Ecort management method and system based on Beidou positioning and video monitoring

Abstract

The invention discloses an escort management method and system based on Beidou positioning and video monitoring, and belongs to the technical field of satellite positioning and video monitoring fusion, and the method comprises the steps: obtaining Beidou dynamic data of an escort vehicle in real time, carrying out the feature extraction, obtaining a vehicle state data set, and judging the type of an abnormal event; generating a corresponding abnormal event trigger signal, selecting a preset region of interest to generate a video calling instruction and a picture focusing instruction, collecting a real-time video picture of the region of interest, and generating a focused video picture and a risk identification level; and performing comprehensive analysis in combination with the vehicle position coordinates, generating a comprehensive risk value, and generating an intelligent scheduling instruction for resource scheduling in combination with preset escorting resource information. According to the invention, multi-source data fusion analysis, dynamic video focusing and environment adaptive evaluation technologies are adopted, an active defensive intelligent escort management and control system can be constructed, and accurate resource scheduling and all-scene risk co-processing are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of satellite positioning and video monitoring fusion, in particular to a method and system for escort management based on Beidou positioning and video monitoring. BACKGROUND

[0002] Satellite positioning and video monitoring technologies are commonly used in the field of safe monitoring of escort vehicles. The Beidou satellite navigation system can provide real-time dynamic data such as position, speed, and heading, and the video monitoring system can collect the surrounding environment of the vehicle through a camera. These two types of technologies are used for trajectory tracking and visual verification, respectively, and constitute the basic guarantee means for safe escort.

[0003] The prior art usually deploys Beidou positioning terminals and vehicle-mounted monitoring devices independently. Positioning data is used for electronic fence crossing alarm, and the video system relies on manual round patrol or fixed angle monitoring. Some solutions use a simple linkage mechanism to call the preset camera screen when the positioning is abnormal, but the selection of the monitoring area lacks event correlation.

[0004] However, the isolated data of the positioning and video systems leads to one-sided risk judgment, and the response to abnormal events relies on manual review, causing delay in disposal. The fixed monitoring angle cannot dynamically adapt to the observation needs of different event types, and there are monitoring blind spots. The positioning drifts seriously when the satellite signal is blocked, and there is a lack of effective auxiliary calibration mechanism, which affects the reliability of decision-making in high-risk areas. SUMMARY

[0005] To solve the above problems, the present application provides a method and system for escort management based on Beidou positioning and video monitoring, which uses multi-source data fusion analysis, dynamic video focusing, and environment adaptive evaluation technology to build an active defense type intelligent escort management and control system, and realizes precise resource scheduling and all-scenario risk collaborative disposal.

[0006] The above objectives can be achieved by the following solutions: The method and system for escort management based on Beidou positioning and video monitoring comprises the following steps: real-time acquisition of Beidou dynamic data of an escort vehicle and feature extraction to obtain a vehicle state data set, wherein the vehicle state data set comprises vehicle position coordinates, vehicle speed change rate, vehicle heading angle and vehicle stay duration; determination of an abnormal event type according to the vehicle state data set to generate a corresponding abnormal event trigger signal; selection of a preset attention area according to the abnormal event type corresponding to the abnormal event trigger signal to generate a video retrieval instruction and a picture focusing instruction; real-time video picture acquisition of the attention area according to the video retrieval instruction and the picture focusing instruction to generate a focused video picture; analysis of the focused video picture to generate a risk identification level; comprehensive analysis of the vehicle position coordinates, the focused video picture and the risk identification level to generate a comprehensive risk value; generation of an intelligent scheduling instruction based on the comprehensive risk analysis result and preset escort resource information; and resource scheduling according to the intelligent scheduling instruction.

[0007] Optionally, the generation of the abnormal event trigger signal comprises: when the deviation degree of the vehicle position coordinates from a preset electronic fence route is greater than a preset deviation degree threshold, a route deviation event is determined and a deviation event trigger signal is generated; when the speed change rate exceeds a preset acceleration threshold, an abnormal motion state event is determined and an abnormal motion trigger signal is generated; when the vehicle heading angle is greater than a preset angle threshold or the vehicle stay duration is greater than a preset time threshold, an emergency alarm event is determined and an emergency alarm trigger signal is generated; wherein the abnormal event trigger signal comprises the deviation event trigger signal, the abnormal motion trigger signal and the emergency alarm trigger signal.

[0008] Optionally, the generation of the video retrieval instruction and the picture focusing instruction comprises: when the abnormal event trigger signal is the deviation event trigger signal, a vehicle two-side area is selected; when the abnormal event trigger signal is the abnormal motion trigger signal, a motion direction sector area is selected; when the abnormal event trigger signal is the emergency alarm trigger signal, a vehicle peripheral panoramic area is selected; generation of the video retrieval instruction and the picture focusing instruction according to the selected area; wherein the attention area comprises the vehicle two-side area, the motion direction sector area and the vehicle peripheral panoramic area.

[0009] Optionally, the generation of the focused video picture comprises: selection of a preset camera group corresponding to the attention area according to the video retrieval instruction and the picture focusing instruction; adjustment of shooting parameters of the camera group according to the vehicle state data set, wherein the shooting parameters comprise a holder angle parameter and a focal length parameter; generation of the focused video picture according to the shooting parameters.

[0010] Optionally, the method further comprises: collecting position coordinates of historical abnormal event occurrences to obtain abnormal coordinate points; dividing a map into regions according to a preset number to obtain basic management regions; setting an alert level for the basic management regions according to the number of abnormal coordinate points of the basic management regions; and adjusting the shooting frame rate of the camera group according to the vehicle position coordinates and the alert level.

