Pedestrian and non-motor vehicle illegal real-time warning method and system

Abstract

The invention relates to the technical field of intelligent traffic, and particularly discloses a pedestrian and non-motor vehicle illegal real-time warning method and system, and the method comprises the steps: collecting the video stream and signal lamp timing information of an intersection in real time; identifying illegal behaviors of travelers and non-motor vehicles from the video stream of the intersection according to the signal lamp timing information of the intersection, and sending an early warning signal; and performing early warning according to the early warning signal. According to the pedestrian and non-motor vehicle law violation real-time warning method provided by the invention, traffic illegal behaviors can be effectively reduced, the traffic accident rate is reduced, the intelligent level and efficiency of urban traffic management are improved, and the life and property safety of the masses is guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation technology, and more specifically, to a real-time warning method for pedestrians and non-motor vehicles violating the law and a real-time warning system for pedestrians and non-motor vehicles violating the law. Background Art

[0002] With the acceleration of urbanization, the urban population and the number of vehicles have increased dramatically, and urban traffic problems have become increasingly prominent. As an important part of urban traffic, pedestrians and non-motor vehicles frequently commit traffic violations, which poses a serious threat to urban traffic order and safety. Traditional traffic management methods mainly rely on on-site law enforcement by traffic police, but due to factors such as insufficient police force and limited law enforcement time periods, it is impossible to achieve all-weather and all-time supervision, making it difficult to effectively curb traffic violations.

[0003] Traffic lights (traffic lights), as traffic control facilities, play a vital role in urban traffic. Violations of pedestrians and non-motor vehicles often lead to traffic congestion and affect road traffic efficiency. Through governance, the traffic behavior of pedestrians and non-motor vehicles can be regulated, traffic congestion can be reduced, and road traffic capacity can be improved. Traditional traffic law enforcement methods cannot achieve real-time and comprehensive supervision of pedestrian and non-motor vehicle violations. Therefore, an efficient and intelligent traffic management system is needed to meet these challenges, realize real-time detection, identification and warning of pedestrian and non-motor vehicle violations, and flexibly adjust according to real-time traffic flow and pedestrian crossing needs to improve traffic efficiency and safety. Summary of the invention

[0004] In order to solve the deficiencies in the prior art, the present invention provides a real-time warning method and system for pedestrian and non-motor vehicle violations to solve the problems existing in the existing traffic management system, improve the intelligence level and efficiency of urban traffic management, and protect the safety of people's lives and property.

[0005] As a first aspect of the present invention, a method for real-time warning of violations of pedestrians and non-motor vehicles is provided, comprising the following steps: Step S1: real-time acquisition of the video stream and traffic light timing information of the intersection; Step S2: Identify illegal behaviors of pedestrians and non-motor vehicles from the video stream of the intersection according to the signal light timing information of the intersection, and send a warning signal; Step S3: issuing an early warning according to the early warning signal.

[0006] Furthermore, the real-time collection of the video stream and the traffic light timing information of the intersection also includes: Collect the video stream of the intersection in real time through the camera; The signal light timing information at the intersection is collected in real time through the signal controller.

[0007] Furthermore, the step of identifying illegal behaviors of pedestrians and non-motor vehicles from the video stream of the intersection according to the signal light timing information of the intersection and sending an early warning signal also includes: When the straight-ahead signal light and the left-turn signal light on the main road turn red, real-time monitoring of pedestrians or non-motor vehicles in the video stream is performed; when it is detected that pedestrians or non-motor vehicles cross the line and wait during the red light period, an early warning signal is sent to the illegal warning device at the intersection to control the illegal warning device at the intersection to issue an early warning; wherein, when the straight-ahead signal light on the main road turns green, if the left-turning non-motor vehicle in the video stream crosses the line and waits, an early warning signal is sent to the illegal warning device at the intersection to control the illegal warning device at the intersection to issue an early warning; When the straight-ahead signal light on the main road is about to turn red, real-time monitoring of pedestrians or non-motor vehicles in the video stream is performed; when pedestrians or non-motor vehicles are detected in the zebra crossing, an early warning signal is sent to the intersection violation warning device to control the intersection violation warning device to issue an early warning; wherein, during the early warning, if the pedestrian is detected to be still in the middle of the zebra crossing, the green light time of the straight-ahead signal light is extended until all pedestrians pass through the zebra crossing; When a pedestrian or non-motor vehicle is identified from the video stream at the intersection as illegally entering the motor vehicle area, a warning signal is sent to the intersection violation warning device to control the intersection violation warning device to issue a warning; When a pedestrian or non-motor vehicle is identified from the video stream of the intersection waiting in front of the motor vehicle lane, the red light time of the signal light of the motor vehicle lane is temporarily extended, and a warning signal is sent to the illegal warning device of the intersection to control the illegal warning device of the intersection to issue a warning; at the same time, the signal light of the sidewalk is controlled to turn on the green light to guide pedestrians and non-motor vehicles to leave the motor vehicle lane and reach a safe position; When a possible conflict between a right-turning motor vehicle and a straight-moving pedestrian or straight-moving non-motor vehicle is identified from the video stream at the intersection, a warning signal is sent to the intersection violation warning device to control the intersection violation warning device to issue a warning.

