“Significant breakthroughs have been made in the field of lane detection. In response to the problem of existing algorithms struggling to balance detection accuracy and speed, researchers propose a new lane detection algorithm based on ARM embedded platform. This algorithm achieves improved detection speed while ensuring detection accuracy by designing a lightweight semantic segmentation network, combining it with a Kalman filter lane tracking model, reconstructing the encoder, and utilizing TensorRT to accelerate inference. The experimental results show that the algorithm can cope with various complex traffic scenarios, with a detection accuracy of up to 98.03% and a detection speed of 50 FSP, meeting the requirements of real-time detection. This research achievement provides a new solution for the field of lane detection, which has important theoretical and practical value.”