3-D reconstruction of man-made objects based on watershed segmentation and 3-D grouping

This paper presents a method for automatic 3-D reconstruction of man-made objects using high resolution aerial imagery. The proposed method consists of three parts. First, by using mean curvature diffusion, noise is eliminated and edges are preserved. Next, images are segmented by a watershed algorithm. Finally, man-made objects are modeled as a set of planar polygonal patches. A roof of man-made object is described as a set of planar polygonal patches, each of which is extracted by watershed-based image segmentation, line segment matching and coplanar grouping. The coplanar grouping process can be restricted to the 3-D line segments from the region boundary. Coplanar grouping and polygonal patch formation is performed per region by selecting 3-D line segments that are matched using epipolar geometry and flight information. The algorithm has been applied to high resolution aerial images and the results show accurate reconstruction of man-made objects.