Vehicle violation event detection method

Abstract

This application provides a vehicle violation detection method, a violation detection system, an electronic device, and a computationally readable storage medium. The vehicle violation detection method utilizes the collaborative control of a cloud server, an edge computing device, and a front-end intelligent device to detect violations in video images of the vehicle under test, and effectively analyzes the detection results. This ensures the security of the original data, improves data utilization, increases the detection capability for different types of vehicle violations, and ultimately achieves effective detection of violation types.

Description

Technical Field

[0001] This application relates to the field of vehicle violation detection, and in particular to a method for detecting vehicle violations, a violation detection system, an electronic device, and a computer-readable storage medium. Background Technology

[0002] In recent years, with the accelerating pace of urbanization, the demand for urban transportation has increased year by year, leading to a continuous rise in the number of different types of vehicles and consequently, an increase in safety hazards and traffic violations. These include private cars using counterfeit license plates, hit-and-run accidents, and vehicle theft; and dump trucks failing to cover their cargo, spilling it into drains, and operating without permits. Different types of vehicles exhibit a wide variety of violations. Therefore, identifying target vehicles and conducting real-time detection and control of their violations is crucial.

[0003] Current designs primarily rely on video and image data collected by various front-end intelligent devices, which are then used by neural network models to continuously train and learn event types to obtain recognition results. Compared to traditional manual methods that rely on captured images and surveillance videos to identify and verify illegal vehicles, this method improves both accuracy and detection efficiency. However, it suffers from high training costs and computational demands, requiring a large amount of sample data to train the neural network. The training and convergence of the neural network place extremely high demands on computing power and storage resources. Meanwhile, cloud servers possess powerful computing capabilities and abundant storage resources, providing computational support and the ability to store massive amounts of data. However, cloud server computing is susceptible to network connectivity issues, and when communication quality cannot be guaranteed, it can easily lead to the leakage or loss of raw data. Summary of the Invention

[0004] This application provides a method, system, electronic device, and computationally readable storage medium for detecting vehicle violations, enabling effective detection of violation types.

[0005] To address the aforementioned technical problems, this application adopts the following technical solution: A vehicle violation detection method is provided, applied to a violation detection system. The violation detection system includes a front-end intelligent device, an edge computing device, and a cloud server. The vehicle violation detection method includes: the front-end intelligent device acquiring video images of a vehicle under test; the front-end intelligent device performing violation detection on the video images and generating a first detection result, the first detection result including a first violation event identification result and its first credibility; the edge computing device acquiring the video images from the front-end intelligent device and performing violation detection on the video images, generating a second detection result; when the first credibility of the first violation event identification result is greater than or equal to a first preset threshold, the cloud server receives the first detection result from the front-end intelligent device and the second detection result from the edge computing device, and generates a violation event detection result based on the first and second detection results; when the first credibility of the first violation event identification result is less than the first preset threshold, the cloud server receives the second detection result from the edge computing device, and generates a violation event detection result based on the second detection result.

[0006] The edge computing device performs violation event detection on the video image and generates a second detection result, including: the edge computing device acquiring a first detection result; when the first confidence level of the first violation event identification result is less than a first preset threshold, the edge computing device performs violation event detection on the video image based on the first violation event and the second violation event and generates a second detection result, wherein the first violation event and the second violation event are different.

[0007] The process includes, after the front-end intelligent device performs violation event detection on the video image and generates a first detection result, the following steps are also included: the front-end intelligent device sends violation event detection results with a first confidence level greater than or equal to a first preset threshold from the first detection result to the edge computing device; the edge computing device acquires the video image from the front-end intelligent device and performs violation event detection on the video image to generate a second detection result, including: the edge computing device acquires the video image and the violation event detection result from the front-end intelligent device; the edge computing device performs violation event detection on the video image for a third violation event based on the violation event detection result, and generates a second detection result; wherein the third violation event is a violation event other than the violation event detection result.

[0008] The process of detecting violations in video images and generating a second detection result further includes: the edge computing device sending the violation detection result with a second confidence level greater than or equal to a second preset threshold from the second detection result to the cloud server.

[0009] The edge computing device includes a first edge computing device and a second edge computing device. The edge computing device acquires video images from a front-end intelligent device, performs violation event detection on the video images, and generates a second detection result. This includes: the first edge computing device acquires video images from the front-end intelligent device, performs violation event detection on the video images, and generates a third detection result; the second edge computing device acquires the video images from the first edge computing device and the third detection result, performs violation event detection on the video images based on the third detection result, and generates a second detection result.

