CN-116030381-B - Method for recognizing parking standard of apron vehicle based on video analysis

CN116030381BCN 116030381 BCN116030381 BCN 116030381BCN-116030381-B

Abstract

The invention relates to a video analysis-based method for identifying the reversing standard of a vehicle on an apron, which is used for identifying the reversing standard of the vehicle on the apron through the actions of an edge algorithm module and a comprehensive processing module, wherein the edge algorithm module is used for acquiring real-time video stream data of a camera in a flight area, identifying target information of vehicle personnel in the video stream data, carrying out structural processing on the target information and pushing the target information to the comprehensive processing module, and the comprehensive processing module is used for organizing the target structural data, carrying out comprehensive judgment, generating a violation event and sending the violation event to a user side for the user to check. The airport safety supervision system has the advantages that after the airport field real data training is carried out, the accuracy of identifying the airport vehicles and personnel is high, and by means of judgment, the reversing behavior of the vehicles and whether people command the vehicles can be identified, and by means of the method, the airport safety supervision depth and breadth are remarkably improved, the working efficiency of supervision personnel is remarkably improved, and the integral safety management level of an airport is improved.

Inventors

LIU QING
CHEN HAN
SONG YIBO
MENG XIANGLIANG
QU SHIQIANG
SUN RUICHENG
SUN JIE
LIU XIAOJIANG

Assignees

青岛民航凯亚系统集成有限公司

Dates

Publication Date: 20260505
Application Date: 20221221

Claims (3)

1. The method is characterized in that the method is used for identifying the reverse standard of the vehicle of the apron through the actions of an edge algorithm module and a comprehensive processing module, wherein the edge algorithm module is used for acquiring real-time camera video stream data of a flight area, identifying target information of vehicle personnel, carrying out structural processing on the target information, and pushing the target information to the comprehensive processing module; The comprehensive processing module is used for organizing the target structured data, comprehensively judging whether the vehicle is backing, commanding whether personnel exist or not, and if the illegal backing behavior is identified, generating an illegal event, and sending the illegal event to a user side for the user to check; Firstly traversing the vehicle dynamic list to obtain vehicle point set information, and dividing a polygon formed by a hypothesis point set into a plurality of finite triangles T1, T2, T3 and T4 according to randomly selected points P (px, py), wherein the geometric centers of the triangles are respectively marked as TC1, TC2, TC3 and TC4, the geometric centers of the triangles are respectively marked as TA1, TA2, TA3 and TA4, and the areas of the triangle are respectively marked as TA1, TA2, TA3 and TA 4; Analyzing the current video frame to obtain geometric center coordinates (cx 1, cy 1) of the vehicle point sets, then, when analyzing the next video frame, recalculating the geometric centers of the vehicle point sets, marking the geometric centers as (cx 2, cy 2), and solving an angle trigonometric function sin theta of an included angle between the connecting line of the geometric center points of the two vehicle point sets and the north direction as a first reference value; Taking all points in the current vehicle point set and geometric centers (cx 2, cy 2) to sequentially perform a second sin theta calculation, wherein the geometric center points (cx 2, cy 2) are used as first calculation points, and comparing all obtained values with a first reference value, then considering that the point closest to the reference value is close to intersect with the geometric center line of the vehicle point set and is marked as (Ix 1, iy 1); When the sin theta of the point in the current vehicle point set is calculated to be closest to two geometric centers, the point is considered to be another intersection point, and is marked as (Ix 2 and Iy 2), and the distances L1 and L2 between the two intersection points and the geometric centers are solved: , 。
2. The method for identifying the reverse specification of the apron vehicle based on video analysis according to claim 1, wherein when calculating the triangle coordinates: setting the triangle coordinates as (x 1, y 1), (x 2, y 2), (x 3, y 3), and calculating the TCn coordinates (TCx, TCy) by the following steps: ; The TAn area calculation method is as follows: ; Solving the geometric center coordinates (cx 1, cy 1) of the vehicle point set: ; storing the position of the solving geometric center into each piece of vehicle information, re-calculating the geometric center of the vehicle when the next video frame analysis is carried out, recording as (cx 2, cy 2), solving the angle trigonometric function sin theta, 。
3. The method for identifying the standard reversing of the apron vehicle based on video analysis according to claim 2 is characterized in that the direction of the head and the tail of the vehicle is judged by judging the sizes of L1 and L2 according to the configuration of the vehicle and the internal configuration, if the vehicle is in a reversing state at present according to the fact that the internal configuration L1 is large as the tail of the vehicle, the vehicle is recorded in a reversing behavior analysis list, personnel are judged next, personnel geometric center points (pcx 1 and pcy 1) are solved, sin theta values of the geometric center points of the personnel and the geometric center points of the vehicle are calculated and are used for judging whether the personnel conduct command in front of the vehicle, when the sin theta difference value between the sin theta value and the geometric center points of the vehicle is smaller than a set constant Y, the personnel are considered to be in front of the vehicle, the vehicle is considered to be in reversing, and when no personnel are judged, the vehicle is considered to be in illegal reversing, a reversing illegal reversing event of the vehicle is generated and is sent to a user side.

Description

Method for recognizing parking standard of apron vehicle based on video analysis Technical Field The invention relates to a video analysis-based standard recognition method for reverse driving of a vehicle on an apron, belonging to the field of airport safety precautions. Background The number of the apron working vehicles is large, the service association is complex, and the requirements on safety are very high because of the related guarantee of flights, and in airport service standards, the vehicles are required to be commanded by special persons in the reversing process, so that the reversing behavior is ensured to be safe and reliable. In traditional supervision, manual patrol is carried out by relying on-site patrol supervision personnel, and a vehicle reversing violation event is found, so that the mode is low in efficiency, meanwhile, the number and the range of the violation phenomena which can be found are too small, a large safety management drain exists, and safety problems are easy to generate. According to the method, a plurality of cameras are arranged on the apron, real-time video popularity analysis is performed, whether the reversing of the vehicle is standard or not is analyzed, the occurrence probability of related illegal behaviors is remarkably improved, and the airport safety management level is effectively improved. Disclosure of Invention In order to overcome the defects of the prior art, the invention provides a method for identifying the reversing standard of a vehicle on an apron based on video analysis, which comprises the following steps: the method is characterized in that the method is used for identifying the parking standard of the vehicle on the apron through the actions of an edge algorithm module and a comprehensive processing module, wherein the edge algorithm module is used for acquiring real-time camera video stream data of a flight area, identifying target information of vehicle personnel in the video stream data, carrying out structural processing on the related information, and pushing the information to the comprehensive processing module; The comprehensive processing module is used for organizing the target structured data, comprehensively judging whether the vehicle is backing, commanding whether personnel exist or not, and if the illegal backing behavior is identified, generating an illegal event, and sending the illegal event to the user side for the user to check. Firstly traversing the vehicle dynamic list to obtain vehicle point set information, and dividing a polygon formed by a hypothesis point set into a plurality of finite triangles T1, T2, T3 and T4 according to randomly selected points P (px, py), wherein the geometric centers of the triangles are respectively marked as TC1, TC2, TC3 and TC4, the geometric centers of the triangles are respectively marked as TA1, TA2, TA3 and TA4, and the areas of the triangle are respectively marked as TA1, TA2, TA3 and TA 4; Analyzing the current video frame to obtain geometric center coordinates (cx 1, cy 1) of the vehicle point sets, then, when analyzing the next video frame, recalculating the geometric centers of the vehicle point sets, marking the geometric centers as (cx 2, cy 2), and solving an angle trigonometric function sin theta of an included angle between the connecting line of the geometric center points of the two vehicle point sets and the north direction as a first reference value; Taking all points in the current vehicle point set and geometric centers (cx 2, cy 2) to sequentially perform a second sin theta calculation, wherein the geometric center points (cx 2, cy 2) are used as first calculation points, and comparing all obtained values with a first reference value, then considering that the point closest to the reference value is close to intersect with the geometric center line of the vehicle point set and is marked as (Ix 1, iy 1); When the sin theta of the point in the current vehicle point set is calculated to be closest to two geometric centers, the point is considered to be another intersection point, and is marked as (Ix 2 and Iy 2), and the distances L1 and L2 between the two intersection points and the geometric centers are solved: When calculating triangle coordinates: setting the triangle coordinates as (x 1, y 1), (x 2, y 2), (x 3, y 3), and calculating the TCn coordinates (TCx, TCy) by the following steps: the TAn area calculation method is as follows: TAn = x1y2-x1y3+ x2y3-x2y1+ x3y1-x2y2; Solving the geometric center coordinates (cx 1, cy 1) of the vehicle point set: storing the position of the solving geometric center into each piece of vehicle information, re-calculating the geometric center of the vehicle when the next video frame analysis is carried out, recording as (cx 2, cy 2), solving the angle trigonometric function sin theta, According to the vehicle configuration and internal configuration, the head and tail directions of the vehicle are judged by judging the sizes of L1