Intelligent traffic control system and control method based on YOLO visual perception

Abstract

The present invention provides a control method of an intelligent traffic control system based on YOLO visual perception, and belongs to the technical field of traffic control. An intelligent traffic control system based on YOLO visual perception includes a video acquisition and processing module and a control module; the present invention connects a camera at an intersection through the video acquisition and processing module, captures the captured video image, analyzes and extracts the image feature information and converts it into control information, and at the same time, the video acquisition and processing module is connected to the control module for communication, and the control module is connected to the traffic light at the intersection, so that the control module can receive the control information and adjust the duration of the traffic light at the intersection in real time according to the control information, that is, according to the traffic volume of different lane types in different directions of the intersection, the duration of the traffic light at the corresponding intersection is controlled in real time, and the duration of the traffic light at the intersection is intelligently and dynamically adjusted, which effectively avoids the problem of traffic congestion in one direction and empty traffic in the other direction in actual road conditions, and improves the traffic efficiency of the road.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic control, and in particular to an intelligent traffic control system based on YOLO (English: You Only LookOnce; Chinese: Target Detection Model) visual perception and a control method thereof. Background Art

[0002] In order to ensure the smooth and safe passage at the intersection of urban roads, traffic lights are usually set at the intersection. Through the timed conversion of traffic lights, different traffic signals are switched to control the traffic in different directions to pass through the intersection smoothly. At present, although the signal switching of existing traffic lights on the market adopts an automatic control mode, the intelligence of this automatic control mode is poor. It only sets the interval time in advance and switches the signal at the corresponding time. In fact, it belongs to a timed switching mode. There are two ways of existing traffic lights. One is that the traffic lights evenly distribute the conduction time. However, if the traffic volume in the east-west direction and the north-south direction of the intersection is greatly different, it will cause congestion on one side and empty space on the other side, which will still affect the traffic efficiency. The other is to set the passage time of the main road to be longer than the passage time of the secondary road. Although it can properly solve the above-mentioned problems of congestion on the main road and empty space on the secondary road, in actual road conditions, the traffic volume on the secondary road will be greater than the traffic volume on the main road. If the passage time of the main road is still greater than the passage time of the secondary road, it will also cause empty space and congestion problems, which will also affect the traffic efficiency. Summary of the invention

[0003] In view of the above problems existing in the prior art, the present invention aims to provide an intelligent traffic control system based on YOLO visual perception and a control method thereof, and set a video acquisition and processing module and a control module. Relying on the cameras installed at the intersections, the traffic flow information on different lanes of each intersection is acquired and processed in real time, and the processed traffic flow information is then transmitted to the traffic control system in real time. The traffic control system completes the real-time dynamic control of the traffic lights to improve the road traffic efficiency.

[0004] The specific technical solutions are as follows:

[0005] The intelligent traffic control system based on YOLO visual perception has the following features:

[0006] The video acquisition and processing module is used to connect to the camera installed at the intersection, load the YOLO model, dynamically capture and analyze the video image, analyze and extract the feature information of the video image, and convert the feature information of the video image into control information;

[0007] The control module is used to receive the control information input by the video acquisition and processing module, and complete the dynamic control of the intersection traffic lights according to the input control information.

[0008] The control method of the intelligent traffic control system based on YOLO visual perception includes the following steps:

[0009] Step S1, video acquisition and processing;

[0010] Step S11, the video acquisition and processing module is powered on and starts to enter a working state;

[0011] Step S12, initialize the system and load the YOLO model, and determine whether the YOLO model loading is successful, if successful, execute step S13, otherwise repeat step S12;

[0012] Step S13, creating a video capture thread to obtain the video information of the camera at the intersection in real time, and judging whether the video capture thread is successfully created. If successful, executing step S14, otherwise repeating step S13;

[0013] Step S14, creating a communication thread for the communication connection between the video acquisition and processing module and the control module, and determining whether the communication thread is successfully created. If successful, executing step S15, otherwise repeating step S14;

[0014] Step S15, creating a YOLO processing thread, which is used to receive the video image information from the camera at the intersection transmitted by the video capture thread, and convert the received video image information into control information, and determine whether the YOLO processing thread is successfully created. If successful, execute step S2, otherwise repeat step S15;

[0015] Step S2, analysis and control;

[0016] Step S21, the control module is powered on and starts to enter a working state;

[0017] Step S22, initializing the system, obtaining the setting information of the external device, and setting it as the initial information of the system in the control module, judging the manual or automatic operation state set in the initial information, if it is judged to be the manual operation state, executing step S23, if it is judged to be the automatic operation state, executing step S24;