[0011] Optionally, the analyzing the focused video picture to generate a risk identification level comprises: extracting dynamic target behavior features in the focused video picture; matching the dynamic target behavior features with a preset behavior rule library to generate a behavior abnormality identification according to a matching result; and generating a risk identification level according to the abnormality identification and the abnormal event trigger signal.

[0012] Optionally, the comprehensive analysis of the vehicle position coordinates, the focused video picture, and the risk identification level to generate a comprehensive risk value comprises: extracting geographical identification features from a preset position feature library according to the vehicle position coordinates; extracting environment features from the focused video picture; matching the environment features with the geographical identification features to generate a risk level correction factor according to a matching result; and calculating a comprehensive risk value by using the risk level correction factor and the risk identification level.

[0013] Optionally, the method further comprises: determining the alert level according to the vehicle position coordinates and the basic management regions; and correcting the risk level correction factor according to the alert level.

[0014] Optionally, the method further comprises: monitoring a strength parameter of a Beidou positioning signal to generate a signal strength index; obtaining reference coordinate data according to the geographical identification features when the strength signal is lower than a preset strength threshold; and correcting the vehicle position coordinates according to the reference coordinate data and the shooting parameters.

[0015] Based on the same inventive concept, the application also provides a convoy management system based on Beidou positioning and video monitoring, which comprises: a Beidou positioning module for real-time acquisition of Beidou dynamic data of a convoy vehicle and feature extraction to obtain a vehicle state data set, the vehicle state data set comprising vehicle position coordinates, vehicle speed change rate, vehicle heading angle and vehicle stay time; an anomaly detection module for judging an abnormal event type according to the vehicle state data set and generating a corresponding abnormal event trigger signal; a video guidance module for selecting a preset area of interest according to the abnormal event type corresponding to the abnormal event trigger signal and generating a video retrieval instruction and a picture focusing instruction; an intelligent monitoring module for collecting real-time video pictures of the area of interest according to the video retrieval instruction and the picture focusing instruction and generating focused video pictures; a behavior analysis module for analyzing the focused video pictures and generating a risk identification level; a risk assessment module for comprehensive analysis of the vehicle position coordinates, the focused video pictures and the risk identification level and generating a comprehensive risk value; a scheduling module for generating an intelligent scheduling instruction based on the comprehensive risk analysis result and preset convoy resource information; and an execution module for resource scheduling according to the intelligent scheduling instruction.

[0016] Compared with the prior art, the application has the following advantages: 1. The application realizes multi-source collaborative perception of abnormal events; through real-time cross verification of Beidou dynamic data and video behavior features, the recognition accuracy of risk scenarios such as route deviation, sudden acceleration and long stay is significantly improved, and the single sensor misjudgment probability is reduced; 2. The application optimizes the monitoring resource dynamic allocation mechanism; according to the event type, the intelligent focusing strategy of the vehicle two-side area, the motion direction sector area or the panoramic area is automatically selected to ensure that the key area is not missed, and the video analysis efficiency is greatly improved; 3. The application establishes an environment-adaptive risk assessment model; the risk value is matched and corrected in combination with the geographical feature library and the real-time picture, the determination threshold is automatically adjusted in the historically high-incidence area, and the false alarm rate and the missed alarm rate are effectively balanced; 4. The application enhances the system robustness in complex scenarios; through video reference assisted Beidou positioning calibration, the continuous and reliable positioning in satellite signal shielding areas such as tunnels and underground spaces is ensured, and the whole-process risk control ability is maintained.

[0017] Other features and advantages of the application will be set forth in the specification, and in part will become apparent from the specification, or can be learned by practice of the application. The objectives and other advantages of the application can be realized and attained by the structure particularly pointed out in the specification, claims and drawings. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to make the technical solutions in the embodiments of the present application or the prior art clearer, the accompanying drawings needed in the embodiments or prior art description will be briefly introduced below. Obviously, the accompanying drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort on the basis of these drawings.

[0019] Figure 1 FIG. 1 is a flowchart of a security escort management method based on Beidou positioning and video monitoring according to an embodiment of the present application.

[0020] Figure 2 FIG. 2 is a schematic diagram of a concerned area corresponding to a deviation event trigger signal according to an embodiment of the present application.

[0021] Figure 3 FIG. 3 is a schematic diagram of a concerned area corresponding to an abnormal motion trigger signal according to an embodiment of the present application.

[0022] Figure 4 FIG. 4 is a schematic diagram of a concerned area corresponding to an emergency alarm trigger signal according to an embodiment of the present application.

[0023] Figure 5 FIG. 5 is a schematic diagram of a security alert level division according to an embodiment of the present application.

[0024] Figure 6 FIG. 6 is a comparison diagram of a revised risk level correction factor according to an embodiment of the present application.

[0025] Figure 7 FIG. 7 is a structural schematic diagram of a security escort management system based on Beidou positioning and video monitoring according to an embodiment of the present application. DETAILED DESCRIPTION

[0026] In order to make the technical solutions in the embodiments of the present application or the prior art clearer, the accompanying drawings needed in the embodiments or prior art description will be briefly introduced below. Obviously, the accompanying drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort on the basis of these drawings.

[0027] With reference to Figure 1 , one embodiment of the present application proposes a security escort management method based on Beidou positioning and video monitoring, which adopts multi-source data fusion analysis, dynamic video focusing and environment adaptive evaluation technology, can construct an active defense type intelligent security escort management and control system, and realize accurate resource scheduling and full-scene risk collaborative disposal.

