Edge detection in range images with multiple window operators

A hybrid approach toward the detection of edges in range images is presented. It combines region growing and one-dimensional orthogonal polynomial fitting techniques for the detection of edges in range images. Edges detected by this method possess good localization properties. The incorporation of the region growing process eliminates internal microedges and produces a segmentation map. The edge and segmentation maps, when conveyed to the higher-level recognition process, may prove valuable for object identification purposes.<>