[0018] Step S23, switching the signal light at a fixed beat according to the time interval preset in the system of the control module, ignoring the control information transmitted by the communication thread, and repeating step S23 until the manual operation state is exited;

[0019] Step S24, determining whether the communication between the video acquisition and processing module and the control module is normally connected. If not, the signal light is switched at a fixed beat according to a time interval preset in the system of the control module, and step S24 is repeated. Meanwhile, the communication connection status between the video acquisition and processing module and the control module is continuously monitored until the communication connection status is normally connected, and then the process turns to step S25;

[0020] Step S25, continuously monitoring the control information transmitted by the video acquisition and processing module, if no new control information is transmitted, repeating step S25, otherwise executing step S26;

[0021] Step S26, analyzing the new control information and determining the type of the new control information. If it is fault information, executing step S27, otherwise executing step S28;

[0022] Step S27, determining that the video acquisition and processing module fails, sending an automatic reset instruction to the video acquisition and processing module, and at the same time, performing a fixed beat timed switching of the signal light according to a pre-set time interval in the system of the control module, and returning to step S25;

[0023] Step S28, counting the vehicle flow, analyzing the transmitted control information, calculating the maximum value of the vehicle flow of lanes of different lane types in the north-south and east-west directions, and obtaining the vehicle flow information of different lanes in the north-south and east-west directions;

[0024] Step S29, determining a state indicator according to the traffic flow information of different lanes in the north-south and east-west directions obtained in step S28;

[0025] Step S3, allocating the duration of the signal light;

[0026] According to the status identifier obtained in step S29, the duration of the traffic lights in different lanes in the north-south and east-west directions is allocated, and the process returns to step S25.

[0027] In the above-mentioned control method of the intelligent traffic control system based on YOLO visual perception, in step S12, a YOLO model loading threshold is pre-set. If the consecutive number of loading of the YOLO model is greater than the set loading threshold when step S12 is repeatedly executed, it is determined that the YOLO model loading fails, the system is exited, and a prompt message indicating that the YOLO model loading failed is issued.

[0028] In the above-mentioned control method of the intelligent traffic control system based on YOLO visual perception, in step S14, a communication threshold of the communication thread is pre-set. If the consecutive number of times the communication thread is created is greater than the set communication threshold when step S14 is repeatedly executed, it is determined that the communication thread creation has failed, the system is exited, and a prompt message of communication connection failure is issued.

[0029] In the above-mentioned control method of the intelligent traffic control system based on YOLO visual perception, in step S22, the external device includes a touch screen or a computer, and the control module system has a built-in timing interval control program for setting the time interval and performing fixed-beat timing switching of the traffic light.

[0030] The above-mentioned control method based on YOLO visual perception intelligent traffic control system, wherein, in step S28, when determining the traffic flow information of different lanes in the north-south and east-west directions, first obtain the maximum value of the traffic flow on the lanes of different lane types in the forward direction of the north-south or east-west direction, if there are multiple lanes on one lane type, then obtain the maximum value of the traffic flow of these multiple lanes as the traffic flow of this lane type in the forward direction, similarly, obtain the maximum value of the traffic flow on the lanes of different lane types in the reverse direction of the north-south or east-west direction, if there are multiple lanes on one lane type, then obtain the maximum value of the traffic flow of these multiple lanes as the traffic flow of this lane type in the reverse direction, finally, compare the maximum value of the traffic flow on the lanes of different lane types in the forward direction of the north-south direction with the maximum value of the traffic flow on the lanes of different lane types in the reverse direction, and take the maximum value of the two as the traffic flow of different lanes in the north-south direction, similarly, compare the maximum value of the traffic flow on the lanes of different lane types in the forward direction of the east-west direction with the maximum value of the traffic flow on the lanes of different lane types in the reverse direction, and take the maximum value of the two as the traffic flow of different lanes in the east-west direction.

[0031] In the above control method of the intelligent traffic control system based on YOLO visual perception, in step S29, the state identification includes category VIII,

[0032] Class I status indicates that there are no through vehicles in the east-west and north-south directions, or there are only left-turning vehicles in the east-west and north-south directions;

[0033] Category II status indicates that there are vehicles going straight in the east-west direction, but no vehicles going straight in the north-south direction;

[0034] Category III status indicates that there are through vehicles in both the east-west and north-south directions, and the through vehicle flow in the east-west direction is greater than that in the north-south direction, and the difference between the two is greater than the preset value;

[0035] Category IV status indicates that there are through vehicles in the north-south direction, but no through vehicles in the east-west direction;

[0036] Class V status indicates that there are through vehicles in both the north-south direction and the east-west direction, and the through vehicle flow in the north-south direction is greater than that in the east-west direction, and the difference between the two is greater than the preset value;

[0037] Category VI status indicates that there are through vehicles in both the north-south direction and the east-west direction, and the through vehicle flow in the north-south direction and the east-west direction is equal;

[0038] Category VII status indicates that there are only left-turning vehicles in the east-west direction or only left-turning vehicles in the north-south direction and there are vehicles in the other corresponding direction;

[0039] Class VIII status indicates that the traffic volume in the east-west direction and the north-south direction are not equal, and the difference between the two does not exceed the preset value.