[0008] Furthermore, the step of providing an early warning according to the early warning signal further includes: The intersection violation warning device includes a warning light, an IP network sound column and an electronic screen. When illegal behaviors of pedestrians and non-motor vehicles are identified, red or yellow light warning information is emitted through the warning light, voice warning information is emitted through the IP network sound column, and text warning information is broadcast through the electronic screen.

[0009] Furthermore, the real-time warning method for pedestrian and non-motor vehicle violations also includes: The illegal monitoring cloud platform collects traffic flow, pedestrian crossing times and weather data at intersections in real time; The early warning strategy is adaptively adjusted based on the traffic flow at the intersection, the number of pedestrian crossings and meteorological data, as well as the historical violation data of pedestrians and non-motor vehicles at the intersection.

[0010] As a second aspect of the present invention, a real-time warning system for pedestrian and non-motor vehicle violations is provided, comprising: Information collection unit, used to collect video streams and traffic light timing information of intersections in real time; The edge computing unit is used to identify illegal behaviors of pedestrians and non-motor vehicles from the video stream of the intersection based on the traffic light timing information at the intersection, and send a warning signal; The illegal warning unit is used to issue an early warning according to the early warning signal.

[0011] Furthermore, the edge computing unit includes an edge computing device and an AI violation recognition model. The edge computing device is deployed in a transformer box at an intersection and is used to run a modern neural network to process a video stream at the intersection. The AI ​​violation recognition model is installed in a system of the edge computing device and identifies the illegal behaviors of pedestrians and non-motor vehicles from the processed video stream in combination with the signal light timing information at the intersection, and sends a warning signal to the violation warning unit. Among them, when the straight-ahead traffic light on the main road turns green, the AI ​​violation recognition model can predict the movement trajectory of non-motor vehicles from the processed video stream to determine the non-motor vehicles that need to turn left. When the non-motor vehicle that needs to turn left in the video stream crosses the line and waits, an early warning signal is sent to the intersection violation warning device to control the intersection violation warning device to issue an early warning.

[0012] Furthermore, the violation warning unit includes an intersection violation warning device, which is deployed on a pedestrian crossing signal pole at an intersection. The intersection violation warning device includes a warning light, an IP network sound column and an electronic screen. When the edge computing unit identifies violations by pedestrians and non-motor vehicles, it controls the warning light to emit red or yellow light warning information, controls the IP network sound column to emit voice warning information, and controls the electronic screen to broadcast text warning information to remind pedestrians and non-motor vehicles of their violations.

[0013] Furthermore, the information collection unit includes a camera and a signal controller, the camera collects the video stream of the intersection in real time, the signal controller collects the signal light timing information of the intersection in real time, and the information collection unit transmits the video stream of the intersection to the edge computing unit through RTSP streaming technology.

[0014] Furthermore, the violation monitoring cloud platform is used to collect real-time traffic flow, pedestrian crossing times and meteorological data at the intersection; the edge computing unit is used to adaptively adjust the early warning strategy based on the traffic flow, pedestrian crossing times and meteorological data at the intersection and the historical violation data of pedestrians and non-motor vehicles at the intersection.

[0015] The real-time warning method for pedestrian and non-motor vehicle violations provided by the present invention has the following advantages: by real-time processing of existing camera (electronic police) video data at road intersections, artificial intelligence technology is used to identify pedestrian and non-motor vehicle violations in real time, and instant warnings are given in various forms such as sound, warning lights, text and images, thereby effectively reducing traffic violations, reducing the incidence of traffic accidents, improving the intelligence level and efficiency of urban traffic management, and protecting the lives and property of the people. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the present invention, but do not constitute a limitation of the present invention.

