CN109978982A - A kind of quick painting methods of point cloud based on inclination image - Google Patents
A kind of quick painting methods of point cloud based on inclination image Download PDFInfo
- Publication number
- CN109978982A CN109978982A CN201910262805.6A CN201910262805A CN109978982A CN 109978982 A CN109978982 A CN 109978982A CN 201910262805 A CN201910262805 A CN 201910262805A CN 109978982 A CN109978982 A CN 109978982A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- image
- characteristic point
- characteristic
- point
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T15/00—3D [Three Dimensional] image rendering
- G06T15/50—Lighting effects
- G06T15/55—Radiosity
Landscapes
- Engineering & Computer Science (AREA)
- Computer Graphics (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
- Image Generation (AREA)
- Processing Or Creating Images (AREA)
Abstract
The invention discloses a kind of quick painting methods of point cloud based on inclination image, comprising the following steps: the image data that object under test is obtained using oblique photograph technology, as target image；Using the characteristic point in second-order partial differential coefficient operator extraction target image；Characteristic point is matched using light stream constant principle, rebuilds three-dimensional point cloud；Three-dimensional point cloud is mapped using S type function, is converted to RGB color multimedia message.The invention avoids laser probe is used, it is greatly lowered the cost for obtaining laser point cloud data；Meanwhile the Time & Space Complexity of three-dimensional reconstruction algorithm is greatly reduced using rapid characteristic points detection algorithm and light stream constant match algorithm；In addition, being mapped by sigmoid function three-dimensional point cloud coordinate realizes quick colouring to point cloud data, the efficiency for improving later period classification and showing.
Description
Technical field
The present invention relates to fortune technical field of data processing more particularly to a kind of point cloud based on inclination image quickly to paint
Method.
Background technique
Currently, the technology of mainstream is to obtain the three-dimensional of target object by laser radar in three-dimensional reconstruction and ranging industry
Coordinate information；Laser radar usually carries on board the aircraft, and for aircraft when by above object under test, laser radar is with certain
Frequency emit electromagnetic wave, radar gets the point cloud data about object under test, the point cloud data of acquisition by echo data
Attribute includes intensity, the direction of echo of coordinate information (x, y, z) and echo etc. of the object in geographical world coordinate system.
Problem of the existing technology is: the complicated for operation of laser radar apparatus is not easy in universal business scenario
It promotes the use of a large area；Laser probe belongs to close instrument, and use cost is relatively high；Laser point cloud data amount is big, thus is not easy to
It stores and transmits；Laser point cloud data cannot can get determinand due to being limited by scan frequency as picture
The fine local feature of body；Laser point cloud data is unfavorable for the application in later period without any colouring information.
The technologies such as three-dimensional reconstruction, ranging and classification based on point cloud data have been widely applied how high in production practice
The high-precision laser point cloud data that gets of effect low cost is a urgent problem to be solved.
Summary of the invention
In view of the deficiencies of the prior art, the present invention provides a kind of quick painting methods of point cloud based on inclination image, solves
The above problem in the prior art.
To achieve the above object, the present invention provides technical solution below:
A kind of quick painting methods of point cloud based on inclination image, comprising the following steps:
The image data that object under test is obtained using oblique photograph technology, as target image；
Using the characteristic point in second-order partial differential coefficient operator extraction target image；
Characteristic point is matched using light stream constant principle, rebuilds three-dimensional point cloud；
Three-dimensional point cloud is mapped using S type function, is converted to RGB color multimedia message.
Optionally, the step: using the characteristic point in second-order partial differential coefficient operator extraction image, comprising:
Feature texture extraction is carried out using the target image as data source, the mathematical description that the feature texture extracts is such as
Under:
Wherein:Height, width are pixel x, the height and width of neighborhood window at y,
Gradx, grady are gradient image of the original image on horizontal and vertical；
The characteristic value of each characteristic point is calculated to (5) and gradient image in conjunction with formula (1).
Optionally, the step: using in the characteristic point in second-order partial differential coefficient operator extraction target image, further includes:
The characteristic point for meeting the following conditions is filtered out from the characteristic point extracted: characteristic value is greater than preset threshold value
And it is located in area-of-interest.
Optionally, the step: matching characteristic point using the constant principle of light stream, rebuilds three-dimensional point cloud, comprising:
The process of Feature Points Matching is as follows: J (x)=I (x-d)+n (x), and wherein n is noise.
Optionally, the step: mapping three-dimensional point cloud using S type function, is converted to RGB color multimedia message, comprising:
Three coordinate values of cloud are normalized, mathematical formulae description are as follows:
Wherein x, y, z are coordinate of the point cloud in world coordinate system, and R, G, B is the pixel corresponding to display color system
Value.
