CN107560592A - A kind of precision ranging method for optronic tracker linkage target - Google Patents
A kind of precision ranging method for optronic tracker linkage target Download PDFInfo
- Publication number
- CN107560592A CN107560592A CN201710718983.6A CN201710718983A CN107560592A CN 107560592 A CN107560592 A CN 107560592A CN 201710718983 A CN201710718983 A CN 201710718983A CN 107560592 A CN107560592 A CN 107560592A
- Authority
- CN
- China
- Prior art keywords
- target
- optronic
- point
- matching
- video camera
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
- Y02A10/40—Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping
Landscapes
- Measurement Of Optical Distance (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of precision ranging method for optronic tracker linkage target, by being demarcated to multiple optronic tracker video cameras；Moving object detection is carried out to each optronic tracker video image, and extracts the SIFT feature of object region；Different optronic tracker extracting target from video image characteristic points are matched；According to camera parameter and the characteristic point of matching to carrying out three-dimensional reconstruction；Finally, average distance that Three-dimensional Gravity is laid foundations to video camera is calculated as target range.Compared with prior art, the present invention uses passive ranging method, using the station target acquisition of land and sea border defense video monitoring as background, based on binocular stereo vision principle, ranging is carried out using multiple optronic trackers collection video image at land and sea border defense video monitoring station, there is the characteristics of cost is low, be easy to carry, strong antijamming capability, easy and effective method.
Description
Technical field
The present invention relates to technical field of video monitoring, and in particular to a kind of accurate survey for optronic tracker linkage target
Away from method.
Background technology
In the struggle such as modern military operation, anti-terrorism, anti-smuggling, fast and accurately identifying and position target of investication turns into people
The matter of utmost importance be concerned about, and the accurate estimation of target range is an important step, and there is important military and civilian to be worth.
The method species of object ranging is a lot, according to whether need externally to measurand apply energy can be divided into initiative range measurement and by
Dynamic two kinds of ranging.White light projection target of the initiative range measurement method using wave beams such as laser or with certain texture structure, passes through analysis
The propagation time of light beam or the texture deformation of reflected light determine target range, such as laser ranging, Structure light method, Moire fringe
Method, phase comparator etc..Its advantage is that accuracy is higher；Shortcoming is that the various ripples of transmitting may cause injury target or exposure itself
Target, while equipment complexity, not Portable belt, cost height etc..Passive ranging method then directly passes through the video image information of target
To estimate target range.Passive ranging method includes stereoscopic vision, motion telemetry, monocular ranging etc..
China's boundary line on land and coastline have established all standing, round-the-clock, the video prison of 24 hours substantially at present
Control network.If key sections on road, the video monitoring website of marine key area can be built, then integrated use white light, it is infrared,
A variety of detection means such as radar, laser, it will substantially increase land and sea border defense information obtaining ability, have to strengthening land and sea border defense management and control
Highly important meaning.
The content of the invention
To solve drawbacks described above, it is an object of the invention to provide a kind of precision ranging for optronic tracker linkage target
Method, input cost is low, and strong antijamming capability, method is easy and effective, and results contrast is accurate.
To reach above-mentioned purpose, the technical solution adopted by the present invention is：A kind of essence for optronic tracker linkage target
True distance-finding method, it is characterised in that comprise the following steps：
1）Demarcate optronic tracker video camera：Rower is entered using Zhang Zhengyou methods respectively to optronic tracker video camera C1 and C2
It is fixed；According to national forest park in Xiaokeng principle, the relation established between camera review location of pixels and scene point location；According to shooting
Machine model, the model parameter of video camera is solved by the image coordinate and world coordinates of known scaled point, obtains video camera C1 and C2
Inside and outside parameter；
2）IMAQ and moving object detection：The video image I1 and I2 of two optronic trackers of synchronous acquisition, using background
Calculus of finite differences, motion detection is carried out to I1 and I2, and current frame image and background image are carried out into mathematic interpolation, obtain error image,
Background and moving target are split by establishing the mixed Gauss model of background again, extract motion target area A1 and
A2；
3）Extract region A1 and A2 feature：Using SIFT feature extraction algorithm, motion target area A1 and A2 are extracted respectively
SIFT feature point set P1 and P2；A series of steps are determined by metric space extremum extracting, the positioning of key point, key point direction
Suddenly, the Feature Descriptor of 128 dimensional vectors is generated；
4）Characteristic matching：Using basis matrix as model, using RANSAC parameter estimation algorithms, by slightly matching and essence matching two
The individual stage completes characteristic matching, obtains the matching double points set S in P1 and P2.
