CN117727182A

CN117727182A - Method, device, equipment and medium for detecting traffic flow of intersection

Info

Publication number: CN117727182A
Application number: CN202311802873.XA
Authority: CN
Inventors: 任晓意
Original assignee: Wuxi Mingda Transportation Technology Consulting Co ltd
Current assignee: Wuxi Mingda Transportation Technology Consulting Co ltd
Priority date: 2023-12-25
Filing date: 2023-12-25
Publication date: 2024-03-19

Abstract

The application relates to the technical field of traffic flow detection, in particular to a traffic flow detection method, device, equipment and medium for an intersection, wherein the method comprises the following steps: the lane classification is carried out based on lane characteristics, the lane category corresponding to each lane in the intersection is determined, preprocessing is carried out based on the video monitoring image, a vehicle flow detection area is selected in the video monitoring image, and the vehicle flow detection area is used for intensively processing and analyzing the part really related to the vehicle flow, so that the accuracy of vehicle flow detection is improved. And detecting the traffic flow based on the lane type corresponding to each lane in the target branch and the traffic flow detection area in the video monitoring image, determining the branch traffic flow of the target branch, and selecting a proper traffic flow detection method aiming at different lane types when detecting the traffic flow so as to improve the accuracy of the branch traffic flow and the traffic flow detection efficiency. And finally, synthesizing the branch road traffic of each target branch road in the intersection to obtain the total traffic flow of the intersection.

Description

Method, device, equipment and medium for detecting traffic flow of intersection

Technical Field

The present disclosure relates to the field of traffic flow detection technologies, and in particular, to a method, an apparatus, a device, and a medium for detecting traffic flow at an intersection.

Background

When urban traffic network traffic jams, it is extremely important to analyze traffic flow at an intersection, and further, a corresponding traffic signal control scheme is provided based on the traffic flow at the intersection, and the traffic flow is manually calculated through a traffic video monitoring system by observing through a signal control engineer in a traditional mode.

Therefore, in most of today, video detection technology is used to automatically count the traffic flow at an intersection, and when the video detection technology is used to automatically detect the traffic flow, the traffic flow is used as a static parameter measurement, i.e. the number of vehicles passing through the intersection per unit time. However, such a detection method cannot distinguish the number of vehicles in different driving directions in the same lane, for example, when the lane passes through the left turn and the vehicle is executed at the same time, the detection method in the related art cannot accurately distinguish the vehicles in the two different driving directions, so that the accuracy of the vehicle flow detection is not high.

Therefore, how to provide a traffic flow detection method for an intersection with high accuracy is a problem to be solved by those skilled in the art.

Disclosure of Invention

The purpose of the application is to provide a method, a device, equipment and a medium for detecting traffic flow of an intersection, which are used for solving at least one technical problem.

The above object of the present application is achieved by the following technical solutions:

in a first aspect, the present application provides a method for detecting traffic flow at an intersection, which adopts the following technical scheme:

a traffic flow detection method for an intersection, comprising:

obtaining lane characteristics of a target branch in an intersection, dividing lanes based on the lane characteristics, and determining lane categories corresponding to each lane in the intersection, wherein the lane categories comprise: the target branch road is any branch road in the intersection;

acquiring a video monitoring image of the target branch, preprocessing the video monitoring image, and selecting a traffic flow detection area from a middle frame of the video monitoring image;

detecting the traffic flow based on the lane category corresponding to each lane in the target branch and the traffic flow detection area in the video monitoring image, and determining the branch traffic flow of the target branch;

And synthesizing the branch traffic flow of each target branch in the intersection to obtain the total traffic flow of the intersection.

By adopting the technical scheme, lane classification is carried out based on lane characteristics, lane categories corresponding to each lane in an intersection are determined, meanwhile, preprocessing is carried out based on video monitoring images, a traffic flow detection area is selected in the video monitoring images in a frame mode, and the frame-selected traffic flow detection area is used for intensively processing and analyzing parts really related to traffic flow, so that false detection and missed detection are reduced, accuracy of traffic flow detection is improved, and computing resources and storage resources of electronic equipment are saved. And then, detecting the traffic flow based on the lane type corresponding to each lane in the target branch and the traffic flow detection area in the video monitoring image, determining the branch traffic flow of the target branch, and selecting an adaptive traffic flow detection method aiming at different lane types when detecting the traffic flow so as to improve the accuracy of the branch traffic flow and the traffic flow detection efficiency. And finally, synthesizing the branch road traffic of each target branch road in the intersection to obtain the total traffic flow of the intersection. In order to improve the accuracy of the traffic flow of the intersection, traffic flow detection is performed on each target branch in the intersection to obtain the branch traffic flow of the target branch, and then the branch traffic flows are summarized to obtain the total traffic flow of the intersection.

The present application may be further configured in a preferred example to: the determining the branch traffic flow of the target branch based on the traffic flow detection area in the video monitoring image and the traffic flow class corresponding to each lane in the target branch includes:

vehicle identification is carried out based on the vehicle flow detection area in the video monitoring image, and a vehicle identification frame corresponding to each vehicle is marked;

based on the traffic flow detection area in the video monitoring image, carrying out vehicle tracking on each vehicle identification frame corresponding to the multi-directional lane, and determining a first direct traffic flow and a steering traffic flow in the multi-directional lane;

based on the traffic flow detection area in the video monitoring image, carrying out vehicle tracking on each vehicle identification frame corresponding to the unidirectional straight-through lane, and determining a second straight-through traffic flow in the unidirectional straight-through lane;

determining a straight traffic flow of the target leg based on the sum of the first straight traffic flow and the second straight traffic flow, wherein the leg traffic flow comprises: the straight traffic flow and the steered traffic flow.

