Multiple Plane Detection Method from Range Data of Digital Imaging System for Moving Robot Applications

Abstract

Imaging system using CCD sensors for automatic navigation of intelligent robot is a central element to recognize and interact with the surrounding environment. Specifically, finding a planar surface on 3D space is very important for efficient and safe operation of a mobile robot. In this chapter, a noise rejection filter is introduced for defining planar surfaces to reduce the voting of noisy data. We test the normal directions of two arbitrary planes in a small region, which are determined by three vertexes of a triangle and its rotation. If the angle of two normal directions is lower than a given threshold, it is voted into the Hough parameter space. This method is similar to a noise rejection filter to verify the planarity of local planes. We can get accurate parameters of the plane in RHT because most noises and nonplanar data cannot vote into the Hough parameter space. We use a scan window to vote locally. The scan window explores all regions by changing the window size. The window operation improves the accuracy of plane detection because the plane is locally consistent and increases the search speed for finding planes. Finally, the performance of the algorithm for real range data obtained from a stereo imaging system has been verified.