Mobile robot navigation method based on improved optical flow algorithm

The invention discloses a mobile robot navigation method based on an improved optical flow algorithm and belongs to the field of machine vision application. The method comprises defining a new data item by using bilateral filtering and local weighting in combination with grayscale conservation and gradient conservation assumptions, establishing an optical flow calculation model, and processing the large-displacement motion of an object by using a multi-resolution hierarchical refining method; then, acquiring a relative depth TTC from the optical flow field, and constructing the gravitational potential field function of the target object through the coordinates of the target and the end pose of the robot by using an artificial potential field model; constructing the repulsive potential field function of an obstacle by the TTC, and achieving a purpose of obstacle avoidance and motion towards the target of the robot under the common effect of the two potential fields. The method improves a traditional optical fl