Data processing method and system of vehicle-mounted monitoring system

Abstract

The invention discloses a data processing method and system of a vehicle-mounted monitoring system. The vehicle-mounted monitoring system comprises a plurality of automobile data recorders and a plurality of edge computing terminals. One automobile data recorder can communicate and interact with a plurality of edge computing terminals; the edge computing terminal is used for implementing the data processing method. The data processing method comprises the following steps: acquiring to-be-detected video data uploaded by an automobile data recorder; identifying a first target object from the to-be-detected video data, and generating a detection frame of the first target object; wherein the first target object is a motor vehicle; tracking the first target object by adopting a multi-target tracking algorithm; extracting lane information from the to-be-detected video data; and on the basis of the lane information, detecting illegal driving behaviors of a first target object.

Description

Technical Field

[0001] The present invention relates to the field of data processing information technology, and in particular to a data processing method and system for a vehicle-mounted monitoring system. Background Art

[0003] At present, the supervision of traffic violations is mainly carried out by setting up violation photo-taking cameras on road sections for monitoring. However, relying solely on cameras at fixed points to take photos of violations and manual supervision is not able to cover a large area.

[0004] To this end, the present invention provides a vehicle-mounted monitoring system that can enhance traffic supervision while saving costs and is easy to maintain to detect traffic violations. Summary of the invention

[0005] In order to overcome the above technical defects, the present invention provides a data processing method and system for a vehicle-mounted monitoring system.

[0006] In order to solve the above problems, the present invention is implemented according to the following technical solutions:

[0007] In a first aspect, the present invention provides a data processing method for a vehicle-mounted monitoring system, wherein the vehicle-mounted monitoring system comprises a plurality of driving recorders and a plurality of edge computing terminals; one driving recorder can communicate and interact with a plurality of edge computing terminals; the edge computing terminal is used to implement the data processing method, and the data processing method comprises the following steps:

[0008] Obtain the video data to be detected uploaded by the driving recorder;

[0009] Identify a first target object from the video data to be detected, and generate a detection frame of the first target object; wherein the first target object is a motor vehicle;

[0010] Tracking the first target object using a multi-target tracking algorithm;

[0011] Extracting lane information from the video data to be detected;

[0012] Based on the lane information, illegal driving behavior of the first target object is detected.

[0013] Preferably, the detection of illegal driving behavior of the first target object specifically includes the step of detecting a motor vehicle lane change violation on a solid line:

[0014] Based on the detection frame of the first target object, generating a left fitting point and a right fitting point; the left fitting point and the right fitting point are used to represent the position of the first target object in the video data;

[0015] Based on the left fitting point and the right fitting point of the first target object and the lane information, it is determined whether the first target object has a lane change violation of a solid line.

[0016] Preferably, the left fitting point is the coordinate of the lower left corner of the detection frame, and the right fitting point is the coordinate of the lower right corner of the detection frame;

[0017] The calculation formula of the left fitting point is:

[0018] The calculation formula of the left fitting point is:

[0019] Among them, (cx, cy) represents the coordinates of the center point of the detection box, (w, h) represents the width and height of the detection box, (x1, y1) represents the coordinates of the left fitting point, and (x2, y2) represents the coordinates of the right fitting point.

[0020] Preferably, the detection of illegal driving behavior of the first target object specifically includes the step of detecting the motor vehicle driving in the wrong direction:

[0021] identifying a right lane in the lane information;

[0022] Identify a first target object whose detection frame is located in the right lane;

[0023] The front of the first target object is identified to determine whether the first target object has violated the traffic regulations by driving in the wrong direction.

[0024] Preferably, the data processing method further includes:

[0025] When the first target object is detected to have illegal driving behavior;

[0026] Identify the license plate information of the first target object;

[0027] The video data to be detected, the violation detection result and the license plate number information are sent to the target device.

[0028] Preferably, the data processing method further includes:

[0029] Identify a first target object and a second target object in the video data to be detected, and generate corresponding first target detection frames and second target detection frames respectively; wherein the first target object is a motor vehicle and the second target object is a non-motor vehicle;

[0030] Using a multi-target tracking algorithm to track the first target object and the second target object respectively;

[0031] Using a trajectory prediction model to predict the trajectories of the first target object and the second target object respectively, to obtain a first predicted trajectory and a second predicted trajectory;

[0032] Calculating whether there is an accident conflict point between a plurality of first predicted trajectories and between the first predicted trajectories and the second predicted trajectories;

[0033] When there is an accident conflict point, generate road accident warning information;

[0034] Based on the road accident warning information, instruct the corresponding driving recorder to capture potential accident video data;

[0035] Send road accident warning information, video data to be detected and potential accident video data to the target device.

