CN116564083A - Expressway traffic jam detection method based on improved CrowdDet algorithm - Google Patents

Expressway traffic jam detection method based on improved CrowdDet algorithm Download PDF

Info

Publication number
CN116564083A
CN116564083A CN202310508001.6A CN202310508001A CN116564083A CN 116564083 A CN116564083 A CN 116564083A CN 202310508001 A CN202310508001 A CN 202310508001A CN 116564083 A CN116564083 A CN 116564083A
Authority
CN
China
Prior art keywords
traffic
vehicle
congestion
speed
algorithm
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
Application number
CN202310508001.6A
Other languages
Chinese (zh)
Inventor
陈玉婷
王池社
陈迪逢
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Anhui University of Science and Technology
Original Assignee
Anhui University of Science and Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Anhui University of Science and Technology filed Critical Anhui University of Science and Technology
Priority to CN202310508001.6A priority Critical patent/CN116564083A/en
Publication of CN116564083A publication Critical patent/CN116564083A/en
Pending legal-status Critical Current

Links

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
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0108—Measuring and analyzing of parameters relative to traffic conditions based on the source of data
    • G08G1/0116—Measuring and analyzing of parameters relative to traffic conditions based on the source of data from roadside infrastructure, e.g. beacons
    • 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
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/52—Scale-space analysis, e.g. wavelet analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
    • G06V10/806—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level of extracted features
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
    • G06V20/54—Surveillance or monitoring of activities, e.g. for recognising suspicious objects of traffic, e.g. cars on the road, trains or boats
    • 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
    • G08G1/0129—Traffic data processing for creating historical data or processing based on historical data
    • 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
    • G08G1/0145—Measuring and analyzing of parameters relative to traffic conditions for specific applications for active traffic flow control
    • 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
    • G08G1/048—Detecting movement of traffic to be counted or controlled with provision for compensation of environmental or other condition, e.g. snow, vehicle stopped at detector
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • Software Systems (AREA)
  • Medical Informatics (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Databases & Information Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Molecular Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Traffic Control Systems (AREA)

Abstract

The invention belongs to the technical field of computer vision and intelligent traffic, and discloses a method for detecting traffic jam of a highway based on an improved CrowdDet algorithm. The invention uses the target detection knowledge to detect the vehicles of the expressway under the congestion scene, and uses the detected vehicle traffic parameters to evaluate the traffic state. The method comprises the following steps: (1) Installing a camera on the expressway, collecting a monitoring video on the expressway through the camera, and performing frame extraction, screening, labeling and other works on the video to obtain images for training and testing; (2) Detecting vehicles in the traffic jam scene by using an improved CrowdDet target detection algorithm; (3) Based on the detected vehicle traffic parameters, the vehicle running speed is divided by combining a highway service Level (LOS) and a traffic classification speed method, and then the traffic state is evaluated. Although there are many scenarios in which highway congestion is caused, and the duration of congestion is also different, the proposed method is applicable to most scenarios.

