CN112216119A - Method for identifying traffic vehicle passing event on highway - Google Patents
Method for identifying traffic vehicle passing event on highway Download PDFInfo
- Publication number
- CN112216119A CN112216119A CN202010679157.7A CN202010679157A CN112216119A CN 112216119 A CN112216119 A CN 112216119A CN 202010679157 A CN202010679157 A CN 202010679157A CN 112216119 A CN112216119 A CN 112216119A
- Authority
- CN
- China
- Prior art keywords
- congestion
- area
- vehicle
- target
- threshold
- 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
- 238000000034 method Methods 0.000 title claims abstract description 28
- 238000001514 detection method Methods 0.000 claims abstract description 40
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 5
- 238000005070 sampling Methods 0.000 claims description 18
- 238000001914 filtration Methods 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 claims description 3
- 238000005286 illumination Methods 0.000 abstract description 2
- 230000000007 visual effect Effects 0.000 abstract description 2
- 238000004364 calculation method Methods 0.000 abstract 1
- 206010039203 Road traffic accident Diseases 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/017—Detecting movement of traffic to be counted or controlled identifying vehicles
- G08G1/0175—Detecting movement of traffic to be counted or controlled identifying vehicles by photographing vehicles, e.g. when violating traffic rules
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/065—Traffic control systems for road vehicles by counting the vehicles in a section of the road or in a parking area, i.e. comparing incoming count with outgoing count
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Traffic Control Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a method for identifying a traffic incident of a highway, which belongs to the technical field of visual image detection data processing, judges by using data acquired by a highway camera, and comprises the following steps: vehicle detection, vehicle track tracking, vehicle flow calculation, congestion judgment, driving accident judgment, illegal parking and driving judgment; the method can automatically abstract the image characteristics of the input convolutional neural network layer by layer, reduce the workload of artificially marking the characteristics and save the labor cost and the time cost; the target positioning of the vehicle can be quickly realized, and the type of the vehicle (car, truck, bus, etc.) can be identified; the method has good adaptability to different environments, illumination intensity and weather conditions; the customization requirement is flexible (the event identification parameters such as a detection boundary line, a vehicle speed threshold value and the like can be adjusted in real time); the problem of shielding can be handled to a certain extent in the crowded environment of large-flow vehicles, and higher detection accuracy is guaranteed.
Description
Technical Field
The invention relates to the technical field of visual image detection data processing, in particular to a method for identifying a traffic incident of a highway.
Background
With the continuous increase of the vehicle occupancy in China, the driving conditions of roads are more and more, which seriously affects the traveling life of people, and particularly, the congestion or the violation accidents of regulations on expressways bring huge economic losses to the nation. If the running condition of the current road traffic can be accurately judged, the traffic can be effectively dredged and managed.
At present, when a video detection technology is used for detecting traffic driving events, two modes are available, namely a mode of transmitting video images to a monitoring center; the other method is that after traffic parameters such as flow, road occupancy, speed, inter-vehicle distance, queuing length and the like are obtained, a plurality of traffic state parameters are selected, and a predefined congestion judging method is utilized to judge the traffic congestion.
The first mode generally adopts a manual processing method, so that the efficiency is low and more traffic roads cannot be processed; the second method is not very accurate in obtaining various parameters, so that the final processing result is not accurate, and the method has no good expansion capability. In general, the prior art has the problem that the road traffic state cannot be accurately and effectively judged when the road congestion condition is judged based on the video technology.
Disclosure of Invention
It is an object of the present invention to overcome the above-mentioned background difficulties and to provide a method of identifying highway vehicle traffic events.
