CN102201167A - Video-based driveway automatic identification method - Google Patents
Video-based driveway automatic identification method Download PDFInfo
- Publication number
- CN102201167A CN102201167A CN2011100863093A CN201110086309A CN102201167A CN 102201167 A CN102201167 A CN 102201167A CN 2011100863093 A CN2011100863093 A CN 2011100863093A CN 201110086309 A CN201110086309 A CN 201110086309A CN 102201167 A CN102201167 A CN 102201167A
- Authority
- CN
- China
- Prior art keywords
- point
- road
- rotation
- track
- starting 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
Images
Landscapes
- Traffic Control Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a video-based driveway automatic identification method, which relates to a road identification method of accumulated frame-difference technology. The method mainly introduces a neural network as a classifier and uses the neural network and the frame-difference technology in combination to identify driveway positions and detect vehicles. Compared with the traditional road identification method which has the problems of complicated process, susceptibility to interference and unsatisfactory identification effect, the method provided by the invention uses learning, association and fault tolerance of the neural network to eliminate disturbances in frame-difference images, improves the accuracy of driveway position identification and provides a basis to automatically mark detection-line positions for the following vehicle detections.
Description
Technical field
The present invention relates to traffic intelligent system field, in particular for the lane recognition method that the traffic flow of road, traffic information etc. are monitored in real time, specifically a kind of automobile track automatic identifying method based on video, this method can in time be extracted telecommunication flow information (vehicle flowrate, lane occupancy ratio, the speed of a motor vehicle etc.), when traffic congestion does not take place, in time take measures, relieve traffic congestion, prevent congested in traffic the generation.When traffic congestion takes place, can in time release news, notify follow-up vehicle in time to adjust travel route, congested traffic is alleviated as early as possible.
Background technology
The real-time collection of transport information and processing are one of gordian techniquies of intelligent transportation system ITS.Advanced traffic information acquisition system not only can be held current traffic circulation situation and following traffic of prediction, and can improve traffic efficiency greatly, realizes the intensive style development of communications and transportation.Traditional traffic data collection method is to bury inductive coil underground to carry out data acquisition under the road surface, and sensing range is little, the construction maintenance trouble, and the failure rate height, especially the destructiveness to high grade pavement is big.
And, adopt the video detecting method of the video information vehicle flowrate data of traffic flow along with the development of computer hardware technique, have incomparable superiority and high performance-price ratio.The main method of road Identification is to use edge detecting technology to detect each bar markings of road in image at present, but because various interference are (as the vehicle that passes through, the building of road both sides, trees, and building, trees stay the shade on road etc. under illumination) influence, it is not ideal enough make to detect effect.For example, because the surface gathered water that rains, cause reflective seriously, reflective place markings are difficult to differentiate, because vehicle blocks, lane boundary line also is difficult to effective extraction simultaneously.
Summary of the invention
The objective of the invention is problem undesirable at existing detection effect in the present road Identification, that lane boundary line also is difficult to effective extraction, a kind of automobile track automatic identifying method based on video is proposed, this method can in time be extracted telecommunication flow information (vehicle flowrate, lane occupancy ratio, the speed of a motor vehicle etc.), when traffic congestion does not take place, in time take measures, relieve traffic congestion, prevent congested in traffic the generation.When traffic congestion takes place, can in time release news, notify follow-up vehicle in time to adjust travel route, congested traffic is alleviated as early as possible.
