A Method for Lunar Obstacle Detection Based on Multi-scale Morphology Transformation

Lunar obstacle detection is the key technique for Lunar lander landing, the main step to detect obstacles is to separate and extract the obstacle regions. This paper puts forward a watershed method based on multi-scale mathematical morphology operation to separate the obstacles of the lunar. The method put up multi-scale open and close operator to morphological gradient image, and wipe off the local valley produced by noise and quantification error of the gradient image. The experiment proves that using watershed transformation adopting the method to process the image can wipe off the local valley produced by over-segmentation and separate the dark edge efficiently, in other words, detect the obstacles of the lunar. The effect is much better than the process of the traditional watershed transformation.