5）Three-dimensional reconstruction：After obtaining correct matching double points set S, by calibrated camera model, using SVD side
Method carries out the three-dimensional reconstruction of characteristic point, rebuilds three-dimensional point set corresponding to matching double points set；
6）Distance calculates：After completing three-dimensional reconstruction, a three-dimensional point set Q on moving target is obtained, calculate in Q each point to imaging
The average distance of machine C1 or C2 photocentre is accurate distance of the target to video camera.
Preferably, step 3）In SIFT feature extraction lazy weight in the case of, can add
Harris characteristic points make reconstructed results more prominent in detail, are more nearly real-world object as supplement.
The present invention uses passive ranging method, using the station target acquisition of land and sea border defense video monitoring as background, based on binocular solid
Visual theory, video image is gathered to carry out ranging using multiple optronic trackers at land and sea border defense video monitoring station, there is cost
It is low, be easy to carry, strong antijamming capability, easy and effective method the characteristics of.
Compared with prior art, target ranging method proposed by the present invention has the advantages that：
1）Passive ranging method directly estimates target range by the video image information of target, has hidden, no electromagnetic wave
It is the advantages that interference, safe in strategic point defence field；
2）Compared with existing Technique of Initiative Range Measurement, on the premise of ranging distance and precision meet to require, use cost is lower,
Method is more easy and effective；
3）The three-dimensional reconstruction of binocular stereo vision, the visual processes scenery mode of the mankind is simulated, can be in a variety of conditions
The lower steric information for perceiving three-dimensional scenic, easy to operate, measurement distance accuracy is high, can be widely applied to strategic point defence, security protection
The fields such as monitoring.
Brief description of the drawings
Below according to drawings and Examples, the feature of the present invention is further described.
Fig. 1 is the workflow schematic diagram of the present invention.
Fig. 2 is the data flowchart that optronic tracker is demarcated in the present invention.
Fig. 3 is the data flowchart of object ranging in the present invention.
Embodiment
It is a kind of embodiment of the present invention referring to accompanying drawing 1, Fig. 2 and Fig. 3.
The present embodiment carries out object ranging, particular technique implementation bag using the three-dimensional reconstruction principle of binocular stereo vision
Include following steps：
1）The camera calibration stage：Rower is entered using Zhang Zhengyou plane references method respectively to optronic tracker video camera C1 and C2
It is fixed.Detailed process：Camera plane target, video camera and plane target drone can move freely through video camera from different perspectives, mark
Video camera internal reference is kept constant during fixed；Detect the characteristic point in image；It is public according to the general principle of Zhang Zhengyou standardizations
Formula calculates the intrinsic parameter of video camera and outer parameter；Camera lens have distortion, and the parameter calculated is entered as initial value
Row Optimizing Search, obtains distortion factor, so as to calculate the exact value of all parameters, shown in specific data flowchart accompanying drawing 2；
2）The three-dimensional reconstruction stage：After camera calibration, the match point simultaneous equations of extraction at least two images, Three-dimensional Gravity is carried out
Build, mainly including the steps such as IMAQ, target detection, SIFT feature extraction, characteristic matching, characteristic point three-dimensional reconstruction, data
Flow chart is as shown in Figure 3.Detailed process：
A. moving target is obtained using background subtraction：Using the mathematic interpolation of adjacent two field picture, then by establishing the mixed of background
Close Gauss model to split background and moving target, realize the detection of Moving Targets in Sequent Images；
B. the characteristic point of image is obtained using SIFT feature extraction algorithm：Metric space extremum extracting is first passed through, according to X-Y scheme
Picture and the convolution of Gaussian kernel obtain the space representation under different scale, establish image DOG pyramids；Then characteristic point is determined
Position, by each characteristic point extracted compared with 26 characteristic points around it, Local Extremum is extracted, then by fitting three-dimensional two
Secondary function, the position and yardstick of characteristic point are accurately positioned, while reject skirt response point and the low point of contrast, enhancing matching
Stability；Meanwhile to ensure the rotational invariance of characteristic point, choose ladder of each characteristic point in the range of 3 σ in Gaussian image
Degree and direction, the statistics of column hisgram is entered to it, determines characteristic point direction；Finally, in metric space where characteristic point
The gradient information in its 8 directions of 4*4 window calculation, the Feature Descriptor of 128 dimensional vectors can be generated.