Technical scheme of the present invention is:
A kind of automobile track automatic identifying method based on video, it may further comprise the steps:
(a). input is used to discern the video sequence of road, adopt frame difference method to handling the frame difference image that obtains consecutive frame between per two the consecutive frame images in the video sequence, this frame difference image is carried out the binary image that smoothing processing and Threshold Segmentation obtain the consecutive frame difference image, white pixel region representation vehicle region wherein, black-pixel region is represented the background area;
(b). binary image is located the left rotation and right rotation point by scanning from left to right and from right to left, be that the center obtains boundary line, the road left and right sides by rotation sweep with the left rotation and right rotation point respectively again, demarcate starting point at the crossover location of boundary line, the left and right sides simultaneously and be used for ensuing rotation sweep;
(c). with the starting point is the point of rotation, boundary line, the road left and right sides is the start and end position, with the certain number of degrees of each rotation is step-length, scan whole road area, number of track-lines according to this road is divided equally into N corresponding section to the sweep trace zone, in each section from the visual field by being the image distance camera position as far as near direction by be divided equally into M zone as far as near steering handle sweep trace, set corresponding continuity threshold value according to the size in each zone, add up the white line hop count of each regional continuation degree greater than threshold value, the data of statistics are during as the learning sample of BP neural network, the continuation degree in M zone all as the network input value, judges manually whether current scanning position is that the track is as the network objectives output valve greater than the white line hop count of threshold value in each section;
(d). set up the BP neural network and also train, can be used as after the network convergence and judge whether current location is the sorter in track with the learning sample data;
(e). during actual identification road, the white pixel of the current scanning position of input statistics count and continuous white pixel point hop count normalization after data, the network output valve be exactly current scanning position whether be the court verdict in track.
Step of the present invention (b) for the binary image that obtains, is set corresponding threshold value according to the distance of image distance camera position afterwards, goes the operation of isolated point.
Be the point of rotation with the starting point in the step of the present invention (c), boundary line, the road left and right sides rotates around starting point respectively that to extend out the 0-5 degree be the start and end position.
A kind of automobile track automatic identifying method based on video, it may further comprise the steps:
(a). input is used to discern the video sequence in track, adopt frame difference method to handling the frame difference image that obtains consecutive frame between per two the consecutive frame images in the video sequence, this frame difference image is carried out the binary image that smoothing processing and Threshold Segmentation obtain the consecutive frame difference image, white pixel region representation vehicle region wherein, black-pixel region is represented the background area;
(b). binary image is located the left rotation and right rotation point: at first, adopt sweep trace to scan from left to right meeting first white pixel point is left-handed turning point; Equally, adopt sweep trace to scan from right to left meeting first white pixel point is the dextrorotation turning point;
Be the center with the left rotation and right rotation point respectively, two sweep traces are rotated scanning, and intersecting at a point until left and right sides sweep trace is starting point; The connecting line of starting point and left-handed turning point is boundary line, the road left side, and the connecting line of starting point and dextrorotation turning point is boundary line, road the right;
(c). the number of track-lines according to road to be identified laterally is divided equally into N corresponding section to the sweep trace zone, in each section, be divided equally into M zone from the visual field by separatrix as far as near each section of steering handle, with the starting point is the point of rotation, the left and right boundary line of road is the start and end position, with the certain number of degrees of each rotation is step-length, scans the road area of whole binary image;
In M * N zone, according to the visual field by as far as the corresponding continuity threshold value of near direction setting, adding up each regional continuation degree is the continuous white pixel point white line hop count greater than threshold value greater than the white line hop count of threshold value, the data of statistics are during as the learning sample of BP neural network, with the continuation degree in M zone in each section greater than the white line hop count of threshold value all as the network input value, judge manually whether current scanning position is that the track is as the network objectives output valve;
(d). set up the BP neural network and also train with the learning sample data, network convergence is afterwards as judging whether current location is the sorter in track;
(e). during reality identification road, the white pixel of current scanning position of statistics counted and continuous white pixel dotted line hop count or the data input category device after the normalization, the network output valve of sorter be exactly current scanning position whether be the court verdict in track.
The present invention for the binary image that obtains, sets corresponding threshold value according to the distance of image distance camera position in step (b) afterwards, goes the operation of isolated point.
The present invention is the point of rotation with the starting point in step (c), and boundary line, the road left and right sides rotates around starting point respectively that to extend out the 0-5 degree be the start and end position.
The present invention is the point of rotation with the starting point in step (c), and boundary line, the road left and right sides rotates around starting point respectively and extends out 3 degree and be the start and end position.
The sector number of the present invention's N section in step (c) is: 10-16..
The number of regions in the present invention M zone in step (c) is carried out corresponding division according to track length, is 5-6.