[0036] In a second aspect, the present invention provides a data processing system for a vehicle-mounted monitoring system, wherein the vehicle-mounted monitoring system includes a plurality of driving recorders and a plurality of edge computing terminals; one driving recorder can communicate and interact with a plurality of edge computing terminals; the data processing system is arranged in the edge computing terminal, and the data processing system includes:

[0037] An acquisition module is used to acquire the video data to be detected uploaded by the driving recorder;

[0038] an identification module, which is used to identify a first target object from the video data to be detected and generate a detection frame of the first target object; wherein the first target object is a motor vehicle;

[0039] A tracking module, which is used to track the first target object using a multi-target tracking algorithm;

[0040] An extraction module, used for extracting lane information from the video data to be detected;

[0041] A detection module is used to detect illegal driving behavior of the first target object based on the lane information.

[0042] Preferably, the detection module detects the illegal driving behavior of the first target object, specifically including the step of detecting the illegal lane change of a motor vehicle on a solid line:

[0043] Based on the detection frame of the first target object, generating a left fitting point and a right fitting point; the left fitting point and the right fitting point are used to represent the position of the first target object in the video data;

[0044] Based on the left fitting point and the right fitting point of the first target object and the lane information, it is determined whether the first target object has a lane change violation of a solid line.

[0045] Preferably, the left fitting point is the coordinate of the lower left corner of the detection frame, and the right fitting point is the coordinate of the lower right corner of the detection frame;

[0046] The calculation formula of the left fitting point is:

[0047] The calculation formula of the left fitting point is:

[0048] Among them, (cx, cy) represents the coordinates of the center point of the detection box, (w, h) represents the width and height of the detection box, (x1, y1) represents the coordinates of the left fitting point, and (x2, y2) represents the coordinates of the right fitting point.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] The present invention provides a data processing method for a vehicle-mounted monitoring system, wherein the vehicle-mounted monitoring system includes multiple driving recorders and multiple edge computing terminals; one driving recorder can communicate and interact with multiple edge computing terminals; and the edge computing terminal is used to implement the data processing method. The data processing method includes the following steps: obtaining video data to be detected uploaded by the driving recorder; identifying a first target object from the video data to be detected and generating a detection frame for the first target object; wherein the first target object is a motor vehicle; tracking the first target object using a multi-target tracking algorithm; extracting lane information from the video data to be detected; and detecting illegal driving behavior of the first target object based on the lane information.

[0051] The vehicle monitoring system of the present invention has a driving recorder as an important component, which is responsible for capturing video data during vehicle driving, and the video data is then uploaded to the edge computing terminal for further processing. The present invention implements a vehicle violation detection system for driving recorder videos.

[0052] The present invention installs a driving recorder on a vehicle to capture the illegal driving information of the vehicle, thereby solving the limitation that the traditional violation capture system is fixed in location and can only capture the illegal driving information of the vehicle at the fixed location where the violation capture system is installed. The present invention also overcomes the problems of damaging the buried coils on the highway and weak anti-interference in the traditional monitoring mode, reduces the cost of the monitoring system, and improves the monitoring efficiency. The present invention not only improves the flexibility and accuracy of violation detection, but also reduces the monitoring cost, improves road safety, and has important practical application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The specific embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings, wherein:

[0054] Figure 1 is a three-dimensional schematic diagram of a mounting bracket of the present invention;

[0055] Figure 2 It is a schematic diagram of the assembly of a mounting bracket of the present invention;

[0056] Figure 3 is a cross-sectional schematic diagram of the second clamping arm of the present invention;

[0057] Figure 4 It is a flow chart of a data processing method of the present invention;

[0058] Figure 5 It is a schematic diagram for identifying a vehicle traveling in reverse according to the present invention;

[0059] In the figure:

[0060] 10-Connector;

[0061] 20-first clamping arm, 21-clamping claw;

[0062] 30-second clamping arm, 31-counterweight;

[0063] 40- rubber parts;

[0064] 50 - placement board, 51 - flexible pad, 52 - plastic base. DETAILED DESCRIPTION

[0065] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0066] At present, the supervision of traffic violations is mainly carried out by setting up violation photo-taking cameras on road sections for monitoring. However, relying solely on cameras at fixed points to take photos of violations and manual supervision is not able to cover a large area.

[0067] refer to Figure 4 , an embodiment of the present invention provides a flow chart of a data processing method of a vehicle-mounted monitoring system. In the vehicle-mounted monitoring system of the present invention, the driving recorder is an important component responsible for capturing video data during vehicle driving, and these video data are then uploaded to the edge computing terminal for further processing. The present invention implements a vehicle violation detection system for driving recorder videos.

[0068] In the present invention, the vehicle-mounted monitoring system includes multiple driving recorders and multiple edge computing terminals; one driving recorder can communicate and interact with multiple edge computing terminals; and the edge computing terminal is used to implement the data processing method.

[0069] It is understandable that the edge computing terminal is set in the road, such as in the control box of a smart street lamp. The dashcam can communicate with the edge computing terminal through the vehicle. For example, the V2X platform built on TMEC (Tencent Edge Computing Platform) allows vehicles to communicate directly with roadside wireless base stations, enabling data exchange between the dashcam and the edge computing terminal through the vehicle.

