The Research Of Automatic Detection Technology For Crowding In Bus

In this study,the video files collected by the camera in both forepart and backpart of the bus are taken as the subjects to study the passenger crowding condition through technique of video image processing.The passenger crowding condition detecting system is designed to collect the passenger flow information in the bus efficiently and improve the conventional bus service scientifically.Firstly the current situation of the ITS is systematically investigated to study the structure of the ITS.Based on the explicit analyses of ITS's common function,the mainstream ITS programs are taken as the references to study the development of ITS.The conventional bus in Xi'an is taken as the example to automatically detect the passenger crowding condition in the bus by extending the use of the current monitoring system.Then,the passenger crowding criteria in the bus is set based on the clear definition of bus passenger crowding.The video images collected by the monitoring camera in the bus are used to design an efficient and accurate algorithm,automatically detecting the passenger crowding condition in the bus.Two types of detecting algorithm(Video Perceptual Hashing and Edge Difference)are combined to improve accuracy and efficiency of the improved hybrid algorithm.Finally,the hybrid algorithm is used to analyze the image data from the camera in both forepart and backpart of the bus.The result integrates both two types of data to improve the accuracy of the algorithm.Through the calculating experiments of 12 video files(including 8664 video images),the results show that the average accuracy is up to 92.3% and the detecting speed is about 0.3/frame,thus proving accuracy and efficiency of the improved hybrid algorithm and the system designed in this study.MATLAB GUI is designed to visualize the interface to facilitate the users,providing the users some useful functions,such as window display,file input and results output.