Beneficial effect of the present invention:
The present invention from the detection of the angle research moving object of computer vision, the method for road Identification, has theory significance to the intelligent transportation research based on video aspect theoretical.On the other hand, method is by research intermediary mathematical theory, Flame Image Process, pattern-recognition etc., for its popularization in intelligent transportation provides theoretical foundation.Aspect practical application, for intelligent transportation system, the Computer Vision Detection technology realizes real time information sampling, just is widely used.Road Identification and moving vehicle detection based on video that this project is studied have good effect, effectively remove the interference in the frame difference image, and the recognition accuracy height does not need a large amount of input pictures, has good real-time performance.Guaranteeing to improve the completeness that detects target information under the real-time condition, for vehicle flowrate, the isoparametric acquisition of the speed of a motor vehicle in the actual traffic system are laid a good foundation.
The present invention also has the following advantages:
The first, classic method at first needs to obtain the image that the view picture picture does not all have vehicle, and new method does not then need;
The second, classic method can be subjected to the edge of buildings and trees, and building and trees stay the influence at the edge of the shade on road under illumination, and new method only detects moving object, therefore can not be subjected to the influence of these static factors;
The 3rd, new method has adopted accumulation frame difference technology, can get rid of the influence of most of accidental noise spot that occurs in the image.
Description of drawings
Fig. 1 is a process flow diagram of the present invention.
Fig. 2 is the scanning road area synoptic diagram of step of the present invention (b).
Fig. 3 is experimentation of the present invention and figure as a result.
Wherein, Fig. 3-a is the video image of identification road; Fig. 3-b is a frame difference binary image,
Fig. 3-c is the track line chart after the neural network identification; Fig. 3-d is that adjacent right lane line and left-lane line are got intermediate value.
Embodiment
The present invention is further illustrated below in conjunction with drawings and Examples.
The present invention considered in long period, each bar track of road all can have abundant vehicle to pass through, and the place that has vehicle to pass through also must be road, in view of the above, proposition reaches the thought of identification stationary body by identification moving object, and designs the roads recognition method of using accumulation frame difference technology.
On this basis, by introducing neural network, utilize the characteristic of neural network to get rid of interference in the frame difference image as sorter, proposes a kind ofly to discern the site of road method based on neural network and frame difference technology.Experimental result shows that the method effect is more satisfactory, can satisfy the needs of practical application.
Adopt the method for neural network and frame difference, the concrete grammar step is as follows:
(1) variable and parameter are set:
If it is W * H (W and H are respectively the width and the height of every two field picture) that being used to of input discerned the size of the every two field picture of video sequence of road, n two field picture g
nPixel (x, gray-scale value y) they are g
n(x, y).When n 〉=2, the frame difference image d of n two field picture and n-1 two field picture
nPixel (x, gray-scale value y) they are d
n(x, y).Dt
n(x y) is frame difference image d
n(x, the temporary variable of y) using in the processing procedure.
The threshold value of dividing vehicle and background is T1.
Whole image is divided into the R section from top to bottom, sets and to remove isolated point threshold value TG[R], and according to the image distance camera position by as far as the ascending threshold value of near direction setting.
Current scan line length is DL, and sweep trace is divided M section, array LC[M equally] be used to write down the qualified continuous white line hop count of each section.
If sweep trace line initial point position place represents the white line section of vehicle and occupies the about y of length
1Individual pixel, white line section mid point is x to the starting point distance
1, sweep trace terminating point position represents that the white line section of vehicle occupies the about y of length
2Individual pixel, white line section mid point is x to the starting point distance
2Can computing center put starting point thus apart from being for the vehicle of x occupies length y:
Then whether qualified threshold value is TL=ky to central point apart from the white line section for x to starting point.K is scale-up factor (k＜=1).
(2) the BP neural network of a M dimension of initialization input 1 dimension output is trained with the learning sample that filters out.
(3) initialization, n=2.
