A2两个子区域，第二个激光雷达的扫描点集B聚类后生成B1、B2、B3三个子区域，第三个激光雷达的扫描点集C聚类后生成C1、C2两个子区域。最终，根据A1、A2、B1、B2、B3、C1、C2所有子区域进行融合，首先将每个激光雷达聚类后的子区域的位置信息变换到同一坐标系，并根据所有子区域的边界划分出X、Y、 Z三个子区域，根据每个子区域对应的激光雷达的波长信息融合确定每个子区域对不同波长的反射强度信息，将地面环境分割成具有不同激光反射特征的子区域X、Y、Z。
进一步的，本发明实施例中所涉及的地面环境的检测方法与采用传统方法的结合。比如，可以通过多个摄像头采集周围地面环境的图像数据，在对多波长的各激光雷达扫描得到的数据融合外，还增加对摄像头采集到的图像数据的处理和融合，例如，对摄像头采集到的图像数据进行处理，识别地面环境类型，与通过多波长激光雷达确定的各子区域的地面环境类型进行对比，进一步确认地面环境类型，增加了地面环境检测系统的鲁棒性。
上文结合图1、图2、图3a、图3b，对本发明实施例提供的地面环境检测方法进行了详细的介绍。图4示出了本申请所涉及的地面环境检测设备的一种可能的结构示意图。该检测设备可以实现上述图2中方法实施例中地面环境检测设备的功能，本实施例中未定义的术语及实现细节可以参考上述图2的方法实施例。如图4所示，该地面环境检测设备40包括激光扫描单元41，数据采集单元42，扫描数据处理单元43和地面环境确定单元44。其中，激光扫描单元41用于采用不同工作波长激光探测信号扫描周围地面环境，接收地面环境针对探测信号反射回来的反射信号；数据采集单元42用于根据激光扫描单元41接收的反射信号确定地面环境的每个扫描点的扫描点信息，这里的扫描点信息包含扫描点相对于激光雷达的方向角、距离以及激光反射强度；扫描数据处理单元43用于根据每个扫描点信息确定每个扫描点的空间坐标信息和激光反射特征，将地面环境分割成具有不同激光反射特征的子区域，这里的激光反射特征包括对不同波长激光的反射率；地面环境确定单元44用于确定各子区域的地面环境类型。
可选的，激光扫描单元41可以是机械旋转式激光雷达，也可以是固态激光雷达。可以是具有不同工作波长的多个激光雷达，也可以是具有多个工作波长的激光雷达。
本实施例提供的地面环境检测设备，采用不同波长的激光探测信号扫描周围地面环境，并根据地面环境反射的反射信号，确定每个扫描点对不同波长激光的反射率，将地面环境分割成具有不同激光反射特征的子区域，确定 各子区域的地面环境类型。由于利用了不同波长的激光扫描地面，并依据地面环境在不同波长激光下的激光反射强度判断地面环境类型，提高了对复杂地面环境的感知效果，更好地确定可通行路面。
可选的，地面环境确定单元44具体用于根据每种类型地面环境的激光反射特征，确定各子区域的地面环境类型。这里的每种类型地面环境的激光反射特征可以通过预定义的公式进行表示。
可选的，地面环境确定单元44具体用于将各子区域的激光反射特征输入神经网络，获取神经网络输出的各子区域的地面环境类型。这里所采用的神经网络通过采用不同地面环境类型对不同波长激光的反射率数据作为输入，地面环境类型作为输出的一组数据作为样本数据集进行训练，得到用于区分不同地面环境类型的数据模型，该神经网络通过该数据模型，对于地面环境确定单元44输入的各子区域的激光反射特征，输出的各子区域的地面环境类型。
进一步的，如图4所示，扫描数据处理单元43包括融合子单元431和区域分割子单元432。其中，融合子单431用于依据每个扫描点的扫描点信息和各激光雷达的安装位置，将各扫描点的位置信息变换到同一坐标系，对通过各激光雷达获取的各扫描点的扫描点信息进行融合，确定每个扫描点的空间坐标信息和激光反射特征；区域分割子单元432用于根据每个扫描点的空间坐标信息和激光反射特征进行区域分割，将地面环境分割成具有不同激光反射特征的子区域。即，融合子单元431先将各激光雷达扫描得到的扫描点按照位置信息进行融合，区域分割子单元再根据每个融合后的扫描点的位置信息和对不同波长激光的反射强度信息进行区域分割，获得具有不同激光反射特征的子区域。相关将各激光雷达扫描得到的扫描点按照位置信息进行融合以及根据将融合后的扫描点进行区域分割的方法和细节描述与前面方法实施例中对于步骤204采用图3a所示的实施方式的说明基本相同，这里不再赘述。
可选的，扫描数据处理单元43包括融合子单元431和区域分割子单元432。其中，区域分割子单元432用于将各激光雷达获取的扫描点根据扫描点信息分别进行区域分割，生成每个激光雷达具有不同激光反射强度的聚类后 子区域；融合子单元431用于将每个激光雷达的聚类后子区域的位置信息变换到同一坐标系，对各激光雷达聚类后子区域内的各扫描点的扫描点信息按照变换后子区域的位置信息进行融合，将地面环境分割成具有不同激光反射特征的子区域。即，在接收到来自多个激光雷达的扫描点信息后，区域分割子单元432首先将各激光雷达的扫描数据分别聚类进行分割，然后融合子单元431对不同激光雷达的分割后的子区域的数据依据位置信息进行融合，将地面环境分割成具有不同激光反射特征的子区域。相关将各激光雷达的扫描数据分别聚类分割成子区域以及对分割后的子区域的数据依据位置信息进行融合的方法和细节描述与前面方法实施例中对于步骤204采用图3b所示的实施方式的说明基本相同，这里不再赘述。
可选的，区域分割子单元432采用区域生长方式或者K-means方式进行区域分割。
图5示意性地示出了本发明实施例另一地面环境检测设备50。如图5所示，地面环境检测设备50包括激光扫描单元51和数据处理单元52。其中，激光扫描单元51用于采用不同工作波长激光探测信号扫描周围地面环境，接收所述地面环境针对所述探测信号反射回来的反射信号；数据处理单元52包括处理器521和存储器522。存储器522用于存储计算机执行指令，处理器521执行存储器522中存储的计算机执行指令，用于根据反射信号确定地面环境的每个扫描点的扫描点信息，根据每个扫描点信息确定每个扫描点的空间坐标信息和激光反射特征，将地面环境分割成具有不同激光反射特征的子区域，确定各子区域的地面环境类型。这里的扫描点信息包含扫描点相对于激光雷达的方向角、距离以及激光反射强度，激光反射特征包括对不同波长激光的反射率。
处理器521可以采用通用的中央处理器(Central Processing Unit，CPU)，微处理器，应用专用集成电路(Application Specific Integrated Circuit，ASIC)，或者一个或多个集成电路，用于执行相关程序，以实现本发明实施例所提供的技术方案。
