Abstract LiDAR (light detection and ranging) is one of the most important techniques used in remote sensing and photogrammetry to extract high quality features, objects, and models. In this paper, we present an innovative and simple method for ground point filtering from LiDAR point clouds. This method is inspired by nature, where tornadoes destroy buildings and other manmade objects and uproot trees. This simulates the removing of points of building, trees and other objects from LiDAR point clouds. We have imposed that the tornado is a vertical cone with a down vertex and up base. The vertical cone moves on the ground surface so that a network of points is selected as vertex points for the vertical cone. Points that are located inside the cone are removed as non ground points. In other words, the distances between the points and the cone axis are measured, then these distances are compared with calculated distances relative to the cone equation, finally, the points within the cone are removed. To evaluate the proposed method, the datasets provided by the International Society for Photogrammetry and Remote Sensing (ISPRS) are used. The results show that this method can adapt to complex environments and achieve high filtering accuracy. Furthermore, by comparing our method with other methods tested by the ISPRS, the proposed method achieves an average total error of 6.63%, which is smaller than that of the most methods.

中文翻译：

---

激光雷达点云地面点滤波的 Tornado 方法

摘要 LiDAR（光检测和测距）是遥感和摄影测量中用于提取高质量特征、对象和模型的最重要技术之一。在本文中，我们提出了一种从 LiDAR 点云中过滤地面点的创新且简单的方法。这种方法的灵感来自大自然，龙卷风会摧毁建筑物和其他人造物体并连根拔起树木。这模拟了从 LiDAR 点云中移除建筑物、树木和其他物体的点。我们已经强加了龙卷风是一个垂直的圆锥体，具有向下的顶点和向上的底部。垂直锥体在地面上移动，从而选择一个点网络作为垂直锥体的顶点。位于锥体内部的点将作为非地面点被移除。换句话说，测量点与锥轴之间的距离，然后将这些距离与相对于锥方程的计算距离进行比较，最后去除锥内的点。为了评估所提出的方法，使用了国际摄影测量和遥感学会 (ISPRS) 提供的数据集。结果表明，该方法能够适应复杂环境，实现较高的滤波精度。此外，通过将我们的方法与 ISPRS 测试的其他方法进行比较，所提出的方法的平均总误差为 6.63%，小于大多数方法。使用国际摄影测量和遥感学会 (ISPRS) 提供的数据集。结果表明，该方法能够适应复杂环境，实现较高的滤波精度。此外，通过将我们的方法与 ISPRS 测试的其他方法进行比较，所提出的方法的平均总误差为 6.63%，小于大多数方法。使用国际摄影测量和遥感学会 (ISPRS) 提供的数据集。结果表明，该方法能够适应复杂环境，实现较高的滤波精度。此外，通过将我们的方法与 ISPRS 测试的其他方法进行比较，所提出的方法的平均总误差为 6.63%，小于大多数方法。