(4) ask the frame difference image d of n two field picture and n-1 two field picture
n,
d
n(x，y)＝|g
n(x，y)-g
n-1(x，y)|，(x＝0，1，...，W-1；y＝0，1，...，H-1)。(3-8)
(5) the mode smoothed frame difference image d of employing Gauss Filtering Processing
n, the eliminating noise of trying one's best.
The Gauss template is
Frame difference image d
nBe filtered into through Gauss
(x＝0，1，...，W-1；y＝0，1，...，H-1)(3-9)
d
n(x，y)＝dt
n(x，y)。(x＝0，1，...，W-1；y＝0，1，...，H-1)。(3-10)
(6) use process of iteration to try to achieve optimal threshold T1, frame difference image d
nIn gray values of pixel points belong to the pixel of vehicle greater than being judged as of T1, gray-scale value belongs to the pixel of background less than being judged as of T1.Use threshold value T1 to cut apart frame difference image d
nWith its binaryzation, the pixel assignment that gray-scale value is lower than T1 is 0, and the pixel assignment that is higher than T1 is 255,
(7) according to isolated point threshold value TG[R], remove in the binary image every section isolated point, white pixel UNICOM region area is less than TG[R in every section image] removal.
(8) scan location left rotation and right rotation point from left to right and from right to left, obtain boundary line, the road left and right sides by rotation sweep again, demarcate starting point at the crossover location of rotating trace simultaneously.Process is shown in Fig. 3-3 (a).
(9) with the starting point be the point of rotation, boundary line, the road left and right sides rotates around starting point respectively and extends out 3 degree for the start and end position, is step-length with each rotation 1 degree, scans whole road area.Process is shown in Fig. 3-3 (b).
Dividing current scan line equally is the M section, begins to search the white pixel section from starting point, finds white pixel section computing center to put starting point apart from x, pixel fragment length L Count, if the pixel fragment central point is arranged in i section, and LCount＞TL, white line hop count LC (i) then ++.
With current LC[M] import the BP network of succeeding in school, obtain the output result of BP network.When last sweep trace result of the output of BP network begins continuous becoming for 3 times when being road for non-road from this, the position of record current scan line is boundary line, the left side, track.Continuous when becoming non-road three times for to be road begin from this as last sweep trace result of the output of BP network, the position of record current scan line is boundary line, the right, track.Process as shown in the figure.
(10) if been scanned, the lane line that identifies is boundary line, 4 pairs of left and right sides; The centre position of getting adjacent boundary line, the right and boundary line, the left side from left to right successively is the separatrix in adjacent two tracks, finishes;
Otherwise n++ changes (4).
Embodiment one:
According to the roads recognition method of above introduction, the video of identification road is an example with 352 * 288 gray level image, i.e. W=352, and H=288, the sample frequency of image was 10 frame/seconds.Consider the situation of modal two-way Four-Lane Road in the present urban road.
Manually from 10 frame difference images, filter out 200 learning samples to learning sample according to method and thought as neural network.The threshold value T1 of dividing vehicle and background is obtained automatically by process of iteration when program run.Image segmentation R=6, TG[R]={ 1*1,2*2,3*3,4*4,5*5,6*6}.Sweep trace divides section M=8.Y1=4, y2=80, x1=2, x2=180, scale-up factor k=0.5.
Net result by experiment can find out more intuitively that lane line and real road lane line that experiment extracts are identical substantially, and ratio of precision is higher, and tests and only used 2 two field pictures, and efficient improves obviously.About 5min of the learning time of neural network wherein, the about 200ms of recognition time, because neural network once learns and can repeatedly use, so the time overhead of each identification road is exactly about 200ms during system's long-time running, the effect of experiment is more satisfactory.
The part that the present invention does not relate to prior art that maybe can adopt all same as the prior art is realized.