[0017] Figure 1 This is a flow chart of the real-time warning method for pedestrian and non-motor vehicle violations provided by the present invention.

[0018] Figure 2 This is a structural diagram of the real-time warning system for pedestrian and non-motor vehicle violations provided by the present invention. DETAILED DESCRIPTION

[0019] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation method, structure, characteristics and effects of the real-time warning method for pedestrian and non-motor vehicle violations proposed by the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments. Obviously, the described embodiments are part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] In this embodiment, a real-time warning method for pedestrian and non-motor vehicle violations is provided. Figure 1 As shown, the real-time warning method for pedestrian and non-motor vehicle violations includes: Step S1: real-time acquisition of the video stream and traffic light timing information of the intersection; Preferably, the real-time acquisition of the video stream and traffic light timing information of the intersection also includes: Collect the video stream of the intersection in real time through the camera; The signal light timing information at the intersection is collected in real time through the signal controller.

[0021] Step S2: Identify illegal behaviors of pedestrians and non-motor vehicles from the video stream of the intersection according to the signal light timing information of the intersection, and send a warning signal; Preferably, the step of identifying illegal behaviors of pedestrians and non-motor vehicles from a video stream at an intersection according to the signal light timing information at the intersection and sending an early warning signal further includes: (1) Warning for pedestrians and non-motor vehicles crossing the line: When the straight-ahead signal light and the left-turn signal light on the main road turn red, the signal control unit will quickly transmit the red light signal to the edge computing unit, which will then start the recognition function and conduct real-time monitoring of pedestrians or non-motor vehicles in the video stream. When it is detected that pedestrians or non-motor vehicles have crossed the line and are waiting during the red light period, the edge computing unit will immediately send instructions to the nearest IP network sound column to control it to make a voice broadcast: "Please return if you cross the line" to promptly remind and correct the violations of pedestrians and non-motor vehicles. In particular, when a vehicle is detected crossing the line in a certain direction, the electronic screen will display "Watch out for pedestrians (non-motor vehicles) turning left (going straight, turning right)" to remind motor vehicles to pay attention to potential accident risks. Among them, when the straight-ahead signal light on the main road turns green, if the left-turning non-motor vehicle in the video stream crosses the line and is waiting, a warning signal is sent to the illegal warning device at the intersection to control the illegal warning device at the intersection to issue a warning. Specifically, in the AI ​​violation recognition model, the trajectory prediction algorithm is used to determine the driving direction of pedestrians and non-motor vehicles going straight, turning left, or turning right, so as to distinguish between left-turn crossing and right-turn crossing in combination with the signal light timing, and prevent the AI ​​violation recognition model from making misidentifications or missed identifications (i.e., identifying left-turn crossings when the straight-ahead green light is on). At the same time, in order to more effectively identify pedestrians and non-motor vehicles traveling in other directions crossing the line during the green light period in the current direction, a small rectangular area is divided in front of the stop line. When pedestrians or non-motor vehicles stay in this area for more than 3 seconds, a voice broadcast of "crossing the line, please return" is also made. This measure can effectively prevent left-turning non-motor vehicles from crossing the stop line during the straight-ahead green light without a voice broadcast.

[0022] (2) Warning for pedestrians crossing a zebra crossing and running a red light: When the straight traffic light on the main road is about to turn red, the signal controller will also send the red light signal to the edge computing unit. The edge computing unit monitors pedestrians or non-motor vehicles in the video stream in real time, and pays special attention to the zebra crossing area. When pedestrians or non-motor vehicles are detected in the zebra crossing, the corresponding IP network sound column is sent a command to control it to make a voice broadcast: "The red light is about to turn red, please pass quickly"; in which, during the early warning, if the pedestrian is detected in the middle of the zebra crossing, the green light time of the straight traffic light is extended until all pedestrians pass through the zebra crossing, so as to effectively warn and guide pedestrians to comply with traffic rules. At the same time, the electronic screen displays "Watch out for pedestrians" to remind motor vehicle drivers to pay attention to potential accident risks.

