Building change detection based on object extraction in dense urban areas

Abstract

This study presents a novel approach for building change detection from digital surface models (DSMs), which are generated from the images acquired by a multi-line digital airborne sensor ADS40. Our approach is based on building extraction, which is one of the most challenging research fields. A scheme is proposed that allows efficient integration of a local surface normal angle transform (LSNAT) method and a marker controlled watershed segmentation (MCWS) method for building extraction in dense urban areas mainly from DSMs, and subsequently, performs change detection based on the results of building extraction and the height difference of DSMs. The merits are that really changed buildings are detected, and false-detection can be decreased considerably compared to some other change detection methods. The proposed approach presents wonderful results for building extraction and acceptable results for change detection.