Claims (9)
1. automobile track automatic identifying method based on video is characterized in that it may further comprise the steps:
(a). input is used to discern the video sequence of road, adopt frame difference method to handling the frame difference image that obtains consecutive frame between per two the consecutive frame images in the video sequence, this frame difference image is carried out the binary image that smoothing processing and Threshold Segmentation obtain the consecutive frame difference image, white pixel region representation vehicle region wherein, black-pixel region is represented the background area;
(b). binary image is located the left rotation and right rotation point by scanning from left to right and from right to left, be that the center obtains boundary line, the road left and right sides by rotation sweep with the left rotation and right rotation point respectively again, demarcate starting point at the crossover location of boundary line, the left and right sides simultaneously and be used for ensuing rotation sweep;
(c). with the starting point is the point of rotation, boundary line, the road left and right sides is the start and end position, with the certain number of degrees of each rotation is step-length, scan whole road area, number of track-lines according to this road is divided equally into N corresponding section to the sweep trace zone, in each section from the visual field by being the image distance camera position as far as near direction by be divided equally into M zone as far as near steering handle sweep trace, set corresponding continuity threshold value according to the size in each zone, add up the white line hop count of each regional continuation degree greater than threshold value, the data of statistics are during as the learning sample of BP neural network, the continuation degree in M zone all as the network input value, judges manually whether current scanning position is that the track is as the network objectives output valve greater than the white line hop count of threshold value in each section;
(d). set up the BP neural network and also train, can be used as after the network convergence and judge whether current location is the sorter in track with the learning sample data;
(e). during actual identification road, the white pixel of the current scanning position of input statistics count and continuous white pixel point hop count normalization after data, the network output valve be exactly current scanning position whether be the court verdict in track.
2. the automobile track automatic identifying method based on video according to claim 1, it is characterized in that afterwards in step (b), for the binary image that obtains, set corresponding threshold value according to the distance of image distance camera position, go the operation of isolated point.
3. the automobile track automatic identifying method based on video according to claim 1 is characterized in that in step (c) with the starting point being the point of rotation, and boundary line, the road left and right sides rotates around starting point respectively that to extend out the 0-5 degree be the start and end position.
4. automobile track automatic identifying method based on video is characterized in that it may further comprise the steps:
(a). input is used to discern the video sequence in track, adopt frame difference method to handling the frame difference image that obtains consecutive frame between per two the consecutive frame images in the video sequence, this frame difference image is carried out the binary image that smoothing processing and Threshold Segmentation obtain the consecutive frame difference image, white pixel region representation vehicle region wherein, black-pixel region is represented the background area;
(b). binary image is located the left rotation and right rotation point: at first, adopt sweep trace to scan from left to right meeting first white pixel point is left-handed turning point; Equally, adopt sweep trace to scan from right to left meeting first white pixel point is the dextrorotation turning point;
Be the center with the left rotation and right rotation point respectively, two sweep traces are rotated scanning, and intersecting at a point until left and right sides sweep trace is starting point; The connecting line of starting point and left-handed turning point is boundary line, the road left side, and the connecting line of starting point and dextrorotation turning point is boundary line, road the right;
(c). the number of track-lines according to road to be identified laterally is divided equally into N corresponding section to the sweep trace zone, in each section, be divided equally into M zone from the visual field by separatrix as far as near each section of steering handle, with the starting point is the point of rotation, the left and right boundary line of road is the start and end position, with the certain number of degrees of each rotation is step-length, scans the road area of whole binary image;
In M * N zone, according to the visual field by as far as the corresponding continuity threshold value of near direction setting, adding up each regional continuation degree is the continuous white pixel point white line hop count greater than threshold value greater than the white line hop count of threshold value, the data of statistics are during as the learning sample of BP neural network, with the continuation degree in M zone in each section greater than the white line hop count of threshold value all as the network input value, judge manually whether current scanning position is that the track is as the network objectives output valve;
(d). set up the BP neural network and also train with the learning sample data, network convergence is afterwards as judging whether current location is the sorter in track;
(e). during reality identification road, the white pixel of current scanning position of statistics counted and continuous white pixel dotted line hop count or the data input category device after the normalization, the network output valve of sorter be exactly current scanning position whether be the court verdict in track.
5. the automobile track automatic identifying method based on video according to claim 4, it is characterized in that afterwards in step (b), for the binary image that obtains, set corresponding threshold value according to the distance of image distance camera position, go the operation of isolated point.