[0028] The method according to the embodiment specifically includes: Real-time acquisition of Beidou dynamic data of the escort vehicle and feature extraction to obtain a vehicle state data set, the vehicle state data set including vehicle position coordinates, vehicle speed change rate, vehicle heading angle and vehicle stay duration; Judging an abnormal event type according to the vehicle state data set to generate a corresponding abnormal event trigger signal; According to the abnormal event type corresponding to the abnormal event trigger signal, a preset attention area is selected to generate a video retrieval instruction and a picture focusing instruction; According to the video retrieval instruction and the picture focusing instruction, real-time video pictures of the attention area are collected to generate focused video pictures; Analyzing the focused video pictures to generate a risk identification level; Comprehensively analyzing the vehicle position coordinates, the focused video pictures and the risk identification level to generate a comprehensive risk value; Based on the comprehensive risk analysis result and preset escort resource information, an intelligent scheduling instruction is generated; Resource scheduling is performed according to the intelligent scheduling instruction.

[0029] Specifically, a closed-loop risk management and control system is constructed by fusing Beidou positioning and video monitoring technologies. Firstly, dynamic feature data such as vehicle position coordinates, speed change rate, heading angle and stay duration are extracted in real time, and three types of events, i.e., route deviation, abnormal motion and emergency alarm, are automatically identified based on preset rules. According to the event type, the vehicle's two sides, the motion direction sector or the surrounding panoramic area are selected as the monitoring focus, and the optical parameters of the camera group are dynamically adjusted to obtain the focused pictures. Through video behavior analysis, a risk identification level is generated, and risk correction is performed by matching geographical environmental features. Finally, a quantitative risk value is generated by comprehensively analyzing the vehicle position, video pictures and risk level, which drives the intelligent scheduling of escort resources. Multi-dimensional risk collaborative perception and active defense in the escort process are realized. Through real-time linkage of positioning data and video monitoring, the accuracy of abnormal event identification and response speed are significantly improved; the intelligent focusing mechanism optimizes the allocation of monitoring resources, ensuring that key areas are not missed; the dual verification mechanism of environmental features and behavior analysis effectively reduces the false alarm rate; the intelligent scheduling strategy based on dynamic risk assessment enhances the emergency handling capability, comprehensively improves the safety and efficiency of escort, and reduces the cost of human monitoring.

[0030] Optionally, the generation of the abnormal event trigger signal comprises: When the deviation degree of the vehicle position coordinates from a preset electronic fence route is greater than a preset deviation degree threshold, a route deviation event is determined and a deviation event trigger signal is generated; When the speed change rate exceeds a preset acceleration threshold, an abnormal motion state event is determined and an abnormal motion trigger signal is generated; When the vehicle heading angle is greater than a preset angle threshold or the vehicle stay duration is greater than a preset time threshold, then the emergency alarm event is determined and an emergency alarm trigger signal is generated; The abnormal event trigger signal includes the deviation event trigger signal, the abnormal motion trigger signal and the emergency alarm trigger signal.

[0031] Specifically, first, the vehicle position coordinates, the vehicle speed change rate, the vehicle heading angle and the vehicle stay duration are extracted from the vehicle state data set, which are obtained in real time by the Beidou positioning module; second, different abnormal event types are judged. For the route deviation event, the preset electronic fence route is a pre-set path coordinate sequence that the vehicle should follow, and the deviation degree is defined as the shortest perpendicular distance from the vehicle position coordinates to the preset electronic fence route. For the deviation degree , , wherein is the real-time value of the vehicle position coordinates, obtained by the Beidou positioning data, is the coordinate of the nearest point on the preset electronic fence route, obtained by querying the preset route database, and the deviation degree threshold is a preset maximum deviation distance value. If the calculated deviation degree is greater than the deviation degree threshold, the route deviation event is determined and a deviation event trigger signal is generated. For the abnormal motion state event, the speed change rate is the acceleration. For the speed change rate , , wherein is the current speed value, is the previous speed value, both of which are obtained from the vehicle speed change rate field of the vehicle state data set, is a time interval, which is obtained by system clock calculation, and the acceleration threshold is a preset maximum allowed acceleration value, if the speed change rate is greater than the acceleration threshold, it is determined that the abnormal motion state event occurs and the abnormal motion trigger signal is generated. For the emergency alarm event, the vehicle heading angle is an angle value of the vehicle forward direction, which is directly obtained from the vehicle state data set, the angle threshold is a preset maximum allowed heading angle deviation value, the vehicle stay duration is the duration when the vehicle speed is 0, which is obtained from the vehicle state data set, the time threshold is a preset maximum allowed stay time value, if the vehicle heading angle is greater than the angle threshold or the vehicle stay duration is greater than the time threshold, it is determined that the emergency alarm event occurs and the emergency alarm trigger signal is generated; finally, the abnormal event trigger signal includes the deviation event trigger signal, the abnormal motion trigger signal and the emergency alarm trigger signal. Based on real-time monitoring of vehicle dynamic data, the abnormal event signal is automatically triggered by the preset threshold, combined with subsequent video analysis and resource scheduling, the risk rapid response is realized, the potential risk can be automatically identified, the manual intervention is reduced, and the safety and efficiency of the escort process are improved.