[0070] The data processing method of the present invention can be executed by a data processing system, which can be implemented in the form of hardware and / or software. The system can be configured in an electronic device, such as an edge computing terminal in a vehicle-mounted monitoring system. Figure 4 As shown, the data processing method includes:

[0071] S100: Obtain the video data to be detected uploaded by the driving recorder.

[0072] S200: Identify a first target object from the video data to be detected, and generate a detection frame of the first target object; wherein the first target object is a motor vehicle.

[0073] S300: Tracking the first target object using a multi-target tracking algorithm.

[0074] S400: Extracting lane information from the video data to be detected.

[0075] S500: Based on the lane information, detect illegal driving behavior of the first target object.

[0076] The present invention installs a driving recorder on a vehicle to capture the vehicle's illegal driving information, thereby overcoming the limitation that the traditional violation capture system is fixed in location and can only capture the violation driving information of vehicles at a fixed location where the violation capture system is installed. The present invention also overcomes the problems of damaging the buried coils on the highway and weak anti-interference in the traditional monitoring mode, reduces the cost of the monitoring system, and improves the monitoring efficiency. The present invention not only improves the flexibility and accuracy of violation detection, but also reduces the monitoring cost, improves road safety, and has important practical application value. Specifically, the present invention provides a detailed description of each step of the data processing method.

[0077] S100: Obtain the video data to be detected uploaded by the driving recorder.

[0078] In the specific implementation, the controller integrated in the dashcam is used to collect and upload video data. Specifically, the dashcam transmits the original video data to the vehicle terminal system through the vehicle WiFi, and the vehicle terminal system transmits and interacts with the edge computing terminal. The video data is transmitted to the edge computing terminal for real-time processing.

[0079] S200: Identify a first target object from the video data to be detected, and generate a detection frame of the first target object; wherein the first target object is a motor vehicle.

[0080] In this step, different types of traffic participants can be accurately identified and distinguished, providing precise target information for subsequent tracking and trajectory prediction.

[0081] In a specific implementation, deep learning algorithms such as YOLO (You Only Look Once) or SSD (Single Shot MultiBox Detector) can be used to analyze the video stream in real time to identify non-motor vehicles and motor vehicles. After detecting the target object, the algorithm will mark the location of the target object in the video frame and generate a bounding box to identify these objects. The system will classify the detected target objects into non-motor vehicles (first target object) and motor vehicles (second target object) according to the preset classification criteria.

[0082] In a specific implementation, the target detection model is a Yolov5s model, which has been used for the first target object and the second target object. In the present invention, the first object and the second object are a motor vehicle and a non-motor vehicle respectively.

[0083] S300: Tracking the first target object using a multi-target tracking algorithm.

[0084] In this step, the multi-target tracking algorithm can provide the motion information of the target objects over time, which is crucial for understanding the traffic flow dynamics and vehicle trajectories.

[0085] In the specific implementation, select a suitable multi-target tracking algorithm, such as DeepSORT or MOT (Multi-Object Tracking) algorithm, which can handle problems such as target occlusion, intersection and reappearance. The algorithm will track each target object and maintain the continuity of tracking even when the target temporarily leaves the field of view or reappears after being occluded. The system will update the position and motion status of each target object in real time to provide data support for trajectory prediction.

[0086] In a specific implementation, during the target detection process, non-motor vehicles and motor vehicles in each frame will be detected. In order to unify the information of the same target, it is necessary to give the same target a constant ID in different frames. The DeepSORT multi-target tracking algorithm is used. It adds DeepAssociationMetric and appearance image information to the target tracking of the original Sort algorithm to solve the problem of tracking occluded targets for a long time. It is a tracking framework based on deep appearance feature model and motion information model.

[0087] S400: Extracting lane information from the video data to be detected.

[0088] In a specific implementation, the lane information includes the number of lanes and the type of lane line (solid line, dashed line). Specifically, the implementation method is as follows:

[0089] S410: Edge detection and preprocessing: Use image processing techniques such as grayscale, Gaussian blur, and Canny edge detection to extract lane edge information.

[0090] The edges of the lane lines are extracted by graying and Gaussian blurring the video frames and then applying the Canny edge detection algorithm.

[0091] S420: Region of Interest (ROI) Mask: By defining a Region of Interest (ROI) mask, only the portion of the image containing the lane line is focused on, and other irrelevant regions are ignored.

[0092] In this step, a ROI mask is defined to focus only on the area below the image, since lane lines are usually located in the lower half of the road.

[0093] S430: Line detection: Detect straight lines in the image using Hough Transform, where the straight lines represent lane lines.

[0094] In the specific implementation, Hough transform is applied to detect straight lines in the image, and the straight lines are classified as solid or dashed lines according to their length and continuity. For example, if a straight line is long and continuous, we classify it as a solid line; if a straight line is short and discontinuous, we classify it as a dashed line.