In order to achieve the purpose, the technical scheme is as follows: a method for identifying traffic accident on highway, which uses data collected by highway camera to judge, includes following steps:
s1 vehicle detection: according to the video data of the sampling area, performing type detection on the vehicle in the video based on yolov3, and specifying the vehicle as a target;
s2 vehicle trajectory tracking: performing target tracking by adopting KCF full-symmetric kernel correlation filtering, and defining an entity class which instantiates an object for each detected object; initializing a KCF tracker when an object is instantiated, receiving the coordinate positions of a frame and a target by the KCF tracker, loading the latest frame through an update () function, and calculating the position of the target in the current frame by using the KCF tracker;
and S3 calculating the traffic flow: dynamically defining a section in the sampling area, and when a vehicle object is monitored to completely pass through the section, considering the total traffic flow;
and S4 congestion judgment: firstly, in a sampling area, judging whether a certain vehicle target is a congestion unit, namely, tracking the stay time T of a target vehicle C in the sampling areaduring≥TthresholdIf so, the target is considered as a congestion unit; residence time TduringEqual to the current video frame time TcurSubtracting the initial time C of tracking the target vehicle Cinit_time(ii) a Secondly, detecting a congestion area if N exists in the same lane in the congestion detection areathreshold_perloadIndividual congestion units, or total presence of N in the areathreshold_totalThe congestion unit judges that the area is a congestion area; thirdly, judging the congestion in multiple stages, and if the current congestion detects that the area is blocked or N existsthresholdAnd entering the next congestion detection area for congestion judgment according to the congestion target, wherein the set of congestion areas needing to be reported is P = { area =iI < n }; wherein, areaiIndicating an ith congestion detection area, numbering the congestion detection areas from front to back according to the driving direction, and taking n as the number of the congestion area at the last detected current position;
s5 judging the traffic accident: in the sampling area, target vehicles meeting the congestion condition and target vehicles not meeting the congestion condition exist at the same time, or the number of the target vehicles is small and the congestion condition exists, and the target vehicles are judged to meet the accident condition;
s6 illegal parking and driving judgment: in the sampling area, dynamically observing the monitoring area at the cloud end, defining an area as a forbidden area and a forbidden area, and judging that the illegal conditions are met when a monitored target enters the area; when it is monitored that a target enters the area for a certain time, the condition of illegal parking is judged to be met.
Further, the convolutional neural network in step S1 is based on yolov3 and optimized for vehicle type detection, and is designed and built for vehicle type detection.
Further, T in the step S4thresholdThe unit is a preset adjustable value in seconds, and the value range is more than or equal to 1.
Further, N in the step S4threshold_perloadAnd Nthreshold_totalThe value range is more than or equal to 1 for preset adjustable values.
Further, N in the step S4thresholdThe value range is more than or equal to 1 for the preset adjustable value.
The beneficial effect who adopts above-mentioned scheme does: the method can accurately judge the traffic condition of the current road, give results such as flow, vehicle track tracking, traffic jam, traffic accident, illegal parking, illegal driving and the like, and is favorable for traffic dispersion and supervision. The method can automatically process the currently acquired road traffic picture, judges the current traffic road condition, has better applicability and robustness, and provides reliable judgment basis for road traffic supervision. The method can automatically abstract the image characteristics of the input convolutional neural network layer by layer, reduce the workload of artificially marking the characteristics and save the labor cost and the time cost; the target positioning of the vehicle can be quickly realized, and the type of the vehicle (car, truck, bus, etc.) can be identified; the method has good adaptability to different environments, illumination intensity and weather conditions; the customization requirement is flexible (the event identification parameters such as a detection boundary line, a vehicle speed threshold value and the like can be adjusted in real time); the problem of shielding can be handled to a certain extent in the crowded environment of large-flow vehicles, and higher detection accuracy is guaranteed.
Detailed Description
The technical solution of the present invention will be clearly and completely described below with reference to the specific embodiments of the present invention. The described embodiments are only some, not all embodiments of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1
The method utilizes data collected by a highway camera to judge, and comprises the following steps:
vehicle detection: according to video data of a sampling area, a convolutional neural network for vehicle type detection is designed and constructed by optimizing vehicle type detection based on yolov3, the detection method is high in running speed block and detection precision, real-time and efficient, and a detected vehicle is designated as a target.
