CN108665448B - Obstacle detection method based on binocular vision - Google Patents
Obstacle detection method based on binocular vision Download PDFInfo
- Publication number
- CN108665448B CN108665448B CN201810392691.2A CN201810392691A CN108665448B CN 108665448 B CN108665448 B CN 108665448B CN 201810392691 A CN201810392691 A CN 201810392691A CN 108665448 B CN108665448 B CN 108665448B
- Authority
- CN
- China
- Prior art keywords
- graph
- point set
- dimensional point
- coefficient
- road
- 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.)
- Expired - Fee Related
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 32
- 230000004888 barrier function Effects 0.000 claims description 9
- 230000008030 elimination Effects 0.000 claims description 9
- 238000003379 elimination reaction Methods 0.000 claims description 9
- 230000003287 optical effect Effects 0.000 claims description 6
- 238000000034 method Methods 0.000 abstract description 4
- 230000000007 visual effect Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000002431 foraging effect Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
- G06T7/0008—Industrial image inspection checking presence/absence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20228—Disparity calculation for image-based rendering
Abstract
The invention provides a binocular vision-based obstacle detection method. Obtaining a road image through a binocular camera, and obtaining a road parallax image through signal processing; the binocular vision module constructs a three-dimensional point set through the road parallax image, randomly takes four points from the three-dimensional point set to calculate the coefficient of a road plane model, and counts the number of local points in the road parallax image; repeatedly counting the number of local points in the road parallax image for multiple times, and taking the coefficient of the road plane model corresponding to the maximum value of the number of the local points as the coefficient of the obstacle detection; removing points lower than the road surface from the three-dimensional point set according to the coefficient of the obstacle detection, and calculating a first actual distance and a second actual distance according to the pixel value of the removed three-dimensional point set point and the resolution interval of the binocular camera; and calculating the relative speed of the obstacle and the vehicle according to the first actual distance and the second actual distance. Compared with the prior art, the method improves the visual processing speed and has strong real-time performance.
Description
Technical Field
The invention relates to the field of image processing and automation, in particular to a binocular vision-based obstacle detection method.
Background
Currently, binocular stereo vision is used as a main means for obtaining three-dimensional position information of a target. The common method is to obtain a stereo image pair synchronously by two cameras, and then calculate the depth information of each point in the image public view according to the principle of stereo vision. However, the operation amount of the process is large, and the real-time performance of the equipment is difficult to ensure. The stereo vision requirement can solve each corresponding point in a pair of stereo images, and the images contain rich information, so that the speed of corresponding and matching the stationary points is very low, and the calculation problem is easy to generate. Although a certain constraint is added under the regularization frame to solve the problem, the problem is converted into the optimal solution of a plurality of conditional functions, the calculation difficulty and the calculation amount are greatly increased, the visual processing speed is difficult to improve, the real-time requirement cannot be met, and the application field is limited.
Disclosure of Invention
In order to solve the technical problem, the invention provides a binocular vision-based obstacle detection method.
The technical scheme of the invention is a binocular vision-based obstacle detection method, which specifically comprises the following steps:
step 1: obtaining a road image through a binocular camera, and obtaining a road parallax image through signal processing of a binocular vision module;
step 2: the binocular vision module constructs a three-dimensional point set through the road parallax image, four points are randomly selected from the three-dimensional point set to calculate the coefficient of a road plane model, and the number of local points in the road parallax image is counted according to the coefficient of the road plane model;
and step 3: repeatedly executing the step 2 for multiple times, wherein the coefficient of the road plane model corresponding to the maximum value of the number of the local points in the repeated execution for multiple times is used as the coefficient of the obstacle detection;
and 4, step 4: removing points lower than the road surface from the three-dimensional point set according to the coefficient of the obstacle detection, and calculating a first actual distance and a second actual distance according to the pixel value of the removed three-dimensional point set point and the resolution interval of the binocular camera;
and 5: and calculating the relative speed of the obstacle and the vehicle according to the first actual distance and the second actual distance.