[0032] Exemplarily, assuming that a vehicle is in the process of driving, the real-time obtained vehicle position coordinates longitude and latitude are , the preset electronic fence route nearest point is , the deviation degree is calculated as 0.14 kilometers, and the deviation degree threshold is preset as 0.1 kilometers. Since the deviation degree is greater than the deviation degree threshold, it is determined that the route deviation event occurs and the deviation event trigger signal is generated; the video monitoring can be immediately triggered to focus on the vehicle side area, and whether there is external interference is checked, so that the emergency response is quickly started, the loss of goods caused by hijacking or route error is effectively prevented, and the safety and efficiency of the escort are improved.

[0033] Optionally, the generating video calling instruction and picture focusing instruction comprises: When the abnormal event trigger signal is the deviation event trigger signal, the vehicle side area is selected; When the abnormal event trigger signal is the abnormal motion trigger signal, the motion direction sector area is selected; When the abnormal event trigger signal is the emergency alarm trigger signal, the vehicle peripheral panoramic area is selected; According to the selected area, the video calling instruction and the picture focusing instruction are generated; Wherein, the concerned area includes the vehicle side area, the motion direction sector area and the vehicle peripheral panoramic area.

[0034] Specifically, first, an abnormal event trigger signal is received, which is specifically one of a deviation event trigger signal, an abnormal motion trigger signal, or an emergency alarm trigger signal; then, according to the type of the specific abnormal event trigger signal received, a corresponding focus area type is selected from the preset focus area mapping rules; when the abnormal event trigger signal is a deviation event trigger signal, the areas on both sides of the vehicle are selected as focus areas, such as Figure 2 As shown in , the area is defined as a rectangular strip space with a preset width W extending to the left and right of the vehicle centerline as the symmetry axis; when the abnormal event trigger signal is an abnormal motion trigger signal, the motion direction fan-shaped area is selected as the focus area, such as Figure 3 As shown, the area is defined as the current position of the vehicle as the vertex, along the current heading angle direction, and the opening angle is the preset angle value. , the radius is the preset length value When the abnormal event trigger signal is an emergency alarm trigger signal, the panoramic area around the vehicle is selected as the focus area, such as Figure 4 As shown, the area is defined as the vehicle as the center and the radius is the preset safety distance value. circular space; finally, specific video retrieval instructions and picture focus instructions are generated according to the selected area of ​​interest type and its spatial parameters. The video retrieval instruction contains the camera group number to be activated, and the picture focus instruction contains the spatial coordinate range parameters of the target area; the area of ​​interest type strictly corresponds to the three preset types of areas on both sides of the vehicle, fan-shaped areas in the direction of movement, and panoramic areas around the vehicle; through intelligent matching of event types and monitoring areas, efficient and accurate deployment of monitoring resources is achieved.

[0035] Optionally, generating a focused video picture includes: Selecting a preset camera group corresponding to the area of ​​interest according to the video retrieval instruction and the picture focus instruction; Adjusting the shooting parameters of the camera group according to the vehicle status data set, the shooting parameters including pan / tilt angle parameters and focal length parameters; A focused video image is generated according to the shooting parameters.

[0036] Specifically, receive the video call command and the picture focus command, parse the focus area type and spatial parameters contained in the command; query the preset camera configuration mapping table according to the focus area type, select the camera group of the corresponding area, and the mapping table stores the correspondence between different area types and camera group numbers; obtain the vehicle position coordinates and vehicle heading angle in the vehicle status data set; calculate the pan / tilt angle parameters of each camera in the camera group, and use the azimuth angle to calculate the pan / tilt angle parameters for the areas on both sides of the vehicle. The pan / tilt angle parameters are equal to Degrees, for the fan-shaped area in the direction of motion, the pan / tilt angle parameter is equal to A pan-tilt angle parameter of a vehicle peripheral panoramic area A focal length parameter is calculated , wherein is a coordinate of a boundary point of a region of interest, is a coordinate of a camera position; the focal length parameter is calculated A focal length parameter is calculated , wherein is a camera installation height, is a distance from the vehicle to the boundary of the region of interest, is an optical conversion coefficient, and obtained from a three-dimensional map database, determined by camera optical parameters; the pan-tilt angle parameter and the focal length parameter are sent to a camera group; the camera group adjusts the optical lens and the pan-tilt direction according to the parameters, and collects real-time video streams; electronic image stabilization and region cropping are performed on the video streams, and focused video pictures within the region of interest are output.

[0037] Exemplarily, when the system generates focused video pictures of the regions on both sides of the vehicle for the escort vehicle in the tunnel, the lateral camera group numbered CT01-CT04 is selected, the vehicle heading angle is equal to 285 degrees, the left camera pan-tilt angle parameter is calculated to be equal to 195 degrees, and the right camera pan-tilt angle parameter is calculated to be equal to 375 degrees, i.e. 15 degrees, according to the camera installation height equal to 5 meters, the distance from the vehicle to the boundary of the region of interest equal to 1.5 meters, and the camera optical conversion coefficient equal to 120, the focal length c number is calculated to be equal to 400 pixel units, and the adjusted camera captures clear pictures of the tunnel side wall; the problem of positioning drift in a closed space is overcome, the influence of insufficient light and limited view angle in the tunnel is eliminated through dynamic optical parameter adjustment, continuous clear monitoring of suspicious targets on both sides of the vehicle is ensured, the risk of hidden approach specific to the tunnel environment is effectively prevented, and the safety protection capability in special terrain is improved.

[0038] Optionally, the method further comprises: collecting position coordinates of historical abnormal events to obtain abnormal coordinate points; dividing the map into regions according to a preset number to obtain basic management regions; setting an alert level for the basic management region according to the number of abnormal coordinate points of the basic management region; adjusting the shooting frame rate of the camera group according to the vehicle position coordinate and the alert level.