[0095] S440: Lane classification: Classify the detected straight line as a solid line or a dashed line based on the length and continuity of the straight line. A long continuous straight line is identified as a solid line, while a short, discontinuous straight line may be identified as a dashed line.

[0096] S450: Lane number and position estimation: By using a clustering algorithm, specifically clustering based on discrete Fréchet distance, the lane trajectory is divided into lane trajectory segments and a road trajectory segment cluster is generated. The road trajectory segment cluster is fitted using a least squares estimation algorithm to extract estimated road feature parameters. Based on the expectation maximization algorithm and the mixed Gaussian model, the number of lanes and the target lane information corresponding to each lane are determined.

[0097] S500: Based on the lane information, detect illegal driving behavior of the first target object.

[0098] In the specific implementation, the lane information and the tracking data of the target object are combined to determine whether there is a violation by analyzing whether the vehicle crosses the lane line and other behaviors.

[0099] In a specific implementation of the present invention, the first target object performs detection of illegal driving behavior, specifically including the step of detecting a motor vehicle lane change violation on a solid line:

[0100] S511: Generate a left fitting point and a right fitting point based on the detection frame of the first target object; the left fitting point and the right fitting point are used to indicate the position of the first target object in the video data.

[0101] Specifically, the left fitting point is the coordinate of the lower left corner of the detection box, and the right fitting point is the coordinate of the lower right corner of the detection box.

[0102] In a specific implementation, for each detected motor vehicle, the coordinates of the lower left corner of its detection box are extracted as the left fitting point, and the coordinates of the lower right corner are extracted as the right fitting point. These coordinates are used to represent the position of the motor vehicle in the video data.

[0103] In a specific implementation, the calculation formula of the left fitting point is:

[0104] The calculation formula of the left fitting point is:

[0105] Among them, (cx, cy) represents the coordinates of the center point of the detection box, (w, h) represents the width and height of the detection box, (x1, y1) represents the coordinates of the left fitting point, and (x2, y2) represents the coordinates of the right fitting point.

[0106] S512: Based on the left fitting point and the right fitting point of the first target object and the lane information, determine whether the first target object has violated the lane change rule by using a solid line.

[0107] In a specific implementation, the left fitting point and the right fitting point of the motor vehicle are compared with the detected lane line. If the detection frame of the motor vehicle moves from one lane line to the other lane line in consecutive video frames and crosses the solid line, it can be determined as a solid line lane change.

[0108] Furthermore, by analyzing the position change of the motor vehicle in the video frame and combining the position of the lane line, it is determined whether the vehicle has crossed the solid line. If the left fitting point or the right fitting point of the vehicle crosses the solid line, this can be regarded as a violation of the solid line lane change.

[0109] It is understandable that due to the particularity of vehicles and the accuracy of violation identification, it is not possible to directly use the detection frame of the detection network for vehicle targets in the video, because the detection frame in each frame is always larger than the area of ​​the car itself, and for violation identification, some judgments are based on whether the car's wheels are on or over the line. Take the example of a car changing lanes on a solid line. Since the car's detection frame is still at a certain height in the picture, simply looking at the picture will cause the car body to have exceeded the lane line, but the car's wheels have not crossed the line. If the boundary of the detection frame or the center point of the detection frame is simply used for judgment, the result will be very inaccurate and it is easy to make a misjudgment.

[0110] In such a case, it is found through data analysis that the wheel position is the best as the vehicle identification position, but it is difficult to detect the wheel. Since the identification is based on a two-dimensional image, the present invention uses two fitted points, one on the left and one on the right, to represent the position of a vehicle. These two points are the lower left and lower right corners of the detection frame, respectively, to reduce the occurrence of misjudgment as much as possible.

[0111] Through the above technology, the on-board monitoring system can effectively detect and determine whether a motor vehicle has violated the regulations by changing lanes on a solid line, thereby improving road safety and traffic management efficiency.

[0112] In a preferred implementation, the detection of illegal driving behavior of the first target object specifically includes the step of detecting the motor vehicle driving in the wrong direction:

[0113] S521: Identify the right lane in the lane information.

[0114] In a specific implementation, the right lane line is identified by:

[0115] S5211: Image preprocessing: Convert video frames to grayscale images to reduce computational complexity and focus on brightness information. This is because lane lines are easier to detect in grayscale images.

[0116] S5212: Edge Detection: Apply the Canny edge detection algorithm to identify edges in an image. The Canny algorithm is a popular edge detection method that reduces noise and preserves edge information.

[0117] S5213: Lane Detection: Use Hough Transform to detect straight lines in an image. Hough Transform can detect straight line segments from an image, even if these straight line segments are interfered by noise or partially blocked.

[0118] S5214: Lane Classification: Distinguish the left lane and the right lane based on the slope of the detected lines. Usually, the slope of the left lane is negative, while the slope of the right lane is positive. By calculating the slope of each line, they can be classified as left or right lane.