Preferably, in step 2, the parallax image is a graph, the pixel values of the ith row and the jth column in the parallax image are graphs (i, j), i belongs to [1, M ], j belongs to [1, N ], and the parallax image graph is an image of M × N;
the three-dimensional point set in the step 2 is as follows:
G＝{(j，i，graph(i，j))|i∈[1，M]，j∈[1，N]}
the three-dimensional point set G comprises M multiplied by N points;
in the step 2, the arbitrary four points taken from the three-dimensional point set G are respectively:
(j+j1，i+i1，graph(i+i1，j+j1))，(j+j2，i+i2，graph(i+i2，j+j2))
(j+j3，i+i3，graph(i+i3，j+j3))，(j+j4，i+i4，graph(i+i4，j+j4))
wherein, (i + i)1)∈[1，M],(i+i2)∈[1，M],(i+i3)∈[1，M],(i+i4)∈[1，M],
(j+j1)∈[1，N],(j+j2)∈[1，N],(j+j3)∈[1，N],(j+j4)∈[1，N]；
In the step 2, the road plane model is as follows:
[abcd][j+j1i+i1graph(i+i1，j+j1)1]T＝0
[abcd][j+j2i+i2graph(i+i2，j+j2)1]T＝0
[abcd][j+j3i+i3graph(i+i3，j+j3)1]T＝0
[abcd][j+j4i+i4graph(i+i4，j+j4)1]T＝0
wherein a is a column coefficient of the road plane model, b is a row coefficient of the road plane model, c is a pixel coefficient of the road plane model, and d is an offset coefficient of the road plane model;
in step 2, the number of local interior points in the road parallax image is counted according to the coefficient of the road plane model as follows:
if a/d + j + b/d + i + c/d + graph (i, j) +1< t in the three-dimensional point set G, regarding the (j, i, graph (i, j)) as an inner point, otherwise, regarding the (j, i) as an outer point; wherein t is a local interior point threshold;
traversing the three-dimensional point set G, and counting the number of local points in the three-dimensional point set G as R;
preferably, the number of times of repeatedly executing step 2 in step 3 is K, and the number of local points of repeatedly executing step 2 for multiple times is respectively:
R1,R2,...,RK
wherein R is1,R2,...,RKThe medium maximum value is RSS is more than or equal to 1 and less than or equal to K, and the coefficient a of the corresponding road plane modelSColumn coefficients for obstacle detection, bSLine coefficients for obstacle detection, cSPixel coefficient for obstacle detection, dSA bias coefficient for obstacle detection;
preferably, the step 4 of removing from the three-dimensional point set is:
if a in the three-dimensional point set GS/dS*j+bS/dS*i+cS/dS*graph(i，j)+1<0, (j, i, graph (i, j)) is removed from the three-dimensional point set G;
the three-dimensional point set after elimination in the step 4 is as follows:
G*＝{(j*，i*，graph(i*，j*))|i*∈[1，M*]，j*∈[1，N*]}
wherein, the three-dimensional point set G after elimination*In which contains M*×N*Point;
in step 4, the resolution interval of the binocular camera is as follows:
[1,2),[2,3),…,[2B-2,2B-1)
wherein, B is the digit of the binocular camera, and the resolution of the binocular camera is 2B；
Three-dimensional point set G after statistical elimination*The number of pixels of the midpoint in the resolution interval is respectively:
the first actual distance in step 4 is:
wherein N isXTo be composed of
Comparing with barrier threshold value g one by one, the first one is in accordance with NXIf the distance is greater than g, f is the camera focal length of the binocular camera, and T is the distance between the optical centers of the two cameras of the binocular camera;
the second actual distance in step 4 is:
wherein N isYTo be N1,N2,…,N2B-2Comparing with barrier threshold g one by one, the second one is in accordance with NYIf the distance is greater than g, f is the camera focal length of the binocular camera, and T is the distance between the optical centers of the two cameras of the binocular camera;
preferably, the relative speed in step 5 is:
wherein Δ h is an obstacle detection period.
Drawings
FIG. 1: is a method flow diagram of the present invention;
FIG. 2: a binocular distance measurement schematic diagram is shown;
FIG. 3: is a schematic view of the present invention.
Detailed Description
In order to facilitate the understanding and implementation of the present invention for those of ordinary skill in the art, the present invention is further described in detail with reference to the accompanying drawings and examples, it is to be understood that the embodiments described herein are merely illustrative and explanatory of the present invention and are not restrictive thereof.