[0010] The third detection result includes the second violation event identification result and its third credibility; violation event detection is performed on the video image based on the third detection result, and a second detection result is generated, including: the second edge computing device determines the detected violation events with a third credibility greater than or equal to a third preset threshold based on the third detection result; the second edge computing device performs violation event detection on the video image for undetected violation events based on the detected violation events, and generates a second detection result.

[0011] This also includes: cloud servers performing error analysis on violation detection results to obtain error analysis results; and edge computing devices performing correction processing and model training on violation detection results based on error analysis results.

[0012] To address the aforementioned technical problems, this application adopts the following technical solution: A violation detection system is provided, comprising: a front-end intelligent device for acquiring video images of a vehicle under test, further configured to perform violation event detection on the video images and generate a first detection result, the first detection result including a first violation event identification result and its first credibility; an edge computing device connected to the front-end intelligent device, configured to acquire the video images from the front-end intelligent device, perform violation event detection on the video images, and generate a second detection result; and a cloud server connected to both the front-end intelligent device and the edge computing device. When the first credibility of the first violation event identification result is greater than or equal to a first preset threshold, the cloud server receives the first detection result from the front-end intelligent device and the second detection result from the edge computing device, and generates a violation event detection result based on the first and second detection results; when the first credibility of the first violation event identification result is less than the first preset threshold, the cloud server receives the second detection result from the edge computing device, and generates a violation event detection result based on the second detection result.

[0013] To solve the above-mentioned technical problems, the technical solution adopted in this application is: to provide an electronic device, including a processor and a memory connected to the processor, the memory storing program instructions, and the processor executing the program instructions to implement the vehicle violation event detection method described above.

[0014] To solve the above-mentioned technical problems, the technical solution adopted in this application is: to provide a computer-readable storage medium storing program instructions, which, when executed by a processor, can implement the vehicle violation event detection method described above.

[0015] The beneficial effects of the embodiments of this application are as follows: The vehicle violation event detection method of this application, through the collaborative control of cloud server, edge computing device and front-end intelligent device, performs violation event detection on video images of the vehicle under test, and effectively analyzes the detection results, thereby ensuring the security of the original data, improving the utilization rate of data, increasing the detection capability of vehicle violation event types, and thus achieving effective detection of violation event types. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the steps of an embodiment of the vehicle violation detection method of this application;

[0017] Figure 2 This application Figure 1 A flowchart illustrating step S300 of one embodiment;

[0018] Figure 3 This application Figure 1 A flowchart illustrating another embodiment of step S300;

[0019] Figure 4 This application Figure 1 A flowchart illustrating another embodiment of step S300;

[0020] Figure 5 This is a flowchart illustrating the steps of another embodiment of the vehicle violation detection method of this application;

[0021] Figure 6 This is a schematic diagram of the structure of an embodiment of the violation detection system of this application;

[0022] Figure 7 This is a schematic diagram of another embodiment of the violation detection system of this application;

[0023] Figure 8 This is a schematic diagram of the structure of an embodiment of the computer-readable storage medium of this application. Detailed Implementation

[0024] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0025] The terms "first" and "second" in this application are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise expressly specified. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the steps or units listed, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such processes, methods, products, or apparatus.

[0026] This application provides a method for detecting vehicle violations, wherein the method is applied to a violation detection system, the violation detection system including a front-end intelligent device, an edge computing device, and a cloud server. Figure 1 As shown, Figure 1 This is a schematic flowchart illustrating the steps of an embodiment of the vehicle violation detection method of this application; the vehicle violation detection method can... Figure 1 The method steps shown are implemented, specifically including steps S100 to S400.

[0027] Step S100: The front-end intelligent device acquires video images of the vehicle under test.

[0028] The front-end intelligent device is an intelligent device used to collect video images of the vehicle under test. In this embodiment, the front-end intelligent device includes devices such as intelligent cameras and electronic dogs installed on the road.

[0029] Specifically, in this embodiment, multiple front-end intelligent devices are set up at road checkpoints and other locations within the detection range. Each intelligent front-end device acquires video images of the vehicle to be tested at the corresponding road checkpoint and performs violation detection on the vehicle to be tested.

[0030] Step S200: The front-end intelligent device performs violation event detection on the video image and generates a first detection result, which includes the first violation event identification result and its first credibility.