[0023] (3) Warning for pedestrians and non-motor vehicles entering the motor vehicle lane: The edge computing unit has comprehensive road monitoring capabilities. When the edge computing unit identifies pedestrians or non-motor vehicles illegally entering the motor vehicle area from the video stream at the intersection, it quickly sends instructions to the nearby IP network sound column to control it to make a voice broadcast: "The motor vehicle lane is dangerous", so as to promptly remind pedestrians and non-motor vehicles to stay away from the motor vehicle lane and ensure road traffic safety. And if pedestrians or non-motor vehicles are identified from the video stream at the intersection waiting in front of the motor vehicle lane, the red light time of the motor vehicle lane signal light will be temporarily extended, and a voice broadcast will be made: "The motor vehicle lane is dangerous, please leave quickly". In order to attract the attention of motor vehicles, "Pay attention to pedestrians and non-motor vehicles on the motor vehicle lane" will be displayed on the electronic screen to warn motor vehicle drivers to observe their surroundings and pay attention to safety. At the same time, the signal lights on the sidewalk are controlled to turn green to guide pedestrians and non-motor vehicles to leave the motor vehicle lane and reach a safe location; (4) Right-turning motor vehicle conflict warning: In order to deal with possible conflicts between right-turning motor vehicles and pedestrians and non-motor vehicles, the edge computing unit will analyze the driving trajectory of right-turning motor vehicles and the location information of pedestrians and non-motor vehicles in real time. When a possible conflict between a right-turning motor vehicle and a straight-moving pedestrian or non-motor vehicle is identified from the video stream at the intersection, a warning signal is sent to the intersection violation warning device to control the intersection violation warning device to issue a warning.

[0024] Specifically, when a motor vehicle crosses the line in the non-motor vehicle right turn lane, and there are non-motor vehicles or pedestrians in the straight area of ​​the non-motor vehicle lane, a voice broadcast is made. When a potential conflict scenario is predicted, the edge computing unit will immediately trigger the corresponding IP network sound column to make a voice broadcast: "Watch out for right-turning vehicles", to remind the surrounding pedestrians and non-motor vehicles to avoid, thereby reducing the risk of traffic accidents. At the same time, the electronic screen displays "Watch out for right-turning pedestrians (non-motor vehicles)", reminding motor vehicle drivers to pay attention to observe potential accident risks.

[0025] Step S3: issuing an early warning according to the early warning signal.

[0026] Preferably, the step of issuing an early warning according to the early warning signal further includes: The intersection violation warning device includes a warning light, an IP network sound column and an electronic screen. When illegal behaviors of pedestrians and non-motor vehicles are identified, red or yellow light warning information is emitted through the warning light, voice warning information is emitted through the IP network sound column, and text warning information is broadcast through the electronic screen.

[0027] Preferably, the real-time warning method for pedestrian and non-motor vehicle violations also includes: The illegal monitoring cloud platform collects traffic flow, pedestrian crossing times and weather data at intersections in real time; The early warning strategy is adaptively adjusted based on the traffic flow at the intersection, the number of pedestrian crossings and meteorological data, as well as the historical violation data of pedestrians and non-motor vehicles at the intersection.

[0028] As another embodiment of the present invention, Figure 2 As shown, a real-time warning system for pedestrian and non-motor vehicle violations is provided, comprising: Information collection unit, used to collect video streams and traffic light timing information of intersections in real time; The edge computing unit is used to identify illegal behaviors of pedestrians and non-motor vehicles from the video stream of the intersection based on the traffic light timing information at the intersection, and send a warning signal; The illegal warning unit is used to issue an early warning according to the early warning signal.

[0029] Preferably, the edge computing unit includes an edge computing device and an AI violation recognition model. An edge computing device is deployed at each intersection. The edge computing device is deployed in a transformer box at the intersection and is used to run a modern neural network to process the video stream of the intersection. The edge computing device has the characteristics of high performance, compact appearance, and rich input / output interfaces, and is connected to a camera, a signal control machine, a sound column, an electronic screen, and a warning light. The AI ​​violation recognition model is installed in the system of the edge computing device, and identifies the illegal behaviors of pedestrians and non-motor vehicles from the processed video stream in combination with the signal light timing information of the intersection, and sends a warning signal to the violation warning unit. Among them, when the straight-ahead traffic light on the main road turns green, the AI ​​violation recognition model can predict the movement trajectory of non-motor vehicles from the processed video stream to determine the non-motor vehicles that need to turn left. When the non-motor vehicle that needs to turn left in the video stream crosses the line and waits, an early warning signal is sent to the intersection violation warning device to control the intersection violation warning device to issue an early warning.