The present application may be further configured in a preferred example to: the step of tracking the vehicle for each vehicle identification frame corresponding to the multi-directional lane based on the traffic flow detection area in the video monitoring image, and determining the first direct traffic flow and the steering traffic flow in the multi-directional lane includes:

based on the traffic flow detection area in the video monitoring image, carrying out vehicle tracking on each vehicle identification frame corresponding to the multidirectional lane, and determining a rotation direction angle set corresponding to each vehicle identification frame;

and determining the first direct traffic flow and the steering traffic flow in the multidirectional lane based on the rotation direction angle set corresponding to each vehicle identification frame.

The present application may be further configured in a preferred example to: the method further includes, after determining the branch traffic flow of the target branch, performing traffic flow detection based on the lane category corresponding to each lane in the target branch and the traffic flow detection area in the video monitoring image:

predicting the traffic flow of the target branch by using a traffic flow prediction model to obtain the predicted traffic flow corresponding to the target branch;

Performing vehicle flow comparison analysis based on the predicted vehicle flow corresponding to the target branch and the branch vehicle flow, and determining a comparison analysis result;

and when the comparison and analysis result is normal, taking the branch traffic flow of the target branch as a training data set, and retraining the traffic flow prediction model by utilizing the training data set to obtain an updated traffic flow prediction model.

The present application may be further configured in a preferred example to: the method further comprises the steps of:

and when the comparison and analysis result is abnormal, detecting the traffic flow of the target branch again by using a positioning linkage mode to obtain the accurate branch traffic flow, wherein the positioning linkage mode is used for accurately detecting the traffic flow through a positioning system of the vehicle.

The present application may be further configured in a preferred example to: the method for collecting the lane characteristics of the target branch comprises the following steps:

when the lane characteristic acquisition period is reached, acquiring a road image of the target branch in a traffic valley period, and extracting characteristics based on the road image to obtain the current lane characteristic of the target branch;

And acquiring the stored lane characteristics, comparing the characteristics based on the current lane characteristics and the stored lane characteristics, and when the characteristics are different, updating the stored lane characteristics by utilizing the current lane characteristics and recording the updated lane characteristics as the lane characteristics of the target branch.

In a second aspect, the present application provides a traffic flow detection device for an intersection, which adopts the following technical scheme:

an intersection traffic flow detection device comprising:

the lane dividing module is used for acquiring lane characteristics of a target branch in the intersection, dividing lanes based on the lane characteristics, and determining lane categories corresponding to each lane in the intersection, wherein the lane categories comprise: the target branch road is any branch road in the intersection;

the detection area dividing module is used for acquiring a video monitoring image of the target branch, preprocessing the video monitoring image, and selecting a traffic flow detection area from the video monitoring image through a frame;

the traffic flow detection module is used for detecting traffic flow based on the lane category corresponding to each lane in the target branch and the traffic flow detection area in the video monitoring image, and determining the branch traffic flow of the target branch;

And the traffic flow summarizing module is used for synthesizing the branch traffic flow of each target branch in the intersection to obtain the total traffic flow of the intersection.

In a third aspect, the present application provides an electronic device, which adopts the following technical scheme:

at least one processor;

a memory;

at least one application program, wherein the at least one application program is stored in the memory and configured to be executed by the at least one processor, the at least one application program configured to: the traffic flow detection method of the intersection described above is executed.

In a fourth aspect, the present application provides a computer readable storage medium, which adopts the following technical scheme:

a computer-readable storage medium having stored thereon a computer program that, when executed in a computer, causes the computer to perform the intersection traffic flow detection method described above.

In summary, the present application includes at least one of the following beneficial technical effects:

1. the lane classification is carried out based on lane characteristics, the lane category corresponding to each lane in the intersection is determined, meanwhile, preprocessing is carried out based on the video monitoring image, the traffic flow detection area is selected in the video monitoring image in a frame mode, and the frame-selected traffic flow detection area is used for intensively processing and analyzing the part really related to the traffic flow, so that false detection and missed detection are reduced, the accuracy of traffic flow detection is improved, and the computing resources and the storage resources of electronic equipment are saved. And then, detecting the traffic flow based on the lane type corresponding to each lane in the target branch and the traffic flow detection area in the video monitoring image, determining the branch traffic flow of the target branch, and selecting an adaptive traffic flow detection method aiming at different lane types when detecting the traffic flow so as to improve the accuracy of the branch traffic flow and the traffic flow detection efficiency. And finally, synthesizing the branch road traffic of each target branch road in the intersection to obtain the total traffic flow of the intersection. In order to improve the accuracy of the traffic flow of the intersection, detecting the traffic flow of each target branch in the intersection to obtain the branch traffic flow of the target branch, and then summarizing the branch traffic flows to obtain the total traffic flow of the intersection;

2. And (5) carrying out vehicle identification based on the traffic flow detection area in the video monitoring image, and marking a vehicle identification frame corresponding to each vehicle. And then, based on the traffic flow detection area in the video monitoring image, carrying out vehicle tracking on each vehicle identification frame corresponding to the multi-directional lane, and determining the first direct traffic flow and the steering traffic flow in the multi-directional lane. And at the same time, based on the traffic flow detection area in the video monitoring image, carrying out vehicle tracking on each vehicle identification frame corresponding to the unidirectional straight-through lane, and determining the second straight-through traffic flow in the unidirectional straight-through lane. And finally, summing the first straight traffic flow and the second straight traffic flow to determine the straight traffic flow of the target branch. When the traffic flow is detected, the corresponding traffic flow detection method is selected according to different lane categories, so that the accuracy of the bypass traffic flow and the traffic flow detection efficiency are improved.