Description

Expressway traffic jam detection method based on improved CrowdDet algorithm
Technical Field
The invention relates to the field of computer vision and intelligent traffic, in particular to a method for detecting traffic jam of a highway based on an improved CrowdDet algorithm.
Background
Transportation is an indispensable part of the modern development of cities, and is closely related to us in many important aspects. The expressway is used as a backbone frame of a traffic system and plays a key role in the process of urbanization. In recent years, the construction of expressways in China has achieved great achievements, and the scale of expressways is continuously enlarged, but the problem of increasingly outstanding traffic jam still cannot be solved. The occurrence of traffic jam can reduce the service level of a road network, and the reduction of LOS can bring direct or indirect cost to society, such as problems of low traffic running efficiency, environmental pollution, economic LOSs, road safety and the like. If the congested road section is found in time, the traffic operation management department can quickly take solving measures to dredge the congested vehicles, solve the traffic congestion event in a certain range and avoid the occurrence of traffic congestion with larger area by chain reaction.
At present, important research results are obtained in the aspect of traffic jam distinguishing research. Traffic jam discrimination techniques can be classified into three categories according to the difference of extracted traffic parameters: sensor-based traffic congestion determination techniques, vehicle-mounted ad hoc network (VANET) -based traffic congestion determination techniques, and machine vision-based traffic congestion determination techniques. Most sensor-dependent discrimination methods typically require a fixed device, which calculates the average speed over a specific time and then compares it to a predefined threshold, which is time consuming and laborious. The vehicle-mounted self-organizing network forms a self-organizing, structurally open inter-vehicle communication network providing a centerless, data transmission capability supporting multi-hop forwarding for collecting and aggregating real-time speed and location information associated with individual vehicles. But VANET is very expensive to deploy, vehicles must be equipped with on-board units and roadside units must be installed along the road.
Machine vision based congestion discrimination techniques are an attractive and cost effective solution. The camera is cheap, easy to maintain and can provide high-quality video sequences, and with the wide popularization of highway cameras and the continuous development of computer vision technology, a foundation is laid for judging traffic jam by using an image detection technology. At present, many image-based traffic jam judgments are oriented to urban roads, and due to the sealing property of expressways, the traffic jam becomes more and more serious, and a monitoring camera on the expressways generally has a larger visual angle range, and the generated image has a larger background area, so that whether the expressway jam judgments can be directly judged by using an urban road-based monitoring image method or not is also verified.
Disclosure of Invention
Aiming at the problems, the invention provides a highway traffic jam detection method based on an improved CrowdDet algorithm, which concretely adopts the following technical scheme:
1. the method comprises the steps of installing a camera on a highway, collecting a monitoring video on the highway through the camera, and performing frame extraction, screening, labeling and other works on the video to obtain images for training and testing.
1.1 Mounting a monitoring camera on the expressway, and collecting a road monitoring video;
1.2 The collected monitoring video is subjected to frame extraction, screening, labeling and other works, and in order to prevent roadside interference information from affecting the detection effect, the original video image is cut and the region of interest is manually divided.
2. And detecting the vehicle in the traffic jam scene by using an improved CrowdDet target detection algorithm.
2.1 We choose the CrowdDet algorithm as the baseline detector, which was originally designed to solve the problem of dense pedestrian detection, but the occlusion of vehicles under traffic congestion is similar to the problem of dense pedestrian detection, so we use CrowdDet as the baseline detector;
2.2 Because the expressway camera generally has a larger visual angle range, a long-distance small target vehicle exists, and because the small target vehicle has a small coverage area in a single frame of picture, the image is blurred, and the carried characteristic information is relatively less. The fifth layer of the feature extraction backbone network ResNet-50 is replaced by Involution to overcome the problem of remote vehicle detection;
2.3 In addition, vehicles on the expressway are different in size, and when traffic jam occurs, the vehicles are relatively dense and are easy to cause shielding problems, so that the model needs to capture information in different multi-scale contexts to accurately obtain the number of the vehicles, the density of the vehicles and other related traffic parameters. In order to solve the problem of multi-scale detection of vehicles in a congestion state, the feature extraction performance of objects with different scales is improved, biFPN is introduced to learn the importance of different input features, and multi-scale feature fusion from top to bottom and from bottom to top is repeatedly applied at the same time, and the improved algorithm is named IBCNet.
3. Based on the detected traffic parameters of the vehicles, the traffic state is evaluated by combining the high-speed highway LOS and the traffic classification speed method of China.
3.1 According to traffic theory, dividing traffic flow into three phase states, namely free flow, synchronous flow and congestion state. Before the free flow state of the road section changes to the congestion state, a larger flow is reached, which is called traffic capacity, and when congestion occurs, rapid reduction is generated, and the reduction is called traffic capacity reduction. The decrease in traffic capacity can be measured by a decrease in vehicle speed;
3.2 Grading and prescribing the expressway service level according to the Highway engineering technical standard in China, and defining the congestion degree of expressway traffic by using a speed index;
3.3 Automatically identifying by using a multi-target tracking technology, then calculating the travelling distance of an object in two continuous frames, dividing the travelling distance by the time between the two frames to finish speed estimation, drawing a vehicle flow speed time sequence chart by using the obtained vehicle speed result, dividing the whole process of vehicle congestion into five traffic states according to the speed fluctuation amplitude, and judging the time when the traffic congestion is established and dissipated according to the speed change intensity and the speed threshold value;
3.4 State a is steady state and the vehicle flow is free to run. State B is a congestion formation state, which is most pronounced as the speed drops sharply in a short period of time. The state C1 is a serious congestion state, the state C2 is a light congestion state, and if the speed of the vehicle is lower than a threshold value, the traffic flow is considered to enter the congestion state, and the duration of the congestion state is different according to the actual traffic flow. State D is a congestion dispersion state manifested as a rapid rise in speed;
3.5 State C1, 23.9km/h lower than the threshold of C1, severe congestion, stop-and-go phenomenon, extremely small vehicle distance and extremely poor driving comfort. The speed time sequence diagram accords with a state C2, is slightly congested below a C2 threshold value of 38.5km/h, has a speed which is obviously lower than a speed limiting standard, has very small vehicle distance, has poor driving comfort, has great influence on the mobility of the vehicle, and can still move forward.