[0030] It should be noted that the edge computing unit, as the core processing part, is embedded with an AI violation recognition model built on a mature ANN framework. This AI violation recognition model is trained and optimized for common pedestrian and non-motor vehicle violations at road intersections, and integrates traffic light signal timing information and ground line information to monitor the driving trajectories of pedestrians and non-motor vehicles in real time. When the AI ​​violation recognition model detects a violation, the edge computing unit immediately triggers the intersection violation warning device for an immediate warning.

[0031] Specifically, the edge computing unit uses DeepStream video analysis technology to obtain video data from the RTSP stream, and accelerates inference deployment through TensorRT to decode and preprocess the video data. The preprocessed data is then input into the artificial neural network ANN pre-trained model for target detection, identifying pedestrians, non-motor vehicles and other targets in the video, and performing object tracking.

[0032] Preferably, the violation warning unit includes an intersection violation warning device, and four violation warning devices are deployed at each intersection, respectively deployed on the pedestrian crossing signal light poles on the right side of the east entrance, south entrance, west entrance, and north entrance. The intersection violation warning device includes but is not limited to a warning light, an IP network sound column, and an electronic screen. When the edge computing unit identifies the illegal behavior of pedestrians and non-motor vehicles, it controls the warning light to emit red or yellow light warning information, controls the IP network sound column to emit voice warning information, and controls the electronic screen to broadcast text warning information to remind pedestrians and non-motor vehicles of illegal behavior.

[0033] Preferably, the information collection unit includes a camera and a signal controller, which are deployed at a road intersection. The camera collects the video stream of the intersection in real time, and the signal controller collects the signal light timing information of the intersection in real time. The information collection unit transmits the video stream of the intersection to the edge computing unit through RTSP streaming technology.

[0034] Preferably, the real-time warning system for pedestrian and non-motor vehicle violations also includes a violation monitoring cloud platform, which is used to collect real-time traffic flow, pedestrian crossing times and meteorological data at the intersection; the edge computing unit is used to adaptively adjust the warning strategy based on the traffic flow, pedestrian crossing times and meteorological data at the intersection and the historical violation data of pedestrians and non-motor vehicles at the intersection.

[0035] Specifically, the illegal monitoring cloud platform includes illegal monitoring software, which is used to comprehensively monitor the traffic order at multiple intersections and multiple areas. Remote monitoring and management are carried out through the illegal monitoring cloud platform. The illegal monitoring software provided by the illegal monitoring cloud platform can record and analyze illegal behavior data in real time, providing decision-making support and law enforcement basis for traffic management departments.

[0036] In the embodiment of the present invention, the illegal monitoring cloud platform continuously collects illegal data of the current intersection and can classify people of different age groups. When the illegal monitoring cloud platform shows that the pedestrian flow of a certain intersection has increased significantly during the morning and evening peak hours, or the current age group has a high proportion of illegal behaviors, or historical data shows that illegal behaviors occur frequently during this period, the system will automatically adjust the warning strategy of the intersection, increase the frequency of voice warnings, and enable more eye-catching warning lights and text image warnings to strengthen the warning effect. For example, during peak hours or in bad weather conditions, the system automatically increases the frequency and intensity of warnings; when there are many pedestrians crossing the street, the time of voice warnings is appropriately extended, etc. At the same time, if there are bad weather conditions such as rainy days or foggy days, the system will automatically increase the warning intensity according to the monitoring data to ensure that pedestrians and non-motor vehicles can receive warning information in time even when the line of sight is poor, so as to more effectively comply with traffic rules and ensure traffic safety.

[0037] The present invention proposes an innovative joint warning mechanism of sound, light, electricity and signal lights. When an illegal behavior is detected, the system activates the warning light deployed on the pedestrian crossing signal light pole, and forms a strong visual warning effect through red or yellow light and different flashing frequencies; at the same time, through the IP network speaker, concise and clear voice warnings such as "Please return if you cross the line" and "The red light is about to turn red, please pass quickly" are issued to remind pedestrians and non-motor vehicles to pay attention to and correct illegal behaviors; in addition, text warning information is broadcast in real time on the electronic screen, mainly to attract the attention of motor vehicles and prevent accidents between pedestrians and motor vehicles caused by illegal behaviors. The electronic screen is deployed in a conspicuous position, such as under the signal light pole, and can be clearly seen, so that the joint action of the warning light, IP network sound column and electronic screen can achieve a multi-sensory, multi-directional warning and safety protection effect.