3）After the SIFT feature point set P1 and P2 that extract two images target area A1 and A2, calculated using improved RANSAC
Method carries out characteristic matching：To ensure matching precision, completed using thick matching and essence two stages of matching.
Thick matching stage：Using the arest neighbors matching algorithm based on Euclidean distance（Nearest Neighbor-NN）：
Index construction is established using kd-tree to feature point set P2；Then, to each, kd-tree is utilized in P2
Scan for, search arest neighbors characteristic vector b1 and secondary neighbour characteristic vector b2 corresponding with a, if arest neighbors divided by it is secondary closely
Adjacent distance ratio is less than some given proportion threshold value, i.e.,（T value in experiment
For 0.6）, then b1 is a matching；Otherwise it is assumed that a is not matched in P2, may be matched a little to collection by thick matching process
Close S1；
Smart matching stage：Using Epipolar geometry restriction relation, reject the false matches in S1, due to Same Scene is shot two
Any one matching double points in width imageMeet Epipolar geometry（Epipolar Geometry）Restriction relation, and to extremely several
What restriction relation can mathematically use fundamental matrix（Fundamental matrix）To represent.It is basic between two images
Matrix F is the unusual square formation of order 2 of one 3 × 3, has 7 frees degree；To the matching double points (a, b) of two images, meet aTFb
=0；A linear equation of the unknown element on F can be provided due to each group of matching double points, if there is foot in two images
Enough match points (﹥ 7), it is possible to unknown fundamental matrix F is calculated using their Simultaneous Equations, solves basis matrix sheet
Invention is using improved 8 methods.
Wherein the criterion of false matches is in RANSAC algorithms：The basis matrix estimated according to 8 methods, calculate special
The vertical range of sign point to corresponding polar curve is used as constraint（Less than 1.5 pixels）, the matching of restriction relation will be unsatisfactory for, be considered as void
Vacation matching, also referred to as outlier（Correct matching double points are referred to as inlier）.RANSAC algorithms are thought, as long as carrying out N groups at random
Sampling（N is sufficiently large）, just can be in certain confidence level P（Typically take 0.99）It is lower guarantee at least one group of sampling be not by
Pollution, minimum frequency in sampling N meets following condition：
In formulaIt is data error rate, k is to solve for the minimal data needed for model.
False matches rejecting process is in RANSAC algorithms：
I) according to confidence level and data error rate, the sampling number N for needing to carry out is calculated；
II) 8 groups of matching double points are extracted uniformly randomly from thick matching double points collection S1, using improved 8 methods, calculated near
Like fundamental matrix Fran；
III) Fran is verified with each group of corresponding points not being drawn into, if polar curve is apart from sufficiently small, then it is assumed that this, which is corresponded to, is
The correct matching double points consistent with Fran, recalculate data error rate and update N；
IV) the IIth and the IIIth step is repeated, the processing until completing the sampling of N groups；
V) according to maximum correct match point logarithm and minimal error variance, selected from multiple approximate fundamental matrix Fran optimal
Parameter model FR；
VI) correct matching double points set S is calculated using FR, and final fundamental matrix F is recalculated according to S.
4）After obtaining correct matching double points set S, due to two optronic tracker video cameras C1 and C2 inside and outside parameter
It has been demarcated that, according to Feature Points Matching result, three-dimensional point set corresponding to reconstruction.
Three dimensional field sight spotWith the picture point in two-dimentional image planeMeet central projection relation, as shown in following formula：
Wherein：K is referred to as camera Intrinsic Matrix.Because parameter is only relevant with the internal structure of camera in K；[R | t] it is video camera
Outer parameter matrix, describes the position relationship between camera coordinate system and world coordinate system, and P=K [R | t] is is referred to as projection square
Battle array.
