The work mainly uses the related technology of video image processing, obtains the relevant parameters of calculating the traffic light time switching through the processing of the video collected at the intersection, and calculates the traffic time of the traffic light through the statistical traffic parameters to realize the intelligent control of the traffic light. It mainly includes: system initialization (including lane detection and traffic sign recognition), traffic parameter statistics (traffic flow statistics, vehicle speed statistics and vehicle occupancy ratio calculation), and finally realize the intelligent traffic signal control system. The key technologies include: (1) lane line detection: according to the characteristics of the road lane line, the Hough transform is used to detect the lane line in the area of interest on the road; (2) the extraction and recognition of traffic signs (turning and straight) on the ground of the intersection: the ground signs are corrected by the inverse perspective transformation of the image, and the traffic signs are extracted and recognized by combining the detection results of the lane line; (3) Statistics of traffic flow and vehicle speed: establish a background model for the collected video, extract moving vehicle targets by background difference method during the green light period, and calculate the traffic flow and vehicle speed during the green light period; (4) Vehicle occupancy ratio calculation: use projection method and vehicle contour area to calculate the vehicle occupancy ratio during the green light period and the red light period; (5) Calculate the green light passage time through statistical traffic parameters, realization of intelligent traffic signal lamp control system;