Video-based Method for Bus Passenger Counting

In order to overcome the limitations of real-time processing,large amount of data and anti-jamming for the traditional optical-flow or frame-difference methods to extract moving objects from image sequences.The paper proposes a passenger counting system which adopts a row space-time image method.The system includes the generation of row space-time images,moving objects extraction,recognition,matching and counting.The experiment results show that the system has high recognition accuracy,and meets the demand of real-time processing.The new characteristic is robust recognition and low computational complexity.The hardware capability required is not so high that the costs of hardware and software are easy to control.