[0038] The present invention makes full use of the existing electronic police surveillance cameras and transmits video data to the edge computing unit for processing through RTSP streaming technology. This highly integrated design not only saves the cost of front-end equipment deployment, but also simplifies the construction and installation process of the intersection. At the same time, video data transmission does not involve traffic police intranet data, ensuring network data security, which is conducive to the large-scale promotion and application of the model. At the same time, the efficient processing capability of the edge computing unit also ensures the real-time and stability of the system.

[0039] The present invention innovatively introduces a mechanism for intelligent adaptive adjustment of warning strategies. Through the illegal monitoring cloud platform, the system records and analyzes the traffic flow, weather conditions (including accessed meteorological data), pedestrian crossing frequency and historical illegal data at each intersection in real time. The system can automatically adjust the warning strategy and intensity based on these data. For example, during rush hours in the morning and evening or under severe weather conditions, the warning frequency at specific intersections can be automatically increased, and more intuitive warning lights and text and image warnings can be enabled, thereby achieving flexible and precise control of the warning strategy.

[0040] The real-time warning method for pedestrian and non-motor vehicle violations provided by the present invention utilizes artificial intelligence and Internet of Things technologies to perform real-time detection, identification and warning of pedestrian and non-motor vehicle violations at road intersections, which can improve the intelligence level and efficiency of urban traffic management and protect the lives and property of the people.

[0041] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technician familiar with this profession can make some changes or modifications to equivalent embodiments of equivalent changes using the technical contents disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the scope of the technical solution of the present invention.

Claims

1. A real-time warning method for pedestrian and non-motor vehicle violations, characterized in that: The following steps are involved: Step S1: real-time acquisition of the video stream and traffic light timing information of the intersection; Step S2: Identify illegal behaviors of pedestrians and non-motor vehicles from the video stream of the intersection according to the signal light timing information of the intersection, and send a warning signal; Step S3: issuing an early warning according to the early warning signal.

2. The real-time warning method for pedestrian and non-motor vehicle violations according to claim 1 is characterized in that: The real-time acquisition of the video stream and the signal light timing information at the intersection also includes: Collect the video stream of the intersection in real time through the camera; The signal light timing information at the intersection is collected in real time through the signal controller.

3. The real-time warning method for pedestrian and non-motor vehicle violations according to claim 1 is characterized in that: The method of identifying illegal behaviors of pedestrians and non-motor vehicles from the video stream of the intersection according to the signal light timing information of the intersection and sending an early warning signal also includes: When the straight-ahead signal light and the left-turn signal light on the main road turn red, real-time monitoring of pedestrians or non-motor vehicles in the video stream is performed; when it is detected that pedestrians or non-motor vehicles cross the line and wait during the red light period, an early warning signal is sent to the illegal warning device at the intersection to control the illegal warning device at the intersection to issue an early warning; wherein, when the straight-ahead signal light on the main road turns green, if the left-turning non-motor vehicle in the video stream crosses the line and waits, an early warning signal is sent to the illegal warning device at the intersection to control the illegal warning device at the intersection to issue an early warning; When the straight-ahead signal light on the main road is about to turn red, real-time monitoring of pedestrians or non-motor vehicles in the video stream is performed; when pedestrians or non-motor vehicles are detected in the zebra crossing, an early warning signal is sent to the intersection violation warning device to control the intersection violation warning device to issue an early warning; wherein, during the early warning, if the pedestrian is detected to be still in the middle of the zebra crossing, the green light time of the straight-ahead signal light is extended until all pedestrians pass through the zebra crossing; When a pedestrian or non-motor vehicle is identified from the video stream at the intersection as illegally entering the motor vehicle area, a warning signal is sent to the intersection violation warning device to control the intersection violation warning device to issue a warning; When a pedestrian or non-motor vehicle is identified from the video stream of the intersection waiting in front of the motor vehicle lane, the red light time of the signal light of the motor vehicle lane is temporarily extended, and a warning signal is sent to the illegal warning device of the intersection to control the illegal warning device of the intersection to issue a warning; at the same time, the signal light of the sidewalk is controlled to turn on the green light to guide pedestrians and non-motor vehicles to leave the motor vehicle lane and reach a safe position; When a possible conflict between a right-turning motor vehicle and a straight-moving pedestrian or straight-moving non-motor vehicle is identified from the video stream at the intersection, a warning signal is sent to the intersection violation warning device to control the intersection violation warning device to issue a warning.