The invention has the advantages that: according to the invention, the problems of small vehicle size, vehicle shielding and the like in a highway monitoring scene are considered, so that the vehicle omission ratio is high, and the traffic jam detection accuracy is low. Aiming at the problem, an improved CrowdDet algorithm (IBCDet) is provided by introducing an inventory network and a BiFPN module, and the algorithm improves the vehicle detection accuracy in a highway congestion scene through remote information interaction and multi-scale feature fusion, so that higher traffic congestion detection accuracy is realized. On the basis, the tracking algorithm of the IBCDet algorithm is used for calculating the running speed of the vehicle, and the congestion level is divided according to the LOS criterion of the expressway of China, so that the traffic congestion detection is realized. Compared with a common target detection algorithm, the method has the advantage that the best effect is shown in the aspects of vehicle detection accuracy and traffic jam detection accuracy.
Drawings
FIG. 1 traffic congestion detection architecture
FIG. 2 overall network architecture diagram
FIG. 3 is a schematic view of the resolution structure
FIG. 4BiFPN module schematic
Fig. 5IBCNet network structure diagram
FIG. 6 is a flow velocity timing diagram
Detailed Description
The invention is further described below with reference to specific embodiments and illustrations in order to make the technical means, the creation features, the achievement of the purpose and the effect achieved by the invention easy to understand.
The following description of the embodiments of the present invention will be made more fully hereinafter with reference to the accompanying drawings, in which embodiments of the invention are shown, and in which embodiments of the invention are shown, by way of illustration only, and not all embodiments in which the invention may be practiced. All other embodiments, which are obtained by a person of ordinary skill in the art without any inventive effort, are within the scope of the present invention based on the embodiments of the present invention.
Aiming at the problems, the invention provides a highway traffic jam detection method based on an improved CrowdDet algorithm, which concretely adopts the following technical scheme:
1. the method comprises the steps of installing a camera on a highway, collecting a monitoring video on the highway through the camera, and performing frame extraction, screening, labeling and other works on the video to obtain images for training and testing.
1.1 Mounting a monitoring camera on the expressway, and collecting a road monitoring video;
1.2 The collected monitoring video is subjected to frame extraction, screening, labeling and other works, in order to prevent roadside interference information from affecting the detection effect, the original video image is cut and the region of interest is manually divided, and the overall architecture of specific traffic jam detection is shown in figure 1.
2. Detecting vehicles in the traffic jam scene by using an improved CrowdDet target detection algorithm;
2.1 We choose the CrowdDet algorithm as the baseline detector, which was originally designed to solve the problem of dense pedestrian detection, but the occlusion of the vehicle under traffic congestion is similar to the problem of dense pedestrian detection, so we use CrowdDet as the baseline detector, as shown in FIG. 2;
2.2 Because the expressway camera generally has a larger visual angle range, a long-distance small target vehicle exists, and because the small target vehicle has a small coverage area in a single frame of picture, the image is blurred, and the carried characteristic information is relatively less. The fifth layer of the feature extraction backbone network ResNet-50 is replaced by Involution to overcome the problem of remote vehicle detection, and the Involution structure is shown in FIG. 3;
2.3 In addition, vehicles on the expressway are different in size, and when traffic jam occurs, the vehicles are relatively dense and are easy to cause shielding problems, so that the model needs to capture information in different multi-scale contexts to accurately obtain the number of the vehicles, the density of the vehicles and other related traffic parameters. In order to solve the problem of multi-scale detection of vehicles in a congestion state, the feature extraction performance of objects with different scales is improved, biFPN is introduced to learn the importance of different input features, and multi-scale feature fusion from top to bottom and from bottom to top is repeatedly applied at the same time, as shown in FIG. 4;
2.4 By introducing an inventory network and a BiFPN module into a CrowdDet algorithm, the vehicle detection accuracy in the expressway congestion scene can be improved, and the improved network is called IBCNet, as shown in fig. 5.
3. Adding a tracking technology to the improved algorithm, and measuring the running speed of the vehicle;
3.1 Calculating the travel distance of the tracked vehicle in two consecutive frames from the vehicle trajectory, divided by the two frame time difference to complete the speed estimation. The specific formula is shown as (1):
wherein the vehicle is at time t 1 In position (x) 1 ,y 1 ) At t 2 In position (x) 2 ,y 2 )；
3.2 In order to more accurately reflect the running speed of the vehicle in the video, we will calculate the average value of all the vehicle speeds at a specific moment, and the specific formula is shown in (2):
wherein N is the number of vehicles in the video detected by IBCDet algorithm, V i Representing the speed at which the ith vehicle is traveling. The higher the vehicle detection rate, the higher the V to a certain extent, reflected by the formula s The higher the accuracy of (2).
4. Performing traffic state assessment by using time sequence information based on the detected vehicle traffic parameters;
4.1 According to traffic theory, dividing traffic flow into three phase states, namely free flow, synchronous flow and congestion state. Before the free flow state of the road section changes to the congestion state, a larger flow is reached, which is called traffic capacity, and when congestion occurs, rapid reduction is generated, and the reduction is called traffic capacity reduction. The decrease in traffic capacity can be measured by a decrease in vehicle speed;
4.2 Grading and prescribing the highway LOS according to the highway engineering technical standard in China, and defining the congestion degree of the highway traffic according to the speed index;
4.3 Automatically identifying by using a multi-target tracking technology, then calculating the travelling distance of an object in two continuous frames, dividing the travelling distance by the time between the two frames to finish speed estimation, drawing a vehicle flow speed time sequence chart by using the obtained vehicle speed result, dividing the whole process of vehicle congestion into five traffic states according to the speed fluctuation amplitude, and judging the time when the traffic congestion is established and dissipated according to the speed change intensity and the speed threshold value;
4.4 As shown in fig. 6, wherein state a is steady state, the vehicle flow is free running. State B is a congestion formation state, which is most pronounced as the speed drops sharply in a short period of time. The state C1 is a serious congestion state, the state C2 is a light congestion state, and if the speed of the vehicle is lower than a threshold value, the traffic flow is considered to enter the congestion state, and the duration of the congestion state is different according to the actual traffic flow. State D is a congestion dispersion state manifested as a rapid rise in speed;
4.5 State C1, 23.9km/h lower than the threshold of C1, severe congestion, stop-and-go phenomenon, extremely small vehicle distance and extremely poor driving comfort. The speed time sequence diagram accords with a state C2, is slightly congested below a C2 threshold value of 38.5km/h, has a speed which is obviously lower than a speed limiting standard, has very small vehicle distance, has poor driving comfort, has great influence on the mobility of the vehicle, and can still move forward.
The foregoing is a preferred embodiment of the present application and is not intended to limit the present application, and various modifications and variations may be made to the present application by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims (5)

