Non-motor vehicle red light running identification method and system

Abstract

The present application discloses a method and system for identifying non-motor vehicles running red lights, the system comprising a positioning tag arranged on the non-motor vehicle, a server, and a UWB base station and a video acquisition device arranged within a preset road section, the video acquisition device comprising a monitoring camera and a processor, the positioning tag comprising a first positioning channel; the UWB base station sends UWB positioning information obtained by communicating with the first positioning channel of the positioning tag to the server; the monitoring camera is used to collect non-motor vehicle driving videos in the crossing-line area within the preset road section, the processor is used to intercept video slices of the red light time period from the non-motor vehicle driving videos, and transmit the video slices to the server; the server is used to determine whether the non-motor vehicle has run a red light. The present application can accurately identify whether a non-motor vehicle has run a red light, conveniently and quickly.

Description

Technical Field

[0001] The present application belongs to the field of target recognition, and in particular relates to a method and system for identifying non-motor vehicles running red lights. Background Art

[0002] Traditional red light violation detection for non-motor vehicles relies primarily on manual enforcement, with traffic police observing at intersections and intercepting and penalizing vehicles that run red lights. Alternatively, red light violation detection technology based on induction coils has been attempted for non-motor vehicles, but with less than satisfactory results. For example, induction coil technology relies on changes in inductance caused by the passage of a vehicle to detect the vehicle. However, non-motor vehicles are mostly made of materials such as plastic and rubber, with relatively few metal components. This makes it difficult for induction coils to sensitively and accurately detect non-motor vehicle traffic, resulting in frequent misjudgments or even failure to detect non-motor vehicle red light violations.

[0003] Video detection technology also faces challenges in identifying non-motor vehicles running red lights. Although cameras can capture images of non-motor vehicles, the wide variety of non-motor vehicles, including bicycles, electric vehicles, and tricycles, varies greatly in appearance and has random riding postures. This, coupled with complex intersection environments such as changing light (strong direct light, backlighting, shadows, etc.) and background interference (billboards, trees, etc.), makes it difficult for video image-based recognition algorithms to accurately distinguish non-motor vehicles from complex scenes and accurately determine whether they have run a red light. Furthermore, in severe weather conditions such as heavy rain, fog, and dust, video image quality is severely impaired, and the recognition of non-motor vehicles running red lights is almost paralyzed.

[0004] Therefore, the current non-motor vehicle red light running recognition technology still has defects. Summary of the Invention

[0005] The purpose of this application is to provide a method for identifying non-motor vehicles running red lights to solve at least one of the above problems.

[0006] On one hand, the present application discloses a non-motor vehicle red light running identification system, the system comprising a positioning tag provided on the non-motor vehicle, a server, and a UWB base station and a video acquisition device provided within a preset road section, the video acquisition device comprising a surveillance camera and a processor, the positioning tag comprising a first positioning channel;

[0007] The UWB base station sends UWB positioning information obtained by communicating with the first positioning channel of the positioning tag to the server;

[0008] The surveillance camera is used to collect non-motor vehicle driving videos in the crossing-line area within the preset road section range, and the processor is used to intercept video slices during the red light period from the non-motor vehicle driving videos and transmit the video slices to the server;

[0009] The server is used to determine the positioning trajectory of the non-motor vehicle based on the UWB positioning information, and to determine whether the non-motor vehicle has run a red light based on the positioning trajectory and the video slice.

[0010] Optionally, the positioning tag further includes a second positioning channel;

[0011] The server receives GNSS positioning information based on the second positioning channel transmitted by an external positioning system;

[0012] The server is further configured to determine whether there is an abnormality in the positioning trajectory based on the UWB positioning information and the GNSS positioning information.

[0013] Optionally, the server is also used to determine the positioning trajectory and driving trajectory of the non-motor vehicle based on the UWB positioning information and the GNSS positioning information respectively, and determine whether there is any abnormality in the positioning trajectory of the non-motor vehicle based on the overlapping area of ​​the positioning trajectory and the driving trajectory.

[0014] Optionally, the driving road of the non-motor vehicle includes a preset road section range at a traffic light intersection and a driving area connected between the preset road section range;

[0015] The positioning tag closes the second positioning channel when the first positioning channel is enabled after the non-motor vehicle enters the preset road section range, and opens the second positioning channel when the first positioning channel is closed after the non-motor vehicle leaves the preset road section range.

[0016] Optionally, the server is also used to determine the positioning trajectory and driving trajectory of the non-motor vehicle based on the UWB positioning information and the GNSS positioning information respectively, and determine whether there is any abnormality in the positioning trajectory of the non-motor vehicle based on the regional connection part of the positioning trajectory and the driving trajectory.

[0017] Optionally, the server is also used to predict the crossing distance of non-motor vehicles running red lights based on traffic data of non-motor vehicles running red lights within a preset road section and a trained range prediction model, and determine the crossing area based on the crossing distance and the red light no-entry line within the preset road section.

[0018] Optionally, the server is used to determine an abnormal position in the positioning trajectory of the non-motor vehicle based on the positioning trajectories of other non-motor vehicles and the positioning trajectory of the non-motor vehicle, the abnormal position being a position where the difference between the change rate of the positioning trajectory of the non-motor vehicle and the average change rate of the positioning trajectories of other non-motor vehicles is higher than a set threshold; intercepting the timestamp corresponding to the abnormal position, and configuring an interception period based on each timestamp and the difference in change rate, wherein the timestamp is within the interception period; and analyzing whether the non-motor vehicle trajectory runs a red light based on the video data in each interception period in the video slice.

[0019] Optionally, for each abnormal position, if the number of other non-motor vehicles located in front of the non-motor vehicle is greater than the number located behind the non-motor vehicle, the server is also used to move the interception time period forward by a set length of time, otherwise move it backward by a set length of time.

[0020] Optionally, the video slice also includes unlabeled non-motor vehicles; the server is further used to select multiple frames of images from the video slice, the multiple frames of images including at least two frame images corresponding to abnormal positions; the actual displacement and image displacement of the non-motor vehicle identified from at least two frame images corresponding to abnormal positions; based on the actual displacement and the image displacement, combined with the distance between the road marking in one frame of the image and the non-motor vehicle in the image, the crossing area is generated; based on the multiple frames of images, each unlabeled non-motor vehicle and the driving trajectory of the non-motor vehicle are fitted and generated; based on each unlabeled non-motor vehicle, the crossing area and the driving trajectory of the non-motor vehicle, it is determined whether each unlabeled non-motor vehicle runs a red light.