Drawings

FIG. 1 is a flow chart of a method for detecting traffic flow at an intersection according to one embodiment of the present application;

FIG. 2 is a schematic flow chart of determining a bypass traffic flow of a target bypass according to one embodiment of the present disclosure;

FIG. 3 is a schematic view of an apparatus for detecting traffic flow at an intersection according to one embodiment of the present application;

Fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

The present application is described in further detail below in conjunction with fig. 1-4.

The present embodiment is merely illustrative of the present application and is not intended to be limiting, and those skilled in the art, after having read the present specification, may make modifications to the present embodiment without creative contribution as required, but is protected by patent laws within the scope of the present application.

For the purposes of making the objects, technical solutions and advantages of the embodiments of the present application more clear, the technical solutions of the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is apparent that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art based on the embodiments herein without making any inventive effort, are intended to be within the scope of the present application.

In addition, the term "and/or" herein is merely an association relationship describing an association object, and means that three relationships may exist, for example, a and/or B may mean: a exists alone, A and B exist together, and B exists alone. In this context, unless otherwise specified, the term "/" generally indicates that the associated object is an "or" relationship.

Embodiments of the present application are described in further detail below with reference to the drawings attached hereto.

The embodiment of the application provides a traffic flow detection method of an intersection, which is executed by electronic equipment, wherein the electronic equipment can be a server or terminal equipment, and the server can be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server for providing cloud computing service. The terminal device may be a smart phone, a tablet computer, a notebook computer, a desktop computer, or the like, but is not limited thereto, and the terminal device and the server may be directly or indirectly connected through a wired or wireless communication manner, which is not limited herein, and as shown in fig. 1, the method includes steps S101, S102, S103, and S104, where:

step S101: obtaining lane characteristics of a target branch in an intersection, dividing lanes based on the lane characteristics, and determining lane categories corresponding to each lane in the intersection, wherein the lane categories comprise: the target branch road is any branch road in the intersection.

For the embodiment of the application, the intersection is formed by intersecting a plurality of branches, each branch is provided with vehicles to enter and exit, in order to improve the accuracy of the traffic flow of the intersection, the calculated amount of the subsequent traffic flow detection of the electronic equipment is reduced, the traffic flow detection is carried out on each target branch in the intersection, the branch traffic flow of the target branch is obtained, and then the branch traffic flows are summarized to obtain the total traffic flow of the intersection.

Specifically, lane features of a target branch in an intersection are obtained, where the lane features include, but are not limited to: lane line position, lane width, lane markings, wherein lane markings refer to markings on the target leg for indicating direction, function, restriction, etc. of the lane. Then, lane division is carried out based on lane marks in lane features of the target branch, lane categories corresponding to each lane in the intersection are determined, and the lane categories comprise: the vehicle is driven by the vehicle to travel along the straight line, and the vehicle is driven by the vehicle to travel along the straight line. For vehicles on unidirectional straight lanes, the direction of travel is determined; for vehicles on a multidirectional lane, the driving direction is uncertain, namely, whether the vehicle is straight or turns, so that the two types of lanes are preferentially divided, and a proper traffic flow detection method is selected for different lane types, so that the traffic flow on various lanes is accurately and efficiently detected.

Step S102: and acquiring a video monitoring image of the target branch, preprocessing based on the video monitoring image, and selecting a traffic flow detection area from the video monitoring image.

For the embodiment of the application, in order to improve the accuracy of traffic flow detection, an image acquisition device is arranged for each target branch, and the image acquisition device is used for acquiring panoramic video images passing through an intersection on the target branch. The area covered by the panoramic video image acquired by the image acquisition device is relatively large, if the panoramic video image is directly subjected to traffic flow detection, high computing resources are required to be consumed, and more irrelevant interference is introduced. The size and the position of the traffic flow detection area can be set by a user according to the own requirements, and the embodiment of the application is not limited any more. The frame-selected traffic flow detection area is used for intensively processing and analyzing the part really related to the traffic flow, removing the area irrelevant to the traffic flow detection, such as the part far away from the intersection.

Step S103: detecting the traffic flow based on the lane category corresponding to each lane in the target branch and the traffic flow detection area in the video monitoring image, and determining the branch traffic flow of the target branch;

for the embodiment of the application, the traveling direction is determined for the vehicle on the one-way straight lane; for vehicles on a multidirectional lane, the running direction is uncertain, namely, whether the vehicle runs straight or turns is uncertain, so that when the vehicle flow is detected, the adaptive vehicle flow detection method is selected for different lane types, and the accuracy of the bypass vehicle flow and the vehicle flow detection efficiency are improved.

Specifically, vehicle identification is performed based on a vehicle flow detection area in the video monitoring image, and a vehicle identification frame corresponding to each vehicle is marked. For a multi-directional lane, based on a traffic flow detection area in a video monitoring image, vehicle tracking is carried out on each vehicle identification frame corresponding to the multi-directional lane, and a first direct traffic flow and a steering traffic flow in the multi-directional lane are determined, wherein when the vehicles in the multi-directional lane are tracked, a vehicle track analysis method is adopted to respectively count traffic flows in different directions. For a unidirectional straight-through lane, based on a traffic flow detection area in a video monitoring image, vehicle tracking is carried out on each vehicle identification frame corresponding to the unidirectional straight-through lane, and a second straight-through traffic flow in the unidirectional straight-through lane is determined, wherein when the vehicles in the unidirectional straight-through lane are tracked, the second traffic flow is determined by counting the number of vehicles passing in a unit time. Finally, summing the first straight traffic flow and the second straight traffic flow to determine the straight traffic flow of the target branch, wherein the branch traffic flow comprises: the traffic flow is the total number of traffic flows on all unidirectional traffic lanes in the target branch.

Step S104: and synthesizing the branch road traffic flow of each target branch road in the intersection to obtain the total traffic flow of the intersection.