[0040] In the above-mentioned intelligent traffic control system based on YOLO visual perception and control method thereof, in step S3, the relationship between the state identification and the signal light allocation duration is:

[0041] When the status is marked as Class I, Class VI, Class VII, and Class VIII, the east-west and north-south directions are assigned equal durations;

[0042] When the status mark is Class II, the east-west traffic lights are assigned to be always green, and the north-south traffic lights are assigned to be always red;

[0043] When the status mark is Class IV, the north-south traffic lights are assigned to be always green, and the east-west traffic lights are assigned to be always red;

[0044] When the status is marked as Class III, the green light duration of the east-west traffic light is longer than that of the north-south traffic light;

[0045] When the status is identified as Class V, the green light duration of the traffic light in the north-south direction is longer than the green light duration of the traffic light in the east-west direction.

[0046] In the above-mentioned control method of the intelligent traffic control system based on YOLO visual perception, when the status identification is Class I, Class VI, Class VII and Class VIII, the duration of the traffic lights in the east-west direction and the north-south direction under different types of status identification is different.

[0047] The above-mentioned control method of the intelligent traffic control system based on YOLO visual perception, wherein, when the state is identified as Class III, the green light duration of the traffic light in the east-west direction is m times the green light duration of the traffic light in the north-south direction, and when the state is identified as Class V, the green light duration of the traffic light in the north-south direction is n times the green light duration of the traffic light in the east-west direction, and both m and n are real numbers.

[0048] The positive effects of the above technical solution are:

[0049] The above-mentioned YOLO visual perception-based intelligent traffic control system and control method thereof are provided with a video acquisition and processing module and a control module. The video image taken by the camera arranged at the intersection is captured by the video acquisition and processing module, and the feature information of the video image is analyzed and extracted and converted into control information. Then, the control module completes the real-time intelligent dynamic control of the traffic light according to the control information, that is, the duration of the traffic light at the corresponding intersection is controlled in real time according to the traffic volume of different lane types in different directions, which effectively avoids the problem of congested traffic in one direction and empty traffic in the other direction in actual road conditions, thereby improving the traffic efficiency of the road. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a structural diagram of an embodiment of an intelligent traffic control system based on YOLO visual perception of the present invention;

[0051] Figure 2 It is a flow chart of the control method of the YOLO visual perception intelligent traffic control system of the present invention.

[0052] In the attached figure: 1. Video acquisition and processing module; 2. Control module; 3. Camera; 4. Signal light; 5. External device. DETAILED DESCRIPTION

[0053] In order to make the technical means, creative features, objectives and effects of the present invention easier to understand, the following embodiments are combined with the attached Figure 1 To Attachment Figure 2 The technical solution provided by the present invention is described in detail, but the following content is not intended to limit the present invention.

[0054] Figure 1 FIG. 1 is a structural diagram of an embodiment of the YOLO visual perception intelligent traffic control system of the present invention. Figure 1 As shown, the YOLO visual perception-based intelligent traffic control system provided in this embodiment includes: a video acquisition and processing module 1 and a control module 2, which realize connection with a camera 3 and a traffic light 4 at an intersection.

[0055] Specifically, the video acquisition and processing module 1 is used to connect to the camera 3 installed at the intersection. The system is loaded with the YOLO model, which is used to dynamically capture and analyze the video images taken by the camera 3, and analyze and extract the feature information of the video images, obtain the traffic flow of different directions and lane types at the current moment, and convert it into control information input to the subsequent control module 2, which provides conditions for realizing dynamic monitoring of the traffic flow at the intersection and subsequent reasonable control of the duration of the traffic light 4.

[0056] Specifically, the control module 2 is used to receive the control information input by the video acquisition and processing module 1, complete the reception of the control information, and complete the dynamic control of the intersection signal light 4 according to the input control information, that is, as the main structure for controlling the duration of the signal light 4 at the intersection, it ensures that the duration of the signal light 4 at the intersection can correspond to the traffic flow of different lane types in different directions, avoiding the problem of traffic congestion in one direction and empty traffic in the other direction in the north-south and east-west directions, thereby ensuring the traffic efficiency of the road.