[0119] In one specific implementation, straight lines with a slope greater than 0 are identified. In an image, these straight lines are usually located on the right side of the image, and their x-coordinates increase as the image moves downward. For the identified right lane line, the exact position of the lane line can be determined by a fitting algorithm (such as the least squares method).

[0120] S522: Identify a first target object whose detection box is located in the right lane.

[0121] S523: Identify the front of the first target object to determine whether the first target object has violated the traffic rules by driving in the wrong direction.

[0122] It is understandable that when the first target object is identified as the front of the vehicle, it can be determined that the first target object has violated the traffic rules. In the specific implementation, based on the AI ​​visual intelligent analysis algorithm, the system can identify the front or rear of the target vehicle. When it is identified as the front of the vehicle, it is determined that the traffic rules are violated.

[0123] In a preferred implementation, the data processing method further includes:

[0124] S600: When the first target object is detected to have illegal driving behavior, the license plate number information of the first target object is identified.

[0125] In the specific implementation, the PyTesseract OCR (Optical Character Recognition) tool is used to recognize the text on the license plate, which can convert the text in the image into a machine-readable text format. Before OCR, the license plate image needs to be preprocessed, including denoising, binarization, morphological operations, etc., to improve the recognition accuracy.

[0126] S700: Send the video data to be detected, violation detection results and license plate number information to the target device.

[0127] In the specific implementation, the detailed information of the violation event is sent to the monitoring center of the traffic management department in real time so that corresponding law enforcement actions can be taken. The violation detection results and license plate number information are integrated with the original video data to form a complete violation event record. The detailed information of the violation event is sent to the monitoring center of the traffic management department in real time so that corresponding law enforcement actions can be taken.

[0128] In a preferred implementation, the data processing method further includes:

[0129] S1000: Identify a first target object and a second target object in the video data to be detected, and generate corresponding first target detection frames and second target detection frames respectively; wherein the first target object is a motor vehicle, and the second target object is a non-motor vehicle.

[0130] In this step, different types of traffic participants can be accurately identified and distinguished, providing precise target information for subsequent tracking and trajectory prediction.

[0131] In a specific implementation, deep learning algorithms such as YOLO (You Only Look Once) or SSD (Single Shot MultiBox Detector) can be used to analyze the video stream in real time to identify non-motor vehicles and motor vehicles. After detecting the target object, the algorithm will mark the location of the target object in the video frame and generate a bounding box to identify these objects. The system will classify the detected target objects into non-motor vehicles (second target objects) and motor vehicles (first target objects) according to the preset classification criteria.

[0132] S2000: using a multi-target tracking algorithm to track the first target object and the second target object respectively.

[0133] In this step, the multi-target tracking algorithm can provide information about the motion of target objects over time, which is crucial for understanding traffic flow dynamics and predicting potential conflicts.

[0134] In the specific implementation, select a suitable multi-target tracking algorithm, such as DeepSORT or MOT (Multi-Object Tracking) algorithm, which can handle problems such as target occlusion, intersection and reappearance. The algorithm will track each target object and maintain the continuity of tracking even when the target temporarily leaves the field of view or reappears after being occluded. The system will update the position and motion status of each target object in real time to provide data support for trajectory prediction.

[0135] In a specific implementation, during the target detection process, pedestrians, non-motor vehicles and motor vehicles in each frame will be detected. In order to unify the information of the same target, it is necessary to give the same target a constant ID in different frames. The DeepSORT multi-target tracking algorithm is used, which adds

[0136] DeepAssociationMetric and appearance image information are used to solve the problem of tracking occluded targets for a long time. It is a tracking framework based on deep appearance feature model and motion information model.

[0137] S3000: using a trajectory prediction model to respectively predict the trajectories of the first target object and the second target object to obtain a first predicted trajectory and a second predicted trajectory.

[0138] In a specific implementation, machine learning or deep learning methods, such as recurrent neural networks (RNNs) or long short-term memory networks (LSTMs), are used to predict the future position and speed of each target object based on historical trajectory data.

[0139] In the specific implementation, the LSTM (Long Short-Term Memory) network prediction model is used to predict the future position and speed of each target object based on the historical trajectory data. The position and speed data obtained by the tracking algorithm are input into the trajectory prediction model. The model will output the future trajectory of each target object, namely the first predicted trajectory (motor vehicle) and the second predicted trajectory (non-motor vehicle).

[0140] In a specific implementation, the trajectory prediction model uses a Kalman filter, and may also use an extended Kalman filter predictor.

[0141] In a specific implementation, when the target object is detected, the trajectory of the object is predicted by a Kalman filter, and then the trajectory data of the predicted frame is combined and matched with the trajectory detections of the current frame using the Hungarian algorithm, and then the Kalman filter is updated.