For the embodiment of the present application, there are various ways to integrate the traffic flow of each target branch, and the embodiment of the present application is not limited any more, and the user can set the traffic flow by himself. In one implementation manner, time interval division is performed based on time interval information of the branch traffic flows to obtain the branch time interval traffic flows arranged according to time sequence, then format unification is performed on the branch time interval traffic flows of the same time interval information in each target branch, and at least one traffic flow direction corresponding to the branch time interval traffic flows and the traffic flow corresponding to the traffic flow direction are determined. For any period, summarizing traffic flow directions corresponding to traffic flow directions based on the branch period traffic flow of all target branches to obtain total traffic flow of the intersection, wherein the total traffic flow comprises: and the total traffic flow of each corresponding time period in the plurality of time periods.

It can be seen that, in the embodiment of the application, lane division is performed based on lane features, lane types corresponding to each lane in an intersection are determined, meanwhile, preprocessing is performed based on video monitoring images, a traffic flow detection area is selected in a frame in the video monitoring images, and the frame-selected traffic flow detection area is used for intensively processing and analyzing parts really related to traffic flow, so that false detection and missed detection are reduced, accuracy of traffic flow detection is improved, and computing resources and storage resources of electronic equipment are saved. And then, detecting the traffic flow based on the lane type corresponding to each lane in the target branch and the traffic flow detection area in the video monitoring image, determining the branch traffic flow of the target branch, and selecting an adaptive traffic flow detection method aiming at different lane types when detecting the traffic flow so as to improve the accuracy of the branch traffic flow and the traffic flow detection efficiency. And finally, synthesizing the branch road traffic of each target branch road in the intersection to obtain the total traffic flow of the intersection. In order to improve the accuracy of the traffic flow of the intersection, traffic flow detection is performed on each target branch in the intersection to obtain the branch traffic flow of the target branch, and then the branch traffic flows are summarized to obtain the total traffic flow of the intersection.

Further, in order to improve the accuracy and the efficiency of detecting the traffic flow of the branch road, in this embodiment of the present application, as shown in fig. 2, the traffic flow detection is performed based on the lane category corresponding to each lane in the target branch road and the traffic flow detection area in the video monitoring image, and the determining the branch road traffic flow of the target branch road includes: step S1031-step S1034, wherein:

step S1031: and (5) carrying out vehicle identification based on the traffic flow detection area in the video monitoring image, and marking a vehicle identification frame corresponding to each vehicle.

For the embodiments of the present application, the vehicle identification is performed by using a computer vision technology, that is, the vehicle identification is performed on any frame of image of the traffic flow detection area by using a target detection algorithm, where the target detection algorithm includes, but is not limited to: and identifying all vehicles in the traffic flow detection area by using YOLO, SSD, fast R-CNN and the like, and marking each vehicle in the traffic flow detection area to obtain a vehicle identification frame corresponding to each vehicle, wherein the vehicle identification frame needs to be used for selecting the vehicles by all frames and marking the positions and the sizes of the vehicles.

Step S1032: and based on the traffic flow detection area in the video monitoring image, tracking the vehicle of each vehicle identification frame corresponding to the multi-directional lane, and determining the first direct traffic flow and the steering traffic flow in the multi-directional lane.

For the embodiment of the application, based on the traffic flow detection area in the video monitoring image, the vehicle tracking is performed on each vehicle identification frame corresponding to the multi-directional lane, namely, the vehicle tracking algorithm is used for performing the vehicle tracking on each vehicle identification in the multi-directional lane, wherein the vehicle tracking algorithm predicts and updates the vehicle identification frame of each vehicle according to the position and appearance characteristics of the vehicle between the continuous frames. The method for determining the straight traffic flow and the steering traffic flow in the multidirectional lane by means of vehicle tracking is various, the embodiment of the application is not limited any more, in one implementation manner, vehicle tracking is performed on each vehicle identification frame corresponding to the multidirectional lane based on a traffic flow detection area in a video monitoring image, and a rotation direction angle set corresponding to each vehicle identification frame is determined, wherein the rotation direction angle set is formed by rotation direction angles corresponding to the vehicle identification frames in different video frames, and the rotation direction angles are included angles formed between a vehicle driving direction and a lane direction. Then, a first direct traffic flow and a steered traffic flow in the multi-directional lane are determined based on the set of rotational direction angles corresponding to each vehicle identification frame. In another implementation, a target tracking algorithm, for example SORT, deepSORT, is used to track the vehicle identification frames in the multi-directional lanes, establish a motion trail of the vehicle, analyze the vehicle driving direction based on the motion trail of each vehicle, and calculate the first direct traffic flow and the steering traffic flow in the multi-directional lanes based on the analyzed vehicle driving directions.

Step S1033: and based on the traffic flow detection area in the video monitoring image, carrying out vehicle tracking on each vehicle identification frame corresponding to the unidirectional straight-through lane, and determining the second straight-through traffic flow in the unidirectional straight-through lane.

For the embodiment of the application, since the vehicle in the unidirectional straight-running lane can only run straight and cannot run in other directions, analysis of the running direction of the vehicle on the unidirectional straight-running lane is not needed, and the calculation resources of the electronic equipment are saved. Therefore, when each vehicle identification frame corresponding to the unidirectional straight-through lane is tracked based on the vehicle flow detection area in the video monitoring image, the target tracking algorithm is only used for tracking the vehicle of each vehicle identification frame corresponding to the unidirectional straight-through lane, and when the disappearance of the vehicle identification frame of the vehicle in the vehicle flow detection area is detected, the vehicle flow counter is controlled to be increased by one so as to obtain the second straight-through vehicle flow in the unidirectional straight-through lane.

Step S1034: and determining the straight traffic flow of the target branch based on the summation of the first straight traffic flow and the second straight traffic flow, wherein the branch traffic flow comprises: straight traffic flow and steering traffic flow.