存储器522可以是只读存储器(Read Only Memory，ROM)，静态存储设备，动态存储设备或者随机存取存储器(Random Access Memory，RAM)。在 通过软件或者固件来实现本发明实施例提供的技术方案时，用于实现本发明实施例提供的技术方案的程序代码保存在存储器522中，并由处理器521来执行。
具体地，存储器522可以用于存储计算机执行指令，也可以用于存储各种信息，例如，每种类型地面环境的激光反射特征公式。处理器521可以读取该存储器522存储的信息，或者将收集的信息存储至存储器522。
可选的，处理器521用于根据每种类型地面环境的激光反射特征，确定各子区域的地面环境类型。
可选的，处理器521用于将各子区域的激光反射特征输入神经网络，获取神经网络输出的各所述子区域的地面环境类型，这里的神经网络用于区分不同地面环境类型的数据模型通过采用不同地面环境类型对不同波长激光的反射率数据作为输入，地面环境类型作为输出的一组数据作为样本数据集进行训练得到。
可选的，处理器521用于依据每个扫描点的扫描点信息和各激光雷达的安装位置，将各扫描点的位置信息变换到同一坐标系，对通过各激光雷达获取的各扫描点的扫描点信息进行融合，确定每个扫描点的空间坐标信息和激光反射特征，根据每个扫描点的空间坐标信息和激光反射特征进行区域分割，将地面环境分割成具有不同激光反射特征的子区域。
可选的，处理器521用于将各激光雷达获取的扫描点根据扫描点信息分别进行区域分割，生成每个激光雷达具有不同激光反射强度的聚类后子区域，将每个激光雷达的聚类后子区域的位置信息变换到同一坐标系，对各激光雷达聚类后子区域内的各扫描点的扫描点信息按照变换后子区域的位置信息进行融合，将地面环境分割成具有不同激光反射特征的子区域。
当然的，处理器521同样可以采用区域生长方式或者K-means方式进行区域分割。
尽管图5所示的数据处理单元仅仅示出了处理器521、存储器522，但是在具体实现过程中，本领域的技术人员应当明白，还包含实现正常运行所必须的其他器件。如通信接口和总线，其中的通信接口可以采用例如但不限于收发器一类的收发装置，用于实现数据处理单元52与激光扫描单元51之间 的通信。总线可包括一个通路，在处理器521和存储器522之间传送信息。总线可以是外设部件互连标准(peripheral component interconnect，简称PCI)总线或扩展工业标准结构(extended industry standard architecture，简称EISA)总线等。所述总线可以分为地址总线、数据总线、控制总线等。为便于表示，同时，根据具体需要，本领域的技术人员应当明白，图5所示的地面环境检测设备50还可包含实现其他附加功能的硬件器件。
本领域普通技术人员可以意识到，结合本文中所公开的实施例中描述的各方法步骤和单元，能够以电子硬件、计算机软件或者二者的结合来实现，为了清楚地说明硬件和软件的可互换性，在上述说明中已经按照功能一般性地描述了各实施例的步骤及组成。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。本领域普通技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本发明的范围。
所属领域的技术人员可以清楚地了解到，为了描述的方便和简洁，上述描述的装置和单元的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
在本申请所提供的几个实施例中，应该理解到，所揭露的系统、装置和方法，可以通过其它的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如，所述单元/模块的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另外，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口、装置或单元的间接耦合或通信连接，也可以是电的，机械的或其它的形式连接。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本发明实施例方案的目的。
另外，在本发明各个实施例中的各功能单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以是两个或两个以上单元集成在一个 单元中。上述集成的单元既可以采用硬件的形式实现，也可以采用软件功能单元的形式实现。
所述集成的单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中或作为计算机可读介质上的一个或多个指令或代码进行传输。计算机可读介质包括计算机存储介质和通信介质，其中通信介质包括便于从一个地方向另一个地方传送计算机程序的任何介质。存储介质可以是计算机能够存取的任何可用介质。以此为例但不限于：计算机可读介质可以包括RAM、ROM、EEPROM、CD-ROM或其他光盘存储、磁盘存储介质或者其他磁存储设备、或者能够用于携带或存储具有指令或数据结构形式的期望的程序代码并能够由计算机存取的任何其他介质。此外。任何连接可以适当的成为计算机可读介质。例如，如果软件是使用同轴电缆、光纤光缆、双绞线、数字用户线(DSL)或者诸如红外线、无线电和微波之类的无线技术从网站、服务器或者其他远程源传输的，那么同轴电缆、光纤光缆、双绞线、DSL或者诸如红外线、无线和微波之类的无线技术包括在所属介质的定义中。如本发明所使用的，盘(Disk)和碟(disc)包括压缩光碟(CD)、激光碟、光碟、数字通用光碟(DVD)、软盘和蓝光光碟，其中盘通常磁性的复制数据，而碟则用激光来光学的复制数据。上面的组合也应当包括在计算机可读介质的保护范围之内。基于这样的理解，本发明的技术方案本质上或者说对现有技术做出贡献的部分，或者该技术方案的全部或部分可以存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本发明各个实施例所述方法的全部或部分步骤。
最后应说明的是：以上实施例仅用以说明本发明的技术方案，而非对其限制；尽管参照前述实施例对本发明进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本发明各实施例技术方案的保护范围。

