Building extraction using airborne LiDAR data and very high resolution imagery over a complex urban area

Abstract

The occurrence of shadow and diverse building roofs in complex urban areas makes it difficult to extract building automatically using very high resolution ( VHR ) imagery over these areas. In order to solve these two problems, this paper proposed a novel method for building extraction using airborne LiDAR data and VHR imagery. The buildings were initially extracted by thresholding the normalized difference vegetation index ( NDVI) image and LiDAR height data. The initially obtained result was then refined by using NDVI image over shadow areas, image texture and morphological filtering. The proposed method was quantitatively evaluated and compared with existing methods using airborne LiDAR data and QuickBird image of Nanjing City, China. The results indicated that the proposed method effectively reduced the extraction errors caused by shadow and diverse building roof and significantly improved the accuracy of building extraction.