抄録
Obstacle detection and avoidance are essential for achieving a successful soft landing on the moon and the planets. In this paper, we propose an image processing method for obstacle detection and avoidance on the moon and the planets. The proposed method consists of two stages. The first stage is a coarse classification of all points in an image into two regions: the safe landing region and the obstacle region. The discrimination function for the classification is composed of local and semi-global statistics of image intensity. This function is devised so as to avoid the opposition effect of the sun and to be independent of the direction of the sun. In the second stage, a point in the safe landing region that has the maximum distance from the obstacle region is selected as a landing point. Although the classification of the first stage is coarse and there is some miss classification, the procedure of the second stage ensures the selection of a safe landingg point. The ability of the proposed method is evaluated by using CAD images constructed by fractal modeling.