6. the automobile track automatic identifying method based on video according to claim 4 is characterized in that in step (c) with the starting point being the point of rotation, and boundary line, the road left and right sides rotates around starting point respectively that to extend out the 0-5 degree be the start and end position.
7. the automobile track automatic identifying method based on video according to claim 4 is characterized in that in step (c) with the starting point being the point of rotation, and boundary line, the road left and right sides rotates around starting point respectively and extends out 3 degree and be the start and end position.
8. the automobile track automatic identifying method based on video according to claim 4 is characterized in that the sector number of N section in step (c) is: 10-16.
9. the automobile track automatic identifying method based on video according to claim 4 is characterized in that the number of regions in M zone in step (c) is carried out corresponding division according to track length, is 5-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201110086309 CN102201167B (en) | 2010-04-07 | 2011-04-07 | Video-based driveway automatic identification method |
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010140653.1 | 2010-04-07 | ||
| CN201010140653 | 2010-04-07 | ||
| CN 201110086309 CN102201167B (en) | 2010-04-07 | 2011-04-07 | Video-based driveway automatic identification method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102201167A true CN102201167A (en) | 2011-09-28 |
| CN102201167B CN102201167B (en) | 2013-03-06 |
Family
ID=44661814
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 201110086309 Expired - Fee Related CN102201167B (en) | 2010-04-07 | 2011-04-07 | Video-based driveway automatic identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102201167B (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102495638A (en) * | 2011-12-01 | 2012-06-13 | 王鹏勃 | Multiple degree of freedom robot dinosaur capable of being accurately recognized and controlled and achieving method thereof |
| CN102521566A (en) * | 2011-11-25 | 2012-06-27 | 中国科学院光电技术研究所 | Low-power consumption real-time road automatic identification device for ground autonomous equipment |
| CN105046235A (en) * | 2015-08-03 | 2015-11-11 | 百度在线网络技术（北京）有限公司 | Lane line recognition modeling method and apparatus and recognition method and apparatus |
| CN105184240A (en) * | 2015-08-27 | 2015-12-23 | 广西师范学院 | Scan line clustering-based security video road automatic identification algorithm |
| CN105206107A (en) * | 2014-06-24 | 2015-12-30 | 丰田自动车株式会社 | Lane boundary estimation device and lane boundary estimation method |
| CN105282498A (en) * | 2014-06-18 | 2016-01-27 | 富士重工业株式会社 | Image processing apparatus |
| CN105740831A (en) * | 2016-02-02 | 2016-07-06 | 大连楼兰科技股份有限公司 | Stop line detection method applied to intelligent drive |
| CN106415602A (en) * | 2016-08-25 | 2017-02-15 | 深圳市锐明技术股份有限公司 | Method and device for detecting paired lane lines |
| CN107169402A (en) * | 2016-03-08 | 2017-09-15 | 福特全球技术公司 | Vehicle lane is positioned |
| CN107203738A (en) * | 2016-03-17 | 2017-09-26 | 福特全球技术公司 | Vehicle lane boundary alignment |
| CN108335507A (en) * | 2017-01-17 | 2018-07-27 | 星克跃尔株式会社 | Utilize the driving guiding providing method and device of the filmed image of camera |
| CN109204308A (en) * | 2017-07-03 | 2019-01-15 | 上海汽车集团股份有限公司 | The control method and system that the determination method of lane keeping algorithm, lane are kept |
| CN111104839A (en) * | 2018-10-26 | 2020-05-05 | 斯特拉德视觉公司 | Method and apparatus for detecting the presence of an object |
| CN112329631A (en) * | 2020-11-05 | 2021-02-05 | 浙江点辰航空科技有限公司 | Method for carrying out traffic flow statistics on expressway by using unmanned aerial vehicle |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006031365A (en) * | 2004-07-15 | 2006-02-02 | Mitsubishi Electric Corp | Traffic lane recognition image processing apparatus |