Claims (18)

  1. 一种地面环境检测的方法,包括：
    采用不同工作波长激光探测信号扫描地面环境；
    接收所述地面环境针对所述探测信号反射回来的反射信号；
    根据所述反射信号确定所述地面环境的每个扫描点的扫描点信息，所述扫描点信息包含所述扫描点相对于激光雷达的方向角、距离以及激光反射强度；
    根据每个扫描点信息确定每个扫描点的空间坐标信息和激光反射特征，将所述地面环境分割成具有不同激光反射特征的子区域，所述激光反射特征包括对不同波长激光的反射率；
    确定各所述子区域的地面环境类型。
  2. 根据权利要求1所述的方法，其特征在于，所述确定各所述子区域的地面环境类型，包括：
    根据每种类型地面环境的激光反射特征，确定各所述子区域的地面环境类型。
  3. 根据权利要求1所述的方法，其特征在于，所述确定各所述子区域的地面环境类型，包括：
    将所述各子区域的激光反射特征输入神经网络，获取所述神经网络输出的所述各子区域的地面环境类型；
    其中，所述神经网络用于区分不同地面环境类型的数据模型通过采用不同地面环境类型对不同波长激光的反射率数据作为输入，地面环境类型作为输出的一组数据作为样本数据集进行训练得到。
  4. 根据权利要求1至3任一所述的方法，其特征在于，根据每个扫描点信息确定每个扫描点的空间坐标信息和激光反射特征，将所述地面环境分割成具有不同激光反射特征的子区域，包括：
    依据每个扫描点的扫描点信息和各激光雷达的安装位置，将各扫描点的位置信息变换到同一坐标系，对通过各激光雷达获取的各扫描点的扫描点信息进行融合，确定每个扫描点的空间坐标信息和激光反射特征；
    根据每个扫描点的空间坐标信息和激光反射特征进行区域分割，将所述 地面环境分割成具有不同激光反射特征的子区域。
  5. 根据权利要求1至3任一所述的方法，根据每个扫描点信息确定每个扫描点的空间坐标信息和激光反射特征，将所述地面环境分割成具有不同激光反射特征的子区域，包括：
    将各激光雷达获取的扫描点根据扫描点信息分别进行区域分割，生成每个激光雷达具有不同激光反射强度的聚类后子区域；
    将每个激光雷达的聚类后子区域的位置信息变换到同一坐标系，对各激光雷达聚类后子区域内的各扫描点的扫描点信息按照变换后子区域的位置信息进行融合，将所述地面环境分割成具有不同激光反射特征的子区域。
  6. 根据权利要求4或5所述的方法，其特征在于，采用区域生长方式或者K-means方式进行区域分割。
  7. 一种地面环境检测设备,包括：
    激光扫描单元，用于采用不同工作波长激光探测信号扫描地面环境，接收所述地面环境针对所述探测信号反射回来的反射信号；
    数据采集单元，用于根据所述反射信号确定所述地面环境的每个扫描点的扫描点信息，所述扫描点信息包含所述扫描点相对于激光雷达的方向角、距离以及激光反射强度；
    扫描数据处理单元，用于根据每个扫描点信息确定每个扫描点的空间坐标信息和激光反射特征，将所述地面环境分割成具有不同激光反射特征的子区域，所述激光反射特征包括对不同波长激光的反射率；
    地面环境确定单元，用于确定各所述子区域的地面环境类型。
  8. 根据权利要求7所述的检测设备，其特征在于，所述地面环境确定单元，用于：
    根据每种类型地面环境的激光反射特征，确定各所述子区域的地面环境类型。
  9. 根据权利要求7所述的检测设备，其特征在于，所述地面环境确定单元，用于：
    将所述各子区域的激光反射特征输入神经网络，获取所述神经网络输出的所述各子区域的地面环境类型；
    其中，所述神经网络用于区分不同地面环境类型的数据模型通过采用不 同地面环境类型对不同波长激光的反射率数据作为输入，地面环境类型作为输出的一组数据作为样本数据集进行训练得到。