[0021] This application also discloses a method for identifying non-motor vehicles running red lights, comprising:

[0022] Receiving UWB positioning information sent by the UWB base station and obtained by communicating through a first positioning channel between the UWB base station and the positioning tag on the non-motor vehicle;

[0023] Receive a video of a non-motor vehicle traveling in a cross-line area within a preset road section captured by a surveillance camera, and extract video slices of a red light period from the non-motor vehicle traveling video;

[0024] Identify the non-motor vehicle in the video slice, determine the positioning trajectory of the non-motor vehicle based on the UWB positioning information of the non-motor vehicle, and determine whether the non-motor vehicle has run a red light based on the positioning trajectory and the video slice.

[0025] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0026] The non-motor vehicle red light running recognition system of the present application not only collects the non-motor vehicle driving video of the crossing line area in the preset road section through the monitoring camera of the video acquisition device to assist in determining whether the non-motor vehicle has run a red light, but also collects the UWB positioning information of the non-motor vehicle in the preset road section through the positioning tag set on the non-motor vehicle. The UWB positioning information and video slices are combined to jointly determine whether the non-motor vehicle has run a red light, thereby improving the accuracy and efficiency of the determination of non-motor vehicle red light running behavior and reducing the impact of non-motor vehicle type and weather environment on the determination of non-motor vehicle red light running behavior. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments or descriptions of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0028] Figure 1 This is an application environment diagram of a non-motor vehicle red light running recognition system provided by this application;

[0029] Figure 2 A flowchart of a method for identifying a non-motor vehicle running a red light provided in this application;

[0030] Figure 3 A schematic diagram of the structure of a computer device suitable for implementing the method of the embodiment of the present application is provided in the embodiment of the present application. DETAILED DESCRIPTION

[0031] In the following description, specific details such as specific system structures and techniques are provided for purposes of illustration rather than limitation to facilitate a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, circuits, and methods are omitted to avoid obscuring the description of the present application with unnecessary detail.

[0032] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, integers, steps, operations, elements and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0033] It will also be understood that the term "and / or" used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0034] As used in this specification and the appended claims, the term "if" can be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "upon determination" or "in response to determining" or "upon detection of [described condition or event]" or "in response to detecting [described condition or event]," depending on the context.

[0035] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0036] References to "one embodiment" or "some embodiments" in this specification mean that a particular feature, structure, or characteristic described in conjunction with that embodiment is included in one or more embodiments of the present application. Thus, phrases such as "in one embodiment," "in some embodiments," "in other embodiments," and "in other embodiments" appearing in various places in this specification do not necessarily refer to the same embodiment, but rather mean "one or more but not all embodiments," unless otherwise specifically emphasized. The terms "including," "comprising," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0037] It should be understood that the size of the serial numbers of each step in this embodiment does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of this application.

[0038] Figure 1 A schematic diagram of the implementation scenario of this application is shown in FIG. Figure 1 As shown, the direction of the arrow in the figure is the direction of travel of the non-motor vehicle. The figure shows a crossroads, and the cross-line area is set within the road markings. At the crossroads in the figure, the preset road section range covers the area extending a certain distance (for example, 50 meters) outward from each entrance and exit road of the intersection. Multiple UWB base stations are reasonably arranged along the periphery of the road section. The UWB base station can receive the UWB positioning information broadcast by the first positioning channel of the positioning tag configured on the non-motor vehicle in real time. Of course, in actual applications, those skilled in the art can determine the preset road section range and the range of the cross-line area and the location of the UWB base station according to actual needs, and this application does not limit this.

[0039] Optionally, the overall range of positioning information receivable by the multiple UWB base stations needs to cover at least the range of the cross-line area so that positioning information of all non-motor vehicles within the cross-line range can be obtained. Preferably, the overall range of positioning information receivable by the multiple UWB base stations at least covers the range of the preset road section to expand the range of obtainable non-motor vehicle positioning information. More non-motor vehicle positioning information can be obtained to assist in determining whether the non-motor vehicle has run a red light, thereby improving the accuracy of the determination result.

[0040] It should be noted that in the embodiment of the present application, the surveillance camera is installed at a high place similar to the traffic light gantry. At present, some non-motor vehicles use the YOLO5 neural network model to identify non-motor vehicles running red lights through video recognition. However, this red light running recognition requires the full amount of video captured by the surveillance camera to be calculated and analyzed by the neural network model to determine whether there is a non-motor vehicle running a red light. This judgment method requires a large amount of data calculation, which is computationally intensive and inefficient. Video recognition is greatly affected by video quality and may have inaccurate recognition problems. Moreover, this judgment requires obtaining all videos captured by the surveillance camera, and the amount of data transmitted is large. Too many invalid videos occupy network resources.

[0041] It should be understood that the UWB base station of the present application can obtain the UWB positioning information of non-motor vehicles with positioning tags in real time, thereby forming a positioning track by continuously obtaining the UWB positioning information of non-motor vehicles. Obtaining the positioning track through positioning information is a conventional technical means in this field, and this application will not elaborate on this.

[0042] It is further necessary to understand that the traffic information server such as the red light period in the embodiment of the present application can obtain it in real time from the corresponding cloud platform. Of course, in actual applications, traffic information can also be obtained through other means, and the present application does not limit this.

[0043] In a first aspect, embodiments of the present application provide a system for identifying non-motorized vehicles running red lights. The system includes a positioning tag mounted on the non-motorized vehicle, a server, a UWB base station located within a predetermined road section, and a video acquisition device. The video acquisition device includes a surveillance camera and a processor, and the positioning tag includes a first positioning channel.

[0044] The UWB base station sends UWB positioning information obtained by communicating with the first positioning channel of the positioning tag to the server.

[0045] The surveillance camera is used to collect non-motor vehicle driving videos in the crossing-line area within the preset road section range, and the processor is used to intercept video slices of the red light time period from the non-motor vehicle driving videos and transmit the video slices to the server.

