3D LiDAR is widely used in unmanned ground vehicles to perceive the surrounding environment. Ground segmentation is one of the essential tasks for unmanned ground vehicles because of its various applicability. For example, ground segmentation can be used as a preprocessing step for dynamic object removal, traversable region estimation, and point cloud registration. Therefore, fast and
robust performance of the ground segmentation method should be guaranteed for further steps. In this study, we propose a fast ground segmentation method for 3D point clouds of various distributions. The proposed method down-samples a 3D point cloud through a voxelization to generate a uniform point distribution regardless of the distribution of the input point cloud. Accordingly, it increases the speed of the algorithm and exhibits similar performance compared to the original algorithm.