Algorithm of speed estimation based on Radon transform from video surveillance

A novel approach for estimating traffic speed by using image process based on Radon transform is proposed.According to the time-space characteristics of vehicle moving tracks,a mapping relation between image and road by using road traffic marking is established,which requires no more image rectification and calibration and simplifies the in situ calibrating procedure.A new image is achieved by using the time stack method from video surveillance sequence and linear feature of the tracking of vehicles is given.The vehicle motion parameters are estimated with line extraction based on Radon transform.It can significantly reduce the computation cost compared with traditional techniques.This approach is robust and practical in the vehicle speed measurement.