[0039] Specifically, the alert level is divided as Figure 5As shown, the points in the figure represent abnormal coordinate points, and the grid represents the basic management area. First, the location coordinates of historical abnormal events are collected, and the records of past abnormal events are extracted from the system database, each record containing latitude and longitude coordinate values. The coordinates are summarized to obtain a set of abnormal coordinate points. Then, the map is divided into regions according to a preset number, which is the total number of grid cells N or a single grid size parameter. The map is divided into N equal-area rectangular grids by a geographic information system, and each grid is defined as a basic management area. Then, the alert level is set according to the number of abnormal coordinate points in the basic management area. The number of abnormal coordinate points n contained in each basic management area is calculated, and the alert level is divided into three levels: low, medium, and high. When n is less than a preset threshold T1, it is set to a low alert level; when n is greater than or equal to T1 and less than T2, it is set to a medium alert level; and when n is greater than or equal to T2, it is set to a high alert level. T1 and T2 are preset positive integer thresholds. Finally, the shooting frame rate of the camera group is adjusted according to the vehicle location coordinates and the alert level. The vehicle location coordinates are obtained in real time, and the basic management area where the vehicle is located is determined through coordinate matching. The alert level of the area is queried, and the shooting frame rate adjustment value is calculated based on the alert level : , wherein is the preset value of the basic frame rate of the camera group, is the adjustment coefficient corresponding to the alert level, when the alert level is low equals 0.5, when it is medium equals 1, and when it is high equals 2, and the values are obtained from the camera configuration library. After applying the adjustment value, the instruction is issued to the camera group to realize dynamic control of the frame rate. Through historical data analysis and optimization of real-time monitoring intensity, the monitoring resources can be intelligently allocated. In low-risk areas, energy consumption is reduced to prolong the service life of the equipment, and in high-risk areas, the video capture capability is enhanced to improve the threat recognition rate, thereby realizing sustainable and efficient operation of the escort system.

[0040] Exemplarily, assuming that the preset total number of map division grids N is equal to 100, a single basic management area is 1 square kilometer, a total of 50 historical abnormal coordinate points are collected, a certain area contains n equal to 8 points, T1 is equal to 5 and T2 is equal to 10, because n is greater than T1 and less than T2, the area warning level is set to medium; when the escort vehicle enters the area, the vehicle position coordinate matches the confirmation area, the camera group basic frame rate is equal to 30 frames per second, the adjustment coefficient corresponding to the warning level is equal to 1, and the calculation of the shooting frame rate adjustment value is equal to 30*1 equal to 30 frames per second; the system automatically maintains the standard frame rate monitoring, balances the resource consumption and safety demand, ensures continuous and stable capture of video in the medium risk area, avoids missing details due to too low frame rate or wasting resources due to too high frame rate, significantly optimizes the system response efficiency and reduces the overall operating cost.

[0041] Optionally, the analyzing the focused video picture and generating a risk identification level comprises: extracting dynamic target behavior features in the focused video picture; matching the dynamic target behavior features with a preset behavior rule library, and generating a behavior abnormality identification according to a matching result; generating a risk identification level according to the abnormality identification and the abnormal event trigger signal.

[0042] Specifically, first, the focused video picture output by the intelligent monitoring module is acquired, motion target detection is performed on continuous video frames, a background difference method is used to extract dynamic target contours, and a motion trajectory feature vector of each target is calculated. The vector includes three components of speed size, direction angle, and vehicle distance. The speed size is calculated by dividing the displacement of the target mass center of adjacent frames by the frame interval time, the direction angle is obtained by the inverse tangent function of the displacement vector, and the vehicle distance is converted into an actual distance by combining the pixel position of the target in the image with the camera calibration parameters. Second, a preset behavior rule library is constructed. The library stores abnormal behavior feature vectors and corresponding risk weights. For example, the rule can be that if the speed size is greater than a preset speed threshold and the vehicle distance is less than a preset safety distance, the behavior is marked as high-speed approaching, and the risk weight is equal to 0.7. Then, the extracted dynamic target behavior features are matched with the behavior rule library piece by piece. For a matching degree There are: , In the formula, is the i-th feature vector component, When the matching degree is less than the preset matching threshold, it is determined that the matching is successful, and if there are multiple matching results, the corresponding behavior abnormality identifier and risk weight are generated according to the dynamic target behavior feature with the minimum matching degree; finally, the final weight value is obtained by multiplying the risk weight by the correction coefficient corresponding to the current abnormal event trigger signal type, the correction coefficient is C1 when the event trigger signal deviates, for example, it can be 1.2, the correction coefficient is C2 when the abnormal motion trigger signal, for example, it can be 1.5, the correction coefficient is C3 when the emergency alarm trigger signal, for example, it can be 2.0, and the risk identification level is divided according to the final weight value, for example, when the final weight value is less than 0.5, it is low risk, 0.5 to 1.0 is medium risk, and greater than 1.0 is high risk.

[0043] Exemplarily, after a certain escort vehicle triggers a route deviation event, the video picture detects a right dynamic target, calculates that the speed size is equal to 8 meters per second, the direction angle is equal to 90 degrees (perpendicular to the approaching vehicle), and the vehicle distance is equal to 15 meters, matches the high-speed approaching behavior rule in the rule library (the speed threshold is set to be equal to 5 meters per second, and the safety distance is equal to 20 meters), because the speed size is greater than the speed threshold and the vehicle distance is less than the safety distance, the matching degree is equal to 0, the behavior abnormality identifier is generated, the risk weight is equal to 0.7, the deviation event correction coefficient C1 is combined, the final weight value is equal to 0.7x1.2 equal to 0.84, and it is determined to be a medium risk level; the system can accurately identify the motion intention of the potential threat target in the video, realize double verification combined with the Beidou abnormal event, timely find the side fast approaching object when the vehicle deviates from the route, significantly improve the prediction ability of the coordinated crime, and provide the scheduling center with a graded risk decision basis.