[0057] In addition, this embodiment also provides a control method for an intelligent traffic control system based on YOLO visual perception. Figure 2 Flow chart of the control method of the YOLO visual perception intelligent traffic control system of the present invention. Figure 2 As shown, the intelligent traffic control system based on YOLO visual perception is the above-mentioned intelligent traffic control system based on YOLO visual perception, and its control method includes the following steps:

[0058] Step S1, video acquisition and processing;

[0059] Step S2, analysis control; and

[0060] Step S3, allocating the duration of the signal light 4.

[0061] More specifically, the video acquisition and processing process in step S1 includes the following five sub-steps:

[0062] Step S11, after power-on, the video acquisition and processing module 1 is powered on and enters a self-starting state, and its system automatically enters a working state;

[0063] Step S12, first perform the system initialization operation, load the YOLO model, and determine whether the loading of the YOLO model is successful at the end of loading. If successful, execute step S13, otherwise repeat step S12. In this process, the system pre-sets a YOLO model loading threshold. If the number of consecutive times the YOLO model is loaded is greater than the set loading threshold when step S12 is repeatedly executed, it is determined that the YOLO model fails to load, exit the system, and issue a prompt message that the YOLO model fails to load. The subsequent steps are no longer executed, and the system is re-initialized and the YOLO model is loaded, which ensures that the YOLO model can be loaded successfully and meets the subsequent dynamic monitoring needs.

[0064] Step S13, after the YOLO model is loaded successfully, a video capture thread is created to obtain the video information of the camera 3 at the intersection in real time. In this process, it is necessary to determine whether the video capture thread is successfully created. If the video capture thread is successfully created, step S14 is executed, otherwise step S13 is repeated, and the video capture thread is continuously waited for to be successfully created to ensure that the video image can be captured in real time during the subsequent work process;

[0065] Step S14, after the video capture thread is successfully created, a communication thread is created for the communication connection between the video acquisition and processing module 1 and the control module 2, which provides conditions for the subsequent video acquisition and processing module 1 to transmit the control information to the control module 2. In this process, it is necessary to determine whether the communication thread is successfully created. If the communication thread is successfully created, step S15 is executed, otherwise step S14 is repeated. It is worth pointing out that the communication threshold of the communication thread is pre-set in the system. If the number of consecutive times of creating the communication thread is greater than the set communication threshold when step S14 is repeatedly executed, it means that the communication thread creation fails for many times, and the communication thread creation fails. At this time, the system automatically exits and sends a prompt message of communication connection failure. The subsequent steps are no longer executed, waiting for the system to be reinitialized and the YOLO model to be loaded again, the video capture thread and the communication thread to be created, to ensure that the subsequent dynamic monitoring and information transmission requirements can be met.

[0066] Step S15, after the communication thread is successfully created, a YOLO processing thread is created to continuously receive the video image information from the camera 3 at the intersection obtained by the video capture thread, and convert the received video image information into control information. Finally, the communication thread sends the control information to the subsequent control module 2. In this process, it is necessary to determine whether the YOLO processing thread is successfully created. If the YOLO processing thread is successfully created, step S2 is executed, otherwise step S15 is repeated, and the YOLO processing thread is continuously waited for to be successfully created, which provides a guarantee for the subsequent formation of control information.

[0067] More specifically, while the video acquisition and processing module 1 forms control information and transmits it through the communication thread, the control module 2 also enters the working state, such as step S2, and the following steps disclosed in the analysis control process are:

[0068] Step S21, after power-on, the control module 2 is powered on and enters the self-starting state, and the system enters the working state;

[0069] Step S22, after the system enters the working state, the system initialization operation is first performed, and after the system initialization operation is completed, the setting information from the external device 5 is obtained, and the setting information of the external device 5 is set as the initial information of the system in the control module 2, and the operating state set in the initial information is determined. At this time, the operating state includes two operating states, manual or automatic. If the operating state set in the initial information is determined to be a manual operating state, step S23 is executed, and if the operating state set in the initial information is determined to be an automatic operating state, step S24 is executed. It is worth pointing out that the external device 5 used to set the initial information can be an easy-to-operate structure such as a touch screen, or a structure with a mouse and keyboard such as a computer, which can meet the use requirements such as parameter setting, and can be an existing control device for controlling traffic lights. In addition, the system of the control module 2 is built with a timed interval control program, which is used to set the time interval and perform a fixed beat timed switch on the signal light 4, which provides an operating basis for the subsequent control of the signal light 4 in the manual operating state or the control of the signal light 4 when a fault occurs in the automatic operating state, and has higher safety assurance.