[0031] After the front-end intelligent device acquires the video image of the vehicle under test through the above steps, it performs the first violation event detection on the video image to obtain the violation events that the front-end intelligent device can identify and their credibility from the video image, as well as the first violation event identification result and its first credibility.

[0032] It is worth noting that the aforementioned "first violation event" refers to violations that can be identified by front-end intelligent devices, and does not specifically refer to a certain type of violation event. Specifically, the front-end intelligent device can identify and judge different numbers of violations of different categories based on its analytical processing capabilities, and obtain the identification result and its first level of credibility for each violation event.

[0033] For example, in this embodiment, multiple front-end intelligent devices are set up at road checkpoints and other locations within the detection range. After the front-end intelligent devices capture video image data of the vehicle to be tested, if it is an ordinary front-end intelligent device, it can only provide basic video image acquisition functions and has no processing and analysis capabilities. Therefore, its first confidence level for the identification result of the first violation event is set to zero. If it is a front-end intelligent device that integrates some detection algorithms, it can directly process and analyze the captured video image data to obtain a set of violation events and its first confidence level set.

[0034] After obtaining the first detection result through step S200, the front-end intelligent device further performs a credibility analysis on the first detection result to determine whether the violation event in the first detection result meets the credibility requirements, and sends the first detection result that meets the credibility requirements to the cloud server.

[0035] Step S300: The edge computing device acquires video images from the front-end smart device, performs violation event detection on the video images, and generates a second detection result.

[0036] Specifically, the front-end intelligent device collects video images of the vehicle under test and sends them to the edge computing device. The edge computing device performs a second violation event detection on the video images to obtain the violation events that the edge computing device can identify and their credibility, i.e., the aforementioned second detection result. The violation events that the edge computing device can identify include those that the front-end intelligent device can identify, i.e., the first violation events. Simultaneously, the violation events that the edge computing device can identify may also include those that the front-end intelligent device cannot identify.

[0037] Optionally, such as Figure 2 As shown, Figure 2 This application Figure 1 A flowchart illustrating one embodiment of step S300. Step S300 can be achieved through... Figure 2 The method steps shown are implemented, specifically including steps S301 to S304.

[0038] Step S301: The edge computing device acquires the first detection result and video image.

[0039] After generating the first detection result, the front-end intelligent device sends the first detection result and video image to the edge computing device. The first detection result sent by the front-end intelligent device to the edge computing device includes the first level of confidence of the first violation event identification result.

[0040] Step S302: The edge computing device determines whether the first credibility of the first violation event identification result is less than the first preset threshold.

[0041] The edge computing device receives the first detection result and further determines whether the first feasibility of the first violation event identification result is less than a first preset threshold, thereby determining whether the detection result of the front-end intelligent device meets the credibility requirements. The first preset threshold is an empirical value for judging whether the detection result of the front-end intelligent device meets the credibility requirements; it can be specifically limited according to actual circumstances, which will be described in detail here.

[0042] Step S303: When the first confidence level of the first violation event identification result is less than the first preset threshold, the edge computing device performs violation event detection on the video image based on the first violation event and the second violation event, and generates a second detection result, wherein the first violation event and the second violation event are different.

[0043] When the credibility of the first violation event identification result is less than a first preset threshold, the edge computing device determines that the front-end intelligent device's identification result of the first violation event does not meet the credibility requirement, that is, the violation event identified by the front-end intelligent device is an unreliable event. Furthermore, when the edge computing device performs violation event detection on video images, it will identify the first violation event while simultaneously identifying the second violation event. Through this method, the edge computing device can re-detect the first violation event in the video image to improve the credibility of the violation event. At the same time, the edge computing device also detects the second violation event in the video image, thereby effectively enhancing the types of violation events detected by the violation detection system.

[0044] It is worth noting that the second violation event refers to a violation event that can be identified by the edge computing device. In this embodiment, the violation events that can be identified by the edge computing device include not only violation events that can be identified by the front-end smart device, but also other types of violation events.

[0045] Step S304: When the first confidence level of the first violation event identification result is greater than or equal to the first preset threshold, the edge computing device performs violation event detection on the video image based on the second violation event and generates a second detection result.

[0046] When the first credibility of the first violation event identification result is greater than or equal to the first preset threshold, the edge computing device determines that the front-end intelligent device's identification result of the first violation event meets the credibility requirement, that is, the violation event identified by the front-end intelligent device is a credible event. Furthermore, the edge computing device will only detect whether there is a second violation event in the video image, without having to repeat the detection of the first violation event, thereby effectively reducing the computational difficulty of the edge computing device.