[0044] Optionally, the comprehensive analysis of the vehicle position coordinates, the focused video picture and the risk identification level includes: extracting a geographic identifier feature from a preset position feature library according to the vehicle position coordinates; extracting an environment feature from the focused video picture; matching the environment feature with the geographic identifier feature, and generating a risk level correction factor according to the matching result; using the risk level correction factor and the risk identification level to calculate a comprehensive risk value.

[0045] Specifically, first, the vehicle position coordinates are obtained from the vehicle state data set, and the pre-set position feature library is queried according to the coordinates. The feature library stores the mapping relationship between geographic coordinates and corresponding geographic identification features, and the geographic identification features include three types of data: terrain type, fixed building distribution (including various fixed objects such as road signs, trees, fire hydrants), and historical accident frequency. Based on the terrain type, the standard road structure reference value of the corresponding terrain type is obtained according to the statistical median of the number of road feature lines of the same terrain type. Based on the fixed building distribution and the corresponding coordinate data, combined with the peak value of people flow and vehicle flow, the average density per unit area is calculated to obtain the standard target density reference value corresponding to the building distribution. According to the environmental lighting parameters (such as the proportion of accidents at night / tunnel / shady rainy days) in the historical accident records, the average value of the illumination in the high-accident period (such as night) is calculated to obtain the standard light reference value corresponding to the historical accident frequency. At the same time, the focused video picture is obtained, and the image segmentation algorithm is used to extract the environmental features, which specifically include the number of road structure lines , moving target density , light intensity value three parameters, wherein obtained by detecting the number of straight lines through Hough transformation, calculated by dividing the moving target detection result by the picture area, converted to brightness value by RGB three-channel mean value; then the environmental feature vector extracted from the video is compared with the geographic identification feature vector in the position feature library, and for the matching result , , wherein is the standard road structure reference value of the corresponding terrain type in the geographic feature library, is the standard target density reference value corresponding to the building distribution, is the standard light reference value corresponding to the historical accident frequency, , , is a pre-set weight coefficient; a risk level correction factor is generated according to the matching result: , wherein the matching result , is a pre-set maximum difference threshold; finally, the quantitative value (numerical representation as low risk 0.3, medium risk 0.7, high risk 1.0) of the risk identification level is combined to calculate the comprehensive risk value : , wherein For the quantitative value of risk identification level, for example, low risk is 0.3, medium risk is 0.7, and high risk is 1.0, so as to complete the comprehensive analysis.

[0046] Exemplarily, an abnormal motion event of a certain bank escort vehicle is triggered in an overpass area, the vehicle position coordinates match the position feature library to show that the terrain type is a multi-layer overpass, Equal to 4 road lines, the building distribution corresponds to the standard target density reference value Equal to 0.02 targets per square meter, the historical accident frequency is high, Equal to high brightness requirement 120 lumens; video picture extraction Equal to 5, Equal to 0.03 targets per square meter, Equal to 100 lumens; set Equal to 0.5, Equal to 0.3, Equal to 0.2, the maximum difference threshold is equal to 1, the calculation matching result is equal to 0.5*0.25+0.3*0.5+0.2*0.17 equal to 0.125+0.15+0.034 equal to 0.309, the risk level correction factor is equal to 1-0.309 equal to 0.691; if the risk R in the behavior analysis output is equal to 0.7, the comprehensive risk value is equal to 0.7*0.691 approximately equal to 0.484; through the cross verification of geographical features and real-time pictures, the difference in the number of road lines in the overpass complex environment is accurately identified as being caused by the real multi-ramp structure rather than human damage, avoiding misjudgment of normal road features as risks, and significantly improving the reliability of risk assessment under complex road conditions, providing accurate basis for dispatching decisions.

[0047] Optionally, the method further comprises: According to the vehicle position coordinates and the basic management area, the alert level is determined; According to the alert level, the risk level correction factor is corrected.

[0048] Specifically, the corrected risk level correction factor is as shown in the comparison chart Figure 6 First, according to the vehicle position coordinates in the vehicle state data set, the basic management area is matched, the vehicle current basic management area number is determined through coordinate mapping; secondly, the preset alert level of the basic management area is inquired, the level is divided into three levels of low, medium and high according to the number of historical abnormal coordinate points; then the original risk level correction factor is obtained; according to the alert level, the correction coefficient is selected, when the alert level is low , it is equal to 1.1, when the alert level is medium , it is equal to 1.0, and when the alert level is high , it is equal to 0.9; finally, the corrected risk level correction factor is calculated : , and input into the comprehensive risk value calculation; wherein the mapping relationship between the alert level and the correction coefficient is that the higher the alert level is, the smaller the value is, reflecting the improvement of the tolerance of the high-risk area to the difference in geographical features; by dynamically adjusting the sensitivity of the environmental feature matching, the system reduces the false alarm rate in the historical high-risk area.

