Obtaining terrain maps and obstacle contours for terrain-recognition tasks

Effective automatic terrain recognition in a natural environment is a challenging task. The work presented in this article, part of the results obtained in a 6-year project devoted to the development of a hexapod robot for humanitarian demining missions, aims at finding a simple, reliable terrain-recognition solution based on a sensor head, a device that is associated with a landmine detector and is employed to gather terrain-related data. The goal is to construct a processed representation of the main terrain features as an obstacle-contour and terrain-elevation map. The methods and algorithms presented herein have been tested in a fully operative robot composed of several subsystems.