For the embodiment of the application, the straight traffic flow is the sum of the first straight traffic flow and the second straight traffic flow, and at the same time, the straight traffic flow is the total number of traffic flows on all unidirectional straight lanes in the target branch. Steering traffic is traffic in the target leg that is not straight, including but not limited to: left turning traffic, right turning traffic, and u-turn traffic.

It can be seen that, in the embodiment of the present application, vehicle identification is performed based on the traffic flow detection area in the video surveillance image, and the vehicle identification frame corresponding to each vehicle is marked. And then, based on the traffic flow detection area in the video monitoring image, carrying out vehicle tracking on each vehicle identification frame corresponding to the multi-directional lane, and determining the first direct traffic flow and the steering traffic flow in the multi-directional lane. And at the same time, based on the traffic flow detection area in the video monitoring image, carrying out vehicle tracking on each vehicle identification frame corresponding to the unidirectional straight-through lane, and determining the second straight-through traffic flow in the unidirectional straight-through lane. And finally, summing the first straight traffic flow and the second straight traffic flow to determine the straight traffic flow of the target branch. When the traffic flow is detected, the corresponding traffic flow detection method is selected according to different lane categories, so that the accuracy of the bypass traffic flow and the traffic flow detection efficiency are improved.

Further, in order to improve accuracy of the traffic flow in the multi-directional lane, in the embodiment of the present application, based on the traffic flow detection area in the video monitoring image, the vehicle tracking is performed on each vehicle identification frame corresponding to the multi-directional lane, and the first direct traffic flow and the steering traffic flow in the multi-directional lane are determined, including:

Based on a traffic flow detection area in the video monitoring image, carrying out vehicle tracking on each vehicle identification frame corresponding to the multidirectional lane, and determining a rotation direction angle set corresponding to each vehicle identification frame;

For the embodiment of the application, since the video monitoring image is composed of continuous multi-frame images, the traffic flow detection area is also composed of continuous multi-frame images, and therefore, based on the continuous multi-frame images of the detection area, the vehicle tracking is performed on the target vehicle identification frame corresponding to the multi-directional lane, so as to obtain the rotation direction angle of the target vehicle identification frame in each frame image, and the rotation direction angles in each frame image arranged according to time sequence jointly form a rotation direction angle set corresponding to the target vehicle identification frame, wherein the target vehicle identification frame is any vehicle identification frame. There are various ways of determining the rotation direction angle of the target vehicle identification frame, and in one possible way, the rotation direction angle is calculated according to the position change of the target vehicle identification frame between successive frame images, that is, the motion trajectory of the target vehicle identification frame is fitted by using a least square method, and then the rotation direction angle is calculated according to the change of the trajectory. Of course, the rotation direction angle can also be estimated more accurately based on the method of the three-dimensional rigid motion model. In another implementation manner, for any frame of image, an included angle between the lane line and the target vehicle identification frame is calculated by using a computer vision technology, and the included angle between the lane line and the target vehicle identification frame is recorded as a rotation direction angle corresponding to the frame of image.

Because of the specificity of the target branch in the intersection, the actual road condition of the target branch determines the steering angle of the steering vehicle, for example, the turning radius and the layout of the road determine the angle change of the vehicle in the turning process, and the width of the lane can influence the steering angle. Therefore, the electronic device stores in advance the steering direction angle range corresponding to the target branch, and the magnitude of the steering direction angle range is measured by a person skilled in the art based on a large number of experimental operations. Therefore, the first direct traffic flow and the steered traffic flow in the multidirectional lane are determined based on the steering direction angle range and the set of rotational direction angles corresponding to each vehicle identification frame. Specifically, when the number of the rotation direction angle sets falling in the steering direction angle range is greater than a preset number, determining that the vehicle identification frame is steering; otherwise, the vehicle identification frame is determined to be straight, wherein the preset number of sizes are determined by one skilled in the art based on a number of experimental operations. And integrating the driving directions of each vehicle identification frame to obtain the first direct traffic flow and the steering traffic flow in the multidirectional lanes.

It can be seen that, in the embodiment of the present application, based on the traffic flow detection area in the video monitoring image, vehicle tracking is performed on each vehicle identification frame corresponding to the multi-directional lane, a rotation direction angle set corresponding to each vehicle identification frame is determined, and then, based on the rotation direction angle set corresponding to each vehicle identification frame, the first direct traffic flow and the steering traffic flow in the multi-directional lane are determined. When the first direct traffic flow and the steering traffic flow of the multi-directional lane are determined through the rotation direction angle of the vehicle, the specificity of a target branch in the intersection is comprehensively considered, and the accuracy of the traffic flow in the multi-directional lane is improved.

Further, in order to enable the updated traffic flow prediction model to better cope with the change of the traffic condition and adjust the prediction in time, in the embodiment of the present application, the method further includes, after determining the branch traffic flow of the target branch, performing traffic flow detection based on the traffic flow detection area in the video monitoring image and the lane category corresponding to each lane in the target branch:

predicting the traffic flow of the target branch by using the traffic flow prediction model to obtain the predicted traffic flow corresponding to the target branch;

performing traffic flow comparison analysis based on the predicted traffic flow and the branch traffic flow corresponding to the target branch, and determining a comparison analysis result;

For the embodiment of the application, because various emergency situations often exist in the road driving process, and various special situations such as situations that errors exist in traffic flow detection due to special reasons such as weather and the like, intersection traffic flow statistical results are not consistent with actual situations, after the branch traffic flow is determined, traffic flow comparison analysis is conducted on the basis of the branch traffic flow and the predicted traffic flow, and the branch traffic flow corresponding to the target branch is adjusted under the condition that the difference is large, so that the traffic situation of the target branch can be more accurately represented by the branch traffic flow.