[0049] Illustratively, when the escort vehicle enters a certain industrial district basic management area, the area is set to a high alert level due to the frequent occurrence of historical abnormal events, the original risk level correction factor equals 0.6, the correction coefficient equals 0.9, and the corrected risk level correction factor is calculated; if the output high risk R equals 1.0, the comprehensive risk value 1.0x0.54 equals 0.54, which is 10% lower than the original 0.6. The system can automatically relax the environmental matching standard based on the historical risk data of the area, avoid repeatedly triggering high-risk alarms due to common environmental feature differences in crime-prone areas, significantly reduce the false alarm frequency while maintaining the sensitivity to real threats, improve the accuracy of dispatching decisions and resource utilization efficiency.

[0050] Optionally, the method further comprises: monitoring the strength parameter of the Beidou positioning signal to generate a signal strength index; when the strength signal is lower than a preset strength threshold, obtaining reference coordinate data according to the geographic identification features; correcting the vehicle position coordinates according to the reference coordinate data and the shooting parameters.

[0051] Specifically, first, the signal strength parameter output by the Beidou positioning module is monitored in real time, which is directly provided by the Beidou receiver chip; when the signal strength parameter is lower than the preset strength threshold, the coordinate correction process is triggered; the current vehicle position coordinates are queried from the location feature library to extract the geographic identification features of all fixed reference objects within a radius of 100 meters of the coordinate point, i.e. the distribution and corresponding coordinates of fixed buildings, including traffic signs, building outlines, and lamp post positions; the surrounding camera group is activated through a video retrieval instruction to obtain the current shooting parameters, including the pan-tilt angle parameters and the focal length parameters; the theoretical pixel coordinates of the reference objects in the video image are calculated : , wherein is the camera intrinsic matrix, is the rotation and translation matrix corresponding to the pan-tilt angle parameters, ​The three-dimensional coordinates of the reference object are obtained through the position feature library; the offset vector is calculated by comparing the theoretical pixel coordinates with the pixel coordinates of the reference object detected in the actual video image ; the three reference objects with the smallest size are selected , and the least square method is used to solve the vehicle position correction amount , and finally the corrected coordinates are obtained .

[0052] For example, when a certain escort vehicle enters an underground garage, the signal strength parameter decreases to 20 dBm, which is lower than the threshold value 25 dBm, the system extracts the coordinates of the three nearby fire hydrants as reference objects, the focal length of the camera is 35 mm, and the pan-tilt angle is 120 degrees, the theoretical pixel coordinates of the fire hydrants are calculated as (150, 300), (320, 280), and (200, 400), the actual detection coordinates are (152, 305), (318, 283), and (203, 398), and the vehicle position correction amount is obtained by least square solution, and the accuracy of the corrected coordinates is improved; the beneficial effect of this verification example is that the spatial geometric relationship between the video reference object and the Beidou positioning is dynamically calibrated, effectively overcoming the problem of positioning drift caused by satellite signal shielding, and ensuring the continuous provision of reliable position data in special scenes such as tunnels and underground spaces, and establishing an accurate spatial reference for risk decision-making.

[0053] Based on the same inventive concept, as shown in Figure 7 , the present application also provides an escort management system based on Beidou positioning and video monitoring, which comprises: a Beidou positioning module for real-time acquisition of Beidou dynamic data of an escort vehicle and feature extraction to obtain a vehicle state data set, the vehicle state data set including vehicle position coordinates, vehicle speed change rate, vehicle heading angle, and vehicle stay time; an anomaly detection module for determining the type of abnormal event based on the vehicle state data set and generating a corresponding abnormal event trigger signal; a video guidance module for selecting a preset area of interest according to the type of abnormal event corresponding to the abnormal event trigger signal, generating a video retrieval instruction and a picture focusing instruction; an intelligent monitoring module for collecting real-time video images of the area of interest according to the video retrieval instruction and the picture focusing instruction, and generating focused video images; a behavior analysis module for analyzing the focused video images and generating a risk identification level; a risk assessment module for comprehensive analysis of the vehicle position coordinates, the focused video images, and the risk identification level, and generating a comprehensive risk value; The scheduling module is configured to generate an intelligent scheduling instruction based on the comprehensive risk analysis result and preset escort resource information. The execution module is configured to perform resource scheduling according to the intelligent scheduling instruction.

[0054] It should be noted that the above formula can be translated into a standard value without unit or a parameter with the same dimension that can be superimposed by the principle of dimensional consistency and mathematical standardization means (for example, normalization processing, dimensionless parameter conversion or unit system unification), so as to eliminate the interference of different dimensions on the operation logic, make the formula retain the original data distribution characteristics, and have mathematical operation rationality and objective law adaptability. It is a conventional technical means, and will not be described here. The electrical connection between the above-mentioned units does not necessarily mean direct connection, and indirect connection can also be used as long as the purpose of the application is achieved. The above-mentioned is only an exemplary embodiment of the application, and cannot limit the scope of the application.

[0055] That is, any equivalent changes and modifications made according to the teachings of the present application are still within the scope of the present application. Other embodiments of the present application will be readily apparent to those skilled in the art upon considering the specification and practice of the true principles disclosed herein. The present application is intended to cover any variations, uses or adaptive changes to the present application following the general principles of the present application and including commonly known or conventional technical means in the art not disclosed by the present application.