[0142] Specifically, the DeepSORT algorithm passes the interaction ratio of the target detection box obtained by the YOLO algorithm and the previously predicted target tracking box as a combination into the Hungarian algorithm, and uses linear allocation to associate the IDs of each frame. It can also add the target's appearance and position information in the process of inter-frame matching calculation to achieve correct matching when the target object is occluded but reappears later, reducing the frequent jumps of the target ID and achieving the purpose of continuous tracking.

[0143] S4000: Calculate whether there are accident conflict points between multiple first predicted trajectories and between the first predicted trajectories and the second predicted trajectories.

[0144] The present invention calculates whether the trajectories of non-motor vehicles and motor vehicles intersect or approach each other, i.e., accident conflict points, by analyzing the predicted trajectories. This step can identify potential collision risks and provide a basis for early warning.

[0145] In the specific implementation, taking the road intersection without traffic signals as an example, collision problems may occur during the dynamic interaction between electric bicycles and non-motor vehicles, electric bicycles and motor vehicles, etc. By predicting whether there are accident conflict points between the trajectories. Specifically, the predicted trajectory of the target object can be matched to the background of the image data (calibrating points along the ground in the image and forming a predicted trajectory line), and then judging whether there are overlapping intersection points between the predicted trajectories to judge whether there are accident conflict points.

[0146] S5000: When there is an accident conflict point, generate road accident warning information.

[0147] In the specific implementation, once the accident conflict point is detected, the system will trigger the early warning logic and generate early warning information. The early warning information includes the location, time, and type of target object involved in the conflict point.

[0148] In a specific implementation, the road accident warning information includes the accident conflict level, the road location information, and the location information of the driving recorder. Specifically, the accident conflict level can be calibrated according to the conflict relationship between the target objects, such as the first-level accident conflict level between electric bicycles and non-motor vehicles; the second-level accident conflict level between electric bicycles and motor vehicles.

[0149] S6000: Based on the road accident warning information, instruct the corresponding driving recorder to capture video data of potential accidents.

[0150] In specific implementation, the driving recorder is instructed according to the early warning information to record the video data of the potential accident scene for subsequent analysis and evidence preservation.

[0151] S7000: Send road accident warning information, video data to be detected and potential accident video data to the target device.

[0152] In specific implementation, through wireless communication technology, such as 4G / 5G network, the warning information and related video data are transmitted in real time to the target device, such as the smart phone of the owner's family, or the monitoring system of the traffic management center. This step ensures that the warning information can be conveyed to the personnel who need to take action in time, improving the accident response and timely rescue.

[0153] Through the S1000-S7000 of the present invention, by real-time monitoring and analysis of road conditions, potential traffic accident risks can be discovered in time and preventive measures can be taken. Through the prediction model and multi-target tracking algorithm, the system can predict and identify possible accident conflict points, issue early warnings, and provide drivers or traffic management departments with enough time to take risk avoidance measures.

[0154] The present invention also provides a data processing system for a vehicle-mounted monitoring system, the vehicle-mounted monitoring system comprising a plurality of driving recorders and a plurality of edge computing terminals; one driving recorder can communicate and interact with a plurality of edge computing terminals; the data processing system is arranged in the edge computing terminal, and the data processing system comprises:

[0155] An acquisition module is used to acquire the video data to be detected uploaded by the driving recorder;

[0156] an identification module, which is used to identify a first target object from the video data to be detected and generate a detection frame of the first target object; wherein the first target object is a motor vehicle;

[0157] A tracking module, which is used to track the first target object using a multi-target tracking algorithm;

[0158] An extraction module, used for extracting lane information from the video data to be detected;

[0159] A detection module is used to detect illegal driving behavior of the first target object based on the lane information.

[0160] Preferably, the detection of illegal driving behavior of the first target object specifically includes the step of detecting a motor vehicle lane change violation on a solid line:

[0161] Based on the detection frame of the first target object, generating a left fitting point and a right fitting point; the left fitting point and the right fitting point are used to represent the position of the first target object in the video data;

[0162] Based on the left fitting point and the right fitting point of the first target object and the lane information, it is determined whether the first target object has a lane change violation of a solid line.

[0163] Preferably, the left fitting point is the coordinate of the lower left corner of the detection frame, and the right fitting point is the coordinate of the lower right corner of the detection frame;

[0164] The calculation formula of the left fitting point is:

[0165] The calculation formula of the left fitting point is:

[0166] Among them, (cx, cy) represents the coordinates of the center point of the detection box, (w, h) represents the width and height of the detection box, (x1, y1) represents the coordinates of the left fitting point, and (x2, y2) represents the coordinates of the right fitting point.

[0167] The present invention also provides a driving recorder, which is used to deploy a vehicle-mounted monitoring system to implement the above-mentioned data processing method for a data processing system.

[0168] The driving recorder of the present invention comprises a housing and a mounting bracket, wherein the housing is equipped with circuit board modules such as a camera and a processor.

