Generation of orthorectified range images for robots using monocular vision and laser stripes

Range images have a key role for surface mapping in robot navigation and control. The robot control system can easily identify objects and obstacles by manipulating range images, however most of these range images are acquired in perspective projection, thus object position may be incorrect due to distortion caused by the perspective effect. This paper proposes an easy and efficient way to acquire range images, with a single CCD camera in conjunction with laser stripes and afterwards these range images are orthorectified to turn out to surface maps. These orthorectified range images are very useful for biped and quadruped robots, orientating them when navigating among obstacles while robot manipulators can use them to find objects by setting up their joints and picking tools.