| CN1945596A (en) * | 2006-11-02 | 2007-04-11 | 东南大学 | Vehicle lane Robust identifying method for lane deviation warning |
| CN101179710A (en) * | 2007-11-30 | 2008-05-14 | 浙江工业大学 | Intelligent video monitoring apparatus of railway crossing |
| JP2009134429A (en) * | 2007-11-29 | 2009-06-18 | Aisin Aw Co Ltd | Image recognition apparatus and image recognition program |
-
2011
- 2011-04-07 CN CN 201110086309 patent/CN102201167B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006031365A (en) * | 2004-07-15 | 2006-02-02 | Mitsubishi Electric Corp | Traffic lane recognition image processing apparatus |
| CN1945596A (en) * | 2006-11-02 | 2007-04-11 | 东南大学 | Vehicle lane Robust identifying method for lane deviation warning |
| JP2009134429A (en) * | 2007-11-29 | 2009-06-18 | Aisin Aw Co Ltd | Image recognition apparatus and image recognition program |
| CN101179710A (en) * | 2007-11-30 | 2008-05-14 | 浙江工业大学 | Intelligent video monitoring apparatus of railway crossing |
Non-Patent Citations (1)
| Title |
|---|
| 林青: "基于视觉的行车道检测系统研究", 《中国优秀硕士学位论文全文数据库》 * |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102521566A (en) * | 2011-11-25 | 2012-06-27 | 中国科学院光电技术研究所 | Low-power consumption real-time road automatic identification device for ground autonomous equipment |
| CN102521566B (en) * | 2011-11-25 | 2013-07-24 | 中国科学院光电技术研究所 | Low-power consumption real-time road automatic identification device for ground autonomous equipment |
| CN102495638A (en) * | 2011-12-01 | 2012-06-13 | 王鹏勃 | Multiple degree of freedom robot dinosaur capable of being accurately recognized and controlled and achieving method thereof |
| CN105282498B (en) * | 2014-06-18 | 2017-11-14 | 株式会社斯巴鲁 | Image processing apparatus |
| CN105282498A (en) * | 2014-06-18 | 2016-01-27 | 富士重工业株式会社 | Image processing apparatus |
| CN105206107A (en) * | 2014-06-24 | 2015-12-30 | 丰田自动车株式会社 | Lane boundary estimation device and lane boundary estimation method |
| CN105046235A (en) * | 2015-08-03 | 2015-11-11 | 百度在线网络技术（北京）有限公司 | Lane line recognition modeling method and apparatus and recognition method and apparatus |
| US10699134B2 (en) | 2015-08-03 | 2020-06-30 | Baidu Online Network Technology (Beijing) Co., Ltd. | Method, apparatus, storage medium and device for modeling lane line identification, and method, apparatus, storage medium and device for identifying lane line |
| CN105046235B (en) * | 2015-08-03 | 2018-09-07 | 百度在线网络技术（北京）有限公司 | The identification modeling method and device of lane line, recognition methods and device |
| CN105184240B (en) * | 2015-08-27 | 2018-05-29 | 广西师范学院 | Security protection video road automatic identification algorithm based on scan line cluster |
| CN105184240A (en) * | 2015-08-27 | 2015-12-23 | 广西师范学院 | Scan line clustering-based security video road automatic identification algorithm |
| CN105740831B (en) * | 2016-02-02 | 2019-06-07 | 大连楼兰科技股份有限公司 | A kind of stopping line detecting method applied to intelligent driving |
| CN105740831A (en) * | 2016-02-02 | 2016-07-06 | 大连楼兰科技股份有限公司 | Stop line detection method applied to intelligent drive |
| CN107169402B (en) * | 2016-03-08 | 2021-10-08 | 福特全球技术公司 | Vehicle lane positioning |
| CN107169402A (en) * | 2016-03-08 | 2017-09-15 | 福特全球技术公司 | Vehicle lane is positioned |
| CN107203738A (en) * | 2016-03-17 | 2017-09-26 | 福特全球技术公司 | Vehicle lane boundary alignment |
| CN107203738B (en) * | 2016-03-17 | 2022-04-12 | 福特全球技术公司 | Vehicle lane boundary positioning |
| CN106415602A (en) * | 2016-08-25 | 2017-02-15 | 深圳市锐明技术股份有限公司 | Method and device for detecting paired lane lines |
| CN108335507B (en) * | 2017-01-17 | 2021-06-29 | 星克跃尔株式会社 | Method and device for providing driving guidance by using camera to shoot image |
| CN108335507A (en) * | 2017-01-17 | 2018-07-27 | 星克跃尔株式会社 | Utilize the driving guiding providing method and device of the filmed image of camera |
| CN109204308A (en) * | 2017-07-03 | 2019-01-15 | 上海汽车集团股份有限公司 | The control method and system that the determination method of lane keeping algorithm, lane are kept |
