Integration of YOLOP and Kalman Filter on ROS2 platform for Enhanced Lane Detection of AI Formula Robot

This study examines the application of Kalman filter for white line detection in the AI Formula robot. By refining YOLOP-generated path points, the filter effectively reduces the influence of noise, enhancing detection accuracy. In this study, we proposed a lane detection system which integrates YOLOP with Kalman filter in the ROS2 platform. We optimized the program code to make the proposed system adaptable for practical requirements. Experimental results demonstrate Kalman filter’s ability to improve YOLOP’s lane detection, offering a reliable solution for AI Formula pathfinding in real-world scenarios.