Research on the Optimization Algorithm of Urban Road Recognition System based on DM6437

In order to improve the robustness and realtime performance of vehicle safety assistant driving system for identifying the front driving environment, an optimized algorithm for road image processing and lane detection is proposed. Aiming at the characteristic structure of road image, we use static partition and image processing in the region of interest. After the grayscale processing, the optimized median filtering algorithm is used to remove the noise. In the process of lane detection and fitting, the Hough algorithm is improved based on the characteristics of lane lines. When Kalman filtering is used to track lane lines, Hough transform and least square method are used to identify lane lines. Finally, the algorithm is transplanted to the DM6437 hardware platform, and the software code is optimized by combining the architecture features of the embedded system. The actual experimental results verify the real-time performance and robustness of the optimized algorithm.