Claims

1. The escort management method based on Beidou positioning and video monitoring is characterized by: The method comprises: Real-time acquisition of Beidou dynamic data of escort vehicles and feature extraction to obtain a vehicle status data set, which includes vehicle position coordinates, vehicle speed change rate, vehicle heading angle, and vehicle dwell time; Determine the type of abnormal event based on the vehicle status data set and generate a corresponding abnormal event trigger signal; Selecting a preset area of ​​interest according to the abnormal event type corresponding to the abnormal event trigger signal, and generating a video call instruction and a picture focus instruction; According to the video retrieval instruction and the picture focusing instruction, the real-time video picture of the focus area is collected to generate a focused video picture; Analyzing the focused video image to generate a risk identification level; Performing a comprehensive analysis on the vehicle position coordinates, the focused video image, and the risk identification level to generate a comprehensive risk value; Generate intelligent dispatch instructions based on the comprehensive risk analysis results and preset escort resource information; Resource scheduling is performed according to the intelligent scheduling instruction.

2. The escort management method based on Beidou positioning and video monitoring according to claim 1 is characterized in that: Generating an abnormal event trigger signal includes: When the deviation between the vehicle position coordinates and the preset electronic fence route is greater than a preset deviation threshold, it is determined to be a route deviation event and a deviation event trigger signal is generated; When the speed change rate exceeds a preset acceleration threshold, it is determined to be an abnormal motion state event and an abnormal motion trigger signal is generated; When the vehicle heading angle is greater than a preset angle threshold or the vehicle stay time is greater than a preset time threshold, it is determined to be an emergency alarm event and an emergency alarm trigger signal is generated; Among them, the abnormal event trigger signal includes the deviation event trigger signal, the abnormal movement trigger signal and the emergency alarm trigger signal.

3. The escort management method based on Beidou positioning and video monitoring according to claim 2 is characterized in that: The generating of the video call instruction and the screen focus instruction includes: When the abnormal event trigger signal is the deviation event trigger signal, selecting areas on both sides of the vehicle; When the abnormal event trigger signal is the abnormal motion trigger signal, selecting a motion direction sector area; When the abnormal event trigger signal is the emergency alarm trigger signal, selecting a panoramic area around the vehicle; Generate video call instructions and screen focus instructions based on the selected area; The focus areas include areas on both sides of the vehicle, fan-shaped areas in the direction of movement, and panoramic areas around the vehicle.

4. The escort management method based on Beidou positioning and video monitoring according to claim 1 is characterized in that: Generating a focused video picture includes: Selecting a preset camera group corresponding to the area of ​​interest according to the video retrieval instruction and the picture focus instruction; Adjusting the shooting parameters of the camera group according to the vehicle status data set, wherein the shooting parameters include a pan / tilt angle parameter and a focal length parameter; A focused video image is generated according to the shooting parameters.

5. The escort management method based on Beidou positioning and video monitoring according to claim 4 is characterized in that: The method further comprises: Collect the location coordinates of historical abnormal events and obtain abnormal coordinate points; Divide the map into regions according to the preset number to obtain basic management areas; Setting a warning level for the basic management area according to the number of abnormal coordinate points in the basic management area; The shooting frame rate of the camera group is adjusted according to the vehicle position coordinates and the alert level.

6. The escort management method based on Beidou positioning and video monitoring according to claim 5 is characterized in that: Analyzing the focused video image to generate a risk identification level includes: Extracting dynamic target behavior features in the focused video picture; Matching the dynamic target behavior characteristics with a preset behavior rule library, and generating a behavior abnormality identifier based on the matching result; A risk identification level is generated according to the abnormal identification and the abnormal event trigger signal.

7. The escort management method based on Beidou positioning and video monitoring according to claim 6 is characterized in that: The comprehensive analysis of the vehicle position coordinates, the focused video image, and the risk identification level to generate a comprehensive risk value includes: Extracting geographic identification features from a preset location feature library according to the vehicle location coordinates; extracting environmental features from the focused video image; Matching the environmental characteristics with the geographical identification characteristics, and generating a risk level correction factor based on the matching results; The comprehensive risk value is calculated using the risk level correction factor and the risk identification level.

8. The escort management method based on Beidou positioning and video monitoring according to claim 7 is characterized in that: The method further comprises: Determining the alert level according to the vehicle location coordinates and the basic management area; The risk level correction factor is corrected according to the alert level.

9. The escort management method based on Beidou positioning and video monitoring according to claim 7 is characterized in that: The method further comprises: Monitor the strength parameters of Beidou positioning signals and generate signal strength indicators; When the intensity signal is lower than a preset intensity threshold, obtaining reference object coordinate data according to the geographic identification feature; The vehicle position coordinates are corrected according to the reference object coordinate data and the shooting parameters.

10. The escort management system based on Beidou positioning and video monitoring is applied to the escort management method based on Beidou positioning and video monitoring as described in any one of claims 1 to 9, characterized in that: The system comprises: Beidou positioning module, used to obtain Beidou dynamic data of escort vehicles in real time and perform feature extraction to obtain a vehicle status data set, which includes vehicle position coordinates, vehicle speed change rate, vehicle heading angle, and vehicle dwell time; An anomaly detection module, configured to determine the type of an abnormal event based on the vehicle status data set and generate a corresponding abnormal event trigger signal; A video guidance module is used to select a preset area of ​​interest according to the abnormal event type corresponding to the abnormal event trigger signal, and generate a video call instruction and a picture focus instruction; An intelligent monitoring module is used to collect real-time video images of the focus area according to the video retrieval instruction and the picture focusing instruction, and generate a focused video image; A behavior analysis module, configured to analyze the focused video image and generate a risk identification level; a risk assessment module, configured to comprehensively analyze the vehicle position coordinates, the focused video image, and the risk identification level to generate a comprehensive risk value; A scheduling module, configured to generate intelligent scheduling instructions based on the comprehensive risk analysis results and preset escort resource information; An execution module is used to perform resource scheduling according to the intelligent scheduling instruction.

Images