An Obstacle Detection System for a Mobile Robot Based on Radar-Vision Fusion

摘要

This paper presents a systematic scheme for obstacle detection using infrared radar and a monocular vision sensor. On the basis of the visual attention method we use fusion of infrared radar and vision data to obtain better detection performance. In the vision module, salient detection is invalid when no obstacle is present in an image. But infrared radar is used to detect and pre-locate objects. Then vision module locate objects by salient location method. Lastly, we test our radar-vision system in an indoor simulation environment, and our experiments show that the method is both rapid and feasible.