Research On Moving Obstacle Detection And Avoidance Strategy For Agricultural Robot Based On Machine Vision

Real agricultural environment is very complex. There are both static obstacles and moving obstacles in this kind of environment. Thus, moving obstacle detection and avoidance becomes an important part in agricultural robot navigation. This paper dealed with the navigation path detection, moving obstacle detection and avoidance strategy of the agriculture robot navigation system based on the vision information from the stereo camera. The main research and creative achievement included following aspects:A new algorithm for moving obstacle detection in unstructured outdoor environment was proposed for the agricultural robot to dectect the moving obstacle online. First, compensated the frame motion caused by the ego-motion of the robot by using the vision information. Second detected the moving obstacle using a frame differencing algorithm. It was found that the proposed approach can detect the moving obstacle online, and it laid a good foundation for the follow-up obstacle avoidance.Comparision among robot navigation strategies was made. Then velocity obstacle method and fuzzy control method were choosen for depth analysis, based on the method feature, the robot and sensors of this paper used the avoidance strategy was made. Outdoor experiments proved the avoidance strategy feasibility.This paper deeply studied moving obstacle detection and avoidance strategy, and provided a feasible mean to achieve autonomous navigation of agricultural machinery.