[0046] The server is used to determine the positioning trajectory of the non-motor vehicle based on the UWB positioning information, and to determine whether the non-motor vehicle has run a red light based on the positioning trajectory and the video slice.

[0047] The non-motor vehicle red light running recognition system of the present application not only collects the non-motor vehicle driving video of the crossing line area in the preset road section through the monitoring camera of the video acquisition device to assist in determining whether the non-motor vehicle has run a red light, but also collects the UWB positioning information of the non-motor vehicle in the preset road section through the positioning tag set on the non-motor vehicle. The UWB positioning information and video slices are combined to jointly determine whether the non-motor vehicle has run a red light, thereby improving the accuracy and efficiency of the determination of non-motor vehicle red light running behavior and reducing the impact of non-motor vehicle type and weather environment on the determination of non-motor vehicle red light running behavior.

[0048] In an optional embodiment, the positioning tag further includes a second positioning channel. The server receives GNSS positioning information based on the second positioning channel transmitted by an external positioning system; and the server is further configured to determine whether there is an anomaly in the positioning trajectory based on the UWB positioning information and the GNSS positioning information.

[0049] Specifically, it is understandable that UWB positioning is mainly based on time measurement and signal strength measurement, including methods such as time of arrival (TOA), time difference of arrival (TDOA), and angle of arrival (AoA). The target position is determined by measuring the time difference or angle of arrival of the signal propagation, such as measuring the propagation time of the signal from the transmitter to the receiver, and then calculating the distance to achieve positioning. UWB positioning is mainly used in scenarios such as indoor positioning and short-range wireless communications. It can achieve high-precision positioning from sub-centimeter to centimeter levels. In indoor environments, even in the presence of interference such as multipath effects, it can maintain a high positioning accuracy, making it suitable for scenarios with extremely high precision requirements.

[0050] GNSS positioning relies primarily on satellite signals. Multiple satellites transmit signals, which are then received by ground-based receivers. The distance between the receiving point and the satellites is calculated, and then, combined with information such as the satellites' orbital positions, geometric algorithms are used to determine the three-dimensional coordinates of the receiving point. In open outdoor environments, GNSS positioning accuracy generally reaches several to tens of meters. However, in areas with dense urban areas, indoor environments, tunnels, and other locations where signals are easily blocked or reflected, accuracy decreases significantly.

[0051] Therefore, UWB positioning has high positioning accuracy, but has high equipment requirements and costs, while GNSS positioning has average positioning accuracy, but has low equipment requirements and costs. In this embodiment, in order to achieve the purpose of improving the recognition accuracy of non-motor vehicles running red lights while taking into account efficiency and cost, the two positioning methods of UWB positioning and GNSS positioning are integrated, and a positioning tag with a first positioning channel and a second positioning channel is set. A UWB base station is set in a preset road section with traffic lights. The UWB base station receives UWB positioning information through the first positioning channel, and transmits GNSS positioning information to the server through the second positioning channel of the positioning tag, so that the server can make a judgment on non-motor vehicles running red lights based on the two positioning information and video slices, thereby improving the judgment efficiency and accuracy.

[0052] In a specific example, the server determines the positioning trajectories of all non-motor vehicles that cross the line area during the red light period based on UWB positioning information, and determines whether there are non-motor vehicles that may run a red light based on the positioning trajectories and the red light period. If so, video slicing is used to identify whether the non-motor vehicle that may run a red light is indeed a non-motor vehicle that has run a red light.

[0053] During this process, the server only uses the received UWB positioning information to determine the non-motor vehicles that may have run a red light and need to be further judged through video slicing. If the UWB positioning information is inaccurate, there may be cases where non-motor vehicles that need further judgment are missed or non-motor vehicles that do not need further judgment are not completely excluded. Therefore, in order to improve the accuracy of the judgment of non-motor vehicles running a red light, this application receives GNSS positioning information by setting a second positioning channel of the positioning tag, verifies the UWB positioning information through the GNSS positioning information, and determines whether there is any abnormality in the UWB positioning information. If the UWB positioning information passes the verification and there is no obvious abnormality, the judgment result of the UWB positioning information is used normally.

[0054] In a specific example, the current location of a non-motor vehicle can be calculated using UWB positioning information obtained through communication between multiple UWB base stations and non-motor vehicle positioning tags. After continuously collecting UWB positioning information, all consecutive current locations of the non-motor vehicle are connected to obtain the non-motor vehicle's positioning trajectory. Then, based on the red light periods of real-time traffic information and the positioning trajectories of all non-motor vehicles, the positioning trajectories of all non-motor vehicles within the crossing line area during all red light periods can be obtained.

[0055] Among them, when determining the positioning trajectory of a non-motor vehicle, a non-motor vehicle whose positioning trajectory completely passes through the cross-line area during the red light period can be selected as a non-motor vehicle based on demand, that is, the positioning trajectory extends from one end of the cross-line area to the other end along the direction indicated by the red light; a non-motor vehicle whose positioning trajectory is partially located in the cross-line area during the red light period can also be selected as a non-motor vehicle. Technical personnel in this field can set it according to demand in actual application, and this application does not limit this.

[0056] After identifying a non-motor vehicle that may have run a red light, the system obtains the GNSS positioning information of the non-motor vehicle obtained through the second positioning channel from the external positioning system. Based on the non-motor vehicle's GNSS positioning information, the system matches the UWB positioning information to determine whether the UWB positioning information's trajectory in the crossing-line area during the red light period overlaps with the GNSS positioning information in that area. If so, it indicates that the UWB positioning information is consistent with the positioning information of the external positioning system, and the UWB positioning information is highly reliable and can be used to determine whether the non-motor vehicle has run a red light. If there is an anomaly, the GNSS positioning information is used to determine whether the non-motor vehicle has run a red light, and the service personnel are notified to troubleshoot the positioning tag of the relevant non-motor vehicle as soon as possible.

[0057] It should be noted that the criteria for determining whether UWB positioning information and GNSS positioning information overlap can be determined based on actual circumstances and are not limited in this application. For example, multiple corresponding positions are taken from the UWB positioning information and the GNSS positioning information, and the deviation between the time and position is calculated. If the deviation is within a preset threshold, then the UWB positioning information and the GNSS positioning information are determined to overlap.

