Research On Detection And Identification Algorithms For Bus Passengers In Images

Due to the lack of real-time and accurate passenger traffic information,the current bus system can not judge the congestion of the bus,so that the operators can not make a reasonable scheduling plan to provide passengers with a reasonable choice of the car.Therefore,intelligent bus system is particularly important in the field of transportation,and automatic passenger counting technology as the basis for realizing intelligent public transport has attracted much attention.With the rapid development of digital images,the automatic passenger counting method based on image processing is gradually used to statistics the passenger traffic information,and the key point of this counting method is the detection of moving target and the target recognition.Therefore,this paper focuses on moving object detection and target recognition in video image processing.The main contents are as follows:An improved Canny edge detection algorithm is proposed to detect the edge of the image.The traditional non maximum value suppression step is omitted and the double threshold edge detection is changed to the fixed threshold detection.The classic edge detection algorithm can avoid missing some useful edge information.A background extraction based on weight value is used.The algorithm,by introducing two parameters of background weight and weight change,can judge the current frame image in different conditions to extract the image background,then separate the moving passenger target,and overcome the phenomenon that the traditional background extraction algorithm can not completely solve the "virtual shadow"phenomenon.An improved Hough circle detection algorithm based on gradient is used to realize bus passenger recognition.It overcomes the problem that the traditional Hough detection algorithm has low accuracy,too many circles and low program efficiency.The algorithm is mainly improved from three aspects:sliding window,radius selection and center of the circle,in order to reduce the number of candidate circles,so as to reduce the running time of the program and improve the efficiency of operation.Finally,the generated candidate circle is selected by gray filter and circular feature filter,and the passenger target is obtained.OpenCV is used to process the collected bus video images.The experimental results show that the improved Canny edge detection algorithm combined with the background extraction algorithm based on weight value can accurately separate the background image and the moving target,and the computational efficiency is high and the anti-interference ability is strong.The running time is less and the operation efficiency is high.