AN ADAPTIVE FILTERING ALGORITHM FOR BUILDING DETECTION FROM LIDAR DATA

Research on lidar data filtering for building detection has continued to flourish in recent years due to the increasing need for 3-dimensional data in urban development and planning. Over the last decade, many filtering algorithms have been developed to classify lidar point clouds. As a result, interpolation-based filters, slope-based filters and morphological filters were widely accepted. Most of the filtering algorithms require 'raw' lidar data to be rasterised i.e. interpolated into grid images. However, rasterisation often causes a significant loss of information after data processing. To overcome the information loss, we developed an adaptive filtering algorithm that classifies lidar data effectively and efficiently into ground and non-ground for the building detection purposes. The digital elevation model generated from the filtered ground was used to detect man-made objects e.g. buildings. The results show that, by using an adaptive window size indicator, the proposed algorithm can effectively classify lidar data at a high accuracy.