  10. 根据权利要求7至9任一所述的检测设备，其特征在于，所述扫描数据处理单元，包括：
    融合子单元，用于依据每个扫描点的扫描点信息和各激光雷达的安装位置，将各扫描点的位置信息变换到同一坐标系，对通过各激光雷达获取的各扫描点的扫描点信息进行融合，确定每个扫描点的空间坐标信息和激光反射特征；
    区域分割子单元，用于根据每个扫描点的空间坐标信息和激光反射特征进行区域分割，将所述地面环境分割成具有不同激光反射特征的子区域。
  11. 根据权利要求7至9任一所述的检测设备，其特征在于，所述扫描数据处理单元，包括：
    区域分割子单元，用于将各激光雷达获取的扫描点根据扫描点信息分别进行区域分割，生成每个激光雷达具有不同激光反射强度的聚类后子区域；
    融合子单元，用于将每个激光雷达的聚类后子区域的位置信息变换到同一坐标系，对各激光雷达聚类后子区域内的各扫描点的扫描点信息按照变换后子区域的位置信息进行融合，将所述地面环境分割成具有不同激光反射特征的子区域。
  12. 根据权利要求10或11所述的检测设备，其特征在于，所述区域分割子单元采用区域生长方式或者K-means方式进行区域分割。
  13. 一种地面环境检测设备,包括：
    激光扫描单元，用于采用不同工作波长激光探测信号扫描地面环境，接收所述地面环境针对所述探测信号反射回来的反射信号；
    数据处理单元，包括处理器和存储器，所述存储器用于存储计算机执行指令，所述处理器执行所述计算机执行指令，用于根据所述反射信号确定所述地面环境的每个扫描点的扫描点信息，根据每个扫描点信息确定每个扫描点的空间坐标信息和激光反射特征，将所述地面环境分割成具有不同激光反射特征的子区域，确定各所述子区域的地面环境类型；所述扫描点信息包含所述扫描点相对于激光雷达的方向角、距离以及激光反射强度，所述激光反射特征包括对不同波长激光的反射率。
  14. 根据权利要求13所述的检测设备，其特征在于，所述处理器用于：
    根据每种类型地面环境的激光反射特征，确定各所述子区域的地面环境类型。
  15. 根据权利要求13所述的检测设备，其特征在于，所述处理器用于：
    将所述各子区域的激光反射特征输入神经网络，获取所述神经网络输出的所述各子区域的地面环境类型；
    其中，所述神经网络用于区分不同地面环境类型的数据模型通过采用不同地面环境类型对不同波长激光的反射率数据作为输入，地面环境类型作为输出的一组数据作为样本数据集进行训练得到。
  16. 根据权利要求13至15任一所述的检测设备，其特征在于，所述处理器用于：
    依据每个扫描点的扫描点信息和各激光雷达的安装位置，将各扫描点的位置信息变换到同一坐标系，对通过各激光雷达获取的各扫描点的扫描点信息进行融合，确定每个扫描点的空间坐标信息和激光反射特征；
    根据每个扫描点的空间坐标信息和激光反射特征进行区域分割，将所述地面环境分割成具有不同激光反射特征的子区域。
  17. 根据权利要求13至15任一所述的检测设备，其特征在于，所述处理器用于：
    将各激光雷达获取的扫描点根据扫描点信息分别进行区域分割，生成每个激光雷达具有不同激光反射强度的聚类后子区域；
    将每个激光雷达的聚类后子区域的位置信息变换到同一坐标系，对各激光雷达聚类后子区域内的各扫描点的扫描点信息按照变换后子区域的位置信息进行融合，将所述地面环境分割成具有不同激光反射特征的子区域。
  18. 根据权利要求16或17所述的检测设备，其特征在于，所述处理器采用区域生长方式或者K-means方式进行区域分割。
PCT/CN2016/113089 2016-12-29 2016-12-29 一种地面环境的检测方法和装置 WO2018119902A1 (zh)