[0047] Optionally, such as Figure 3 As shown, Figure 3 This application Figure 1 A flowchart illustrating another embodiment of step S300. Step S300 can be achieved through... Figure 3 The method steps shown are implemented, specifically including steps S305 to S306.

[0048] Step S305: Obtain trusted event video images from the front-end smart device.

[0049] In this embodiment, after the front-end intelligent device detects violations in the video image and generates a first detection result, it performs credibility classification processing on the first detection result. In other words, the front-end intelligent device compares the first credibility of the first violation event identification result with a first preset threshold. Violation events with a first credibility less than the first preset threshold are unreliable and are classified as unreliable events by the front-end intelligent device. Conversely, violation events with a first credibility greater than or equal to the first preset threshold are reliable and are classified as reliable events by the front-end intelligent device. Further, the front-end intelligent device sends the reliable events to the edge computing device.

[0050] Step S306: The edge computing device performs violation event detection on the video image based on the trusted event for the third violation event, and generates a second detection result; wherein, the third violation event is an event other than the trusted event.

[0051] Specifically, in this embodiment, when the edge computing device performs violation event detection on video images, it directly detects violation events other than trusted events (i.e., it directly detects the third violation event) and uses the detection result of the third violation event as the second detection result. This eliminates the need for excessive calculation and analysis of the feasibility of the front-end intelligent device, thereby effectively improving the detection efficiency of the edge computing device. The third violation event is a violation event that the edge computing device can identify.

[0052] Optionally, the edge computing device includes a first edge computing device and a second edge computing device. The inclusion of a first edge computing device and a second edge computing device can be understood as the violation detection system comprising multiple edge computing devices (i.e., a first edge computing device and a second edge computing device), which are cascaded sequentially. The first edge computing device is directly connected to the front-end intelligent device, while the remaining edge computing devices (e.g., the second edge computing device) are sequentially connected and indirectly connected to the front-end intelligent device through the first edge computing device. The following example, using a violation detection system comprising two edge computing devices (a first edge computing device and a second edge computing device, where the second edge computing device is the end of the cascade), illustrates another embodiment of the method for implementing step S300. Figure 4 As shown, Figure 4 This application Figure 1 A flowchart illustrating another embodiment of step S300. Step S300 can be achieved through... Figure 3 The method steps shown are implemented, specifically including steps S307 to S308.

[0053] Step S307: The first edge computing device acquires video images from the front-end smart device, performs violation event detection on the video images, and generates a third detection result.

[0054] In the cascading relationship of all edge computing devices, the first edge computing device that directly communicates with the front-end intelligent device acquires video images from the front-end intelligent device, performs violation event detection on the video images, and generates a third detection result. For details, please refer to the method of the edge computing device generating the second detection result in steps S301 to S304, or the method of the edge computing device generating the second detection result in steps S305 to S306. These details will not be elaborated here.

[0055] The third detection result is the detection result and credibility of the violation events that the first edge computing device can identify. Optionally, after obtaining the third detection result, the first edge computing device will also perform credibility analysis on the third detection result and send the violation event detection results that meet the credibility requirements in the third detection result to the cloud server, so that the cloud server can generate violation event detection results.

[0056] Step S308: The second edge computing device acquires the video image and the third detection result of the first edge computing device, performs violation event detection on the video image based on the third detection result, and generates a second detection result.

[0057] Specifically, the third detection result includes the second violation event identification result and its third credibility. The second violation event identification result is the identification result of the violation event identified by the previous edge computing device (here referring to the first edge computing device) cascaded with the second edge computing device.

[0058] The second edge computing device acquires video images from the front-end intelligent device via the first edge computing device, and simultaneously acquires the third detection result from the first edge computing device. Based on the third detection result, it performs violation event detection on the video images and generates a second detection result. Specifically, the second edge computing device determines detected violation events with a third confidence level greater than or equal to a third preset threshold based on the third detection result. This can be understood as the second edge computing device determining all violation events detected by its cascaded upstream edge computing devices based on the third detection result, i.e., determining the detected violation events.

[0059] Furthermore, the second edge computing device performs violation detection on the video image based on the detected violations, generating a second detection result. Specifically, the second edge computing device performs credibility analysis on the detected violations and performs violation detection on the video image for violations other than those meeting the credibility requirements, to generate a second detection result. In other words, the edge computing device at the next level in the cascade performs violation detection on the video image based on the violations detected by all the preceding edge computing devices. The second detection result refers to the detection result generated by the edge computing device at the end of the cascade.