Optionally, the numerical value of the height and width is 7.
Optionally, it the step: filters out in the characteristic point for meeting the following conditions from the characteristic point extracted, also wraps
It includes:
When screening characteristic point, characteristic point is sorted according to the order from lower to big first.
Optionally, the preset threshold value is 128.
Optionally, from filtering out the feature for meeting characteristic value and being located at condition in area-of-interest in the characteristic point extracted
When point, it can be described by following formula for extracting the image characteristic point mask of characteristic point in the region of interest:
Compared with prior art, the invention has the following advantages:
The invention proposes a kind of quick painting methods of point cloud based on inclination image, and the method achieve unmanned plane inclinations
Photography obtains high definition image data, generates point cloud by image data and is quickly painted to cloud.Proposed in the present invention
Method avoid made using laser probe obtain laser point cloud data cost be greatly lowered；Meanwhile using rapid characteristic points
Detection algorithm and light stream constant match algorithm greatly reduce the Time & Space Complexity of three-dimensional reconstruction algorithm；In addition, logical
Crossing sigmoid function and being mapped three-dimensional point cloud coordinate realizes quick colouring to point cloud data, improves later period classification and shows
Efficiency.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention without any creative labor, may be used also for those of ordinary skill in the art
To obtain other attached drawings according to these attached drawings.
Fig. 1 is a kind of flow chart of point quick painting methods of cloud based on inclination image provided by the invention；
Fig. 2 is the flow chart of step S2 in a kind of point quick painting methods of cloud based on inclination image provided by the invention.
Specific embodiment
To enable the purpose of the present invention, feature, advantage more obvious and understandable, implement below in conjunction with the present invention
Attached drawing in example, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that reality disclosed below
Applying example is only a part of the embodiment of the present invention, and not all embodiment.Based on the embodiments of the present invention, this field is common
Technical staff's all other embodiment obtained without making creative work belongs to the model that the present invention protects
It encloses.
To further illustrate the technical scheme of the present invention below with reference to the accompanying drawings and specific embodiments.
Incorporated by reference to reference Fig. 1, Fig. 2, the present invention provides a kind of quick painting methods of point cloud based on inclination image, including
Following steps:
S1, the image data that object under test is obtained using oblique photograph technology, as target image.
In order to closely get the high resolution image of target object, object is restored by 3-dimensional reconstruction technology
Real surface information, the present invention using unmanned plane oblique photograph technology obtain image.The technology is carried on unmanned aerial vehicle platform
More image collection sensors, while from the vertical and inclined direction of all directions four, collect object under test five different angles
The image data of degree obtains target image, and the data source extracted using this target image as feature texture in subsequent step.
S2, using the characteristic point in second-order partial differential coefficient operator extraction target image.
In the step, using best feature point extraction algorithm, characteristic point position information and corresponding characteristic value are extracted simultaneously
Characteristic point quickly is selected using the strategy of thresholding, it is of the same name on target image to get dense feature point on initial pictures
The initial value of point coordinate.
Specifically, step S2 includes the following steps:
S201, feature texture extraction is carried out using the target image as data source.
In the step, the mathematical description that feature texture extracts is as follows:
Wherein:Height, width are pixel x, the height and width of neighborhood window at y,
Gradx, grady are gradient image of the original image on horizontal and vertical；
It is image local feature due to calculating herein, the value of height, width not Ying Tai little, if height,
There is too many characteristic point in too small will will lead to of the value of width, and excessive, can aggravate algorithm and the time in characteristic point is being selected to hold
Pin.
In the present embodiment, it is based on above-mentioned consideration, the numerical value of height and width takes 7 in specific experiment, then h is 3, w
It is 3.
It is understood that in specific experiment, as long as other numerical value features other than height and width also desirable 7
It puts quantity and is balanced at two aspects of selection characteristic point, reach the numerical value of the experiment effect in the present embodiment
In protection scope.
S202, the characteristic value for calculating each characteristic point.
In the step, the characteristic value of each characteristic point is calculated to (5) and gradient image in conjunction with formula (1):
And gradient image is combined, to calculate the characteristic value of each pixel in image.Wherein, these characteristic values can also be with
Performance namely characteristic value image by way of image.
S203, qualified characteristic point is filtered out.
In this step, we select full first to by extracting the sequence of characteristic value progress obtained from big to small
The characteristic point of sufficient condition is used for subsequent tracking and analysis.