To any match point, a and b are substituted into above formula respectively, after obtain one contact equation group
Three-dimensional point M corresponding to (a, b) can be solved, because (a, b) is matching double points, equation group has unique solution in theory, but in reality
In the application of border, due to the influence of calibrated error etc., it can may be determined most using the method that SVD is decomposed without the solution determined
Excellent solution；
5）Target range calculation stages：After carrying out three-dimensional reconstruction to each pair match point in matching double points set, obtain moving mesh
The three-dimensional point set put on, obtain the average distance for arriving video camera photocentre a little in three-dimensional point set just obtained target away from
From.
Using pedestrian as moving target in the present embodiment, linkage test is carried out using two optronic trackers similarly configured：
One optronic tracker is demarcated according to the Zhang Zhengyou methods in camera calibration stage, it is insured in video camera in calibration process
Hold constant, obtain the inside and outside parameter of optronic tracker.Then object ranging is carried out, it is assumed that optronic tracker is motionless, to moving mesh
Mark is detected, and captures two width target images, carries out extracting and matching feature points, and the 3D that target is obtained by three-dimensional reconstruction is dilute
Point cloud is dredged, obtains the average distance for arriving video camera photocentre a little in three-dimensional point set.Test result indicates that proposed
Location algorithm is a kind of feasible passive.
Described above is only presently preferred embodiments of the present invention, and above-mentioned specific embodiment is not limitation of the present invention,
Retouching, modification or the equivalent substitution that all one of ordinary skill in the art are made as described above, belong to the guarantor of the present invention
Protect scope.
Claims (2)
1. a kind of precision ranging method for optronic tracker linkage target, coordinated using two optronic trackers, detected
After moving target information simultaneously capture movement target, the 3-D view of moving target is established based on three-dimensional reconstruction principle, is calculated
Three-dimensional point obtains target to the accurate distance of video camera to the average distance of video camera photocentre, it is characterised in that including following step
Suddenly：
1）Demarcate optronic tracker video camera：Rower is entered using Zhang Zhengyou methods respectively to optronic tracker video camera C1 and C2
It is fixed；According to national forest park in Xiaokeng principle, the relation established between camera review location of pixels and scene point location；According to shooting
Machine model, the model parameter of video camera is solved by the image coordinate and world coordinates of known scaled point, obtains video camera C1 and C2
Inside and outside parameter；
2）IMAQ and moving object detection：The video image I1 and I2 of two optronic trackers of synchronous acquisition, using background
Calculus of finite differences, motion detection is carried out to I1 and I2, and current frame image and background image are carried out into mathematic interpolation, obtain error image,
Background and moving target are split by establishing the mixed Gauss model of background again, extract motion target area A1 and
A2；
3）Extract region A1 and A2 feature：Using SIFT feature extraction algorithm, motion target area A1 and A2 are extracted respectively
SIFT feature point set P1 and P2；A series of steps are determined by metric space extremum extracting, the positioning of key point, key point direction
Suddenly, the Feature Descriptor of 128 dimensional vectors is generated；
4）Characteristic matching：Using basis matrix as model, using RANSAC parameter estimation algorithms, by slightly matching and essence matching two
The individual stage completes characteristic matching, obtains the matching double points set S in P1 and P2；
5）Three-dimensional reconstruction：After obtaining correct matching double points set S, by calibrated camera model, entered using SVD methods
The three-dimensional reconstruction of row characteristic point, rebuild three-dimensional point set corresponding to matching double points set；
6）Distance calculates：After completing three-dimensional reconstruction, a three-dimensional point set Q on moving target is obtained, calculate in Q each point to imaging
The average distance of machine C1 or C2 photocentre is accurate distance of the target to video camera.
