 The paper presents a new, easy-to-use filtering method for separating ground and non-ground measurements from airborne LiDAR data to generate digital terrain models. The method uses an inverted LiDAR point cloud and a cloth covering to determine the locations of ground surface nodes, resulting in an average total error of 4.58% on benchmark datasets. This article was authored by Wu Mingzhang, Jian Baqi, Peng Wan, and others.