The Research Of Planetary Rover’s Wheel Sinkage Detection Method Based On Image Processing

The planets exploration missions require the rovers to traverse over more challenging deformable rough terrain. Wheel-soil interaction plays an important role in rovers’ trafficability in deformable rough terrain. Wheel sinkage is the key parameter for estimating and predirecting wheel-soil interaction. The detection of wheel sinkage has great significance for rover mobility optimization control and prevention of wheel sinking.Most of previous researchers build the wheel sinkage estimation formula through the experiments. This paper researches wheel sinkage detection method from wheel-soil image perspective.This paper studies the image enhancement method for wheel-soil image. In view of the wheel-soil image features to improve the traditional image enhancement algorithms, the saturation dynamic piecewise nonlinear enhancement algorithm is porposed. The enhancing image quality evaluated with information entropy, contrast, sharpness and edge sharpness parameters, most of them are bigger than the image enhanced by index enhancement method, histogram equalization and saturation histogram specification. Then, combining with dynamic saturation piecewise nonlinear enhancement algorithm, the color caused by reflection of soil is removed. The color contrast between wheel area and soil area of wheel-soil image is more obvious therough wheel-soil image enhancing. This provides the basic for image segmentation.Wheel sinkage in deformable rough terrain is defined and the sinkage calculation model which is based on wheel-soil boundary feature is built. Wheel-soil boundary characteristic mathematical description is presented through the analysis of the morphological features of wheel-soil boundary. Wheel-soil image is segmentated in order to extract wheel area outline. Combine wheel area outline with wheel-soil boundary morphological features to extract wheel-soil boundary. The sinkage depth, entrance angle and departure angle can be calculated through input wheel-soil boundary to sinkage calculation model.The method has been validated by experiments using wheel-soil interaction test platform with vision collection system. The algorithm precision testing experiments are worked in flat terrain, the results of the experiments indicate that the sinkage depth accuracy is high, the relative errors of the sinkage depth are around 10%. Image adjustment method is validated through changing the camera angle to take image. And wheel sinkage dynamic detection is validated in flat terrain. The method’s applicability has been validated by experiments under various terrain conditions, which are flat terrain, raised terrain, hollow terrain, and rough terrain.