2. the precision ranging method according to claim 1 for optronic tracker linkage target, it is characterised in that：Step
3）In SIFT feature extraction lazy weight in the case of, can add Harris characteristic points as supplement.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710718983.6A CN107560592B (en) | 2017-08-21 | 2017-08-21 | Precise distance measurement method for photoelectric tracker linkage target |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710718983.6A CN107560592B (en) | 2017-08-21 | 2017-08-21 | Precise distance measurement method for photoelectric tracker linkage target |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107560592A true CN107560592A (en) | 2018-01-09 |
| CN107560592B CN107560592B (en) | 2020-08-18 |
Family
ID=60976143
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710718983.6A Active CN107560592B (en) | 2017-08-21 | 2017-08-21 | Precise distance measurement method for photoelectric tracker linkage target |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107560592B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108333591A (en) * | 2018-01-18 | 2018-07-27 | 中国科学院苏州纳米技术与纳米仿生研究所 | A kind of distance measuring method and its system |
| CN109827547A (en) * | 2019-03-27 | 2019-05-31 | 中国人民解放军战略支援部队航天工程大学 | A kind of distributed multi-sensor extraterrestrial target synchronization association method |
| CN110197104A (en) * | 2018-02-27 | 2019-09-03 | 杭州海康威视数字技术股份有限公司 | Distance measuring method and device based on vehicle |
| CN111046842A (en) * | 2019-12-27 | 2020-04-21 | 京东数字科技控股有限公司 | Method and apparatus for changing information |
| CN111080685A (en) * | 2019-12-17 | 2020-04-28 | 北京工业大学 | Airplane sheet metal part three-dimensional reconstruction method and system based on multi-view stereoscopic vision |
| CN113382223A (en) * | 2021-06-03 | 2021-09-10 | 浙江大学 | Underwater three-dimensional imaging device and method based on multi-view vision |
| CN113484864A (en) * | 2021-07-05 | 2021-10-08 | 中国人民解放军国防科技大学 | Unmanned ship-oriented navigation radar and photoelectric pod collaborative environment sensing method |
| CN113811919A (en) * | 2019-05-21 | 2021-12-17 | 国际商业机器公司 | Progressive 3D point cloud partitioning from objects and backgrounds of tracking sessions |
| CN115661702A (en) * | 2022-10-13 | 2023-01-31 | 华中科技大学 | Sea condition real-time estimation method and system based on smart phone |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102589516A (en) * | 2012-03-01 | 2012-07-18 | 长安大学 | Dynamic distance measuring system based on binocular line scan cameras |
| CN102914294A (en) * | 2012-09-10 | 2013-02-06 | 中国南方电网有限责任公司超高压输电公司天生桥局 | System and method for measuring unmanned aerial vehicle electrical line patrol on basis of images |
| CN104677330A (en) * | 2013-11-29 | 2015-06-03 | 哈尔滨智晟天诚科技开发有限公司 | Small binocular stereoscopic vision ranging system |
| CN104819705A (en) * | 2014-04-23 | 2015-08-05 | 西安交通大学 | Binocular measurement apparatus based on variable vision range target search, and use method thereof |
| CN104883556A (en) * | 2015-05-25 | 2015-09-02 | 深圳市虚拟现实科技有限公司 | Three dimensional display method based on augmented reality and augmented reality glasses |
-
2017
- 2017-08-21 CN CN201710718983.6A patent/CN107560592B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102589516A (en) * | 2012-03-01 | 2012-07-18 | 长安大学 | Dynamic distance measuring system based on binocular line scan cameras |
| CN102914294A (en) * | 2012-09-10 | 2013-02-06 | 中国南方电网有限责任公司超高压输电公司天生桥局 | System and method for measuring unmanned aerial vehicle electrical line patrol on basis of images |
| CN104677330A (en) * | 2013-11-29 | 2015-06-03 | 哈尔滨智晟天诚科技开发有限公司 | Small binocular stereoscopic vision ranging system |
| CN104819705A (en) * | 2014-04-23 | 2015-08-05 | 西安交通大学 | Binocular measurement apparatus based on variable vision range target search, and use method thereof |
| CN104883556A (en) * | 2015-05-25 | 2015-09-02 | 深圳市虚拟现实科技有限公司 | Three dimensional display method based on augmented reality and augmented reality glasses |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108333591A (en) * | 2018-01-18 | 2018-07-27 | 中国科学院苏州纳米技术与纳米仿生研究所 | A kind of distance measuring method and its system |