4. The real-time warning method for pedestrian and non-motor vehicle violations according to claim 3 is characterized in that: The step of issuing an early warning according to the early warning signal further includes: The intersection violation warning device includes a warning light, an IP network sound column and an electronic screen. When illegal behaviors of pedestrians and non-motor vehicles are identified, red or yellow light warning information is emitted through the warning light, voice warning information is emitted through the IP network sound column, and text warning information is broadcast through the electronic screen.

5. The real-time warning method for pedestrian and non-motor vehicle violations according to claim 1 is characterized in that: The real-time warning method for pedestrian and non-motor vehicle violations also includes: The illegal monitoring cloud platform collects traffic flow, pedestrian crossing times and weather data at intersections in real time; The early warning strategy is adaptively adjusted based on the traffic flow at the intersection, the number of pedestrian crossings and meteorological data, as well as the historical violation data of pedestrians and non-motor vehicles at the intersection.

6. A real-time warning system for pedestrian and non-motor vehicle violations, used to implement the real-time warning method for pedestrian and non-motor vehicle violations as described in any one of claims 1 to 5, characterized in that: The real-time warning system for pedestrian and non-motor vehicle violations includes: Information collection unit, used to collect video streams and traffic light timing information of intersections in real time; The edge computing unit is used to identify illegal behaviors of pedestrians and non-motor vehicles from the video stream of the intersection based on the traffic light timing information at the intersection, and send a warning signal; The illegal warning unit is used to issue an early warning according to the early warning signal.

7. The real-time warning system for pedestrian and non-motor vehicle violations according to claim 6 is characterized in that: The edge computing unit includes an edge computing device and an AI violation recognition model. The edge computing device is deployed in a transformer box at an intersection and is used to run a modern neural network to process a video stream at the intersection. The AI ​​violation recognition model is installed in a system of the edge computing device and identifies the illegal behaviors of pedestrians and non-motor vehicles from the processed video stream in combination with the signal light timing information at the intersection, and sends a warning signal to the violation warning unit. Among them, when the straight-ahead traffic light on the main road turns green, the AI ​​violation recognition model can predict the movement trajectory of non-motor vehicles from the processed video stream to determine the non-motor vehicles that need to turn left. When the non-motor vehicle that needs to turn left in the video stream crosses the line and waits, an early warning signal is sent to the intersection violation warning device to control the intersection violation warning device to issue an early warning.

8. The real-time warning system for pedestrian and non-motor vehicle violations according to claim 6 is characterized in that: The violation warning unit includes an intersection violation warning device, which is deployed on a pedestrian crossing signal lamp pole at an intersection. The intersection violation warning device includes a warning light, an IP network sound column and an electronic screen. When the edge computing unit identifies illegal behaviors of pedestrians and non-motor vehicles, it controls the warning light to emit red or yellow light warning information, controls the IP network sound column to emit voice warning information, and controls the electronic screen to broadcast text warning information to remind pedestrians and non-motor vehicles of illegal behaviors.

9. The real-time warning system for pedestrian and non-motor vehicle violations according to claim 6 is characterized in that: The information collection unit includes a camera and a signal controller. The camera collects the video stream of the intersection in real time, and the signal controller collects the traffic light timing information of the intersection in real time. The information collection unit transmits the video stream of the intersection to the edge computing unit through RTSP streaming technology.

10. The real-time warning system for pedestrian and non-motor vehicle violations according to claim 6 is characterized in that: The real-time warning system for pedestrian and non-motor vehicle violations also includes a violation monitoring cloud platform, which is used to collect traffic flow, pedestrian crossing times and meteorological data at intersections in real time; the edge computing unit is used to adaptively adjust the warning strategy based on the traffic flow, pedestrian crossing times and meteorological data at the intersection and the historical violation data of pedestrians and non-motor vehicles at the intersection.

Images