Research on Point Cloud Data Preprocessing for Unmanned LiDAR

With the continuous progress of random social technology and economy, autonomous driving technology has become a key direction for the development of modern automotive industry. The environmental perception system, as an important component of autonomous driving technology, directly determines the safety, stability, and comfort of autonomous vehicles. As an active imaging system, LiDAR has the advantages of high accuracy in obtaining clustering information, being less susceptible to external environmental interference, and not limited by light intensity compared to passive imaging systems. In order to enable the vehicle mounted LiDAR to better achieve environmental perception, it is necessary to preprocess point cloud data. This article mainly studies three key steps in the preprocessing process, including point cloud filtering denoising, downsampling, and ground point cloud segmentation, and verifies the algorithm for the above preprocessing methods. The research in this paper provides technical support for autonomous driving to achieve environmental awareness.