| CN109204308B (en) * | 2017-07-03 | 2020-04-07 | 上海汽车集团股份有限公司 | Determination method of lane keeping algorithm, and control method and system of lane keeping |
| CN111104839A (en) * | 2018-10-26 | 2020-05-05 | 斯特拉德视觉公司 | Method and apparatus for detecting the presence of an object |
| CN111104839B (en) * | 2018-10-26 | 2023-12-19 | 斯特拉德视觉公司 | Method and device for detecting existence of object |
| CN112329631A (en) * | 2020-11-05 | 2021-02-05 | 浙江点辰航空科技有限公司 | Method for carrying out traffic flow statistics on expressway by using unmanned aerial vehicle |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102201167B (en) | 2013-03-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102201167B (en) | Video-based driveway automatic identification method | |
| CN102819952B (en) | Method for detecting illegal lane change of vehicle based on video detection technique | |
| CN108230254B (en) | Automatic detection method for high-speed traffic full lane line capable of self-adapting scene switching | |
| CN101799987B (en) | Self-adaptive intelligent traffic light and control method thereof | |
| CN110717433A (en) | Deep learning-based traffic violation analysis method and device | |
| CN105005771B (en) | A kind of detection method of the lane line solid line based on light stream locus of points statistics | |
| CN103914688B (en) | A kind of urban road differentiating obstacle | |
| CN103077617B (en) | Based on crossing intelligent traffic light supervisory systems and the method for computer vision | |
| CN104537841B (en) | Unlicensed vehicle violation detection method and detection system thereof | |
| CN101702263B (en) | Pedestrian crosswalk signal lamp green wave self-adaption control system and method | |
| CN103531031B (en) | It is a kind of that the current control method that blocks by nothing is realized based on the identification of urban traffic trunk line pliable-closure area video detection | |
| CN103985182B (en) | A kind of bus passenger flow automatic counting method and automatic counter system | |
| CN107301776A (en) | Track road conditions processing and dissemination method based on video detection technology | |
| CN105632186A (en) | Method and device for detecting vehicle queue jumping behavior | |
| CN102855759A (en) | Automatic collecting method of high-resolution satellite remote sensing traffic flow information | |
| CN111597905B (en) | Highway tunnel parking detection method based on video technology | |
| CN104392212A (en) | Method for detecting road information and identifying forward vehicles based on vision | |
| CN105740809A (en) | Expressway lane line detection method based on onboard camera | |
| CN102682602B (en) | Collecting method for road traffic parameters based on video technique | |
| CN102147971A (en) | Traffic information acquisition system based on video image processing technology | |
| CN105741559A (en) | Emergency vehicle lane illegal occupation detection method based on lane line model | |
| CN109272482A (en) | A kind of urban road crossing vehicle queue detection system based on sequence image | |
| CN105654073A (en) | Automatic speed control method based on visual detection | |
| CN104700072A (en) | Lane line historical frame recognition method | |
| CN105118305A (en) | Vehicle management platform of outdoor parking lot exit |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20110928 Assignee: NANJING MIND SOFTWARE CO., LTD. Assignor: Gong Ningsheng Contract record no.: 2012320001008 Denomination of invention: Video-based driveway automatic identification method License type: Common License Record date: 20121121 |
|
| LICC | Enforcement, change and cancellation of record of contracts on the licence for exploitation of a patent or utility model | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20130306 Termination date: 20190407 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |