CN103456170A - Vehicle speed and vehicle queue length detection method based on machine vision - Google Patents
Vehicle speed and vehicle queue length detection method based on machine vision Download PDFInfo
- Publication number
- CN103456170A CN103456170A CN2013101381082A CN201310138108A CN103456170A CN 103456170 A CN103456170 A CN 103456170A CN 2013101381082 A CN2013101381082 A CN 2013101381082A CN 201310138108 A CN201310138108 A CN 201310138108A CN 103456170 A CN103456170 A CN 103456170A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- speed
- surveyed area
- queue length
- telescopic
- 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.)
- Pending
Links
Images
Landscapes
- Image Processing (AREA)
- Image Analysis (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention relates to a vehicle speed and vehicle queue length detection method based on machine vision. The method is applied to a city crossroad. The method is characterized in that a detection area is arranged between a stop line of the city crossroad and a pedestrian crosswalk, and whether a vehicle exists is judged in the detection area; if the vehicle exists, the vehicle speed is calculated; when the vehicle speed on a lane is detected to be less than 2km/h, a queue length detection program is started; a plurality of movable virtual loops which are called telescopic windows for short are arranged along the corresponding lane, and when the sum of the number of nonzero pixels of the telescopic windows is detected to be larger than a certain threshold value and the difference of the number of the nonzero pixels in the telescopic windows and the number of all nonzero pixels in a previous time window is smaller than the certain threshold value, the lengths of the telescopic windows are correspondingly increased; finally, pixel values are changed into a practical length value through camera calibration.
Description
Technical field
The traffic parameter that the present invention relates to intelligent transportation system gathers field, the particularly a kind of speed of a motor vehicle based on machine vision of city intersection and detection method of vehicle queue length of being applied to.
Background technology
Along with economic fast development, traffic has also obtained very large development, but a lot of problems but occurred, and as traffic congestion, traffic hazard take place frequently and traffic environment deterioration etc., and these problems have had influence on daily life.Intelligent transportation system is the newest fruits that transport solution transportation problem is explored.
In urban transportation, intersection affects the traffic safety level in whole city.When the average speeds of intersection is down to below 2km/h, vehicle starts to queue up, and traffic congestion occurs., when vehicle, exercise when slow, the exhaust emissions amount can increase simultaneously, and this just makes the air quality in city further worsen.In real time dynamic traffic information collection is the key that realizes intelligent transportation.And the real-time speed of a motor vehicle of road and vehicle queue length are the key foundation data of intelligent transportation system, so be necessary very much the speed of a motor vehicle of city intersection and the detection method of queue length are studied.
Research about the information acquisition of the speed of a motor vehicle and vehicle queue length in real time at present also starts to obtain extensive concern, has obtained certain achievement in research.Especially in real time the speed of a motor vehicle, as one of important parameter of intelligent transportation, is widely used on expressway and through street.
Although it is so ripe that the research of vehicle queue length detection method does not have the speed of a motor vehicle to detect, and along with the quick emergence of graph processing technique and computer vision, obtained development to a certain degree yet.
The inventor, in realizing process of the present invention, finds that at least there is following shortcoming and defect in prior art:
Prior art mainly concentrates on the good vehicle detection of traffic environment such as highway, through street, can't be applied to complicated city intersection.
The judgement complexity of vehicle queue state.
Summary of the invention
Be confined to highway and through street in order to solve prior art, improve accuracy rate, real-time and the robustness of detection method, the embodiment of the present invention provides a kind of speed of a motor vehicle based on machine vision and the detection method of vehicle queue length, and specific embodiments is as follows:
Camera pedestal is located on the other support in crossing, with the oncoming vehicle of the angle shot of overlooking.The traffic video obtained by video camera, comprising and the speed of a motor vehicle and the irrelevant bulk information of vehicle queue, in speed of a motor vehicle testing process, if entire image is processed, certainly will cause calculated amount excessive, be difficult to meet the requirement that detects the vehicle real-time, need to remove the information irrelevant with detecting vehicle, therefore surveyed area need to be set.In video image, surveyed area is arranged between the stop line and People's Bank of China's zebra stripes of intersection, and the interference that this zone is subject to is minimum; Carry out background calculus of differences, shadow removal and morphology area and fill in surveyed area, then the bianry image obtained is carried out to vertical projection, utilize the vertical projection curve map to judge whether vehicle exists; If there is vehicle, calculate vehicle and sail and roll away from surveyed area frame number used into, calculate instantaneous velocity according to the frame per second of video camera and the width of surveyed area, and then obtain the average speed of intersection.
Vehicle occurs that queuing phenomena is generally when red interval or crossing occur during vehicle pass-through slowly is green light that the speed of a motor vehicle is lower, therefore, when the speed of a motor vehicle being detected lower than 2km/h, starts to start the queue length trace routine.
Because during vehicle queue, camera acquisition to the lane width of frame of video the inside along with the increase of queue length, reduce gradually, in order to improve accuracy of detection, this paper adopts a plurality of telescopic windows (being transformable virtual coil) to be detected, the width of telescopic window is set according to the lane line width at place, in order to can effectively reduce error.In the vehicle queue situation, can set a corresponding n telescopic window, calculate n the pixel count that telescopic window is corresponding simultaneously, then the height that utilizes video camera to set up, surveyed area are demarcated video camera apart from the horizontal range of video camera, visual angle width and the angle of depression size of video camera, thereby can derive actual range corresponding to each pixel, and then calculate the length of vehicle queue.
The accompanying drawing explanation:
In order to be illustrated more clearly in inventive embodiments or technical scheme of the prior art, below will the accompanying drawing of required use in embodiment or description of the Prior Art be briefly described, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skills, do not paying under the prerequisite of creating work, can also obtain with reference to the accompanying drawings other accompanying drawing.
Fig. 1 is that embodiment of the present invention surveyed area arranges exemplary plot.
Fig. 2 is the exemplary plot of embodiment of the present invention vehicle queue detection method.
Embodiment:
For making the purpose, technical solutions and advantages of the present invention clearer, below in conjunction with accompanying drawing, embodiment of the present invention is described further in detail.
In the fixed position of each two field picture of video, the surveyed area that covers a plurality of tracks that will monitor is set.The position of surveyed area is arranged between the stop line and People's Bank of China's zebra stripes of city intersection.The position of surveyed area arranges as shown in Figure 1, carries out the background calculus of differences in surveyed area, utilizes current frame image and background image to subtract each other and the binarization segmentation moving region.Then carry out shadow removal, morphology area filling and vertical projection.
Vertical projection method refers to the number of the bianry image in surveyed area being added up from left to right to each row non-zero pixels of this bianry image.If there is car to exist in surveyed area, will obtain the waveform of sealing through vertical projection, the width of this waveform is the width of moving region.When the width of moving region than given threshold value T
wgreatly, judgement has the existence of vehicle.Given threshold value T
wbe set to 90% of common coupe body width.If there is vehicle, calculate vehicle and sail and roll away from surveyed area frame number used into, calculate instantaneous velocity according to the frame per second of video camera and the width of surveyed area, and then obtain the average speed of intersection.The separate processing in each track, the flow process that process in each track is identical.
When the speed of a motor vehicle being detected lower than 2km/h, start to start the queue length trace routine.
Because during vehicle queue, camera acquisition to the lane width of frame of video the inside along with the increase of queue length, reduce gradually, in order to improve accuracy of detection, this paper adopts a plurality of telescopic windows (being transformable virtual coil) to be detected, the width of telescopic window is set according to the lane line width at place, in order to can effectively reduce error.The width of first telescopic window is set to threshold value W
1, in y direction of principal axis length, be N
1(N
1for Vehicle length pixel of place queuing line, the flase drop of avoiding tomography and vehicle space to cause); The width of second telescopic window is set to threshold value W
2, in y direction of principal axis length, be N
2, the rest may be inferred, and the width of n telescopic window is made as W
n, in y direction of principal axis length, be N
n.As shown in Figure 2.
The telescopic window algorithm is as follows:
First telescopic window is placed on stop line place, track, when first telescopic window non-zero pixels and be greater than a certain threshold value being detected, and in telescopic window all non-zero pixels and with the previous moment window in all non-zero pixels and difference be less than a certain threshold value, telescopic window length adds N
1, on the y direction of principal axis, extend N
1individual pixel unit, otherwise length minimizing, but can not be less than a pixel.
The maximal value M that is elongated to setting when first telescopic window
1the time, it is no longer elongation just, at this moment opens second telescopic window.Repeat above-mentioned steps.
By that analogy, until open n telescopic window.
According to above-mentioned detection method, in the vehicle queue situation, can set a corresponding n telescopic window, calculate n the pixel count that telescopic window is corresponding simultaneously, then the height that utilizes video camera to set up, surveyed area are demarcated video camera apart from the horizontal range of video camera, visual angle width and the angle of depression size of video camera, thereby can derive actual range corresponding to each pixel, and then calculate the length of vehicle queue.
The foregoing is only preferred embodiment of the present invention, in order to limit the present invention, within the spirit and principles in the present invention not all, any modification of doing, be equal to replacement, improvement etc., within all should being included in protection scope of the present invention.
Claims (1)
1. one kind is applied to the speed of a motor vehicle based on machine vision of city intersection and the detection method of vehicle queue length, it is characterized in that: at first between the stop line of city intersection and crossing, surveyed area is set, carrying out background calculus of differences, shadow removal and morphology area in surveyed area fills, then the bianry image obtained is carried out to vertical projection, utilize the vertical projection curve map to judge whether vehicle exists; If there is vehicle, calculate vehicle and sail and roll away from surveyed area frame number used into, calculate instantaneous velocity according to the frame per second of video camera and the width of surveyed area, and then obtain the average speed of intersection.Foundation using the speed of a motor vehicle as vehicle queue, when track, the place speed of a motor vehicle being detected lower than 2km/h, start to start the queue length trace routine.In order to improve accuracy of detection, a plurality of mobilizable virtual coils (abbreviation telescopic window) are set along corresponding track, and when telescopic window non-zero pixels number being detected and being greater than a certain threshold value, and in telescopic window all the non-zero pixels numbers and with the previous moment telescopic window in all the non-zero pixels numbers and difference be less than a certain threshold value, telescopic window length correspondingly increases.Finally utilize camera calibration that pixel value is converted to actual length value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013101381082A CN103456170A (en) | 2013-04-22 | 2013-04-22 | Vehicle speed and vehicle queue length detection method based on machine vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013101381082A CN103456170A (en) | 2013-04-22 | 2013-04-22 | Vehicle speed and vehicle queue length detection method based on machine vision |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103456170A true CN103456170A (en) | 2013-12-18 |
Family
ID=49738488
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2013101381082A Pending CN103456170A (en) | 2013-04-22 | 2013-04-22 | Vehicle speed and vehicle queue length detection method based on machine vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103456170A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104240509A (en) * | 2014-10-10 | 2014-12-24 | 南京莱斯信息技术股份有限公司 | Comprehensive video detection system for transport vehicles |
| CN104637323A (en) * | 2015-03-02 | 2015-05-20 | 王刚 | Waiting time self-regulating signal lamp |
| CN105405299A (en) * | 2015-12-17 | 2016-03-16 | 天津中安视通科技有限公司 | Control method of video velocity measurement system |
| CN105513342A (en) * | 2015-11-25 | 2016-04-20 | 南京莱斯信息技术股份有限公司 | Video-tracking-based vehicle queuing length calculating method |
| CN106781493A (en) * | 2016-12-30 | 2017-05-31 | 迈锐数据（北京）有限公司 | A kind of vehicle queue length simulation system, method and apparatus |
| CN107016865A (en) * | 2017-05-25 | 2017-08-04 | 杜艳林 | A kind of one-way road instruction system and road instruction method |
| CN106128121B (en) * | 2016-07-05 | 2018-08-17 | 中国石油大学(华东) | Vehicle queue length fast algorithm of detecting based on Local Features Analysis |
| CN110047297A (en) * | 2019-04-15 | 2019-07-23 | 河北科技大学 | Vehicle speed measuring method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101197080A (en) * | 2006-12-22 | 2008-06-11 | 四川川大智胜软件股份有限公司 | Novel intelligent traffic light control system |
| US20100079306A1 (en) * | 2008-09-26 | 2010-04-01 | Regents Of The University Of Minnesota | Traffic flow monitoring for intersections with signal controls |
| CN101777259A (en) * | 2010-01-22 | 2010-07-14 | 同济大学 | Method for acquiring mean delay of urban road junction |
| CN102867414A (en) * | 2012-08-18 | 2013-01-09 | 湖南大学 | Vehicle queue length measurement method based on PTZ (Pan/Tilt/Zoom) camera fast calibration |
-
2013
- 2013-04-22 CN CN2013101381082A patent/CN103456170A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101197080A (en) * | 2006-12-22 | 2008-06-11 | 四川川大智胜软件股份有限公司 | Novel intelligent traffic light control system |
| US20100079306A1 (en) * | 2008-09-26 | 2010-04-01 | Regents Of The University Of Minnesota | Traffic flow monitoring for intersections with signal controls |
| CN101777259A (en) * | 2010-01-22 | 2010-07-14 | 同济大学 | Method for acquiring mean delay of urban road junction |
| CN102867414A (en) * | 2012-08-18 | 2013-01-09 | 湖南大学 | Vehicle queue length measurement method based on PTZ (Pan/Tilt/Zoom) camera fast calibration |
Non-Patent Citations (4)
| Title |
|---|
| ANTONIO ALBIOL: "《vision-based vehicle queue length estimation》", 《2011 IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION WORKSHOPS(ICCV WORKSHOPS)》 * |
| R.K.SATZODA: "vision-based vehicle queue detection at traffic junctions", 《2012 7TH IEEE CONFERENCE ON INDUSTRIAL ELECTRONICS AND APPLICATIONS(ICIEA)》 * |
| 周世付: "基于高斯混合模型的多车道车流量检测算法", 《计算机仿真》 * |
| 李卫斌: "基于图像处理的车辆排队长度鲁棒性检测算法", 《计算机测量与控制》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104240509A (en) * | 2014-10-10 | 2014-12-24 | 南京莱斯信息技术股份有限公司 | Comprehensive video detection system for transport vehicles |
| CN104637323A (en) * | 2015-03-02 | 2015-05-20 | 王刚 | Waiting time self-regulating signal lamp |
| CN105513342A (en) * | 2015-11-25 | 2016-04-20 | 南京莱斯信息技术股份有限公司 | Video-tracking-based vehicle queuing length calculating method |
| CN105405299A (en) * | 2015-12-17 | 2016-03-16 | 天津中安视通科技有限公司 | Control method of video velocity measurement system |
| CN106128121B (en) * | 2016-07-05 | 2018-08-17 | 中国石油大学(华东) | Vehicle queue length fast algorithm of detecting based on Local Features Analysis |
| CN106781493A (en) * | 2016-12-30 | 2017-05-31 | 迈锐数据（北京）有限公司 | A kind of vehicle queue length simulation system, method and apparatus |
| CN106781493B (en) * | 2016-12-30 | 2020-09-18 | 迈锐数据（北京）有限公司 | Vehicle queuing length simulation system, method and device |
| CN107016865A (en) * | 2017-05-25 | 2017-08-04 | 杜艳林 | A kind of one-way road instruction system and road instruction method |
| CN110047297A (en) * | 2019-04-15 | 2019-07-23 | 河北科技大学 | Vehicle speed measuring method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103456170A (en) | Vehicle speed and vehicle queue length detection method based on machine vision | |
| CN105632186B (en) | Detection vehicle is jumped a queue the method and apparatus of behavior | |
| CN104021378B (en) | Traffic lights real-time identification method based on space time correlation Yu priori | |
| CN105260713B (en) | A kind of method for detecting lane lines and device | |
| CN105005771B (en) | A kind of detection method of the lane line solid line based on light stream locus of points statistics | |
| CN105005778B (en) | A kind of vehicle on highway detection method of view-based access control model background extracting | |
| CN103714363B (en) | A kind of motor vehicle exhaust smoke video identification system | |
| CN104537841B (en) | Unlicensed vehicle violation detection method and detection system thereof | |
| CN108615358A (en) | A kind of congestion in road detection method and device | |
| CN110287905B (en) | Deep learning-based real-time traffic jam area detection method | |
| CN104157143B (en) | Parking offense detection system and detection method thereof | |
| CN103617410A (en) | Highway tunnel parking detection method based on video detection technology | |
| CN103730015B (en) | Intersection traffic flow detecting method and device | |
| CN106327863A (en) | Method, device and system for detection and early warning of vehicle congestion | |
| CN103177586A (en) | Machine-vision-based urban intersection multilane traffic flow detection method | |
| JP6020567B2 (en) | Three-dimensional object detection apparatus and three-dimensional object detection method | |
| CN103366572B (en) | Method for detecting video traffic parameters of intersection | |
| CN103021179B (en) | Based on the Safe belt detection method in real-time monitor video | |
| CN107644528A (en) | A kind of vehicle queue length detection method based on vehicle tracking | |
| CN107240268A (en) | A kind of detection method of high speed Emergency Vehicle Lane vehicles peccancy towards unmanned plane | |
| CN110021171A (en) | Freeway Conditions localization method, apparatus and system | |
| CN107274678B (en) | A kind of night vehicle flowrate and model recognizing method based on Kinect | |
| US20150324651A1 (en) | Three-dimensional object detection device and three-dimensional object detection method | |
| CN110929676A (en) | Deep learning-based real-time detection method for illegal turning around | |
| CN104794907A (en) | Traffic volume detection method using lane splitting and combining |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20131218 |