Research On Obstacle Recognition And Avoidance Technology For Mars Surface

Achieving Mars soft landing safely is a key step in Mars exploration. So it’svery necessary to develop a soft landing method which can implement autonomousobstacle detection and avoidance. In this paper, according to Theory Method ofSpatiotemporal Characterization and Feature Recognition of Planet’s SurfaceTerrain, some detection methods about dangerous obstacle on the surface of Marsare researched, such as craters and rocks.Firstly, a crater detection method based on snake model is developed. Closedcontour of a crater can be detected effectively by this method. Many craters can bedetected simultaneously with this method.Secondly, a detection algorithm based on gray histogram is presented. Thethreshold value need to set manually. Rocks can be extracted from the whole image.Obstacle regions can be detected completely. A rock recognition method whichpossesses the advantages of adaptability based on Otsu is presented. This algorithmcan select a suitable threshold value automatically and then segment the wholeimage. A calculation method of rocks areas is put forward base on geometricFeatures.Thirdly, selection method of safe landing point is studied. A selecting methodof safe landing point based on shadow information is put forward. First selectedlanding point and other spare landing points are selected effectively and quickly.Finally, an obstacle avoidance method based on Lyapunov function is putforward. Obstacle areas are avoided and the lander can reach upper side of thepredetermined landing point.