[0058] It should be noted that the positioning tag containing the first positioning channel and the second positioning channel in this application can be obtained by using existing technologies based on the ultra-wideband (UWB) positioning principle and the global navigation satellite system (GNSS) positioning principle, and this application will not go into details here. For example, the first positioning channel can be achieved by setting a tag based on the UWB positioning principle, and the second positioning channel can be achieved by setting a receiver.

[0059] In an optional embodiment, the server is also used to determine the positioning trajectory and driving trajectory of the non-motor vehicle based on the UWB positioning information and the GNSS positioning information, respectively, and determine whether there is any abnormality in the positioning trajectory of the non-motor vehicle based on the overlapping area of ​​the positioning trajectory and the driving trajectory.

[0060] In an optional embodiment, the driving road of the non-motor vehicle includes a preset road section range at a traffic light intersection and a driving area connected to the preset road section range; the positioning tag closes the second positioning channel when the first positioning channel is enabled after the non-motor vehicle enters the preset road section range, and opens the second positioning channel when the first positioning channel is closed after the non-motor vehicle leaves the preset road section range.

[0061] Specifically, it is understandable that the display label provided on the non-motor vehicle includes a first positioning channel and a second positioning channel. During the driving process of the non-motor vehicle, if the two positioning channels are fully opened, the energy consumption is relatively high. In order to reduce power consumption, in this optional embodiment, in the driving area where positioning cannot be performed by the first positioning channel, the first positioning channel is closed, and the second positioning channel is opened to obtain GNSS positioning information. After entering the preset road section range, the first positioning channel is opened and the second positioning channel is closed at the same time to receive UWB positioning information. Therefore, the present application enables different positioning channels in different ranges. Since the motion trajectory of the non-motor vehicle is continuous, the GNSS positioning information obtained in this way can also be used to verify the UWB positioning information, thereby reducing energy consumption and data transmission while ensuring the accuracy of the judgment of running a red light.

[0062] In a specific example, when an electric bicycle's positioning tag is traveling in an open area other than an intersection on a daily basis, the receiver of the second positioning channel is turned on and the location coordinates are sent to the server every 10 seconds with low power consumption. When the vehicle approaches the intersection and enters within 30 meters of the boundary of the preset road section (entering the cross-line area, the preset road section range is set according to the intersection geo-fence), the low-power Bluetooth module detects the Bluetooth beacon signal set at the intersection and immediately triggers the positioning tag to switch to the first positioning channel of the UWB positioning mode. At this time, the positioning accuracy can be improved to sub-meter level, and multiple real-time location updates are sent to nearby base stations per second to accurately track the movement trajectory of non-motor vehicles at the intersection.

[0063] In an optional embodiment, the server is also used to determine the positioning trajectory and driving trajectory of the non-motor vehicle based on the UWB positioning information and the GNSS positioning information, respectively, and determine whether there is any abnormality in the positioning trajectory of the non-motor vehicle based on the regional connection part of the positioning trajectory and the driving trajectory.

[0064] Specifically, the first and second positioning channels switch when a non-motorized vehicle enters or exits a preset road section. Since the non-motorized vehicle's driving trajectory is continuous, it receives GNSS positioning information within the driving area and UWB positioning information within the preset road section. Therefore, the UWB and GNSS positioning information should be processed in a corresponding manner at the edge of the preset road section. Based on this, the deviation at the junction of the non-motorized vehicle's UWB and GNSS positioning information can be used to determine whether there is an anomaly in the UWB positioning information.

[0065] In practical applications, the positioning trajectory and driving trajectory of the non-motor vehicle can be obtained based on the UWB positioning information and the GNSS positioning information. The criterion for determining whether the UWB positioning information is abnormal based on the positioning trajectory and the driving trajectory can be selected from at least one of the following criteria:

[0066] 1) Determine whether the ends of the positioning track and the driving track can be connected to form a complete driving track;

[0067] 2) Whether the end of the positioning track and the end of the driving track are connected is at the edge of the preset road section range;

[0068] 3) Determine whether the time corresponding to the end where the positioning trajectory and the driving trajectory connect is consistent.

[0069] It should be noted that, in the above-mentioned judgment criteria, the judgment calculation method and the judgment pass criteria can be set according to actual needs, and this application does not limit this.

[0070] In an optional embodiment, the server is further configured to predict the crossing distance of a non-motor vehicle running a red light based on traffic data of non-motor vehicles running red lights within a preset road section and a trained range prediction model, and to determine the crossing area based on the crossing distance and the red light no-go line within the preset road section. The traffic data includes the location trajectories of non-motor vehicles historically identified as having run red lights within the preset road section, and the crossing area where a non-motor vehicle may have run a red light is determined using the traffic data and the trained range prediction model.

[0071] Specifically, it is understandable that the control of non-motor vehicles is usually much looser than that of motor vehicles. Therefore, most non-motor vehicles will not strictly stop at the edge of the preset road section range, and will usually cross the preset road section range. Therefore, in order to avoid introducing the problem of large amount of calculation and low efficiency in determining the positioning trajectory of non-motor vehicles that do not run red lights due to irregular stops, this application obtains a range prediction model based on neural network model training, further delineates the cross-line area within the preset road section range, and determines the positioning trajectory of non-motor vehicles in the cross-line area, so as to reduce the amount of data of the non-motor vehicle positioning information that needs to be analyzed, improve the efficiency of determining whether the non-motor vehicle has run a red light, reduce the amount of calculation, and avoid introducing unnecessary positioning information processing processes.

[0072] In this specific example, the range prediction model can be trained through the following process:

[0073] 11) determining a training data set, the training data set including the positioning trajectories of non-motor vehicles that cross each preset road section and run a red light during a red light period;

[0074] 12) Inputting the training data set into the model built based on the neural network model for training;

[0075] When the training reaches the preset training end condition, a trained range prediction model is obtained. Based on this trained range prediction model, redundant areas within each preset road section that should not be included in the red light running behavior determination can be predicted. The redundancy areas can then be indented inward from the edge of the preset road section to determine the crossing area.

