Accurate extraction of building roofs from airborne light detection and ranging point clouds using a coarse-to-fine approach

摘要

The accurate extraction of building roofs from airborne light detection and ranging (LiDAR) point clouds plays an important role in many applications, such as digital building modeling and disaster assessment. However, this remains a challenging task because of the diversity of building roof structures, irregular distributions of LiDAR points, and mutual disturbances of neighboring points. Most of the existing methods show little capability to detect inconspicuous roofs, i.e., roofs with small sizes or fuzzy boundaries. We present a coarse-to-fine method to accurately extract roofs from airborne LiDAR point clouds. This method first iteratively extracts large roofs by three successive steps with dynamically adjusted parameters during its "coarse" stage, and then extracts small roofs from the remained points using an improved random sample consensus method during the "fine" stage. Experimental results show that the method can significantly improve the accuracy of roof extraction by robustly identifying most of the inconspicuous roofs in LiDAR point clouds. (C) 2018 Society of Photo-Optical Instrumentation Engineers (SPIE)