3-D Vision Techniques for Autonomous Vehicles

Abstract

There are many sensing strategies for perception for mobile robots, including single camera systems, sonars, passive stereo, and laser range finders. Here, we focus on perception algorithms for range sensors that provide 3-D data directly by active sensing. Using such sensors has the advantage of eliminating the calibration problems and computational costs inherent in passive techniques such as stereo. This Chapter describes some of the techniques we have developed in this area of research. The aim of our work is to produce models of the environment, which we call the terrain, for path planning and object recognition.