[0076] In an optional embodiment, it further includes intercepting video slices of the red light time period from the non-motor vehicle driving video, and before transmitting the video slices to the server, the server determines the target non-motor vehicle that may have run a red light based on the positioning trajectory of the non-motor vehicle, and sends the target non-motor vehicle to the processor. The processor intercepts the video slices corresponding to the red light time period and all target non-motor vehicles from the non-motor vehicle driving video based on all target non-motor vehicles in the red light time period and sends them to the server so that the server determines whether the target non-motor vehicle has run a red light based on the positioning trajectory and the video slices. Therefore, the processor only needs to transmit the partial video slices that the server determines may have run a red light through the positioning trajectory to the server without uploading the video slices of the entire red light time period, which greatly reduces the data transmission volume and improves the determination efficiency.

[0077] In an optional embodiment, in order to determine the time when a non-motor vehicle runs a red light, the server is used to determine the abnormal position in the positioning trajectory of the non-motor vehicle based on the positioning trajectories of other non-motor vehicles and the positioning trajectory of the non-motor vehicle, and the abnormal position is the position where the difference between the change rate of the positioning trajectory of the non-motor vehicle and the average change rate of the positioning trajectories of other non-motor vehicles is higher than a set threshold; intercept the timestamp corresponding to the abnormal position, and configure the interception period based on each timestamp and the difference in change rate, and the timestamp is within the interception period; analyze whether the non-motor vehicle trajectory runs a red light based on the video data in each interception period in the video slice.

[0078] In an optional embodiment, for each abnormal position, if the number of other non-motor vehicles located in front of the non-motor vehicle is greater than the number located behind the non-motor vehicle, the processor is also used to move the interception time period forward for a set period of time, otherwise move it backward for a set period of time.

[0079] In an optional embodiment, the video slice also includes unlabeled non-motor vehicles; the server is also used to select multiple frames of images from the video slice during the red light period, and the multiple frames of images include at least two frame images corresponding to abnormal positions; the actual displacement and image displacement of the non-motor vehicle are identified from the at least two frame images corresponding to abnormal positions; based on the actual displacement and the image displacement, combined with the distance between the road marking in one frame of the image and the non-motor vehicle in the image, the crossing area is generated; based on the multiple frames of images, each unlabeled non-motor vehicle and the driving trajectory of the non-motor vehicle are fitted and generated; based on each unlabeled non-motor vehicle, the crossing area and the driving trajectory of the non-motor vehicle, it is determined whether each unlabeled non-motor vehicle runs a red light.

[0080] The cross-line area of ​​this embodiment is as follows Figure 1 As shown, the road marking for running a red light is road marking 1 within a preset road section range, and the crossing area set in this application is offset inward compared to the road marking.

[0081] It can be seen that this application directly judges the red light running behavior of non-motor vehicles through the positioning trajectory and video slices in the crossing area during the red light period, and includes the judgment evidence of the video slices. It does not require a large amount of video data support, avoiding the network construction and operation and maintenance burden of the road section.

[0082] In an embodiment of the present application, after the server determines the positioning trajectory of the non-motor vehicle based on the UWB positioning information, it can preliminarily determine whether the non-motor vehicle has run a red light based on the positioning trajectory, and then, after obtaining the target non-motor vehicle that may have run a red light based on the determination, reversely verify the non-motor vehicle in combination with other non-motor vehicles in the video slice corresponding to the non-motor vehicle.

[0083] It should be noted that, in this embodiment, other non-motor vehicles are also equipped with positioning tags. Specifically, the server is used to determine the abnormal position in the positioning trajectory of the non-motor vehicle to be identified based on the positioning trajectories of other non-motor vehicles and the positioning trajectory of the non-motor vehicle to be identified, and the abnormal position is the position where the difference between the change rate of the positioning trajectory of the non-motor vehicle to be identified and the average change rate of the positioning trajectories of other non-motor vehicles is higher than a set threshold; intercept the timestamp corresponding to the abnormal position, and configure the interception period based on each timestamp and the difference in change rate, and the timestamp is within the interception period; based on the video data of the video slice within each interception period, finally determine whether the non-motor vehicle trajectory runs a red light.

[0084] This embodiment compares the positioning trajectory of the non-motor vehicle with the positioning trajectories of other non-motor vehicles to determine the abnormal position of the positioning trajectory of the non-motor vehicle. It borrows the underlying principle of the majority principle and takes the average change rate of the positioning trajectories of all other non-motor vehicles as a reference. Since non-motor vehicles that run red lights should be a minority in terms of proportion, even if there are other non-motor vehicles that run red lights, the average change rate is still biased towards the direction of not running red lights. Therefore, the abnormal position described in this application is that the difference between the change rate of the positioning trajectory of the non-motor vehicle and the average change rate of the positioning trajectories of other non-motor vehicles is higher than the set threshold. That is, for the non-motor vehicle to be identified, if its trajectory change rate is significantly different from the average change rate of most non-motor vehicles, then there may be driving abnormalities at this time, such as sprinting to run a red light, or avoiding while running a red light.

[0085] When all abnormal positions of the entire positioning trajectory are confirmed, the timestamps corresponding to the abnormal positions are found, and the video data is intercepted. The intercepted video clips are used to confirm whether there is indeed a red light running behavior and quickly obtain instantaneous photos of non-motor vehicles running red lights. In this way, this application only needs a few seconds of video of certain key points to determine whether non-motor vehicles run red lights, which greatly reduces the transmission and processing volume of video data.

[0086] Specifically, the present application provides a method for configuring a capture period. In an embodiment of the present application, the capture period is configured based on each of the timestamps and in combination with the difference in the rate of change, including:

[0087] According to the preset function, combined with the difference of the change rate, the interception period with the timestamp as the starting point of the period is configured, wherein the preset function is T=a×e -bD +c, where D is the difference in the rate of change, T is the length of the intercept period, and a, b, and c are fitting coefficients.

[0088] In this embodiment, when the difference between the trajectory change rate of a non-motor vehicle and the average change rate of other non-motor vehicles increases, it means that the vehicle's driving state is increasingly deviating from its normal pattern. For example, at an intersection, non-motor vehicles waiting for a red light are moving at a steady speed and in a consistent direction, with an average rate of change approaching zero. However, if a vehicle to be identified suddenly accelerates and rushes toward the opposite intersection, its speed value will rapidly increase. At this time, the length of the interception period will approach zero as the difference in the rate of change increases, making the length of the interception period primarily determined by the smaller value, significantly shortening the interception period. As a result, the greater the difference in the rate of change, the more obvious the red light running characteristic is at that time, and the less video data required. Only a single key frame of data is needed to identify the red light running image. This embodiment focuses video capture on the critical few seconds most likely to reflect red light running behavior, avoiding the processing of large amounts of irrelevant video data. This allows the system to quickly identify the abnormal moment and greatly improve recognition efficiency.