[0169] like Figure 1 As shown, the product of the present invention has better versatility and practicality, and can be used on motorcycles, electric bicycles and cars respectively. Specifically, the mounting bracket includes a connecting member 10, a first clamp arm 20, a second clamp arm 30 and an elastic member. The connecting member is used to connect the camera; the first clamp arm is connected to the connecting member; the second clamp arm is hinged to the first clamp arm, and the second clamp arm is located below the first clamp arm. The second clamp arm has a planar structure; the elastic member is connected to the first clamp arm and the second clamp arm, and the elastic member is used to drive the first clamp arm and the second clamp arm to clamp each other. Among them, the mounting bracket can form a clamping mechanism through the first clamp arm and the second clamp arm for clamping on an external carrier; the mounting bracket can also be stably placed on the external carrier through the planar structure of the second clamp arm.

[0170] In the present technology, the connecting piece is the basic part of the mounting bracket, which is used to connect the camera. The elastic piece is used to provide elastic force so that the first clamping arm and the second clamping arm can be clamped together, thereby being fixed to the relevant part of the motorcycle or electric bicycle.

[0171] In a specific implementation, the connecting member is hinged to the first clamp arm, and the connecting member can swing around the hinge. The connecting member is a T-shaped structure, and the first clamp arm is provided with a mounting groove structure adapted to the connecting member.

[0172] This patent forms a clamping mechanism by cooperating with the first clamping arm and the second clamping arm, and under the action of the elastic member, the clamping mechanism can stably clamp on the handlebars of motorcycles and electric bicycles, so as to meet the use of motorcycles / electric bicycles. In order to adapt to the use of family cars, the planar structure of the second clamping arm allows it to be stably placed on the car interior, such as the co-pilot dashboard.

[0173] In a specific implementation, the first clamp arm, the second clamp arm and the connecting member are all plastic parts molded by plastic injection molding.

[0174] The main bodies of the first clamp arm and the second clamp arm are relatively flat. The first clamp wall has two bent clamp claws 21, which are arranged relatively spaced apart, and the ends of the clamp claws extend toward the second clamp arm. The bent clamp claws can form a claw-shaped clamping mechanism, which can adapt to more external carriers and improve the practicality of the mounting bracket.

[0175] In a preferred embodiment, a first rubber piece is disposed at the end of the clamping jaw; a second rubber piece is disposed at the end of the second clamping arm adjacent to the clamping jaw; wherein both the first rubber piece and the second rubber piece are rubber blocks with elliptical cross-sections.

[0176] Through this design, the rubber part with an oval cross-section can provide better friction when compressed, which helps the clamping jaws to grip the structure of the motorcycle or electric bicycle more firmly and prevent it from slipping during use. On the other hand, the rubber part can adapt to different contact surface shapes and maintain good contact.

[0177] In a specific implementation of the present invention, the elastic member is a torsion spring. It is understandable that the first clamp arm and the second clamp arm of the present invention are like clamps, and the torsion spring provides a stable clamping force, so that the first clamp arm and the second clamp arm can achieve a clamping effect. Specifically, using a torsion spring, a first clamp arm and a second clamp arm to form a clamping mechanism is common knowledge in the art and can be easily implemented by those skilled in the art, so no further explanation is given here.

[0178] In a preferred embodiment, the second clamping arm includes a plastic shell and a plurality of counterweights disposed in the plastic shell.

[0179] Among them, the counterweight block increases the weight of the mounting bracket itself, and is mainly used to stabilize the center of gravity of the mounting bracket and increase the weight of the mounting bracket itself to ensure stability when placed on the car interior so that it can remain stable.

[0180] In a preferred implementation, the mounting bracket also includes a placement plate 50, which has a plurality of magnets built therein and is used to be placed on an external carrier; wherein the counterweight block of the second clamping arm is an iron block, and the second clamping arm can be magnetically connected to the placement plate through the counterweight block.

[0181] The magnetic connection allows the second clamp arm and the placement plate to be quickly and easily installed or removed without complicated alignment or fastening operations, which improves the convenience of use. The placement plate is designed to be used in a car, where the placement plate serves as a base for the mounting bracket to provide stable placement. With the placement plate as a base, the placement plate has a larger area for more stable placement on the car interior. The second clamp arm is then magnetically connected to the placement plate to stably place the entire mounting bracket.

[0182] In a specific implementation, the placement plate includes a flexible pad 51 and a plastic base 52 fixedly connected to the flexible pad; wherein the magnet is disposed in the plastic base.

[0183] In the present technology, the flexible pad can adapt to different carrier surfaces, including curved or uneven surfaces, to improve applicability. The surface of the flexible pad usually has good friction properties, which can increase the friction between the placement plate and the carrier to prevent the camera from sliding or shifting.

[0184] For the data processing method of the vehicle-mounted monitoring system described in this embodiment and other structures of the system, refer to the prior art.