[0070] In step S23, the control module 2 performs a fixed-beat timed switching on the signal light 4 according to a time interval pre-set in the system, and automatically ignores the control information transmitted by the communication thread, and repeats step S23 until the manual operation state is exited. That is, after determining that the operation state in the initial information is the manual operation state, the control module 2 only performs a fixed-beat timed switching on the signal light 4 according to a time interval pre-set in its system, which is the same as the operation mode of the existing signal light 4 on the market.

[0071] Step S24, when determining that the operating status in the initial information is the automatic operating status, it is necessary to first determine whether the communication between the video acquisition and processing module 1 and the control module 2 is normally connected. If not, the signal light 4 is switched at a fixed rhythm according to the pre-set time interval in the system of the control module 2, and step S24 is repeated to ensure the normal operation of the signal light 4. At the same time, the communication connection status between the video acquisition and processing module 1 and the control module 2 is continuously monitored until the communication connection status is normally connected, that is, when it is monitored that the above-mentioned communication thread is successfully created, step S25 is executed, which provides conditions for subsequently ensuring that the system can smoothly enter the real-time adjustment mode in the automatic operating state.

[0072] Step S25, when it is determined to be in the automatic operation state and the communication thread is successfully created, the control module 2 continues to monitor the control information transmitted by the video acquisition and processing module 1. If no new control information is transmitted, step S25 is repeated until new control information is detected to be transmitted, otherwise step S26 is executed;

[0073] Step S26, the control module 2 analyzes the transmitted new control information and determines the type of the new control information. If it is fault information, step S27 is executed, otherwise step S28 is executed;

[0074] Step S27, when the new control information transmitted is fault information, it is determined that the video acquisition and processing module 1 has a fault. At this time, the control module 2 sends an automatic reset instruction to the video acquisition and processing module 1 to achieve reloading or creation of related links to ensure the normal operation of the video acquisition and processing module 1. At the same time, the signal light 4 is switched at a fixed beat according to the pre-set time interval in the system of the control module 2, and the process returns to step S25. It is worth noting that the fault information can be the prompt information of the failure to load the YOLO model, the prompt information of the failure of the communication connection, etc.

[0075] Step S28, when the transmitted new control information is non-fault information, that is, normal control information, the vehicle flow is counted, that is, the transmitted control information is analyzed, and the maximum vehicle flow of lanes of different lane types (straight lanes, left turn lanes and / or right turn lanes) in the north-south and east-west directions is calculated to obtain the vehicle flow information of different lanes in the north-south and east-west directions.

[0076] More specifically, when obtaining traffic flow information on different lanes in the north-south and east-west directions, it is necessary to first obtain the maximum traffic flow on lanes of different lane types in the forward direction of the north-south or east-west direction. If one lane type includes multiple lanes, then the maximum traffic flow of these multiple lanes is obtained as the traffic flow of this lane type in the forward direction. Similarly, obtain the maximum traffic flow on lanes of different lane types in the reverse direction of the north-south or east-west direction. If one lane type has multiple lanes, then the maximum traffic flow of these multiple lanes is obtained as the traffic flow of this lane type in the reverse direction. Finally, compare the maximum traffic flow on lanes of different lane types in the forward direction of the north-south direction with the maximum traffic flow on lanes of different lane types in the reverse direction, and take the largest value of the two as the traffic flow of different lanes in the north-south direction. Similarly, The maximum value of the traffic flow on the lanes of different lane types in the forward direction is compared with the maximum value of the traffic flow on the lanes of different lane types in the reverse direction, and the largest value of the two is taken as the traffic flow of different lanes in the east-west direction. That is, when obtaining the maximum value of the traffic flow in the north-south direction, first obtain the maximum value of the traffic flow on the lanes of different lane types in the south-north direction. If there are multiple lanes on a certain lane type, then obtain the maximum value of the traffic flow of these multiple lanes as the traffic flow of this lane type in the south-north direction. The same method is used to obtain the maximum value of the traffic flow on the lanes of different lane types in the north-south direction. Finally, compare the maximum value of the traffic flow on the lanes of different lane types in the south-north direction with the maximum value of the traffic flow on the lanes of different types in the north-south direction, and select the largest value of the two as the traffic flow information of different lanes in the north-south direction. The traffic flow information of different lanes in the east-west direction can also be obtained in the same way.

[0077] Step S29, after obtaining the traffic flow information of different lanes in the east-west and north-south directions, determine the state identification according to the obtained traffic flow information of different lanes in the north-south and east-west directions.

