Kai Hugo Hustoft Endresen

Tracking Objects in 3D Using Stereo Vision: a Real-time Approach Based on Color Segmentation for Use on a Mobile Robot Kai Hugo Hustoft Endresen

Previous work in the area of stereo vision has been mostly restricted to either stationary use, or very slow movement. With the introduction of increasingly faster computing devices, more and more of the things that required considerable processing time before, can now be done in real-time. This book describes a stereo vision system to be used on a mobile robot, originally submitted as a master thesis in computer science. The system is able to determine the positions of cylindrical and spherical objects in a 3D environment in real-time by using images from two cameras. Corresponding regions in the two images are found, which then enables the system to use disparity and triangulation to calculate the relative positions of the objects.