1. A method for highway traffic congestion detection based on an improved CrowdDet algorithm, comprising the steps of:
(1) Installing a camera on the expressway, collecting a monitoring video on the expressway through the camera, and performing frame extraction, screening, labeling and other works on the video to obtain images for training and testing;
(2) Detecting vehicles in the traffic jam scene by using an improved CrowdDet target detection algorithm;
(3) Calculating the average speed of the vehicle based on the detected number of vehicles and a tracking algorithm;
(4) And detecting traffic state by using the time sequence information.
2. The method for highway traffic congestion detection based on the modified crolddet algorithm of claim 1, wherein: the specific method of the step (1) is as follows:
2.1 Mounting a monitoring camera on the expressway, and collecting a road monitoring video;
2.2 The collected monitoring video is subjected to frame extraction, screening, labeling and other works, and in order to prevent roadside interference information from affecting the detection effect, the original video image is cut and the region of interest is manually divided.
3. The method for highway traffic congestion detection based on the modified crolddet algorithm of claim 1, wherein: the specific method of the step (2) is as follows:
3.1 We choose the CrowdDet algorithm as the baseline detector, which was originally designed to solve the problem of dense pedestrian detection, but the occlusion of vehicles under traffic congestion is similar to the problem of dense pedestrian detection, so we use CrowdDet as the baseline detector;
3.2 Because the expressway camera generally has a larger visual angle range, a long-distance small target vehicle exists, and because the small target vehicle has a small coverage area in a single frame of picture, the image is blurred, and the carried characteristic information is relatively less. The fifth layer of the feature extraction backbone network ResNet-50 is replaced by Involution to overcome the problem of remote vehicle detection;
3.3 In addition, vehicles on the expressway are different in size, and when traffic jam occurs, the vehicles are relatively dense and are easy to cause shielding problems, so that the model needs to capture information in different multi-scale contexts to accurately obtain the number of the vehicles, the density of the vehicles and other related traffic parameters. In order to solve the problem of multi-scale detection of vehicles in a congestion state, the feature extraction performance of objects with different scales is improved, biFPN is introduced to learn the importance of different input features, multi-scale feature fusion from top to bottom and from bottom to top is repeatedly applied, and an improved target detection algorithm is named IBCNet.
4. The method for highway traffic congestion detection based on the modified crolddet algorithm of claim 1, wherein: the specific method of the step (3) is as follows:
4.1 Calculating the travel distance of the tracked vehicle in two consecutive frames from the vehicle trajectory, divided by the two frame time difference to complete the speed estimation. The specific formula is shown as (1):
wherein the vehicle is at time t 1 In position (x) 1 ,y 1 ) At t 2 In positionPut (x) 2 ,y 2 )；
4.2 In order to more accurately reflect the running speed of the vehicle in the video, we will calculate the average value of all the vehicle speeds at a specific moment, and the specific formula is shown in (2):
wherein N is the number of vehicles in the video detected by IBCDet algorithm, V i Representing the speed at which the ith vehicle is traveling. The higher the vehicle detection rate, the higher the V to a certain extent, reflected by the formula s The higher the accuracy of (2).
5. The method for highway traffic congestion detection based on the modified crolddet algorithm of claim 1, wherein: the specific method of the step (4) is as follows:
5.1 According to the expressway service level grading rule in the Highway engineering technical standard of China, defining the congestion degree of expressway traffic by a speed index;
5.2 Dividing the whole process of vehicle congestion into five traffic states according to the speed fluctuation range, and judging the time when the traffic congestion is established and dissipated according to the speed change intensity and the speed threshold value. The speed distribution diagram and the speed confidence interval of different road segments are analyzed to determine the congestion threshold, and generally 40% of the split speed is taken as the judgment basis of the light congestion speed threshold, and 25% of the split speed is taken as the judgment basis of the heavy congestion speed threshold.
CN202310508001.6A 2023-05-06 2023-05-06 Expressway traffic jam detection method based on improved CrowdDet algorithm Pending CN116564083A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310508001.6A CN116564083A (en) 2023-05-06 2023-05-06 Expressway traffic jam detection method based on improved CrowdDet algorithm