The binocular vision module adopts an S32V234 chip, and the chip adopts 4 ARM Cortex A53 as a core CPU; the binocular camera selects a MYNT EYE small foraging binocular camera;
embodiments of the present invention will be described below with reference to fig. 1 to 3. The specific steps of the embodiment of the invention are as follows:
step 1: obtaining a road image through a binocular camera, and obtaining a road parallax image through signal processing of a binocular vision module;
step 2: the binocular vision module constructs a three-dimensional point set through the road parallax image, four points are randomly selected from the three-dimensional point set to calculate the coefficient of a road plane model, and the number of local points in the road parallax image is counted according to the coefficient of the road plane model;
in the step 2, the parallax image is a graph, the pixel value of the ith row and the jth column in the parallax image is a graph (i, j), i belongs to [1, M ], j belongs to [1, N ], the parallax image graph is an image of M × N, M is 1242, and N is 375;
the three-dimensional point set in the step 2 is as follows:
G＝{(j，i，graph(i，j))|i∈[1，M]，j∈[1，N]}
the three-dimensional point set G comprises M multiplied by N points;
in the step 2, the arbitrary four points taken from the three-dimensional point set G are respectively:
(j+j1，i+i1，graph(i+i1，j+j1))，(j+j2，i+i2，graph(i+i2，j+j2))
(j+j3，i+i3，graph(i+i3，j+j3))，(j+j4，i+i4，graph(i+i4，j+j4))
wherein, (i + i)1)∈[1，M],(i+i2)∈[1，M],(i+i3)∈[1，M],(i+i4)∈[1，M],
(j+j1)∈[1，N],(j+j2)∈[1，N],(j+j3)∈[1，N],(j+j4)∈[1，N]；
In the step 2, the road plane model is as follows:
[a b c d][j+j1i+i1graph(i+i1，j+j1)1]T＝0
[a b c d][j+j2i+i2graph(i+i2，j+j2)1]T＝0
[a b c d][j+j3i+i3graph(i+i3，j+j3)1]T＝0
[a b c d][j+j4i+i4graph(i+i4，j+j4)1]T＝0
wherein a is a column coefficient of the road plane model, b is a row coefficient of the road plane model, c is a pixel coefficient of the road plane model, and d is an offset coefficient of the road plane model;
in step 2, the number of local interior points in the road parallax image is counted according to the coefficient of the road plane model as follows:
if a/d + j + b/d + i + c/d + graph (i, j) +1< t in the three-dimensional point set G, regarding the (j, i, graph (i, j)) as an inner point, otherwise, regarding the (j, i) as an outer point; wherein t is 0.022, and is a local interior point threshold;
traversing the three-dimensional point set G, and counting the number of local points in the three-dimensional point set G as R;
and step 3: repeatedly executing the step 2 for multiple times, wherein the coefficient of the road plane model corresponding to the maximum value of the number of the local points in the repeated execution for multiple times is used as the coefficient of the obstacle detection;
in step 3, the number of times of repeatedly executing step 2 is K equals to 500 times, and the number of local points repeatedly executing step 2 for multiple times is respectively:
R1,R2,...,RK
wherein R is1,R2,...,RKThe medium maximum value is RSS is more than or equal to 1 and less than or equal to K, and the coefficient a of the corresponding road plane modelSColumn coefficients for obstacle detection, bSLine coefficients for obstacle detection, cSPixel coefficient for obstacle detection, dSA bias coefficient for obstacle detection;
and 4, step 4: removing points lower than the road surface from the three-dimensional point set according to the coefficient of the obstacle detection, and calculating a first actual distance and a second actual distance according to the pixel value of the removed three-dimensional point set point and the resolution interval of the binocular camera;
in the step 4, the three-dimensional point set is removed as follows:
if a in the three-dimensional point set GS/dS*j+bS/dS*i+cS/dS*graph(i，j)+1<0, (j, i, graph (i, j)) is removed from the three-dimensional point set G;
the three-dimensional point set after elimination in the step 4 is as follows:
G*＝{(j*，i*，graph(i*，j*))|i*∈[1，M*]，j*∈[1，N*]}
wherein, the three-dimensional point set G after elimination*In which contains M*×N*Point;
in step 4, the resolution interval of the binocular camera is as follows:
[1,2),[2,3),…,[2B-2,2B-1)
wherein, B-8 is the digit of the binocular camera, and the resolution of the binocular camera is 256;
three-dimensional point set G after statistical elimination*The number of pixels of the midpoint in the resolution interval is respectively:
the first actual distance in step 4 is:
wherein f is 0.004m, T is 0.54m, NXTo be composed of
Comparing with barrier threshold g in sequence, g being 1100, the first one being in line with NXIf the distance is greater than g, f is the camera focal length of the binocular camera, and T is the distance between the optical centers of the two cameras of the binocular camera;
the second actual distance in step 4 is:
wherein N isYTo be N1,N2,…,N2B-2Comparing with barrier threshold g one by one, the second one is in accordance with NYIf the distance is greater than g, f is the camera focal length of the binocular camera, and T is the distance between the optical centers of the two cameras of the binocular camera;
and 5: calculating the relative speed of the obstacle and the vehicle according to the first actual distance and the second actual distance;
in step 5, the relative speed is as follows:
wherein Δ h ═ 0.071s is the obstacle detection period.
