Abstract

LiDAR sensors measure the distance between the sensor and sensed points based on laser pulses. The point cloud gathered by airborne LiDAR sensors includes all natural and manmade features on the earth’s surface. These points are basis for Digital Surface Model (DSM) generation of the scene. In most applications such as designing pipeline routes and geometric plan of the roads digital terrain models (DTM) are needed instead. DTM is created by removing the points that belong to manmade features from DSM. Methods used to create a DTM from LiDAR DSM are called data filtering approaches. In this paper after reviewing existing TIN-based filtering approaches a new method for filtering LiDAR data proposed. Our method uses the slope of triangles in Delaunay triangulation to recognize and modify the height of points belong to manmade features. Our proposed method can filter 75% of non-terrain points and the quality of created DTM is fine.