Vehicle trajectory tracking: KCF (Kernelized correlation filters) is adopted for target tracking, and is called kernel correlation filtering completely, so that the method is a differential tracking method. In the method, a target detector is generally trained in the tracking process, the target detector is used for detecting whether the predicted position of the next frame is a target, and then a new detection result is used for updating a training set so as to update the target detector. While the target detector is trained, the target area is generally selected as a positive sample, and the area around the target is a negative sample, although the area closer to the target is more likely to be a positive sample. We define an entity class that instantiates an object for each detected object. An object instantiates a KCF tracker that is initialized. The KCF tracker accepts a frame and the coordinate position of the target. The KCF tracker is able to calculate where the target is located in the current frame by loading the latest frame through the update () function.
And (3) calculating the traffic flow: and dynamically defining a section in the sampling area, when a vehicle target is monitored to completely pass through the section, the section is regarded as the traffic flow accumulation, the data is reported once in unit time, and the background counts the traffic flow in unit time or in a plurality of unit times.
Example 2
Congestion determination was performed on the basis of example 1:
firstly, in a sampling area, judging whether a certain vehicle target is a congestion unit, namely, tracking the stay time T of a target vehicle C in the sampling areaduring≥TthresholdThen the target is considered to be a congestion unit, where TthresholdThe value is a preset adjustable value, the unit is second, and the value range is more than or equal to 1; residence time Tduring=Tcur-Cinit_time(wherein T iscurFor the current video frame time, Cinit_timeInitial time for tracking target vehicle C);
for example: setting TthresholdIs 2 seconds, tracking the initial time C of the target vehicle Cinit_time11:00:00, current video frame time TcurAt 11:00:05, the residence time Tduring= Tcur-Cinit_time=5 seconds, at which time the dwell time T isduring≥TthresholdThen, the target vehicle C is determined to be a congestion unit.
Secondly, detecting a congestion area if N exists in the same lane in the congestion detection areathreshold_perloadIndividual congestion units, or total presence of N in the areathreshold_totalThe congestion unit judges that the area is a congestion area; wherein N isthreshold_perloadAnd Nthreshold_totalThe preset adjustable values are all larger than or equal to 1;
for example: setting Nthreshold_perloadA value of 50, Nthreshold_totalThe value is 150, and when the number of congestion units in the same lane in the detection area reaches 50 or the total number of congestion units in the detection area reaches 150, the area is judged to be a congestion area.
And (3) multi-stage congestion judgment: detecting area congestion or the presence of N if the current congestion isthresholdAnd entering the next congestion detection area for congestion judgment according to the congestion target, wherein the set of congestion areas needing to be reported is P = { area =iI < n }; wherein, areaiIndicating an ith congestion detection area, numbering the congestion detection areas from front to back according to the driving direction, and taking n as the number of the congestion area at the last detected current position; n is a radical ofthresholdThe value range is more than or equal to 1 for the preset adjustable value.
For example: setting Nthreshold50, when the congestion target of the detection area reaches 50, the area is judged as the first area1A congested area and area1And combining the joining with the P, then entering the next congestion detection area for congestion judgment until n congestion detection areas are detected, joining the detection areas meeting the conditions into a set P, and finally reporting the set P.
Example 3
The driving accident judgment is carried out on the basis of the embodiment 1 and the embodiment 2: in the sampling area, target vehicles meeting the congestion condition and target vehicles not meeting the congestion condition exist at the same time, or the number of the target vehicles is small and the congestion condition exists, and the current accident is stored and reported when the condition is judged to be met.
Example 4
And (3) carrying out illegal parking and driving judgment on the basis of the embodiment 1 and the embodiment 2: in the sampling area, dynamically observing the monitoring area at the cloud end, defining an area as a forbidden area and a forbidden area, and judging that the illegal conditions are met when a monitored target enters the area; when a target entering the area is monitored to reach a certain time, judging that the target meets the condition of illegal parking; both cases are reported.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.