It should be understood that parts of the specification not set forth in detail are well within the prior art.
It should be understood that the above description of the preferred embodiments is given for clarity and not for any purpose of limitation, and that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (3)
1. A barrier detection method based on binocular vision is characterized by comprising the following steps:
step 1: obtaining a road image through a binocular camera, and obtaining a road parallax image through signal processing of a binocular vision module;
step 2: the binocular vision module constructs a three-dimensional point set through the road parallax image, four points are randomly selected from the three-dimensional point set to calculate the coefficient of a road plane model, and the number of local points in the road parallax image is counted according to the coefficient of the road plane model;
and step 3: repeatedly executing the step 2 for multiple times, wherein the coefficient of the road plane model corresponding to the maximum value of the number of the local points in the repeated execution for multiple times is used as the coefficient of the obstacle detection;
and 4, step 4: removing points lower than the road surface from the three-dimensional point set according to the coefficient of the obstacle detection, and calculating a first actual distance and a second actual distance according to the pixel value of the removed three-dimensional point set point and the resolution interval of the binocular camera;
and 5: calculating the relative speed of the obstacle and the vehicle according to the first actual distance and the second actual distance;
in the step 4, the three-dimensional point set is removed as follows:
if a in the three-dimensional point set GS/dS*j+bS/dS*i+cS/dS(j, i, graph (i, j)) is removed from the three-dimensional point set G if graph (i, j) +1< 0;
the three-dimensional point set after elimination in the step 4 is as follows:
G*＝{(j*，i*，graph(i*，j*))|i*∈[1，M*]，j*∈[1，N*]}
wherein, the three-dimensional point set G after elimination*In which contains M*×N*Point;
in step 4, the resolution interval of the binocular camera is as follows:
[1，2)，[2，3)，…，[2B-2，2B-1)
wherein, B is the digit of the binocular camera, and the resolution of the binocular camera is 2B；
Three-dimensional point set G after statistical elimination*The number of pixels of the midpoint in the resolution interval is respectively:
the first actual distance in step 4 is:
wherein N isXTo be composed of
Comparing with barrier threshold value g one by one, the first one is in accordance with NXIf the distance is greater than g, f is the camera focal length of the binocular camera, and T is the distance between the optical centers of the two cameras of the binocular camera;
the second actual distance in step 4 is:
wherein N isYTo be N1，N2，…，N2B-2Comparing with barrier threshold g one by one, the second one is in accordance with NYIf the distance is greater than g, f is the camera focal length of the binocular camera, and T is the distance between the optical centers of the two cameras of the binocular camera;
in step 5, the relative speed is as follows:
wherein Δ h is an obstacle detection period.