Priority Applications (6)

Application Number Priority Date Filing Date Title
KR1020197021867A KR102243118B1 (ko) 2016-12-29 2016-12-29 지상 환경 검출 방법 및 장치
PCT/CN2016/113089 WO2018119902A1 (zh) 2016-12-29 2016-12-29 一种地面环境的检测方法和装置
EP16925037.0A EP3553566B1 (en) 2016-12-29 2016-12-29 Method and apparatus for detecting ground environment
JP2019535919A JP6798032B2 (ja) 2016-12-29 2016-12-29 地面環境検出方法及び機器
CN201680091952.6A CN110114692B (zh) 2016-12-29 2016-12-29 一种地面环境的检测方法和装置
US16/456,057 US11455511B2 (en) 2016-12-29 2019-06-28 Ground environment detection method and apparatus

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2016/113089 WO2018119902A1 (zh) 2016-12-29 2016-12-29 一种地面环境的检测方法和装置

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US16/456,057 Continuation US11455511B2 (en) 2016-12-29 2019-06-28 Ground environment detection method and apparatus

Publications (1)

Publication Number Publication Date
WO2018119902A1 true WO2018119902A1 (zh) 2018-07-05

Family

ID=62710144

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/113089 WO2018119902A1 (zh) 2016-12-29 2016-12-29 一种地面环境的检测方法和装置

Country Status (6)

Country Link
US (1) US11455511B2 (zh)
EP (1) EP3553566B1 (zh)
JP (1) JP6798032B2 (zh)
KR (1) KR102243118B1 (zh)
CN (1) CN110114692B (zh)
WO (1) WO2018119902A1 (zh)

