Complex Lane Line Detection Under Autonomous Driving

In this paper, we propose an improved lane determination algorithm to increase the accuracy and stability of the driving assistance system in an automatic driving system when determining the position of a lane in a complex driving environment. Based on the proposed single-scale Retinex enhancement and Gamma correction fusion image enhancement algorithm, we introduce a Hidden Markov Model improvement road segmentation algorithm to extract lane feature points by separating the region of interest and at the same time, we design a vanishing line detection technology based on improved V-disparity map. Experimental data show that the proposed algorithm has excellent detection capabilities and can detect main lanes in complex road environments such as shadows, traffic congestion, and uneven lighting.