Object following based on SURF for mobile robots

Abstract

This paper presents an object following algorithm based on Speed UP Robust Features (SURF). By using features extracted from video sequences using SURF operator, target object can be tracked from the images. Based on the deviation of the tracked object from the image center and the resizing ratio, suitable control actions can be imposed on the wheel motors to move the robot to closely follow the target object. Experimental results have shown that the proposed object following system can achieve the objective of real-time following of the target object.