The vehicle flow prediction model is obtained by training the convolutional neural network based on the historical vehicle flow training sample of the target branch, so when the relevant information of the target branch is input into the vehicle flow prediction model, the vehicle flow prediction model automatically predicts the vehicle flow to obtain the predicted vehicle flow corresponding to the target branch, namely, when the predicted vehicle flow obtained by the vehicle flow prediction model is utilized, the unique characteristics and influence factors of the target branch are considered, so that the predicted vehicle flow is more accurate, and the actual condition of the target branch can be reflected better. Then, comparing and analyzing the traffic flow based on the predicted traffic flow and the branch traffic flow corresponding to the target branch, and determining that the comparison and analysis result is abnormal when the difference between the predicted traffic flow and the branch traffic flow is larger than a traffic flow threshold; otherwise, determining that the comparison analysis result is normal, wherein the traffic flow threshold value is a reasonable traffic flow difference value which can exist between the predicted condition and the actual condition. And when the comparison and analysis result is normal, the characterization traffic flow detection determines that the branch traffic flow accords with the normal condition of the target branch, and then the branch traffic flow of the target branch is used as a training data set, and the training data set is utilized to retrain the traffic flow prediction model, so that an updated traffic flow prediction model is obtained. Because the traffic condition is dynamically changed, the latest branch traffic flow is brought into the training data set, so that the latest change of the traffic condition can be captured, the updated traffic flow prediction model can better cope with the change of the traffic condition, and the prediction can be adjusted in time.

Therefore, in the embodiment of the application, since various emergency situations often exist in the road driving process and various special situations such as situations where errors exist in traffic flow detection due to special reasons such as weather exist, the statistical result of the traffic flow at the intersection is inconsistent with the actual situation. Therefore, the traffic flow prediction model is utilized to predict the traffic flow of the target branch, the predicted traffic flow corresponding to the target branch is obtained, meanwhile, the traffic flow comparison analysis is carried out based on the predicted traffic flow corresponding to the target branch and the branch traffic flow, and the comparison analysis result is determined. And when the comparison and analysis result is normal, taking the branch traffic flow of the target branch as a training data set, and retraining the traffic flow prediction model by utilizing the training data set to obtain an updated traffic flow prediction model. The latest branch traffic flow is brought into the training data set, so that the latest change of the traffic condition can be captured, the updated traffic flow prediction model can better cope with the change of the traffic condition, and the prediction can be adjusted in time.

Further, in order to improve accuracy and reliability of the traffic flow data, in this embodiment of the present application, the traffic flow comparison analysis is performed based on the predicted traffic flow and the bypass traffic flow corresponding to the target bypass, and after determining the comparison analysis result, the method further includes:

For the embodiment of the application, when the comparison and analysis result is abnormal, the branch vehicle flow determined by the vehicle flow detection is greatly different from the normal condition of the target branch, and various reasons for the situation are included, such as data acquisition errors, special events affecting the normal running of the vehicle, bad weather conditions, road infrastructure changes and the like. In order to improve the accuracy and reliability of traffic flow detection, under the condition that the branch traffic flow determined by the traffic flow detection is greatly different from the normal condition of a target branch, the positioning linkage mode is used as a supplementary means of traffic flow detection, so that the accurate branch traffic flow corresponding to the target branch is obtained, and the accuracy and reliability of traffic flow data are greatly improved.

Specifically, when the comparison analysis result is abnormal, the positioning data of the vehicle running on the target branch is collected through an interface with a vehicle positioning system (such as a GPS), wherein the positioning data includes but is not limited to: vehicle position, vehicle speed, direction of travel, and vehicle ID. The collected positioning data is then preprocessed to obtain preprocessed positioning data, wherein the preprocessing includes, but is not limited to: data cleaning, data filtering, data conversion and data analysis. Further, vehicle tracking is performed based on the preprocessed positioning data, a running track of each vehicle is determined, and traffic flow statistics is performed based on the running tracks corresponding to all vehicles in the target branch, so as to obtain accurate branch traffic flow, wherein the accurate branch traffic flow comprises: and the corresponding direction traffic flow of each driving direction in the target branch.

Therefore, in the embodiment of the application, in order to improve the accuracy and reliability of traffic flow detection, under the condition that the branch traffic flow determined by traffic flow detection is greatly different from the normal condition of the target branch, the positioning linkage mode is used as a supplementary means of traffic flow detection, so that the accurate branch traffic flow corresponding to the target branch is obtained, and the accuracy and reliability of traffic flow data are greatly improved.

Further, in order to improve accuracy and automation degree of lane feature extraction and provide reliable data support for subsequent traffic flow analysis and application, in the embodiment of the present application, the method for acquiring lane features of the target branch includes:

For the embodiment of the application, along with continuous construction of the road, the lane distribution of the target branch in the intersection may change, so that in order to ensure that the lane characteristics can accurately represent the current condition of the target branch, the lane characteristics are collected once again after each interval of the lane characteristic collection period, so as to ensure the accuracy and timeliness of the lane characteristics of the target branch.

Specifically, when the period of collecting the traffic characteristics reaches the period of collecting the road image of the target branch in the traffic low valley period, various modes are available for collecting the road image, for example, a camera installed at the target branch is used for shooting, and an unmanned aerial vehicle is used for collecting the road image, and meanwhile, fewer vehicles on the road can be ensured by selecting the traffic low valley period, interference factors are reduced, and the accurate extraction of the traffic characteristics is facilitated. Then, feature extraction is carried out based on the road image to obtain the current lane feature of the target branch, namely, the feature extraction is carried out by utilizing the computer vision and image processing technology to extract lane line information, lane width, lane mark and the like, and the lane line information, the lane width and the lane mark of the target branch are marked as the current lane feature. And then, acquiring the stored lane characteristics stored in the electronic equipment in advance, wherein the stored lane characteristics are lane characteristics of a target branch stored before, comparing the characteristics based on the current lane characteristics and the stored lane characteristics, judging whether the characteristics are different before the current lane characteristics and the stored lane characteristics, and when the characteristics are different, updating the stored lane characteristics by utilizing the current lane characteristics and recording the stored lane characteristics as the lane characteristics of the target branch so as to facilitate the subsequent use in detecting the traffic flow. The feature extraction method based on the image processing and computer vision technology can improve the accuracy and the automation degree of the lane feature extraction and provide reliable data support for the subsequent traffic flow analysis and application.