[0078] At this time, the status identification includes the following VIII categories, and the traffic flow information of different lanes corresponding to the east-west and north-south directions is as follows:

[0079] Class I status indicates that there are no through vehicles in the east-west and north-south directions, or there are only left-turning vehicles in the east-west and north-south directions;

[0080] Category II status indicates that there are vehicles going straight in the east-west direction, but no vehicles going straight in the north-south direction;

[0081] Category III status indicates that there are through vehicles in both the east-west and north-south directions, and the through vehicle flow in the east-west direction is greater than that in the north-south direction, and the difference between the two is greater than the preset value. It is worth pointing out that the preset value here is a real number set by the controller according to the actual road conditions;

[0082] Category IV status indicates that there are through vehicles in the north-south direction, but no through vehicles in the east-west direction;

[0083] Class V status indicates that there are through vehicles in both the north-south direction and the east-west direction, and the through vehicle flow in the north-south direction is greater than that in the east-west direction, and the difference between the two is greater than the preset value. Similarly, the preset value here is also a real number set by the controller according to the actual road conditions;

[0084] Category VI status indicates that there are through vehicles in both the north-south direction and the east-west direction, and the through vehicle flow in the north-south direction and the east-west direction is equal;

[0085] Category VII status indicates that there are only left-turning vehicles in the east-west direction or only left-turning vehicles in the north-south direction and there are vehicles in the other corresponding direction;

[0086] Class VIII status indicates that the traffic volume in the east-west direction and the north-south direction are not equal, and the difference between the two does not exceed the preset value. Similarly, the preset value here is also a real number set by the control personnel according to the actual road conditions.

[0087] By comparing and analyzing the traffic flow information of different lanes in different directions of the intersection, the state identification of the current intersection is obtained, which provides conditions for controlling the switching time of the traffic light 4 of the current intersection according to different state identifications.

[0088] More specifically, after determining the status identification of the intersection, the control module 2 controls the signal light 4 at the intersection according to the signal light 4 duration allocation in step S3, that is, allocates duration for the signal lights 4 in different lanes in the north-south and east-west directions according to the status identification obtained in step S29, and returns to execute step S25 after the allocation is completed, so as to meet the use requirements of adjusting the duration allocation of the signal light 4 in the next time period and ensure the continuity of the process.

[0089] More specifically, in step S3, the relationship between the state identifier and the duration allocated to the signal light 4 is:

[0090] When the status identification is Class I, Class VI, Class VII and Class VIII, the signal lights 4 in the east-west direction and the north-south direction are assigned equal durations;

[0091] When the status indicator is Class II, the signal light 4 in the east-west direction is assigned to be always green, and the signal light 4 in the north-south direction is assigned to be always red;

[0092] When the status mark is Class IV, the signal light 4 in the north-south direction is assigned to be always green, and the signal light 4 in the east-west direction is assigned to be always red;

[0093] When the status mark is Class III, the green light duration of the east-west direction signal light 4 is longer than the green light duration of the north-south direction signal light 4;

[0094] When the status identification is Class V, the green light duration of the signal light 4 in the north-south direction is greater than the green light duration of the signal light 4 in the east-west direction.

[0095] This ensures that the switching duration of the traffic light 4 at the intersection is adjusted in real time according to the traffic flow information of different lanes in different directions of the intersection, effectively avoiding the problem of traffic congestion in one direction and empty traffic in the other direction due to timed switching, thereby improving road traffic efficiency.

[0096] In addition, it is worth pointing out that when the status identification is determined to be Class I, Class VI, Class VII and Class VIII, the duration of the traffic lights 4 in the east-west direction and the north-south direction under different types of status identifications are different. Since the Class I, Class VI, Class VII and Class VIII status identifications cannot appear at the same time and appear independently, the duration of the traffic lights 4 in the east-west direction and the north-south direction under the individual Class I, Class VI, Class VII and Class VIII status identifications are equal. However, when the duration of the traffic lights 4 in the east-west direction and the north-south direction under the Class I, Class VI, Class VII and Class VIII status identifications are mixed and compared, there are still two situations where the duration of the traffic lights 4 is the same or different. The specific duration of the traffic lights 4 is set according to the actual situation of the intersection.

[0097] More specifically, when the status is determined to be Class III, the green light duration of the east-west traffic light 4 is set to m times the green light duration of the north-south traffic light 4, and m is a real number, which can ensure that the green light duration of the east-west traffic light 4 is much longer than the green light duration of the north-south traffic light 4, which is more in line with the actual road conditions where the traffic volume in the east-west direction is greater than the traffic volume in the north-south direction. Similarly, when the status is determined to be Class V, the green light duration of the north-south traffic light 4 is n times the green light duration of the east-west traffic light 4, and n is a real number, which can ensure that the green light duration of the north-south traffic light 4 is much longer than the green light duration of the east-west traffic light 4, which is more in line with the actual road conditions where the traffic volume in the north-south direction is greater than the traffic volume in the east-west direction.

