Obstacle detection based on image and laser points fusion for a small ground robot

Abstract

Obstacle detection is still full of challenge for a small ground mobile robot with limited load capacity and computing resource that works in complex outdoor environment. This paper proposes an obstacle detection method based on fusion laser points and image. The algorithm employs a fuzzy clustering algorithm to detect obstacle in laser points and trains a SVM classifier with online updating and retraining ability by using multi image features to find obstacles in image. The algorithm calibrates the range finder and camera and mapped the laser points into image to enhance the obstacle detection result. Experiments show the proposed algorithm can improve the obstacle detection result of single sensor in campus environment.