[0089] The embodiment of the present application uses this function to greatly reduce the data transmission and storage burden of the server. If a traditional method is to fully analyze the behavior of non-motor vehicles running red lights, it may be necessary to transmit videos of the entire red light period, which is a huge amount of data. Now, only the key period is intercepted based on the function, and the amount of data transmitted is sharply reduced. Take an intersection with a daily non-motor vehicle traffic of thousands of vehicles as an example. Before the use of this function, the daily video data storage demand was as high as hundreds of GB. After the introduction of function optimization, the storage capacity can be reduced to several GB, saving a lot of storage space. At the same time, it reduces the network bandwidth usage during data transmission, ensuring the smooth operation of the system and providing feasibility support for large-scale intersection monitoring in cities.

[0090] Furthermore, this function lays the foundation for intelligent upgrades to the system. As traffic data continues to accumulate, machine learning algorithms can be used to dynamically optimize the values ​​of a, b, and c, allowing the function to continuously adapt to changes in urban traffic, such as the emergence of new non-motorized vehicles and adjustments to traffic regulations.

[0091] In an optional embodiment, for each abnormal position, if the number of other non-motor vehicles located in front of the non-motor vehicle is greater than the number located behind the non-motor vehicle, configuring the interception period based on each of the timestamps and the difference in the rate of change further includes:

[0092] The interception period is moved forward by a set time length, and vice versa, it is moved backward by a set time length.

[0093] In an embodiment of the present application, when it is found that the number of other non-motor vehicles in front of the non-motor vehicle is greater than the number behind it, it is possible that the vehicle to be identified may be accelerating to overtake the vehicle in front in an attempt to run a red light. At this time, the interception period is moved forward by a set time period, which can accurately capture the starting moment of the vehicle's acceleration, lane change, etc. For example, at an intersection with heavy traffic, if there are multiple vehicles in front of an electric vehicle to be identified that are normally waiting for the red light, and it suddenly speeds up, the rate of change of its trajectory is in sharp contrast to that of the surrounding vehicles. By moving the interception period forward, the system is no longer limited to the conventional period based solely on the difference in rate of change, but instead locks in its preparatory action for running a red light in advance, making the judgment basis more sufficient, greatly reducing the possibility of misjudgment, and improving the accuracy of identifying true red light running behaviors.

[0094] Or, for example, take a busy intersection with an average daily non-motorized vehicle traffic volume of 10,000 vehicles. If the capture period is fixed, a large number of images of vehicles waiting normally or starting slowly may be captured, while the images that truly reflect the core behavior of running a red light account for a small proportion. After adopting this implementation method, through precise time period shifting, the system focuses on high-suspicion moments, so that the video data captured every second is closely centered around possible red light running behaviors, which not only reduces the pressure on subsequent video analysis algorithms, but also improves the overall recognition efficiency, maximizes the effectiveness of limited data processing resources, and ensures the efficient and stable operation of large-scale intersection monitoring systems. It can be seen that compared with fixed time period capture, the dynamic adjustment in the embodiment of the present application avoids wasting resources on invalid or low-relevance video clips.

[0095] Confirmation of red light running for non-motor vehicles without positioning tags:

[0096] In an optional embodiment, it is also possible to confirm that non-motor vehicles that are not equipped with positioning tags have run a red light. It should be noted that the configuration of positioning tags may not be widely popularized in most cases. Therefore, the embodiment of the present application further uses a local coverage method, and only requires the presence of a non-motor vehicle equipped with a positioning tag at a section of intersection to realize the red light running identification of all non-motor vehicles in the road section. At the same time, it does not require a large amount of data and processing. Specifically, the core idea of ​​this embodiment is to perform spatial projection through the trajectory and positioning trajectory recognized by video, and determine the positioning trajectory of the non-motor vehicle by means of spatial projection, so as to determine whether the non-motor vehicle that is not equipped has run a red light.

[0097] Specifically, in this embodiment, the video slice also includes unlabeled non-motor vehicles; the server is also used to select multiple frames of images from the video slice, and the multiple frames of images include at least two frame images corresponding to abnormal positions; the actual displacement and image displacement of the non-motor vehicle are identified from the at least two frame images corresponding to abnormal positions; based on the actual displacement and the image displacement, combined with the distance between the road marking in one frame of the image and the non-motor vehicle in the image, the crossing area is generated; based on the multiple frames of images, each unlabeled non-motor vehicle and the driving trajectory of the non-motor vehicle are fitted and generated; based on each unlabeled non-motor vehicle, the crossing area and the driving trajectory of the non-motor vehicle, it is determined whether each unlabeled non-motor vehicle runs a red light.

[0098] Specifically, as shown in the figure, in the embodiment of the present application, multiple frames of images are first selected from the video slice. It should be noted that the multiple frames of images can be selected according to a preset time interval, such as 10 frames, or can be selected randomly. Of course, the multiple frames of images of the present application should at least include the entire time period of the video slice, such as at least one frame of the first N frames and at least one frame of the last N frames. In this way, although multiple frames of images are selected, the movement time of the non-motor vehicle is covered, so that the data is more accurate.

[0099] The embodiment of the present application cleverly borrows the characteristics of the gantry camera configuration. At this time, the shooting angle of the camera is an oblique angle, so the video data includes height data and depth data. After selecting multiple frames of images, the actual displacement and image displacement of the non-motor vehicle are identified from at least two frames corresponding to abnormal positions. Specifically, the multiple frames of images of the present application include at least two frames corresponding to abnormal positions, that is, the two frames of abnormal position images actually have a large displacement. According to the large displacement, the depth change in the video can be calculated from the video combined with the depth calculation algorithm. Since the video of the present application is shot from top to bottom, the position change in the video is directly reflected in the image, which can be measured by the distance between the two images. At this time, at least two frames with a time difference are selected. Due to the time difference, the position coordinates of different time points can be found from the positioning trajectory to obtain the distance difference L1 on the positioning trajectory. At the same time, a distance difference is formed on the images of the two video frames, so that the correspondence between the image distance and the actual distance can be obtained. Then, combined with the road marking of one of the frames of image, Figure 1 The road marking 1 in the figure generates an off-line area online (for example, the left line of the off-line area is obtained by moving the length of one vehicle body toward the road marking 2 compared to the position of the road marking 1. Similarly, the right line is obtained by moving the length of one vehicle body toward the road marking 1. This application does not elaborate on this).