[0185] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Therefore, any modification, equivalent change and modification made to the above embodiment according to the technical essence of the present invention without departing from the technical solution of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. A data processing method for a vehicle-mounted monitoring system, characterized in that: The vehicle monitoring system includes multiple driving recorders and multiple edge computing terminals; one driving recorder can communicate and interact with multiple edge computing terminals; the edge computing terminal is used to implement the data processing method, and the data processing method includes the following steps: Obtain the video data to be detected uploaded by the driving recorder; Identify a first target object from the video data to be detected, and generate a detection frame of the first target object; wherein the first target object is a motor vehicle; Tracking the first target object using a multi-target tracking algorithm; Extracting lane information from the video data to be detected; Based on the lane information, illegal driving behavior of the first target object is detected.

2. The data processing method of a vehicle monitoring system according to claim 1, characterized in that: The first target object is detected for illegal driving behavior, specifically including the steps of detecting illegal lane changes of motor vehicles on a solid line: Based on the detection frame of the first target object, generating a left fitting point and a right fitting point; the left fitting point and the right fitting point are used to represent the position of the first target object in the video data; Based on the left fitting point and the right fitting point of the first target object and the lane information, it is determined whether the first target object has a lane change violation of a solid line.

3. The data processing method of a vehicle monitoring system according to claim 2, characterized in that: The left fitting point is the coordinate of the lower left corner of the detection frame, and the right fitting point is the coordinate of the lower right corner of the detection frame; The calculation formula of the left fitting point is: The calculation formula of the left fitting point is: Among them, (cx, cy) represents the coordinates of the center point of the detection box, (w, h) represents the width and height of the detection box, (x1, y1) represents the coordinates of the left fitting point, and (x2, y2) represents the coordinates of the right fitting point.

4. The data processing method of a vehicle monitoring system according to claim 1, characterized in that: The first target object is detected for illegal driving behavior, specifically including the steps of detecting the motor vehicle driving in the wrong direction: identifying a right lane in the lane information; Identify a first target object whose detection frame is located in the right lane; The front of the first target object is identified to determine whether the first target object has violated the traffic regulations by driving in the wrong direction.

5. The data processing method of a vehicle monitoring system according to claim 1, characterized in that: The data processing method further includes: When the first target object is detected to have illegal driving behavior; Identify the license plate information of the first target object; The video data to be detected, the violation detection result and the license plate number information are sent to the target device.

6. The data processing method of a vehicle monitoring system according to claim 1, characterized in that: The data processing method further includes: Identify a first target object and a second target object in the video data to be detected, and generate corresponding first target detection frames and second target detection frames respectively; wherein the first target object is a motor vehicle and the second target object is a non-motor vehicle; Using a multi-target tracking algorithm to track the first target object and the second target object respectively; Using a trajectory prediction model to predict the trajectories of the first target object and the second target object respectively, to obtain a first predicted trajectory and a second predicted trajectory; Calculating whether there is an accident conflict point between a plurality of first predicted trajectories and between the first predicted trajectories and the second predicted trajectories; When there is an accident conflict point, generate road accident warning information; Based on the road accident warning information, instruct the corresponding driving recorder to capture potential accident video data; Send road accident warning information, video data to be detected and potential accident video data to the target device.

7. A data processing system for a vehicle-mounted monitoring system, characterized in that: The vehicle monitoring system includes multiple driving recorders and multiple edge computing terminals; one driving recorder can communicate and interact with multiple edge computing terminals; the data processing system is arranged on the edge computing terminal, and the data processing system includes: An acquisition module is used to acquire the video data to be detected uploaded by the driving recorder; an identification module, which is used to identify a first target object from the video data to be detected and generate a detection frame of the first target object; wherein the first target object is a motor vehicle; A tracking module, which is used to track the first target object using a multi-target tracking algorithm; An extraction module, used for extracting lane information from the video data to be detected; A detection module is used to detect illegal driving behavior of the first target object based on the lane information.

8. The data processing system of the vehicle monitoring system according to claim 7, characterized in that: The detection module detects the illegal driving behavior of the first target object, specifically including the step of detecting the illegal lane change of a motor vehicle on a solid line: Based on the detection frame of the first target object, generating a left fitting point and a right fitting point; the left fitting point and the right fitting point are used to represent the position of the first target object in the video data; Based on the left fitting point and the right fitting point of the first target object and the lane information, it is determined whether the first target object has a lane change violation of a solid line.

9. The data processing system of the vehicle-mounted monitoring system according to claim 8, characterized in that: The left fitting point is the coordinate of the lower left corner of the detection frame, and the right fitting point is the coordinate of the lower right corner of the detection frame; The calculation formula of the left fitting point is: The calculation formula of the left fitting point is: Among them, (cx, cy) represents the coordinates of the center point of the detection box, (w, h) represents the width and height of the detection box, (x1, y1) represents the coordinates of the left fitting point, and (x2, y2) represents the coordinates of the right fitting point.

Images