Development Of Intelligent Rail Transit Passenger Flow Detection Equipment

With the development of computer vision technology,we use intelligent video analysis method for passenger flow detection,which has become a smart traffic and computer vision research hotspot.Intelligent passenger flow detection equipment can provide data support for passenger forecasting and early warning;our equipment helps the rail transit planner to design a reasonable traffic route to improve the efficiency of rail transportation.At present,in the pedestrian detection process,passenger flow detection equipment is vulnerable to scenes,light and other environmental factors.Especially in the high population density,serious pedestrian block,complex and changing background environment of the subway scene,the design of efficient passenger flow detection system is particularly difficult.At the same time,the detection efficiency of existing equipment is not high,they cannot achieve real-time detection,and it is difficult to meet the practical application requirements.In order to solve the above problems,this paper designs and develops intelligent rail transit passenger flow detection equipment which can adapt to the subway scene with high accuracy and efficiency.Based on the classical pedestrian detecting and tracking algorithm,this paper adds depth information.We use equipment to capture both the color image and the depth image in the scene.The method of background subtraction and setting height threshold is used to extract pedestrian head and shoulder position,which can effectively reduce the amount of calculation of pedestrian and improve the efficiency of calculation.We extract the HOG of pedestrians' heads and shoulders,and then we use SVM to identify pedestrians.Finally,we use target tracking algorithm which based on Kalman and CamShift to track pedestrians.According to the pedestrian trajectory,the bidirectional passenger flow is calculated by drawing the counting line in the scene.Equipment is continued to enrich system functions.In the embedded front-end equipment,we add the data transmission,video transmission and other functions.We add device management,data receiving,data retention and other functions on the server.After a long time test of many subway stations,the average precision is 96.57%and the running time is 21fps.It is fully proved that the passenger flow detection equipment is able to meet the practical application in accuracy,real-time,stability requirements.