[0100] Afterwards, based on the multiple frames of images, each untagged non-motor vehicle and the driving trajectory of the non-motor vehicle are fitted and generated. Specifically, as can be seen from the above, a distance difference is formed on the images of the two video frames. Therefore, for non-motor vehicles that are not equipped with positioning tags, their actual displacement distance can be calculated through the two frames of images (based on the correspondence between the image distance and the actual distance). At this time, combined with the crossing area and the red light period, it can be determined whether the non-motor vehicle without a positioning tag has run a red light.

[0101] Furthermore, the present application also provides a method for identifying unlabeled non-motor vehicles running red lights. In this embodiment, the method can cope with the situation where the road surface is non-planar. Specifically, the method determines whether each unlabeled non-motor vehicle has run a red light based on each unlabeled non-motor vehicle, the crossing area, and the driving trajectory of the non-motor vehicle, including:

[0102] Projecting the non-motor vehicle's travel trajectory onto the positioning trajectory according to a deflection angle;

[0103] The deflection angle is continuously changed until the degree of overlap between the driving trajectory and the positioning trajectory is higher than a set overlap threshold, and the final deflection angle is output.

[0104] Converting the driving trajectory of each untagged non-motor vehicle into a positioning trajectory of the untagged non-motor vehicle using the deflection angle finally outputted;

[0105] Based on the crossing area and the positioning trajectory of each untagged non-motor vehicle, it is determined whether each untagged non-motor vehicle has run a red light.

[0106] In this embodiment, the video data is accompanied by depth information and height information due to the shooting angle. If the road surface is horizontal, the height information can be offset by multiple frames of images. However, if the road surface height is not horizontal, such as uphill or downhill, the error caused by the height information cannot be eliminated by image calibration. Based on this, the present application further removes the height information by deflection, so as to cope with the situation where the road surface is non-planar.

[0107] In an embodiment of the present application, since both the driving trajectory (the trajectory determined in the video data) and the positioning trajectory (the trajectory identified by the base station) of the non-motor vehicle can be obtained, with the help of the current road surface characteristics (the slope of the road surface is generally uphill or downhill), the height information can be removed by angle correction. By configuring the deflection angle, the driving trajectory is projected onto the positioning trajectory, and through continuous deflection correction, when the overlap between the two is higher than the set threshold (for example, 95%), the height error can be eliminated accordingly. Then, the driving trajectory of each unlabeled non-motor vehicle is converted into the positioning trajectory of the unlabeled non-motor vehicle at the final output deflection angle, so that the positioning trajectory of each unlabeled non-motor vehicle eliminates the error disturbance caused by the height information. Using the converted positioning trajectory, it can be determined whether the unlabeled non-motor vehicle on the uphill and downhill roads has run a red light.

[0108] Based on the same principle, this application also provides a method for identifying non-motor vehicles running red lights. Figure 2 As shown, the method includes:

[0109] S100: Receive UWB positioning information sent by a UWB base station and obtained through a first positioning channel communication between the UWB base station and a positioning tag on a non-motor vehicle.

[0110] S200: Receive a non-motor vehicle driving video of a crossing-line area within a preset road section captured by a surveillance camera, and extract video slices of a red light period from the non-motor vehicle driving video.

[0111] S300: Identify the non-motor vehicle in the video slice, determine the positioning trajectory of the non-motor vehicle based on the UWB positioning information of the non-motor vehicle, and determine whether the non-motor vehicle has run a red light based on the positioning trajectory and the video slice.

[0112] Since the principle of this method is similar to that of the above-mentioned system, the implementation of this method can refer to the implementation of the above-mentioned system and will not be described in detail here.

[0113] An embodiment of the present application also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above method when executing the computer program.

[0114] An embodiment of the present application further provides a computer-readable medium, wherein the computer-readable medium stores a computer program, and the computer program implements the above method when executed by a processor.

[0115] The embodiments of the present application provide a computer program product, which can implement the above-mentioned methods when the computer program product is run on a computer.

[0116] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer programs. The systems, devices, modules, or units described in the above embodiments may be implemented by computer chips or entities, or by products having certain functions. A typical implementation device is a computer device. Specifically, the computer device may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0117] In a typical example, a computer device specifically includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the method executed by the client as described above is implemented, or when the processor executes the program, the method executed by the server as described above is implemented.

[0118] Reference below Figure 3 , which shows a structural diagram of a computer device 600 suitable for implementing an embodiment of the present application.

[0119] like Figure 3 As shown, the computer device 600 includes a central processing unit (CPU) 601, which can perform various appropriate tasks and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage portion 608 into a random access memory (RAM) 603. Various programs and data required for the operation of the computer device 600 are also stored in the RAM 603. The CPU 601, ROM 602, and RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0120] The following components are connected to the I / O interface 605: an input section 606 including a keyboard, a mouse, and the like; an output section 607 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and a speaker; a storage section 608 including devices such as a hard disk; and a communication section 609 including a network interface card such as a LAN card or a modem. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the I / O interface 605 as needed. Removable media 611, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 610 as needed, so that computer programs read therefrom can be installed in the storage section 608 as needed.

[0121] In particular, according to embodiments of the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present application include a computer program product comprising a computer program tangibly embodied on a machine-readable medium, the computer program comprising program code for executing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via the communication portion 609 and / or installed from a removable medium 611.

[0122] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0123] For the convenience of description, the above devices are described as being divided into various units according to their functions. Of course, when implementing this application, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0124] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0125] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0126] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0127] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0128] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0129] The present application may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communications network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0130] The various embodiments in this specification are described in a progressive manner. Similar parts between the various embodiments can be referred to in conjunction with each other. Each embodiment focuses on the differences between the other embodiments. In particular, the system embodiments are generally similar to the method embodiments, so the description is relatively simple. For relevant parts, refer to the description of the method embodiments.

