ScholarWorks@성균관대학교

초록

Recently, as autonomous driving technology develops, the importance of real-time image processing is also being noted. Accordingly, high power consumption of high-performance image processing systems is an issue that needs to be resolved. In this paper, we propose an FPGA-based high-speed image processing autonomous driving system architecture to solve this problem. The proposed system maximizes the lane detection speed through preprocessing such as region of interest (ROI), grayscale transformation, Bird Eye View(BEV), and Sobel filter, and the core parallel Hough Transform(HT) structure. To verify the effectiveness of the system, experiments were conducted using a kid's electric vehicle, an Ultra96V2 board, and an autonomous driving track. The entire system achieved 65 fps and successfully drove two laps of a 12m×16m track in 1 minute and 27 seconds.

키워드