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110441269A (zh) * 2019-08-13 2019-11-12 江苏东交工程检测股份有限公司 标线反光检测方法、装置、设备及存储介质
CN110674292A (zh) * 2019-08-27 2020-01-10 腾讯科技（深圳）有限公司 一种人机交互方法、装置、设备及介质
CN111123278A (zh) * 2019-12-30 2020-05-08 科沃斯机器人股份有限公司 分区方法、设备及存储介质
WO2020142879A1 (zh) * 2019-01-07 2020-07-16 深圳市大疆创新科技有限公司 数据处理方法、探测装置、数据处理装置、可移动平台
CN111724485A (zh) * 2020-06-11 2020-09-29 浙江商汤科技开发有限公司 实现虚实融合的方法、装置、电子设备及存储介质
EP3779501A1 (en) * 2019-08-15 2021-02-17 Volvo Car Corporation Vehicle systems and methods utilizing lidar data for road condition estimation
CN112585656A (zh) * 2020-02-25 2021-03-30 华为技术有限公司 特殊路况的识别方法、装置、电子设备和存储介质
CN113670277A (zh) * 2021-08-25 2021-11-19 广东博智林机器人有限公司 地面装饰安装测绘方法、装置和测绘小车
CN114812529A (zh) * 2021-01-18 2022-07-29 上海理工大学 一种洁净室测点装置及洁净室的测点方法
CN117872354A (zh) * 2024-03-11 2024-04-12 陕西欧卡电子智能科技有限公司 一种多毫米波雷达点云的融合方法、装置、设备及介质

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11124193B2 (en) 2018-05-03 2021-09-21 Volvo Car Corporation System and method for providing vehicle safety distance and speed alerts under slippery road conditions
US10852158B1 (en) 2019-09-27 2020-12-01 Kitty Hawk Corporation Distance sensor test system
CN113040645B (zh) * 2019-12-26 2022-08-02 江苏美的清洁电器股份有限公司 清洁设备、清洁设备控制方法及装置
CN111948669B (zh) * 2020-08-11 2023-01-10 锐驰智光(苏州)科技有限公司 一种基于激光雷达的高光谱数据信息获取系统
JP7338607B2 (ja) 2020-10-29 2023-09-05 トヨタ自動車株式会社 車両位置推定装置
CN112598690A (zh) * 2020-12-11 2021-04-02 浙江大华技术股份有限公司 一种包裹切分方法、装置、设备及介质
CN115248428B (zh) * 2021-04-28 2023-12-22 北京航迹科技有限公司 激光雷达的标定、扫描方法、装置、电子设备及存储介质
CN113436258B (zh) * 2021-06-17 2023-09-12 中国船舶重工集团公司第七0七研究所九江分部 基于视觉与激光雷达融合的海上浮码头检测方法及系统

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2007240314A (ja) * 2006-03-08 2007-09-20 Omron Corp 物体検出装置
CN101536051A (zh) * 2006-09-28 2009-09-16 B.E.A.有限公司 用于存在检测的传感器
CN103776318A (zh) * 2014-01-03 2014-05-07 中国人民解放军陆军军官学院 光电检测环境模拟系统
CN104656101A (zh) * 2015-01-30 2015-05-27 福州华鹰重工机械有限公司 一种障碍物检测方法
CN105094143A (zh) * 2015-08-27 2015-11-25 泉州装备制造研究所 基于无人机的地图显示方法和装置