| CN110197104A (en) * | 2018-02-27 | 2019-09-03 | 杭州海康威视数字技术股份有限公司 | Distance measuring method and device based on vehicle |
| CN109827547A (en) * | 2019-03-27 | 2019-05-31 | 中国人民解放军战略支援部队航天工程大学 | A kind of distributed multi-sensor extraterrestrial target synchronization association method |
| CN109827547B (en) * | 2019-03-27 | 2021-05-04 | 中国人民解放军战略支援部队航天工程大学 | Distributed multi-sensor space target synchronous correlation method |
| CN113811919A (en) * | 2019-05-21 | 2021-12-17 | 国际商业机器公司 | Progressive 3D point cloud partitioning from objects and backgrounds of tracking sessions |
| CN111080685A (en) * | 2019-12-17 | 2020-04-28 | 北京工业大学 | Airplane sheet metal part three-dimensional reconstruction method and system based on multi-view stereoscopic vision |
| CN111046842A (en) * | 2019-12-27 | 2020-04-21 | 京东数字科技控股有限公司 | Method and apparatus for changing information |
| CN111046842B (en) * | 2019-12-27 | 2022-02-15 | 京东科技控股股份有限公司 | Method and apparatus for changing information |
| CN113382223A (en) * | 2021-06-03 | 2021-09-10 | 浙江大学 | Underwater three-dimensional imaging device and method based on multi-view vision |
| CN113484864A (en) * | 2021-07-05 | 2021-10-08 | 中国人民解放军国防科技大学 | Unmanned ship-oriented navigation radar and photoelectric pod collaborative environment sensing method |
| CN115661702A (en) * | 2022-10-13 | 2023-01-31 | 华中科技大学 | Sea condition real-time estimation method and system based on smart phone |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107560592B (en) | 2020-08-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107560592A (en) | A kind of precision ranging method for optronic tracker linkage target | |
| CN107808407B (en) | Binocular camera-based unmanned aerial vehicle vision SLAM method, unmanned aerial vehicle and storage medium | |
| CN103868460B (en) | Binocular stereo vision method for automatic measurement based on parallax optimized algorithm | |
| EP3489898B1 (en) | Method and apparatus for estimating disparity | |
| CN107635129B (en) | Three-dimensional trinocular camera device and depth fusion method | |
| CN106960454B (en) | Depth of field obstacle avoidance method and equipment and unmanned aerial vehicle | |
| CN110689562A (en) | Trajectory loop detection optimization method based on generation of countermeasure network | |
| CN109034018A (en) | A kind of low latitude small drone method for barrier perception based on binocular vision | |
| CN108230392A (en) | A kind of dysopia analyte detection false-alarm elimination method based on IMU | |
| CN106529495A (en) | Obstacle detection method of aircraft and device | |
| CN107657644B (en) | Sparse scene flows detection method and device under a kind of mobile environment | |
| CN108805906A (en) | A kind of moving obstacle detection and localization method based on depth map | |
| CN110956661B (en) | Method for calculating dynamic pose of visible light and infrared camera based on bidirectional homography matrix | |
| CN105654547B (en) | Three-dimensional rebuilding method | |
| CN106033614B (en) | A kind of mobile camera motion object detection method under strong parallax | |
| CN110969667A (en) | Multi-spectrum camera external parameter self-correction algorithm based on edge features | |
| CN112184757A (en) | Method and device for determining motion trail, storage medium and electronic device | |
| CN109191513B (en) | Power equipment stereo matching method based on global optimization | |
| CN111998862B (en) | BNN-based dense binocular SLAM method | |
| CN105258673B (en) | A kind of target ranging method based on binocular synthetic aperture focusing image, device | |
| CN104182968A (en) | Method for segmenting fuzzy moving targets by wide-baseline multi-array optical detection system | |
| CN113643345A (en) | Multi-view road intelligent identification method based on double-light fusion | |
| CN116071424A (en) | Fruit space coordinate positioning method based on monocular vision | |
| CN115359130A (en) | Radar and camera combined calibration method and device, electronic equipment and storage medium | |
| CN116381649A (en) | Combined calibration method, device and storage medium |
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 |