[0060] Through the above methods, edge computing devices achieve a cascading effect, thereby effectively improving the credibility of violation events and increasing the types of violation events detected by the violation detection system.

[0061] Step S400: The cloud server generates violation event detection results based on the detection results of the front-end intelligent devices and edge computing devices.

[0062] In this embodiment, after the front-end intelligent device completes step S200, it performs a credibility analysis on the first detection result. When the first credibility of the first violation event identification result is greater than or equal to the first preset threshold, the cloud server receives the first detection result from the front-end intelligent device and the second detection result from the edge computing device, and generates a violation event detection result based on the first and second detection results. When the first credibility of the first violation event identification result is less than the first preset threshold, the cloud server receives the second detection result from the edge computing device, and generates a violation event detection result based on the second detection result.

[0063] In other words, the front-end intelligent device only sends violation event detection results with a first credibility level greater than or equal to a first preset threshold to the cloud server; violation event detection results with a first credibility level less than the first preset threshold are not sent to the cloud server. After the front-end intelligent device performs credibility analysis, the edge computing device performs credibility analysis on the second detection result after executing step S300. The edge computing device only sends violation event detection results with a second feasibility level greater than or equal to a second preset threshold to the cloud server; violation event detection results with a second credibility level less than the second preset threshold are not sent to the cloud server. In summary, the edge computing device and the front-end intelligent device only send the first and second detection results that meet the feasibility requirements to the cloud server, so that the server can directly obtain the violation events of the vehicle under test based on the first and second detection results sent by the front-end intelligent device and the edge computing device, thereby generating the violation event detection result for the vehicle under test.

[0064] Specifically, the cloud server assimilates diverse data (here, diverse data refers to detection results that meet credibility requirements obtained from edge computing devices and front-end intelligent devices), and provides upper-layer application platforms with functions such as data display, processing, and presentation through a built-in analytical model. Furthermore, managers process events through an information management platform, and the cloud server performs error analysis on the violation event detection results to obtain error analysis results. Further, the cloud server synchronously returns the error-analyzed violation event detection results to the edge computing device, which then corrects the violation event detection results based on the error analysis results. During this correction, the edge computing device can also optimize model training, improving the accuracy of the detection results.

[0065] Unlike existing technologies, the vehicle violation detection method of this application uses the collaborative control of cloud servers, edge computing devices and front-end intelligent devices to detect violations in video images of the vehicle under test, and effectively analyzes the detection results, thereby ensuring the security of the original data, improving the utilization rate of data, increasing the detection capability of vehicle violation types, and thus achieving effective detection of violation types.

[0066] This application also proposes another embodiment of the vehicle violation detection method, such as... Figure 5 As shown, Figure 5 This is a flowchart illustrating the steps of another embodiment of the vehicle violation detection method of this application; the vehicle violation detection method can be... Figure 5 The method steps shown are implemented, specifically including steps S500 to S1400.

[0067] Step S500: The front-end intelligent device acquires video images of the vehicle under test, performs violation event detection on the video images, and generates a first detection result, which includes the first violation event identification result and its first credibility.

[0068] For specific implementation methods, please refer to steps S100 to S200 of the above embodiments, which will not be repeated here.

[0069] Step S600: The front-end intelligent device performs a credibility analysis on the first detection result.

[0070] Step S700: The front-end intelligent device sends the violation event detection results that meet the credibility requirements in the first detection result to the cloud server.

[0071] Steps S600 to S700 will be described together.

[0072] After executing step S500, the front-end intelligent device will perform a credibility analysis on the first detection result. The front-end intelligent device will only send the violation event detection results with a first credibility greater than or equal to the first preset threshold to the cloud server, and will not send the violation event detection results with a first credibility less than the first preset threshold to the cloud server, thereby effectively reducing the computational load of the cloud server and improving the generation efficiency of violation event detection results.

[0073] Step S800: The edge computing device acquires the results of the credibility analysis of the first detection result by the front-end intelligent device and the video image.

[0074] For specific implementation details, please refer to step S301 of the above embodiment, which will not be repeated here.

[0075] Step S900: The edge computing device determines whether the first credibility of the first violation event identification result is less than the first preset threshold.

[0076] For specific implementation details, please refer to step S302 of the above embodiment, which will not be repeated here.

[0077] Step S1000: When the first confidence level of the first violation event identification result is less than the first preset threshold, the edge computing device performs violation event detection on the video image based on the first violation event and the second violation event, and generates a second detection result, wherein the first violation event and the second violation event are different.