Family Cites Families (29)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CA1235773A (en) * 1983-12-23 1988-04-26 Shigeto Nakayama Device for detecting road surface condition
US7630806B2 (en) * 1994-05-23 2009-12-08 Automotive Technologies International, Inc. System and method for detecting and protecting pedestrians
US7840342B1 (en) * 1997-10-22 2010-11-23 Intelligent Technologies International, Inc. Road physical condition monitoring techniques
NL1009364C2 (nl) * 1998-06-10 1999-12-13 Road Ware B V Inrichting voor het bepalen van een profiel van een wegdek.
JP2002156452A (ja) * 2000-11-20 2002-05-31 Hioki Ee Corp レーザレーダシステム
JP2005090974A (ja) * 2003-09-12 2005-04-07 Daihatsu Motor Co Ltd 先行車認識装置
US7026600B2 (en) 2004-02-26 2006-04-11 Rosemount Aerospace Inc. System and method of identifying an object in a laser beam illuminated scene based on material types
JP3955616B2 (ja) * 2005-09-01 2007-08-08 松下電器産業株式会社 画像処理方法、画像処理装置及び画像処理プログラム
CN101806579B (zh) * 2009-02-16 2012-11-21 华为技术有限公司 反射镜位置采样、标定方法及装置和激光器
WO2010124284A1 (en) * 2009-04-24 2010-10-28 Hemant Virkar Methods for mapping data into lower dimensions
CN102142892B (zh) * 2010-06-30 2014-12-17 华为技术有限公司 一种探测脉冲的产生方法和相干光时域反射仪
JP2012189535A (ja) * 2011-03-14 2012-10-04 Ihi Corp 植生検出装置及び植生検出方法
US9155675B2 (en) * 2011-10-12 2015-10-13 Board Of Trustees Of The University Of Arkansas Portable robotic device
JP2013181968A (ja) * 2012-03-05 2013-09-12 Ricoh Co Ltd 光学装置
US9110196B2 (en) 2012-09-20 2015-08-18 Google, Inc. Detecting road weather conditions
DE102013002333A1 (de) * 2013-02-12 2014-08-14 Continental Teves Ag & Co. Ohg Verfahren und Strahlensensormodul zur vorausschauenden Straßenzustandsbestimmung in einem Fahrzeug
US9128190B1 (en) * 2013-03-06 2015-09-08 Google Inc. Light steering device with an array of oscillating reflective slats
CN103198302B (zh) * 2013-04-10 2015-12-02 浙江大学 一种基于双模态数据融合的道路检测方法
JP2015014514A (ja) * 2013-07-04 2015-01-22 パイオニア株式会社 識別装置
US9329073B2 (en) * 2013-12-06 2016-05-03 Honeywell International Inc. Adaptive radar system with mutliple waveforms
CN104463217A (zh) * 2014-12-15 2015-03-25 长春理工大学 基于激光雷达的路面类型识别方法及装置
CN104408443B (zh) 2014-12-15 2017-07-18 长春理工大学 多传感器辅助的基于激光雷达的路面类型识别方法及装置
US9453941B2 (en) * 2014-12-22 2016-09-27 GM Global Technology Operations LLC Road surface reflectivity detection by lidar sensor
CN104850834A (zh) 2015-05-11 2015-08-19 中国科学院合肥物质科学研究院 基于三维激光雷达的道路边界检测方法
JP2016223795A (ja) * 2015-05-27 2016-12-28 国立大学法人名古屋大学 床面状態検出装置および床面状態検出方法
WO2017053415A1 (en) * 2015-09-24 2017-03-30 Quovard Management Llc Systems and methods for surface monitoring
CN105510897A (zh) * 2015-12-01 2016-04-20 中国科学院上海技术物理研究所 基于地物类型卫星激光雷达出射激光波长反射率估算方法
KR20170096723A (ko) * 2016-02-17 2017-08-25 한국전자통신연구원 라이다 시스템 및 이의 다중 검출 신호 처리 방법
US10761195B2 (en) * 2016-04-22 2020-09-01 OPSYS Tech Ltd. Multi-wavelength LIDAR system

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2007240314A (ja) * 2006-03-08 2007-09-20 Omron Corp 物体検出装置
CN101536051A (zh) * 2006-09-28 2009-09-16 B.E.A.有限公司 用于存在检测的传感器
CN103776318A (zh) * 2014-01-03 2014-05-07 中国人民解放军陆军军官学院 光电检测环境模拟系统
CN104656101A (zh) * 2015-01-30 2015-05-27 福州华鹰重工机械有限公司 一种障碍物检测方法
CN105094143A (zh) * 2015-08-27 2015-11-25 泉州装备制造研究所 基于无人机的地图显示方法和装置