Furthermore, it should be understood that although the present description refers to embodiments, not every embodiment may contain only a single embodiment, and such description is for clarity only, and those skilled in the art should integrate the description, and the embodiments may be combined as appropriate to form other embodiments understood by those skilled in the art.
Claims (5)
1. A method for identifying a traffic incident on a highway, characterized by: the method for judging by utilizing the data collected by the expressway camera comprises the following steps:
vehicle detection: according to the video data of the sampling area, performing type detection on the vehicle in the video based on yolov3, and specifying the vehicle as a target;
vehicle trajectory tracking: performing target tracking by adopting KCF full-symmetric kernel correlation filtering, and defining an entity class which instantiates an object for each detected object; initializing a KCF tracker when an object is instantiated, receiving the coordinate positions of a frame and a target by the KCF tracker, loading the latest frame through an update () function, and calculating the position of the target in the current frame by using the KCF tracker;
and (3) calculating the traffic flow: dynamically defining a section in the sampling area, and when a vehicle object is monitored to completely pass through the section, considering the total traffic flow;
and (3) congestion judgment: firstly, in a sampling area, judging whether a certain vehicle target is a congestion unit, namely, tracking the stay time T of a target vehicle C in the sampling areaduring≥TthresholdIf so, the target is considered as a congestion unit; residence time TduringEqual to the current video frame time TcurSubtracting the initial time C of tracking the target vehicle Cinit_time(ii) a Secondly, detecting a congestion area if N exists in the same lane in the congestion detection areathreshold_perloadIndividual congestion units, or total presence of N in the areathreshold_totalThe congestion unit judges that the area is a congestion area; thirdly, judging the congestion in multiple stages, and detecting the blockage or the storage of the area if the current congestion is detectedIn NthresholdAnd entering the next congestion detection area for congestion judgment according to the congestion target, wherein the set of congestion areas needing to be reported is P = { area =iI < n }; wherein, areaiIndicating an ith congestion detection area, numbering the congestion detection areas from front to back according to the driving direction, and taking n as the number of the congestion area at the last detected current position;
judging a driving accident: in the sampling area, target vehicles meeting the congestion condition and target vehicles not meeting the congestion condition exist at the same time, or the number of the target vehicles is small and the congestion condition exists, and the target vehicles are judged to meet the accident condition;
and (3) illegal parking and driving judgment: in the sampling area, dynamically observing the monitoring area at the cloud end, defining an area as a forbidden area and a forbidden area, and judging that the illegal conditions are met when a monitored target enters the area; when it is monitored that a target enters the area for a certain time, the condition of illegal parking is judged to be met.
2. The method of identifying a highway vehicle passing event according to claim 1, wherein: the convolutional neural network in the step S1 is based on yolov3 and optimized for vehicle type detection, and is designed and built for vehicle type detection.
3. The method of identifying a highway vehicle passing event according to claim 1, wherein: t in the step S4thresholdThe unit is a preset adjustable value in seconds, and the value range is more than or equal to 1.
4. The method of identifying a highway vehicle passing event according to claim 1, wherein: n in the step S4threshold_perloadAnd Nthreshold_totalThe value range is more than or equal to 1 for preset adjustable values.