2. The binocular vision based barrier detection method of claim 1, wherein: in the step 2, the parallax image is a graph, the pixel value of the ith row and the jth column in the parallax image is a graph (i, j), i belongs to [1, M ], j belongs to [1, N ], and the parallax image graph is an image of M N;
the three-dimensional point set in the step 2 is as follows:
G＝{(j，i，graph(i，j))|i∈[1，M]，j∈[1，N]}
the three-dimensional point set G comprises M multiplied by N points;
in the step 2, the arbitrary four points taken from the three-dimensional point set G are respectively:
(j+j1，i+i1，graph(i+i1，j+j1))，(j+j2，i+i2，graph(i+i2，j+j2))
(j+j3，i+i3，graph(i+i3，j+j3))，(j+j4，i+i4，graph(i+i4，j+j4))
wherein, (i + i)1)∈[1，M]，(i+i2)∈[1，M]，(i+i3)∈[1，M]，(i+i4)∈[1，M]，(j+j1)∈[1，N]，(j+j2)∈[1，N]，(j+j3)∈[1，N]，(j+j4)∈[1，N]；
In the step 2, the road plane model is as follows:
[a b c d][j+j1 i+i1graph(i+i1，j+j1) 1]T＝0
[a b c d][j+j2 i+i2graph(i+i2，j+j2) 1]T＝0
[a b c d][j+j3 i+i3graph(i+i3，j+j3) 1]T＝0
[a b c d][j+j4 i+i4graph(i+i4，j+j4) 1]T＝0
wherein a is a column coefficient of the road plane model, b is a row coefficient of the road plane model, c is a pixel coefficient of the road plane model, and d is an offset coefficient of the road plane model;
in step 2, the number of local interior points in the road parallax image is counted according to the coefficient of the road plane model as follows:
if a/d + j + b/d + i + c/d graph (i, j) +1< t in the three-dimensional point set G, regarding the (j, i, graph (i, j)) as an inner point, otherwise, regarding the (j, i) as an outer point; wherein t is a local interior point threshold;
and traversing the three-dimensional point set G, and counting the number of the local points in the three-dimensional point set G as R.
3. The binocular vision based barrier detection method of claim 1, wherein: in step 3, the number of times of repeatedly executing step 2 is K, and the number of local points repeatedly executing step 2 for multiple times is respectively:
R1,R2,...,RK
wherein R is1,R2,...,RKThe medium maximum value is RSS is more than or equal to 1 and less than or equal to K, and the coefficient a of the corresponding road plane modelSColumn coefficients for obstacle detection, bSLine coefficients for obstacle detection, cSPixel coefficient for obstacle detection, dSA deviation factor for obstacle detection.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810392691.2A CN108665448B (en) | 2018-04-27 | 2018-04-27 | Obstacle detection method based on binocular vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810392691.2A CN108665448B (en) | 2018-04-27 | 2018-04-27 | Obstacle detection method based on binocular vision |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108665448A CN108665448A (en) | 2018-10-16 |
| CN108665448B true CN108665448B (en) | 2022-05-13 |
Family
ID=63781073
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810392691.2A Expired - Fee Related CN108665448B (en) | 2018-04-27 | 2018-04-27 | Obstacle detection method based on binocular vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108665448B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109601109A (en) * | 2018-12-07 | 2019-04-12 | 江西洪都航空工业集团有限责任公司 | A kind of unmanned grass-cutting vehicle collision-proof method based on binocular vision detection |
| CN114442642B (en) * | 2022-04-02 | 2022-07-15 | 深圳市普渡科技有限公司 | Path planning method, path planning device, computer equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2003269937A (en) * | 2002-03-15 | 2003-09-25 | Sony Corp | Obstruction recognition device, obstruction recognition method, obstruction recognition program and mobile robot device |
| EP2233358A1 (en) * | 2009-03-24 | 2010-09-29 | Aisin Seiki Kabushiki Kaisha | Obstruction detecting apparatus |
| CN105005999A (en) * | 2015-08-12 | 2015-10-28 | 北京航空航天大学 | Obstacle detection method for blind guiding instrument based on computer stereo vision |
| CN105083291A (en) * | 2014-04-25 | 2015-11-25 | 歌乐株式会社 | Driver auxiliary system based on visual line detection |
| CN106970375A (en) * | 2017-02-28 | 2017-07-21 | 河海大学 | A kind of method that building information is automatically extracted in airborne laser radar point cloud |