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3553566A4 *

Cited By (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2020142879A1 (zh) * 2019-01-07 2020-07-16 深圳市大疆创新科技有限公司 数据处理方法、探测装置、数据处理装置、可移动平台
CN110441269A (zh) * 2019-08-13 2019-11-12 江苏东交工程检测股份有限公司 标线反光检测方法、装置、设备及存储介质
US11592566B2 (en) 2019-08-15 2023-02-28 Volvo Car Corporation Vehicle systems and methods utilizing LIDAR data for road condition estimation
EP3779501A1 (en) * 2019-08-15 2021-02-17 Volvo Car Corporation Vehicle systems and methods utilizing lidar data for road condition estimation
CN110674292A (zh) * 2019-08-27 2020-01-10 腾讯科技（深圳）有限公司 一种人机交互方法、装置、设备及介质
CN111123278B (zh) * 2019-12-30 2022-07-12 科沃斯机器人股份有限公司 分区方法、设备及存储介质
CN111123278A (zh) * 2019-12-30 2020-05-08 科沃斯机器人股份有限公司 分区方法、设备及存储介质
CN112585656A (zh) * 2020-02-25 2021-03-30 华为技术有限公司 特殊路况的识别方法、装置、电子设备和存储介质
CN111724485A (zh) * 2020-06-11 2020-09-29 浙江商汤科技开发有限公司 实现虚实融合的方法、装置、电子设备及存储介质
CN111724485B (zh) * 2020-06-11 2024-06-07 浙江商汤科技开发有限公司 实现虚实融合的方法、装置、电子设备及存储介质
CN114812529A (zh) * 2021-01-18 2022-07-29 上海理工大学 一种洁净室测点装置及洁净室的测点方法
CN114812529B (zh) * 2021-01-18 2023-12-08 上海理工大学 一种洁净室测点装置及洁净室的测点方法
CN113670277A (zh) * 2021-08-25 2021-11-19 广东博智林机器人有限公司 地面装饰安装测绘方法、装置和测绘小车
CN117872354A (zh) * 2024-03-11 2024-04-12 陕西欧卡电子智能科技有限公司 一种多毫米波雷达点云的融合方法、装置、设备及介质
CN117872354B (zh) * 2024-03-11 2024-05-31 陕西欧卡电子智能科技有限公司 一种多毫米波雷达点云的融合方法、装置、设备及介质

Also Published As

Publication number Publication date
EP3553566A1 (en) 2019-10-16
EP3553566A4 (en) 2020-01-08
EP3553566B1 (en) 2022-06-22
CN110114692B (zh) 2021-09-17
JP2020504827A (ja) 2020-02-13
US20190317218A1 (en) 2019-10-17
KR102243118B1 (ko) 2021-04-21
CN110114692A (zh) 2019-08-09
JP6798032B2 (ja) 2020-12-09
KR20190098238A (ko) 2019-08-21
US11455511B2 (en) 2022-09-27

Similar Documents

Publication Publication Date Title
WO2018119902A1 (zh) 一种地面环境的检测方法和装置
Nissimov et al. Obstacle detection in a greenhouse environment using the Kinect sensor
Bansal et al. Pointillism: Accurate 3d bounding box estimation with multi-radars
JP6802331B2 (ja) 車線処理方法及び装置
EP3430427B1 (en) Processing method of a 3d point cloud
Bar Hillel et al. Recent progress in road and lane detection: a survey
CN107818557B (zh) 用于自动车辆的增强型相机对象检测
Sivaraman et al. Looking at vehicles on the road: A survey of vision-based vehicle detection, tracking, and behavior analysis
US10852426B2 (en) System and method of utilizing a LIDAR digital map to improve automatic driving
WO2020243962A1 (zh) 物体检测方法、电子设备和可移动平台
US11774554B2 (en) Electronic device, system and method for augmenting image data of a passive optical sensor
CN111868803A (zh) 产生合成雷达信号
CN102915444A (zh) 图像配准
Santos et al. Underwater place recognition using forward‐looking sonar images: A topological approach
CN112444822A (zh) 合成激光雷达信号的产生
CN112639811A (zh) 用于分析处理具有扩展对象识别的传感器数据的方法
Gholami et al. Real-time obstacle detection by stereo vision and ultrasonic data fusion
Lin et al. Density variation-based background filtering algorithm for low-channel roadside lidar data
CN114495045A (zh) 感知方法、感知装置、感知系统及相关设备
CN113988197B (zh) 基于多相机、多激光雷达的联合标定及目标融合检测方法
JP2005141517A (ja) 車両検出方法及び車両検出装置
Mussone et al. An innovative method for the analysis of vehicle movements in roundabouts based on image processing
WO2021059967A1 (ja) 物体認識装置及び物体認識プログラム
CN112766100A (zh) 一种基于关键点的3d目标检测方法
WO2024042607A1 (ja) 外界認識装置及び外界認識方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16925037

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2019535919

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2016925037

Country of ref document: EP

Effective date: 20190710

ENP Entry into the national phase

Ref document number: 20197021867

Country of ref document: KR

Kind code of ref document: A