Therefore, in the embodiment of the application, when the lane feature collection period is reached, the road image of the target branch in the traffic valley period is collected, feature extraction is performed based on the road image, the current lane feature of the target branch is obtained, and after each interval of the lane feature collection period, the lane feature is collected again, so that the accuracy and timeliness of the lane feature of the target branch are ensured. And then, carrying out feature comparison based on the current lane feature and the stored lane feature, and when the features are different, updating the stored lane feature by utilizing the current lane feature and marking the updated lane feature as the lane feature of the target branch. The feature extraction method based on the image processing and computer vision technology can improve the accuracy and the automation degree of the lane feature extraction and provide reliable data support for the subsequent traffic flow analysis and application.

The above embodiments describe a method for detecting traffic flow at an intersection from the viewpoint of a method flow, and the following embodiments describe a device for detecting traffic flow at an intersection from the viewpoint of a virtual module or a virtual unit, specifically the following embodiments.

The embodiment of the application provides a traffic flow detection device for an intersection, as shown in fig. 3, the traffic flow detection device for an intersection may specifically include:

The lane dividing module 210 is configured to obtain lane characteristics of a target branch in the intersection, perform lane division based on the lane characteristics, and determine a lane category corresponding to each lane in the intersection, where the lane category includes: the target branch road is any branch road in the intersection;

the detection area dividing module 220 is configured to obtain a video monitoring image of the target branch, perform preprocessing based on the video monitoring image, and select a traffic flow detection area in a frame in the video monitoring image;

the traffic flow detection module 230 is configured to detect traffic flow based on a lane category corresponding to each lane in the target branch and a traffic flow detection area in the video monitoring image, and determine a branch traffic flow of the target branch;

the traffic flow summarizing module 240 is configured to synthesize the branch traffic flow of each target branch in the intersection to obtain the total traffic flow of the intersection.

For the embodiment of the application, lane classification is performed based on lane characteristics, lane types corresponding to each lane in an intersection are determined, meanwhile, preprocessing is performed based on video monitoring images, a vehicle flow detection area is selected in a frame mode in the video monitoring images, and the frame-selected vehicle flow detection area is used for intensively processing and analyzing parts really related to the vehicle flow, so that false detection and missing detection are reduced, accuracy of vehicle flow detection is improved, and computing resources and storage resources of electronic equipment are saved. And then, detecting the traffic flow based on the lane type corresponding to each lane in the target branch and the traffic flow detection area in the video monitoring image, determining the branch traffic flow of the target branch, and selecting an adaptive traffic flow detection method aiming at different lane types when detecting the traffic flow so as to improve the accuracy of the branch traffic flow and the traffic flow detection efficiency. And finally, synthesizing the branch road traffic of each target branch road in the intersection to obtain the total traffic flow of the intersection. In order to improve the accuracy of the traffic flow of the intersection, traffic flow detection is performed on each target branch in the intersection to obtain the branch traffic flow of the target branch, and then the branch traffic flows are summarized to obtain the total traffic flow of the intersection.

In one possible implementation manner of the embodiment of the present application, when performing traffic flow detection based on the lane type corresponding to each lane in the target branch and the traffic flow detection area in the video monitoring image, the traffic flow detection module 230 is configured to:

vehicle identification is carried out based on a vehicle flow detection area in the video monitoring image, and a vehicle identification frame corresponding to each vehicle is marked;

based on a traffic flow detection area in the video monitoring image, tracking vehicles on each vehicle identification frame corresponding to the multi-directional lane, and determining a first direct traffic flow and a steering traffic flow in the multi-directional lane;

based on a traffic flow detection area in the video monitoring image, carrying out vehicle tracking on each vehicle identification frame corresponding to the unidirectional straight-through lane, and determining a second straight-through traffic flow in the unidirectional straight-through lane;

and determining the straight traffic flow of the target branch based on the summation of the first straight traffic flow and the second straight traffic flow, wherein the branch traffic flow comprises: straight traffic flow and steering traffic flow.

In one possible implementation manner of the embodiment of the present application, when performing vehicle tracking on each vehicle identifier frame corresponding to a multi-directional lane based on the vehicle flow detection area in the video monitoring image, the vehicle flow detection module 230 is configured to:

In one possible implementation manner of the embodiment of the present application, a traffic flow detection device for an intersection further includes:

the comparison analysis module is used for predicting the traffic flow of the target branch by using the traffic flow prediction model to obtain the predicted traffic flow corresponding to the target branch;

and the accurate detection module is used for detecting the traffic flow of the target branch again by using a positioning linkage mode when the comparison and analysis result is abnormal, so as to obtain the traffic flow of the accurate branch, wherein the positioning linkage mode is used for accurately detecting the traffic flow through a positioning system of the vehicle.

the lane feature acquisition module is used for acquiring road images of the target branch in the traffic valley period when the lane feature acquisition period is reached, and extracting features based on the road images to obtain the current lane features of the target branch;

It will be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working process of the traffic flow detection device for an intersection described above may refer to the corresponding process in the foregoing method embodiment, and will not be described herein again.