| CN107341454A (en) * | 2017-06-21 | 2017-11-10 | 海信集团有限公司 | The detection method and device of barrier, electronic equipment in a kind of scene |
-
2018
- 2018-04-27 CN CN201810392691.2A patent/CN108665448B/en not_active Expired - Fee Related
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2003269937A (en) * | 2002-03-15 | 2003-09-25 | Sony Corp | Obstruction recognition device, obstruction recognition method, obstruction recognition program and mobile robot device |
| EP2233358A1 (en) * | 2009-03-24 | 2010-09-29 | Aisin Seiki Kabushiki Kaisha | Obstruction detecting apparatus |
| CN105083291A (en) * | 2014-04-25 | 2015-11-25 | 歌乐株式会社 | Driver auxiliary system based on visual line detection |
| CN105005999A (en) * | 2015-08-12 | 2015-10-28 | 北京航空航天大学 | Obstacle detection method for blind guiding instrument based on computer stereo vision |
| CN106970375A (en) * | 2017-02-28 | 2017-07-21 | 河海大学 | A kind of method that building information is automatically extracted in airborne laser radar point cloud |
| CN107341454A (en) * | 2017-06-21 | 2017-11-10 | 海信集团有限公司 | The detection method and device of barrier, electronic equipment in a kind of scene |
Non-Patent Citations (4)
| Title |
|---|
| A Framework for Applying Point Clouds Grabbed by Multi-Beam LIDAR in Perceiving the Driving Environment;Jian Liu 等;《Sensors 2015》;20150831;21931-21956 * |
| GMSAC——一种鲁棒的基于高斯混合模型的基础矩阵估计算法;徐帆 等;《中国图象图形学报》;20080930;第13卷(第9期);1789-1795 * |
| Unified Stereovision for Ground, Road,and Obstacle Detection;Paolo Lombardi 等;《Intelligent Vehicles Symposium》;20080912;783-788 * |
| 基于双目视觉的道路障碍物识别与检测方法研究;李靖 等;《江苏科技信息》;20170430(第11期);52-55 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108665448A (en) | 2018-10-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110569704B (en) | Multi-strategy self-adaptive lane line detection method based on stereoscopic vision | |
| CN105225482B (en) | Vehicle detecting system and method based on binocular stereo vision | |
| US11295168B2 (en) | Depth estimation and color correction method for monocular underwater images based on deep neural network | |
| CN108564041B (en) | Face detection and restoration method based on RGBD camera | |
| CN107808131B (en) | Dynamic gesture recognition method based on dual-channel deep convolutional neural network | |
| CN105716539B (en) | A kind of three-dimentioned shape measurement method of quick high accuracy | |
| CN105956597A (en) | Binocular stereo matching method based on convolution neural network | |
| CN103996201A (en) | Stereo matching method based on improved gradient and adaptive window | |
| CN103458261B (en) | Video scene variation detection method based on stereoscopic vision | |
| CN110197505B (en) | Remote sensing image binocular stereo matching method based on depth network and semantic information | |
| CN112801074B (en) | Depth map estimation method based on traffic camera | |
| CN108615244A (en) | A kind of image depth estimation method and system based on CNN and depth filter | |
| CN108665448B (en) | Obstacle detection method based on binocular vision | |
| CN111046843A (en) | Monocular distance measurement method under intelligent driving environment | |
| CN113743391A (en) | Three-dimensional obstacle detection system and method applied to low-speed autonomous driving robot | |
| CN111105451B (en) | Driving scene binocular depth estimation method for overcoming occlusion effect | |
| CN115760893A (en) | Single droplet particle size and speed measuring method based on nuclear correlation filtering algorithm | |
| CN111951339A (en) | Image processing method for performing parallax calculation by using heterogeneous binocular cameras | |
| Huang et al. | ES-Net: An efficient stereo matching network | |
| Zheng et al. | Binocular intelligent following robot based on YOLO-LITE | |
| CN107944350A (en) | A kind of monocular vision Road Recognition Algorithm merged based on appearance and geological information | |
| CN102142148B (en) | Video space-time feature extraction method | |
| CN108090920B (en) | Light field image depth stream estimation method | |
| CN112419387B (en) | Unsupervised depth estimation method for solar greenhouse tomato plant image | |
| CN111932584B (en) | Method and device for determining moving object in image |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20220513 |