[0098] The YOLO visual perception-based intelligent traffic control system and control method provided in this embodiment include a video acquisition and processing module 1 and a control module 2. The video acquisition and processing module 1 is connected to a camera 3 at an intersection to capture the captured video image, analyze and extract image feature information and convert it into control information. At the same time, the video acquisition and processing module 1 is communicated with the control module 2, and the control module 2 is connected to the traffic light 4 at the intersection, so that the control module 2 can receive the control information and adjust the duration of the traffic light 4 at the intersection in real time according to the control information, that is, the duration of the traffic light 4 at the corresponding intersection is controlled in real time according to the traffic volume of different lane types in different directions of the intersection, so as to realize intelligent dynamic adjustment of the duration of the traffic light 4 at the intersection, effectively avoiding the problem of traffic congestion in one direction and empty traffic in the other direction in actual road conditions, and improving the traffic efficiency of the road.

[0099] The above are only preferred embodiments of the present invention, and are not intended to limit the implementation methods and protection scope of the present invention. Those skilled in the art should be aware that all solutions obtained by equivalent substitutions and obvious changes made using the description and illustrations of the present invention should be included in the protection scope of the present invention.

Claims

1. A control method for an intelligent traffic control system based on YOLO visual perception, wherein the intelligent traffic control system based on YOLO visual perception includes a video acquisition and processing module, which is used to connect a camera installed at an intersection, load a YOLO model, dynamically capture and analyze video images, analyze and extract feature information of the video images, and convert the feature information of the video images into control information; And a control module, which is used to receive the control information input by the video acquisition and processing module, and complete the dynamic control of the intersection signal light according to the input control information, and the control method based on the YOLO visual perception intelligent traffic control system includes the following steps: Step S1, video acquisition and processing; Step S11, the video acquisition and processing module is powered on and starts to enter a working state; Step S12, initialize the system and load the YOLO model, and determine whether the YOLO model loading is successful, if successful, execute step S13, otherwise repeat step S12; Step S13, creating a video capture thread to obtain the video information of the camera at the intersection in real time, and judging whether the video capture thread is successfully created. If successful, executing step S14, otherwise repeating step S13; Step S14, creating a communication thread for the communication connection between the video acquisition and processing module and the control module, and determining whether the communication thread is successfully created. If successful, executing step S15, otherwise repeating step S14; Step S15, creating a YOLO processing thread, which is used to receive the video image information from the camera at the intersection transmitted by the video capture thread, and convert the received video image information into control information, and determine whether the YOLO processing thread is successfully created. If successful, execute step S2, otherwise repeat step S15; Step S2, analysis and control; Step S21, the control module is powered on and starts to enter a working state; Step S22, initializing the system, obtaining the setting information of the external device, and setting it as the initial information of the system in the control module, judging the manual or automatic operation state set in the initial information, if it is judged to be the manual operation state, executing step S23, if it is judged to be the automatic operation state, executing step S24; Step S23, switching the signal light at a fixed beat according to a time interval preset in the system of the control module, ignoring the control information transmitted by the communication thread, and repeatedly executing step S23 until the manual operation state is exited; Step S24, determining whether the communication between the video acquisition and processing module and the control module is normally connected. If not, switching the signal light at a fixed beat according to a time interval preset in the system of the control module, and repeating step S24. Meanwhile, continuously monitoring the communication connection status between the video acquisition and processing module and the control module until the communication connection status is normally connected, and then turning to step S25; Step S25, continuously monitoring the control information transmitted by the video acquisition and processing module, if no new control information is transmitted, repeating step S25, otherwise executing step S26; Step S26, analyzing the new control information and determining the type of the new control information. If the new control information is fault information, executing step S27, otherwise executing step S28; Step S27, determining that the video acquisition and processing module fails, sending an automatic reset instruction to the video acquisition and processing module, and at the same time, performing a fixed beat timed switching of the signal light according to a pre-set time interval in the system of the control module, and returning to step S25; Step S28, counting the vehicle flow, analyzing the transmitted control information, calculating the maximum value of the vehicle flow of lanes of different lane types in the north-south and east-west directions, and obtaining the vehicle flow information of different lanes in the north-south and east-west directions; Step S29, determining a state indicator according to the traffic flow information of different lanes in the north-south and east-west directions obtained in step S28; Step S3, allocating the duration of the signal light; According to the state identification obtained in step S29, durations are allocated to the traffic lights of different lanes in the north-south and east-west directions, and the process returns to step S25.

2. The control method of the intelligent traffic control system based on YOLO visual perception according to claim 1 is characterized in that: In step S12, a loading threshold of the YOLO model is preset. If the consecutive number of times the YOLO model is loaded is greater than the set loading threshold when step S12 is repeatedly executed, it is determined that the YOLO model loading fails, the system is exited, and a prompt message indicating that the YOLO model loading failed is issued.