5. The method of identifying a highway vehicle passing event according to claim 1, wherein: what is needed isN in the step S4thresholdThe value range is more than or equal to 1 for the preset adjustable value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010679157.7A CN112216119A (en) | 2020-07-15 | 2020-07-15 | Method for identifying traffic vehicle passing event on highway |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010679157.7A CN112216119A (en) | 2020-07-15 | 2020-07-15 | Method for identifying traffic vehicle passing event on highway |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112216119A true CN112216119A (en) | 2021-01-12 |
Family
ID=74058570
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010679157.7A Pending CN112216119A (en) | 2020-07-15 | 2020-07-15 | Method for identifying traffic vehicle passing event on highway |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112216119A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002190013A (en) * | 2000-12-21 | 2002-07-05 | Nec Corp | System and method for detecting congestion by image recognition |
| US8903636B1 (en) * | 2013-12-02 | 2014-12-02 | Abdualrahman Abdullah Mohammad Al Kandari | Accident detection system and method for accident detection |
| CN107742418A (en) * | 2017-09-29 | 2018-02-27 | 东南大学 | A kind of urban expressway traffic congestion status and stifled point position automatic identifying method |
| CN110287905A (en) * | 2019-06-27 | 2019-09-27 | 浙江工业大学 | A kind of traffic congestion region real-time detection method based on deep learning |
| CN110472496A (en) * | 2019-07-08 | 2019-11-19 | 长安大学 | A kind of traffic video intelligent analysis method based on object detecting and tracking |
-
2020
- 2020-07-15 CN CN202010679157.7A patent/CN112216119A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002190013A (en) * | 2000-12-21 | 2002-07-05 | Nec Corp | System and method for detecting congestion by image recognition |
| US8903636B1 (en) * | 2013-12-02 | 2014-12-02 | Abdualrahman Abdullah Mohammad Al Kandari | Accident detection system and method for accident detection |
| CN107742418A (en) * | 2017-09-29 | 2018-02-27 | 东南大学 | A kind of urban expressway traffic congestion status and stifled point position automatic identifying method |
| CN110287905A (en) * | 2019-06-27 | 2019-09-27 | 浙江工业大学 | A kind of traffic congestion region real-time detection method based on deep learning |
| CN110472496A (en) * | 2019-07-08 | 2019-11-19 | 长安大学 | A kind of traffic video intelligent analysis method based on object detecting and tracking |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR101095528B1 (en) | An outomatic sensing system for traffic accident and method thereof | |
| CN108550262B (en) | Urban traffic sensing system based on millimeter wave radar | |
| CN110287905B (en) | Deep learning-based real-time traffic jam area detection method | |
| Hourdos et al. | Real-time detection of crash-prone conditions at freeway high-crash locations | |
| CN104751642B (en) | A kind of advanced road real-time predictor method of traffic flow operation risk | |
| CN110549941A (en) | pedestrian collision graded early warning method based on real-time information | |
| CN110077398B (en) | Risk handling method for intelligent driving | |
| CN110796862B (en) | Highway traffic condition detection system and method based on artificial intelligence | |
| CN112509315B (en) | Traffic accident detection method based on video analysis | |
| CN111613070B (en) | Traffic signal lamp control method, traffic signal lamp control device, electronic equipment and computer storage medium | |
| CN104537841A (en) | Unlicensed vehicle violation detection method and detection system thereof | |
| US20240046787A1 (en) | Method And System For Traffic Clearance At Signalized Intersections Based On Lidar And Trajectory Prediction | |
| CN105513371A (en) | Expressway illegal parking detection method based on kernel density estimation | |
| CN113851017A (en) | Pedestrian and vehicle identification and early warning multifunctional system based on road side RSU | |
| CN111754786A (en) | System for identifying traffic vehicle passing events on highway | |
| TWI774034B (en) | Driving warning method, system and equipment based on internet of vehicle | |
| CN111899514A (en) | Artificial intelligence's detection system that blocks up | |
| KR101133222B1 (en) | System and Method for monitoring traffic and accident based on image processing | |
| CN117079465A (en) | Safety early warning method and system for driving vehicle on expressway | |
| CN112216119A (en) | Method for identifying traffic vehicle passing event on highway | |
| Kamijo et al. | Development and evaluation of real-time video surveillance system on highway based on semantic hierarchy and decision surface | |
| CN115294797A (en) | Intelligent safety control method for highway level crossing | |
| CN108806260A (en) | A kind of multiple target flow statistical method based on radar | |
| JP7276276B2 (en) | Dangerous driving detection device, dangerous driving detection system, and dangerous driving detection program | |
| Morris et al. | Real-time video based highway traffic measurement and performance monitoring |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210112 |