In an embodiment of the present application, as shown in fig. 4, an electronic device 300 shown in fig. 4 includes: a processor 301 and a memory 303. Wherein the processor 301 is coupled to the memory 303, such as via a bus 302. Optionally, the electronic device 300 may also include a transceiver 304. It should be noted that, in practical applications, the transceiver 304 is not limited to one, and the structure of the electronic device 300 is not limited to the embodiment of the present application.

The processor 301 may be a CPU (Central Processing Unit ), general purpose processor, DSP (Digital Signal Processor, data signal processor), ASIC (Application Specific Integrated Circuit ), FPGA (Field Programmable Gate Array, field programmable gate array) or other programmable logic device, transistor logic device, hardware components, or any combination thereof. Which may implement or perform the various exemplary logic blocks, modules, and circuits described in connection with this disclosure. Processor 301 may also be a combination that implements computing functionality, e.g., comprising one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

Bus 302 may include a path to transfer information between the components. Bus 302 may be a PCI (Peripheral Component Interconnect, peripheral component interconnect Standard) bus or an EISA (Extended Industry Standard Architecture ) bus, or the like. Bus 302 may be divided into an address bus, a data bus, a control bus, and the like. For ease of illustration, only one thick line is shown in fig. 4, but not only one bus or type of bus.

The Memory 303 may be, but is not limited to, a ROM (Read Only Memory) or other type of static storage device that can store static information and instructions, a RAM (Random Access Memory ) or other type of dynamic storage device that can store information and instructions, an EEPROM (Electrically Erasable Programmable Read Only Memory ), a CD-ROM (Compact Disc Read Only Memory, compact disc Read Only Memory) or other optical disk storage, optical disk storage (including compact discs, laser discs, optical discs, digital versatile discs, blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer.

The memory 303 is used for storing application program codes for executing the present application and is controlled to be executed by the processor 301. The processor 301 is configured to execute the application code stored in the memory 303 to implement what is shown in the foregoing method embodiments.

Among them, electronic devices include, but are not limited to: mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and the like, and stationary terminals such as digital TVs, desktop computers, and the like. But may also be a server or the like. The electronic device shown in fig. 4 is only an example and should not be construed as limiting the functionality and scope of use of the embodiments herein.

The present application provides a computer readable storage medium having a computer program stored thereon, which when run on a computer, causes the computer to perform the corresponding method embodiments described above. Compared with the related art, the method and the device for detecting the traffic flow of the vehicle have the advantages that lane division is carried out based on lane characteristics, the lane category corresponding to each lane in the intersection is determined, meanwhile, preprocessing is carried out based on the video monitoring image, the traffic flow detection area is selected in the video monitoring image in a frame mode, the frame-selected traffic flow detection area is used for intensively processing and analyzing the part really related to the traffic flow, false detection and missing detection are reduced, accuracy of traffic flow detection is improved, and computing resources and storage resources of electronic equipment are saved. And then, detecting the traffic flow based on the lane type corresponding to each lane in the target branch and the traffic flow detection area in the video monitoring image, determining the branch traffic flow of the target branch, and selecting an adaptive traffic flow detection method aiming at different lane types when detecting the traffic flow so as to improve the accuracy of the branch traffic flow and the traffic flow detection efficiency. And finally, synthesizing the branch road traffic of each target branch road in the intersection to obtain the total traffic flow of the intersection. In order to improve the accuracy of the traffic flow of the intersection, traffic flow detection is performed on each target branch in the intersection to obtain the branch traffic flow of the target branch, and then the branch traffic flows are summarized to obtain the total traffic flow of the intersection.

It should be understood that, although the steps in the flowcharts of the figures are shown in order as indicated by the arrows, these steps are not necessarily performed in order as indicated by the arrows. The steps are not strictly limited in order and may be performed in other orders, unless explicitly stated herein. Moreover, at least some of the steps in the flowcharts of the figures may include a plurality of sub-steps or stages that are not necessarily performed at the same time, but may be performed at different times, the order of their execution not necessarily being sequential, but may be performed in turn or alternately with other steps or at least a portion of the other steps or stages.

The foregoing is only a partial embodiment of the present application and it should be noted that, for a person skilled in the art, several improvements and modifications can be made without departing from the principle of the present application, and these improvements and modifications should also be considered as the protection scope of the present application.

Claims

1. A traffic flow detection method at an intersection, comprising:

2. The method for detecting traffic flow at an intersection according to claim 1, wherein the determining the branch traffic flow of the target branch based on the lane category corresponding to each lane in the target branch and the traffic flow detection area in the video surveillance image includes:

3. The method for detecting traffic flow at an intersection according to claim 2, wherein the determining the first direct traffic flow and the diverted traffic flow in the multi-directional lane by performing vehicle tracking on each vehicle identification frame corresponding to the multi-directional lane based on the traffic flow detection area in the video surveillance image includes:

4. The method for detecting traffic at an intersection according to claim 1, wherein the detecting traffic based on the lane category corresponding to each lane in the target branch and the traffic detection area in the video surveillance image, after determining the branch traffic of the target branch, further comprises:

5. The method according to claim 4, wherein the comparing the traffic flow to the predicted traffic flow and the branch traffic flow corresponding to the target branch, and determining the comparison result, further comprises:

6. The traffic flow detection method of an intersection according to claim 1, wherein the method of collecting the lane characteristics of the target branch includes:

7. An intersection traffic flow detection device comprising:

8. An electronic device, comprising:

at least one processor;

a memory;

at least one application program, wherein the at least one application program is stored in the memory and configured to be executed by the at least one processor, the at least one application program configured to: the traffic flow detection method of the intersection according to any one of claims 1 to 6.

9. A computer-readable storage medium, having stored thereon a computer program which, when executed in a computer, causes the computer to execute the traffic flow detection method of the intersection according to any one of claims 1 to 6.