3. The control method of the intelligent traffic control system based on YOLO visual perception according to claim 1 is characterized in that: In step S14, a communication threshold of the communication thread is preset. If the number of consecutive times of creating a communication thread is greater than the set communication threshold when step S14 is repeatedly executed, it is determined that the communication thread creation fails, the system is exited, and a prompt message of communication connection failure is issued.

4. The control method of the intelligent traffic control system based on YOLO visual perception according to claim 1 is characterized in that: In step S22, the external device includes a touch screen or a computer, and the control module system has a built-in time interval control program for setting the time interval and performing fixed-beat timed switching of the signal light.

5. The control method of the intelligent traffic control system based on YOLO visual perception according to claim 1 is characterized in that: In step S28, when determining the traffic flow information of different lanes in the north-south and east-west directions, first obtain the maximum value of the traffic flow on lanes of different lane types in the forward direction of the north-south or east-west direction. When there are multiple lanes on one lane type, obtain the maximum value of the traffic flow of these multiple lanes as the traffic flow of this lane type in the forward direction. Similarly, obtain the maximum value of the traffic flow on lanes of different lane types in the reverse direction of the north-south or east-west direction. When there are multiple lanes on one lane type, obtain the maximum value of the traffic flow of these multiple lanes as the traffic flow of this lane type in the reverse direction. Finally, compare the maximum value of the traffic flow on lanes of different lane types in the forward direction of the north-south direction with the maximum value of the traffic flow on lanes of different lane types in the reverse direction, and take the largest value of the two as the traffic flow of different lanes in the north-south direction. Similarly, compare the maximum value of the traffic flow on lanes of different lane types in the forward direction of the east-west direction with the maximum value of the traffic flow on lanes of different lane types in the reverse direction, and take the largest value of the two as the traffic flow of different lanes in the east-west direction.

6. The control method of the intelligent traffic control system based on YOLO visual perception according to claim 1 is characterized in that: In step S29, the status identification includes category VIII, Class I status indicates that there are no through vehicles in the east-west and north-south directions, or there are only left-turning vehicles in the east-west and north-south directions; Category II status indicates that there are vehicles going straight in the east-west direction, but no vehicles going straight in the north-south direction; Category III status indicates that there are through vehicles in both the east-west and north-south directions, and the through vehicle flow in the east-west direction is greater than that in the north-south direction, and the difference between the two is greater than the preset value; Category IV status indicates that there are through vehicles in the north-south direction, but no through vehicles in the east-west direction; Class V status indicates that there are through vehicles in both the north-south direction and the east-west direction, and the through vehicle flow in the north-south direction is greater than that in the east-west direction, and the difference between the two is greater than the preset value; Category VI status indicates that there are through vehicles in both the north-south direction and the east-west direction, and the through vehicle flow in the north-south direction and the east-west direction is equal; Category VII status indicates that there are only left-turning vehicles in the east-west direction or only left-turning vehicles in the north-south direction and there are vehicles in the other corresponding direction; Class VIII status indicates that the traffic volume in the east-west direction and the north-south direction are not equal, and the difference between the two does not exceed the preset value.

7. The control method of the intelligent traffic control system based on YOLO visual perception according to claim 6 is characterized in that: In step S3, the relationship between the state identifier and the allocated duration of the signal light is: When the status is marked as Class I, Class VI, Class VII, and Class VIII, the east-west and north-south directions are assigned equal durations; When the status mark is Class II, the east-west traffic lights are assigned to be always green, and the north-south traffic lights are assigned to be always red; When the status mark is Class IV, the north-south traffic lights are assigned to be always green, and the east-west traffic lights are assigned to be always red; When the status is marked as Class III, the green light duration of the east-west traffic light is longer than that of the north-south traffic light; When the status is identified as Class V, the green light duration of the traffic light in the north-south direction is longer than the green light duration of the traffic light in the east-west direction.

8. The control method of the intelligent traffic control system based on YOLO visual perception according to claim 7 is characterized in that: When the status identification is Class I, Class VI, Class VII and Class VIII, the duration of the traffic lights in the east-west direction and the north-south direction under different types of status identifications is different.

9. The control method of the intelligent traffic control system based on YOLO visual perception according to claim 7 is characterized in that: When the status is identified as Class III, the green light duration of the traffic light in the east-west direction is m times the green light duration of the traffic light in the north-south direction. When the status is identified as Class V, the green light duration of the traffic light in the north-south direction is n times the green light duration of the traffic light in the east-west direction, and both m and n are real numbers.

Images