Specifically, in the present embodiment, our condition is as follows in our experiment: the coordinate of characteristic point is interested
Region in and characteristic value be greater than preset threshold value.
From filtered out in the characteristic point extracted meet characteristic value and be located at the characteristic point of condition in area-of-interest when, use
It can be described by following formula in extracting the image characteristic point mask for being located at characteristic point in region of interest:
S3, characteristic point is matched using light stream constant principle, rebuilds three-dimensional point cloud.
In the step, by the constant Kanade-Lucas-Tomasi method of light stream in image procossing to characteristic point into
Row matching generates three-dimensional point cloud from unmanned plane inclination image data.
Specifically, the process of Feature Points Matching is as follows:
J (x)=I (x-d)+n (x) (7)
Wherein n is noise.
Selection herein makes the displacement when residual error minimum defined by double integer window W.
ε=∫W[I(x-d)-J(x)]2ωdx (8)
In this description, ω is a weighting function, for example, ω can be set to 1 in simply example.This
Weighting function ω may rely on the luminance patterns of image, can also be selected as Gauss likelihood function to emphasize the center of window
Domain.When this displacement is much smaller than the size of window, linearization technique processing is most effective.
Shown in the process of Feature Points Matching such as equation (7), Newton iterative has been used to come accurately in specific practice
Solve this equation.The convergence rate of solution is controlled by specifying the number of iteration.It is specific mathematical derivation process below:
The difference of two windows is defined as form by the present embodiment:
X=[x, y]T
D=[dx,dy]T (9)
1 is set by weighting function ω (x) herein in order to simplify operation；Herein by J in a=[ax,ay]TPlace is launched into Thailand
It strangles series and casts out higher order term, linearly turn to following form:
Wherein: ξ=[ξx,ξy]T, Wo Menling:X=a substitutes into above formula available following both direction herein
Partial differential equation:
Therefore:
Wherein:
In order to solve displacement d, we enable (13) formula be equal to 0, it may be assumed that
We obtain after transposition:
∫∫W[J (x)-I (x)] g (x) ω (x) dx=- ∫ ∫WgTDg (x) ω (x) dx=- [∫ ∫WgTg(x)ω(x)dx]d
(15)
We must solve matrix equation below as can be seen from the above equation:
Zd=e (16)
Z is the matrix of a 2*2, and e is the column vector of 2*1:
Z=∫ ∫WgTg(x)ω(x)dx
E=∫ ∫W[J(x)-I(x)]g(x)ω(x)dx (17)
S4, three-dimensional point cloud is mapped using S type function, is converted to RGB color multimedia message.
In the step, based on the Domain relation between the location information of cloud and point cloud, in conjunction with pseudo-colours theory
Realize the quick colouring to cloud.
Wherein, it is to assign RGB face to gray value according to specific criterion that pseudocolor image processing, which is also known as pseudo- colour processing,
The operation of color.The main application of pseudo-colours is in order to which people visually observes and explain the gray scale mesh in piece image or image sequence
Mark.Having in the visual and findspot cloud of engineering staff is not easy to since the point cloud data of inclination video generation lacks general colouring information
Information is imitated, the present invention is used as foundation that Pseudo-color Technique is combined quickly to paint a cloud by the Domain relation between point cloud.
In the present embodiment, three coordinate values of cloud will be normalized herein first, mathematical formulae description
Shown in following formula (18), wherein x, y, z are coordinate of the cloud in world coordinate system, R, G, and B corresponds to display color
The quick colouring of a cloud may be implemented in the pixel value of system, transformation in this way.
Based on the above embodiment, we use texture information to extract matching characteristic point, to this spy in the present invention
The description of sign point only used the characteristic value of 3 floating types；The constant theory of light stream has been used in dense feature point matching process
In Kanade-Lucas-Tomasi principle so that matching algorithm is to restrain rapidly；Cloud will be put finally by Sigmoid function
Three-dimensional world coordinate mapping become rgb value carry out rendering colouring.Therefore entire requirement of the algorithm to hardware store is greatly lowered, and
Algorithms T-cbmplexity is also largely under control, and is able to satisfy again while to rescue reduction to hardware performance in real time
Property and high-precision demand.
The above, the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although referring to before
Stating embodiment, invention is explained in detail, those skilled in the art should understand that: it still can be to preceding
Technical solution documented by each embodiment is stated to modify or equivalent replacement of some of the technical features；And these
It modifies or replaces, the spirit and scope for technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution.