[0078] For specific implementation details, please refer to step S303 of the above embodiment, which will not be repeated here.

[0079] Step S1100: When the first confidence level of the first violation event identification result is greater than or equal to the first preset threshold, the edge computing device performs violation event detection on the video image based on the second violation event and generates a second detection result.

[0080] For specific implementation details, please refer to step S304 of the above embodiment, which will not be repeated here.

[0081] Step S1200: The edge computing device performs a credibility analysis on the second detection result.

[0082] Step S1300: The edge computing device sends the violation event detection results that meet the credibility requirements in the second detection results to the cloud server.

[0083] Steps S1200 to S1300 will be described together.

[0084] After the edge computing device generates the corresponding second detection result through steps S1000 and S1100, the edge computing device performs a credibility analysis on the second detection result. Specifically, the edge computing device only sends violation event detection results with a second feasibility greater than or equal to a second preset threshold to the cloud server, while violation detection results with a second credibility less than the second preset threshold are not sent to the cloud server. This effectively reduces the computational load on the cloud server and improves the generation efficiency of violation event detection results.

[0085] Step S1400: The cloud server generates violation event detection results based on the detection results of the front-end intelligent devices and edge computing devices.

[0086] For specific implementation details, please refer to step S400 of the above embodiment, which will not be repeated here.

[0087] Through the above methods, the violation detection system achieves effective detection of violation event types through the collaborative control of cloud servers, edge computing devices, and front-end intelligent devices. It also fully utilizes the data collection and processing capabilities of front-end intelligent devices and the event analysis capabilities of edge computing devices, progressively improving the credibility of event detection and enhancing the types of violation events detected. At the same time, it leverages the powerful computing and storage capabilities of cloud servers to achieve the fusion, storage, and exposure of diverse data, ensuring the security of raw data, improving the utilization rate of raw data and the timeliness of event detection, and enhancing the ability to detect different types of vehicle violation events.

[0088] This application provides a violation detection system, such as Figure 6 As shown, Figure 6This is a schematic diagram of an embodiment of the violation detection system of this application. The violation detection system 10 includes a front-end intelligent device 100, an edge computing device 200, and a cloud server 300.

[0089] The front-end intelligent device 100 is used to collect video images of the vehicle under test. The front-end intelligent device 100 is also used to detect violations in the video images and generate a first detection result, which includes a first violation event identification result and its first credibility. For example, in this embodiment, the front-end intelligent device 100 includes various common bullet cameras, PTZ cameras, etc., used to capture and collect images and videos of vehicles passing through roads and checkpoints in real time. It also includes a front-end intelligent device 100 integrating some intelligent detection algorithms, such as a city management PTZ camera, which can identify and process the captured information of some specific vehicles to obtain a first-layer detection result and its first credibility set (and the aforementioned first detection result and its first credibility set).

[0090] The edge computing device 200 is connected to the front-end intelligent device 100. The edge computing device 200 is used to acquire video images from the front-end intelligent device 100, perform violation event detection on the video images, and generate a second detection result.

[0091] The cloud server 300 is connected to the front-end intelligent device 100 and the edge computing device 200. When the first credibility of the first violation event identification result is greater than or equal to the first preset threshold, the cloud server 300 receives the first detection result from the front-end intelligent device 100 and the second detection result from the edge computing device 200, and generates a violation event detection result based on the first detection result and the second detection result. When the first credibility of the first violation event identification result is less than the first preset threshold, the cloud server 300 receives the second detection result from the edge computing device 200, and generates a violation event detection result based on the second detection result.

[0092] Specifically, the cloud server 300 is connected via a network to at least one edge computing device 200 and multiple front-end intelligent devices 100 located within the detection range and connected to the at least one edge computing device 200. The edge computing device 200 can connect to the cloud server through various remote communication methods, such as BeiDou satellite communication, Ethernet transmission, and private wireless networks. The edge computing device 200 primarily performs intelligent detection of violations by target vehicles. Based on video image data of the vehicle under test collected by the front-end intelligent devices 100 within the detection range, it performs intelligent identification and early warning of violations related to the vehicle under test and transmits the analysis results to the cloud server 300. The cloud server 300 can be a single server in the cloud or a server cluster composed of multiple servers. The cloud server 300 receives processed data from the edge devices, aggregates and integrates diverse data, and provides global, long-term big data processing and analysis to support real-time intelligent decision-making and execution of local business operations.

