Method of Bicycle Trajectory Extraction with Camera Calibration Technique

Getting the location information of bicycle riders accurately is the basis for studying the characteristics of bicycle traffic flow. This paper introduces an approach that can extract the bicycle trajectory based on the camera calibration technique. The relationship between the world coordinate system and pixel coordinate systems is established, while according to the extracted pixel data of the cyclist, the time-space motion data of the rider can be calculated by Matlab. Compared with the measured data, the computational locations show that the average relative error of trajectory data is 2.67% and the maximum relative error is less than 6%. The results show clearly that the provided approach of bicycle trajectory extraction is correct and effective, which can be used in traffic flow study.