Outdoor scene analysis using range data

1 . Introduction Outdoor scene analysis is a major Dart of an autonomous vehicle. It allows a vehicle to navigate safely through an unknown environment, to build a internal model Of me environment, and to identify landmarks. Several sources of data can be used to build such a vision system: range finders, color cameras. sonars. While a reliable system would probably use all of these Sensors, our work focused on the use of range data for outdoor vision. The advantages of using range data for an autonomous vehicle fall in two categories: first, the measurements are less sensitive to environmental conditions, such BS lighting. thus alleviating shadow or highlight problems. Second, a geometrical description of the observed scene is more easily derived from range data. This property is important in the area of autonomous vehicles in which the output of vision programs must be converted into usable space coordinates for navigation. For our work, we use the ERIM laser range sensor which proviaes reasonable accuracy, a field of view large enough for outdoor applications, and high acquisition speed.