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310508001.6A CN116564083A (en) 2023-05-06 2023-05-06 Expressway traffic jam detection method based on improved CrowdDet algorithm

Publications (1)

Publication Number Publication Date
CN116564083A true CN116564083A (en) 2023-08-08

Family

ID=87495895

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310508001.6A Pending CN116564083A (en) 2023-05-06 2023-05-06 Expressway traffic jam detection method based on improved CrowdDet algorithm

Country Status (1)

Country Link
CN (1) CN116564083A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117315934A (en) * 2023-09-25 2023-12-29 阜阳交通能源投资有限公司 Expressway traffic flow real-time monitoring and congestion prediction system based on unmanned aerial vehicle

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117315934A (en) * 2023-09-25 2023-12-29 阜阳交通能源投资有限公司 Expressway traffic flow real-time monitoring and congestion prediction system based on unmanned aerial vehicle

Similar Documents

Publication Publication Date Title
CN102819952B (en) Method for detecting illegal lane change of vehicle based on video detection technique
CN101510356B (en) Video detection system and data processing device thereof, video detection method
CN100573618C (en) A kind of traffic intersection four-phase vehicle flow detection method
CN103914688A (en) Urban road obstacle recognition system
Fernández-Caballero et al. Road-traffic monitoring by knowledge-driven static and dynamic image analysis
Song et al. Vehicle behavior analysis using target motion trajectories
CN105608431A (en) Vehicle number and traffic flow speed based highway congestion detection method
CN106355884B (en) A kind of vehicle on highway guidance system and method based on vehicle classification
CN103366571A (en) Intelligent method for detecting traffic accident at night
CN102496281B (en) Vehicle red-light violation detection method based on combination of tracking and virtual loop
CN105336169A (en) Method and system for judging traffic jams based on videos
CN103617410A (en) Highway tunnel parking detection method based on video detection technology
KR20100119476A (en) An outomatic sensing system for traffic accident and method thereof
CN105513349A (en) Double-perspective learning-based mountainous area highway vehicle event detection method
CN106327880A (en) Vehicle speed identification method and system based on monitored video
EP2813973B1 (en) Method and system for processing video image
CN111597905A (en) Highway tunnel parking detection method based on video technology
CN116564083A (en) Expressway traffic jam detection method based on improved CrowdDet algorithm
CN102789686B (en) Road traffic flow detecting method based on road surface brightness composite mode recognition
CN114781479A (en) Traffic incident detection method and device
KR102505867B1 (en) Multi-vehicle data analysis and traffic signal control method and system using LiDar sensor
CN110796862A (en) Highway traffic condition detection system and method based on artificial intelligence
CN103310206A (en) Moped detection method based on multi-feature and multi-frame information fusion
CN113313012B (en) Dangerous driving behavior identification method based on convolution generation countermeasure network
CN113392817A (en) Vehicle density estimation method and device based on multi-row convolutional neural network

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