Claims (9)
1. a kind of quick painting methods of point cloud based on inclination image, which comprises the following steps:
The image data that object under test is obtained using oblique photograph technology, as target image；
Using the characteristic point in second-order partial differential coefficient operator extraction target image；
Characteristic point is matched using light stream constant principle, rebuilds three-dimensional point cloud；
Three-dimensional point cloud is mapped using S type function, is converted to RGB color multimedia message.
2. the point cloud quick painting methods according to claim 1 based on inclination image, which is characterized in that the step:
Using the characteristic point in second-order partial differential coefficient operator extraction image, comprising:
Feature texture extraction is carried out using the target image as data source, the mathematical description that the feature texture extracts is as follows:
Wherein:Height, width are pixel x, the height and width of neighborhood window at y, gradx,
Grady is gradient image of the original image on horizontal and vertical；
The characteristic value of each characteristic point is calculated to (5) and gradient image in conjunction with formula (1).
3. the point cloud quick painting methods according to claim 1 based on inclination image, which is characterized in that the step:
Using in the characteristic point in second-order partial differential coefficient operator extraction target image, further includes:
The characteristic point for meeting the following conditions is filtered out from the characteristic point extracted: characteristic value is greater than preset threshold value and position
In in area-of-interest.
4. the point cloud quick painting methods according to claim 1 based on inclination image, which is characterized in that the step:
Characteristic point is matched using light stream constant principle, rebuilds three-dimensional point cloud, comprising:
The process of Feature Points Matching is as follows: J (x)=I (x-d)+n (x), and wherein n is noise.
5. the point cloud quick painting methods according to claim 1 based on inclination image, which is characterized in that the step:
Three-dimensional point cloud is mapped using S type function, is converted to RGB color multimedia message, comprising:
Three coordinate values of cloud are normalized, mathematical formulae description are as follows:
Wherein x, y, z are coordinate of the point cloud in world coordinate system, and R, G, B is the pixel value corresponding to display color system.
6. the point cloud quick painting methods according to claim 1 based on inclination image, which is characterized in that the height
And the numerical value of width is 7.
7. the point cloud quick painting methods according to claim 3 based on inclination image, which is characterized in that the step:
It is filtered out in the characteristic point for meeting the following conditions from the characteristic point extracted, further includes:
When screening characteristic point, characteristic point is sorted according to the order from lower to big first.
8. the point cloud quick painting methods according to claim 3 based on inclination image, which is characterized in that described to set in advance
Fixed threshold value is 128.
9. the point cloud quick painting methods according to claim 3 based on inclination image, which is characterized in that from extracting
Characteristic point in filter out when meeting characteristic value and being located at the characteristic point of condition in area-of-interest, be located at region of interest for extracting
The image characteristic point mask of interior characteristic point can be described by following formula:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910262805.6A CN109978982B (en) | 2019-04-02 | 2019-04-02 | Point cloud rapid coloring method based on oblique image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910262805.6A CN109978982B (en) | 2019-04-02 | 2019-04-02 | Point cloud rapid coloring method based on oblique image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109978982A true CN109978982A (en) | 2019-07-05 |
| CN109978982B CN109978982B (en) | 2023-04-07 |
Family
ID=67082471
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910262805.6A Active CN109978982B (en) | 2019-04-02 | 2019-04-02 | Point cloud rapid coloring method based on oblique image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109978982B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110580703A (en) * | 2019-09-10 | 2019-12-17 | 广东电网有限责任公司 | distribution line detection method, device, equipment and storage medium |
| CN112613107A (en) * | 2020-12-26 | 2021-04-06 | 广东电网有限责任公司 | Method and device for determining construction progress of tower project, storage medium and equipment |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004221635A (en) * | 2003-01-09 | 2004-08-05 | Seiko Epson Corp | Color conversion device, color conversion method, color conversion program, and print control device |
| CN102341825A (en) * | 2009-03-03 | 2012-02-01 | 微软公司 | Multi-modal tone-mapping of images |
| CN102629328A (en) * | 2012-03-12 | 2012-08-08 | 北京工业大学 | Probabilistic latent semantic model object image recognition method with fusion of significant characteristic of color |
