Abstract: As a new remote sensing technology, LiDAR (Light Detection and Ranging) is getting wide attentions from users and researchers for its outstanding features. LiDAR is less affected by weather and shadow and has a certain degree of penetration to the gap between surface features. It gets high-precision 3D point cloud data of the land surface and provides a new method for contour extraction. Since contours can’t pass through the buildings and the water areas, the traditional methods need a lot of manual works to edit the contours in these areas, which costs much time. Therefore a method which could generate contours automatically based on the feature constraints from LiDAR point clouds is proposed in this paper. This method mainly includes the following aspects (1) DEM generation which provides basic data for the generation of contours; (2) Feature extraction based on LiDAR point data, mainly including buildings and waters, etc. (3) Automatic generation of TIN based on feature extraction, and on this basis, extraction of fine contours by tension spline function. Compared with the traditional technology of photogrammetry and stereopair, the method in this paper is more reliable, efficient and precise.

Key words: LiDAR, filter, feature constraint, TIN, contour