Research on the algorithm of the front vehicle speed based on vehicle video

Vehicle speeding is a lead to one of the major causes of highway traffic accidents, the range of traditional fixed camera video velocity measurement system is very limited. In this paper, the front vehicle's driving state is obtained by using the vehicle mounted binocular CCD camera, and a method for measuring the speed of the front vehicle is proposed. The proposed method using the improved Hough transform to identify the lane lines and extract the region of interest. Complete the vehicle detection based the characteristic of vehicle shadow, and the distance between the speed measuring vehicle and the front vehicle is obtained by using the three-dimensional reconstruction distance measurement algorithm. The relative displacement of the vehicle and the corresponding time interval are obtained by selecting the appropriate image frame in the vehicle video image, and the driving speed of the front vehicle can be calculated. The above algorithm is tested, and it can get the high accuracy of the speed measurement results.