[0093] Optionally, such as Figure 7 As shown, Figure 7 This is a schematic diagram of another embodiment of the violation detection system of this application. Depending on different business needs, the edge computing device 200 can be extended to n-1 layers to achieve a cascading effect, improve the types of violation events detected, and enhance the reliability of data processing. For example, in this embodiment, the violation detection system 10 includes multiple edge computing devices 200 (i.e., n-1 edge computing devices 200, such as the first edge computing device 210...the (n-1)th edge computing device 220). Each edge computing device 200 is sequentially connected to the front-end intelligent device 100, forming a cascading relationship. Each edge computing device 200 sequentially performs violation event detection on the video data, thereby achieving a cascading effect, improving the types of violation events detected, and enhancing the reliability of data processing. Specifically, the first edge computing device 210 is an edge computing device 200 directly connected to the front-end intelligent device 100, while the remaining edge computing devices 200 (e.g., the (n-1)th edge computing device 220) are sequentially connected to the front-end intelligent device 100, and indirectly connected to it through the first edge computing device 210. The detection method for violations by the edge computing device 200 after forming a cascaded relationship can be specifically described in steps S307 to S308 of the above-mentioned vehicle violation detection method, and will not be repeated here.

[0094] Optionally, the violation detection system 10 also includes a management platform 400 and an application platform 500. Users can manage the front-end intelligent device 100 through the management platform 400, such as switching the detection target of the front-end intelligent device 100. Users can also observe the detection results of violation events in real time through the application platform 500, and control the cloud server 300 or edge computing device 200 to modify the detection results based on the violation event detection results.

[0095] This application also provides an electronic device, including a processor and a memory connected to the processor. The memory stores program instructions, and the processor executes the program instructions to implement the vehicle violation event detection method of any of the above embodiments.

[0096] This application also proposes a computer-readable storage medium. See [link to application]. Figure 8 , Figure 8 This is a schematic diagram of the structure of an embodiment of the computer-readable storage medium of this application.

[0097] The computer-readable storage medium 30 of this application embodiment stores program instructions 31, which are executed to implement the above-described data download method.

[0098] The program instructions 31 can be formed into a program file and stored in the aforementioned storage medium as a software product, so that an electronic device (which may be a personal computer, server, or network device, etc.) or processor can execute all or part of the steps of the vehicle violation detection method of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks, or terminal devices such as computers, servers, mobile phones, and tablets.

[0099] In this embodiment, the computer-readable storage medium 30 may be, but is not limited to, a USB flash drive, SD card, PD optical drive, portable hard drive, large-capacity floppy drive, flash memory, multimedia memory card, server, etc.

[0100] In one embodiment, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the electronic device to perform the steps in the above-described embodiment of the vehicle violation detection method.

[0101] Furthermore, if the aforementioned functions are implemented as software and sold or used as an independent product, they can be stored in a mobile terminal-readable storage medium. That is, this application also provides a storage device storing program instructions that can be executed to implement the vehicle violation detection method of the above embodiments. This storage device can be, for example, a USB flash drive, optical disc, or server. In other words, this application can be embodied in the form of a software product, which includes several instructions to cause a smart terminal to execute all or part of the steps of the vehicle violation detection method of each embodiment.

[0102] In summary, the vehicle violation event detection method of this application, through the collaborative control of cloud servers, edge computing devices, and front-end intelligent devices, detects violations in video images of the vehicle under test and effectively analyzes the detection results. This ensures the security of the original data, improves data utilization, and increases the detection capability of vehicle violation event types, thereby achieving effective detection of violation event types. Furthermore, the violation detection system, through the collaborative control of cloud servers, edge computing devices, and front-end intelligent devices, not only achieves effective detection of violation event types but also fully utilizes the data acquisition and processing capabilities of front-end intelligent devices and the event analysis capabilities of edge computing devices, progressively improving the credibility of event detection and enhancing the types of violations detected. Simultaneously, leveraging the powerful computing and storage capabilities of cloud servers, it achieves the fusion, storage, and visualization of diverse data, ensuring the security of the original data, improving the utilization rate of the original data and the timeliness of event detection, and enhancing the detection capability of vehicle violation event types.

[0103] It is worth noting that the accompanying drawings are only for illustrating the structural and connection relationships of the product of this invention, and do not limit the specific structural dimensions of the product of this invention.

