A TIN-based classification approach for buildings or vegetation extraction

摘要

In this paper, we propose a new TIN-based classification approach for feature extraction from Light Detection and Ranging (LiDAR) data point clouds. The method builds TIN first and then computes variance (standard deviation) of normal vectors at each node to classify objects from raw LiDAR data point clouds. Experimental results indicate that our method can effectively classify buildings or vegetation via choosing different threshold values of variance (or standard deviation) of normal vectors.