[0131] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A non-motor vehicle red light running recognition system, characterized in that: The system includes a positioning tag installed on the non-motor vehicle, a server, a UWB base station and a video acquisition device installed within a preset road section, the video acquisition device includes a monitoring camera and a processor, and the positioning tag includes a first positioning channel; The UWB base station sends UWB positioning information obtained by communicating with the first positioning channel of the positioning tag to the server; The surveillance camera is used to collect non-motor vehicle driving videos in the crossing-line area within the preset road section range, and the processor is used to intercept video slices during the red light period from the non-motor vehicle driving videos and transmit the video slices to the server; The server is used to determine the positioning trajectory of the non-motor vehicle based on the UWB positioning information, and determine whether the non-motor vehicle has run a red light based on the positioning trajectory and the video slice; The positioning tag also includes a second positioning channel; The server receives GNSS positioning information based on the second positioning channel transmitted by an external positioning system; The server is further configured to determine whether there is an abnormality in the positioning trajectory based on the UWB positioning information and the GNSS positioning information; The non-motor vehicle's driving road includes a preset road section range at a traffic light intersection and a driving area connected to the preset road section range; The positioning tag closes the second positioning channel when the first positioning channel is enabled after the non-motor vehicle enters the preset road section range, and opens the second positioning channel when the first positioning channel is closed after the non-motor vehicle leaves the preset road section range; The video slice also includes an unlabeled non-motor vehicle; the server is further configured to select multiple frames of images from the video slice, the multiple frames of images including at least two frames corresponding to abnormal positions; identify the actual displacement and image displacement of the non-motor vehicle from the at least two frames corresponding to the abnormal positions; generate the crossing-line area based on the actual displacement and the image displacement, combined with the distance between the road marking in one frame of the image and the non-motor vehicle in the image; and generate each unlabeled non-motor vehicle and the non-motor vehicle's driving trajectory by fitting based on the multiple frames of images; Whether each unlabeled non-motor vehicle runs a red light is determined based on each unlabeled non-motor vehicle, the crossing area, and the driving trajectory of the non-motor vehicle.

2. A non-motor vehicle red light running recognition system as claimed in claim 1, characterized in that: The server is also used to determine the positioning trajectory and driving trajectory of the non-motor vehicle based on the UWB positioning information and the GNSS positioning information respectively, and determine whether there is an abnormality in the positioning trajectory of the non-motor vehicle based on the overlapping area of ​​the positioning trajectory and the driving trajectory.

3. The non-motor vehicle red light running recognition system according to claim 1, characterized in that: The server is further configured to determine the positioning trajectory and driving trajectory of the non-motor vehicle based on the UWB positioning information and the GNSS positioning information, respectively, and determine whether there is an abnormality in the positioning trajectory of the non-motor vehicle based on the regional connection part of the positioning trajectory and the driving trajectory.

4. The non-motor vehicle red light running recognition system according to claim 1, characterized in that: The server is also used to predict the crossing distance of non-motor vehicles running red lights based on traffic data of non-motor vehicles running red lights within a preset road section and a trained range prediction model, and determine the crossing area based on the crossing distance and the red light no-entry line within the preset road section.

5. The non-motor vehicle red light running recognition system according to claim 1, characterized in that: The server is configured to determine an abnormal position in the positioning trajectory of the non-motor vehicle based on the positioning trajectories of other non-motor vehicles and the positioning trajectory of the non-motor vehicle, the abnormal position being a position where a difference between a rate of change of the positioning trajectory of the non-motor vehicle and an average rate of change of the positioning trajectories of other non-motor vehicles is greater than a set threshold; intercepting a timestamp corresponding to the abnormal position, and configuring an interception period based on each timestamp and the difference in the rate of change, wherein the timestamp is within the interception period; The non-motor vehicle trajectory is analyzed based on the video data in each intercepted time period in the video slice to determine whether the non-motor vehicle has run a red light.

6. A non-motor vehicle red light running recognition system as claimed in claim 5, characterized in that: For each abnormal position, if the number of other non-motor vehicles located in front of the non-motor vehicle is greater than the number located behind the non-motor vehicle, the server is further configured to move the interception period forward by a set time length, and vice versa.

7. A method for identifying non-motor vehicles running red lights, characterized in that: include: Receiving UWB positioning information sent by the UWB base station and obtained by communicating through a first positioning channel between the UWB base station and the positioning tag on the non-motor vehicle; receiving a video of a non-motor vehicle traveling in a cross-line area within a preset road section captured by a surveillance camera, and extracting a video slice of a red light period from the non-motor vehicle traveling video; Identifying a non-motor vehicle in the video slice, determining a positioning trajectory of the non-motor vehicle based on the UWB positioning information of the non-motor vehicle, and determining whether the non-motor vehicle has run a red light based on the positioning trajectory and the video slice; The positioning tag further includes a second positioning channel; and the method further includes: Receiving GNSS positioning information based on the second positioning channel transmitted by an external positioning system; Determining whether there is an abnormality in the positioning trajectory based on the UWB positioning information and the GNSS positioning information; The non-motor vehicle's driving road includes a preset road section range at a traffic light intersection and a driving area connected to the preset road section range; The positioning tag closes the second positioning channel when the first positioning channel is enabled after the non-motor vehicle enters the preset road section range, and opens the second positioning channel when the first positioning channel is closed after the non-motor vehicle leaves the preset road section range; The video slice also includes a non-motor vehicle without a label; and the method further includes: A plurality of frames of images are selected from the video slices, the plurality of frames of images including at least two frame images corresponding to abnormal positions; the actual displacement and image displacement of the non-motor vehicle are identified from the at least two frame images corresponding to the abnormal positions; the crossing-line area is generated based on the actual displacement and the image displacement, combined with the distance between the road marking in one frame of the image and the non-motor vehicle in the image; each unlabeled non-motor vehicle and the driving trajectory of the non-motor vehicle are fitted and generated based on the plurality of frames of images; and whether each unlabeled non-motor vehicle runs a red light is determined based on each unlabeled non-motor vehicle, the crossing-line area and the driving trajectory of the non-motor vehicle.

Images