Morphology-based building detection from airborne lidar data

The advent of Light Detection and Ranging (lidar) technique provides a promising resource for three-dimensional building detection. Due to the difficulty of removing vegetation, most building detection methods fuse lidar data with multispectral images for vegetation indices and relatively few approaches use only lidar data. However, the fusing process may cause errors introduced by resolution and time difference, shadow and high-rise building displacement problems, and the geo-referencing process. This research presents a morphological building detecting method to identify buildings by gradually removing non-building pixels. First, a ground-filtering algorithm separates ground pixels with buildings, trees, and other objects. Then, an analytical approach removes the remaining non-building pixels using size, shape, height, building element structure, and the height difference between the first and last returns. The experimental results show that this method provides a comparative performance with an overall accuracy of 95.46 percent as in a study site in Austin urban area.