| CN103325108A (en) * | 2013-05-27 | 2013-09-25 | 浙江大学 | Method for designing monocular vision odometer with light stream method and feature point matching method integrated |
| CN105629980A (en) * | 2015-12-23 | 2016-06-01 | 深圳速鸟创新科技有限公司 | Single-camera oblique photography three-dimensional modeling system |
| CN106228609A (en) * | 2016-07-09 | 2016-12-14 | 武汉广图科技有限公司 | A kind of oblique photograph three-dimensional modeling method based on spatial signature information |
| US20170032565A1 (en) * | 2015-07-13 | 2017-02-02 | Shenzhen University | Three-dimensional facial reconstruction method and system |
| WO2018056802A1 (en) * | 2016-09-21 | 2018-03-29 | Universiti Putra Malaysia | A method for estimating three-dimensional depth value from two-dimensional images |
-
2019
- 2019-04-02 CN CN201910262805.6A patent/CN109978982B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004221635A (en) * | 2003-01-09 | 2004-08-05 | Seiko Epson Corp | Color conversion device, color conversion method, color conversion program, and print control device |
| CN102341825A (en) * | 2009-03-03 | 2012-02-01 | 微软公司 | Multi-modal tone-mapping of images |
| CN102629328A (en) * | 2012-03-12 | 2012-08-08 | 北京工业大学 | Probabilistic latent semantic model object image recognition method with fusion of significant characteristic of color |
| CN103325108A (en) * | 2013-05-27 | 2013-09-25 | 浙江大学 | Method for designing monocular vision odometer with light stream method and feature point matching method integrated |
| US20170032565A1 (en) * | 2015-07-13 | 2017-02-02 | Shenzhen University | Three-dimensional facial reconstruction method and system |
| CN105629980A (en) * | 2015-12-23 | 2016-06-01 | 深圳速鸟创新科技有限公司 | Single-camera oblique photography three-dimensional modeling system |
| CN106228609A (en) * | 2016-07-09 | 2016-12-14 | 武汉广图科技有限公司 | A kind of oblique photograph three-dimensional modeling method based on spatial signature information |
| WO2018056802A1 (en) * | 2016-09-21 | 2018-03-29 | Universiti Putra Malaysia | A method for estimating three-dimensional depth value from two-dimensional images |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110580703A (en) * | 2019-09-10 | 2019-12-17 | 广东电网有限责任公司 | distribution line detection method, device, equipment and storage medium |
| CN110580703B (en) * | 2019-09-10 | 2024-01-23 | 广东电网有限责任公司 | Distribution line detection method, device, equipment and storage medium |
| CN112613107A (en) * | 2020-12-26 | 2021-04-06 | 广东电网有限责任公司 | Method and device for determining construction progress of tower project, storage medium and equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109978982B (en) | 2023-04-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110111248B (en) | Image splicing method based on feature points, virtual reality system and camera | |
| CN106327532B (en) | A kind of three-dimensional registration method of single image | |
| KR101554241B1 (en) | A method for depth map quality enhancement of defective pixel depth data values in a three-dimensional image | |
| CN110033475B (en) | A method for detecting and eliminating moving objects in aerial images generated by high-resolution textures | |
| CN110047144A (en) | A kind of complete object real-time three-dimensional method for reconstructing based on Kinectv2 | |
| CN113379661B (en) | Dual-branch convolutional neural network device for infrared and visible light image fusion | |
| JPH0375682A (en) | High frequency signal detecting apparatus | |
| Wu et al. | Densely pyramidal residual network for UAV-based railway images dehazing | |
| WO2019140945A1 (en) | Mixed reality method applied to flight simulator | |
| Peng et al. | PSMD-Net: A novel pan-sharpening method based on a multiscale dense network | |
| CN107329116B (en) | Airborne radar three-dimensional motion scene display method | |
| CN107330964A (en) | A kind of display methods and system of complex three-dimensional object | |
| CN104217461B (en) | A parallax mapping method based on a depth map to simulate a real-time bump effect | |
| CN108364292A (en) | A kind of illumination estimation method based on several multi-view images | |
| CN105719250A (en) | Image inpainting method based on simple background, system and shooting camera | |
| CN112669280A (en) | Unmanned aerial vehicle oblique aerial photography right-angle image control point target detection method based on LSD algorithm | |
| CN109978982A (en) | A kind of quick painting methods of point cloud based on inclination image | |
| CN112465977B (en) | Method for repairing three-dimensional model water surface loophole based on dense point cloud | |
| CN113144613A (en) | Model-based volume cloud generation method | |
| CN114972646B (en) | Method and system for extracting and modifying independent ground objects of live-action three-dimensional model | |
| Bai et al. | Making the Earth clear at night: A high-resolution nighttime light image deblooming network | |
| CN108629742A (en) | True orthophoto shadow Detection and compensation method, device and storage medium | |
| CN107958489B (en) | Curved surface reconstruction method and device | |
| CN110335342A (en) | A real-time generation method of hand model for immersive simulator | |
| CN117593465B (en) | Three-dimensional visualization to achieve virtual display method and system of smart city |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |