The Pattern Recognition Design Of Bus Passenger Count System

Automatic passenger counting technology is a crucial one in intelligent public traffic system, which can be used for achieving real-time bus scheduling, optimization of bus lines and traffic forecasts. Heavy traffic density and the crowded conditions when getting on and off character the common points of public transport in our country. A variety of counting methods applied to China's statistical system, which have been used at home and abroad for passenger counting, are not precision and couldn't give out the result with accuracy. So the pattern recognition design of bus passenger counting system has important practical significance.Based on the analysis of automatic passenger counting technology at present situation, the infrared sensor groups, which can be installed on the first step of front and back doors in practice, are used in opposite radio way. They will collect the step information without affecting the normal getting on or off. Due to powerful capability, C8051F040 single chip could produce pulses of infrared light emitting diode driver, control the infrared light-emitting diodes and infrared receiver diode scanning cycle, and trains the whole data to the host computer for analysis accurately by serial port, which laid a solid foundation for RBF neural network-based footprints pattern recognition.PC for the collected data in the database, image segmentation and build on information database and the data of the characteristic parameters. The characteristic parameters to data are extracted as the RBF neural network input. Then, using of data-processing ability of Matlab as a major development tool, we can prove that RBF neural network-based pattern recognition is to accurately determine the footprints by training and generalization. Through crowded and sparse passenger bus simulation, experimental results show that the recognition rate of footprints is 94%. The concrete relation between the number of footprints and passengers has yet to be verified in practice.The design in this paper, which is innovative, safe and non-contact, can be widely used in these public places with heavy flow density, traffic congestion and difficult to accurately count, such as shopping malls, stadiums, railway stations, and so on.