Detection of Cylindrical and Spherical Surfaces and Determining Their Parameters in Segmentation of Range Images

Abstract

An analytical solution of the problem of estimating the coordinates of the axis of a cylindrical surface and its radius by processing range images is obtained. The proposed methods and algorithms can detect fragments of such surfaces in range images with the prescribed confidence probability. The method is also generalized for finding spherical surfaces in range images. The efficiency of the proposed methods and algorithms is verified by simulation using real-life range images of urban industrial scenes.