As the boost of modern society, research on the extraction of road traffic markings has become increasingly popular. To improve the regional convolutional neural network, improve the road surface cloud segmentation algorithm based on the radius filtering algorithm and area division method, and combine the two algorithms to improve the sparse point cloud road traffic marking extraction algorithm. Finally, the study will integrate the single frame image and road surface cloud data frame by frame, apply the improved road traffic marking extraction algorithm of sparse point cloud to the road surface cloud with single frame image, and construct the road traffic marking extraction algorithm integrating single frame image and sparse point cloud. The effectiveness of the improved regional convolutional neural network algorithm proposed in the study was verified, and it was found that the average recall rate of the algorithm was 0.841, the average accuracy was 85.4%, and the operation speed was 125.6 seconds. Its performance was superior to other algorithms compared. In addition, the study also compared and analyzed the performance of the fusion road traffic marking extraction algorithm, and found that the average extraction edge length difference of the algorithm's road marking extraction was 0.0315m, and the average relative error between the algorithm and the internal verification points was 0.0493, which is better than the comparison algorithm. Based on the comprehensive experimental results, it was found that the performance of the proposed improved regional convolutional neural network algorithm and the traffic marking extraction algorithm that integrates single frame images and sparse point clouds is superior to the comparison algorithm. Meanwhile, the proposed fusion lane marker extraction algorithm has significantly improved the accuracy and precision compared to traditional lane marker extraction algorithms, and has enormous application potential in the field of road traffic.