[0104] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for detecting vehicle violation incidents, characterized in that, The vehicle violation detection method is applied to a violation detection system, which includes a front-end intelligent device, an edge computing device, and a cloud server. The vehicle violation detection method includes: The front-end intelligent device collects video images of the vehicle under test; The front-end intelligent device performs violation event detection on the video image and generates a first detection result, which includes a first violation event identification result and its first credibility. The edge computing device acquires video images from the front-end intelligent device and performs violation event detection on the video images to generate a second detection result. The violation events that the edge computing device can identify include violation events that the front-end intelligent device can identify and violation events that the front-end intelligent device cannot identify. When the first credibility of the first violation event identification result is greater than or equal to the first preset threshold, the cloud server receives the first detection result of the front-end smart device and the second detection result of the edge computing device, and generates a violation event detection result based on the first detection result and the second detection result; When the first confidence level of the first violation event identification result is less than the first preset threshold, the cloud server receives the second detection result from the edge computing device and generates the violation event detection result based on the second detection result.

2. The vehicle violation detection method according to claim 1, characterized in that, The edge computing device performs violation event detection on the video image and generates a second detection result, including: The edge computing device acquires the first detection result; When the first confidence level of the first violation event identification result is less than the first preset threshold, the edge computing device performs violation event detection on the video image based on the first violation event and the second violation event, and generates a second detection result, wherein the first violation event is different from the second violation event.

3. The vehicle violation event detection method of claim 1, wherein After the front-end intelligent device performs violation event detection on the video image and generates a first detection result, it further includes: The front-end intelligent device sends the violation event detection result in the first detection result, where the first confidence level is greater than or equal to the first preset threshold, to the edge computing device. The edge computing device acquires video images from the front-end intelligent device, performs violation event detection on the video images, and generates a second detection result, including: The edge computing device acquires video images from the front-end intelligent device and the violation event detection results; The edge computing device performs a third violation detection on the video image based on the violation detection result, and generates the second detection result; wherein, the third violation is a violation event other than the violation detection result.

4. The vehicle violation event detection method according to claim 2 or 3, characterized by, After performing violation event detection on the video image and generating a second detection result, the method further includes: The edge computing device sends the violation event detection result in the second detection result, where the second confidence level is greater than or equal to the second preset threshold, to the cloud server.

5. The vehicle violation event detection method of claim 1, wherein, The edge computing device includes a first edge computing device and a second edge computing device; The edge computing device acquires video images from the front-end intelligent device, performs violation event detection on the video images, and generates a second detection result, including: The first edge computing device acquires video images from the front-end smart device, performs violation event detection on the video images, and generates a third detection result; The second edge computing device acquires the video image and the third detection result from the first edge computing device, performs violation event detection on the video image based on the third detection result, and generates a second detection result.

6. The vehicle violation detection method according to claim 5, characterized in that, The third detection result includes the second violation event identification result and its third credibility; The step of detecting violations in the video image based on the third detection result and generating a second detection result includes: The second edge computing device determines, based on the third detection result, detected violation events whose third credibility is greater than or equal to a third preset threshold; The second edge computing device performs undetected violation event detection on the video image based on the detected violation events, and generates a second detection result.

7. The vehicle violation event detection method of claim 1, wherein, Also includes: The cloud server performs error analysis on the violation detection results to obtain the error analysis results. The edge computing device corrects the violation detection results and trains the model based on the error analysis results.

8. A violation detection system characterized by, include: A front-end intelligent device is used to collect video images of the vehicle under test. The front-end intelligent device is also used to detect violations in the video images and generate a first detection result, which includes a first violation identification result and its first credibility. An edge computing device is connected to the front-end intelligent device. The edge computing device is used to acquire video images from the front-end intelligent device, perform violation event detection on the video images, and generate a second detection result. The violation events that the edge computing device can identify include violation events that the front-end intelligent device can identify and violation events that the front-end intelligent device cannot identify. A cloud server is connected to the front-end intelligent device and the edge computing device. When the first credibility of the first violation event identification result is greater than or equal to a first preset threshold, the cloud server receives the first detection result of the front-end intelligent device and the second detection result of the edge computing device, and generates a violation event detection result based on the first detection result and the second detection result. When the first confidence level of the first violation event identification result is less than the first preset threshold, the cloud server receives the second detection result from the edge computing device and generates the violation event detection result based on the second detection result.

9. An electronic device, comprising: The device includes a processor and a memory connected to the processor, the memory storing program instructions, and the processor executing the program instructions to implement the vehicle violation event detection method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The system stores program instructions that, when executed by a processor, implement the vehicle violation detection method as described in any one of claims 1-7.

Images