Obstacle avoidance for mobile robot navigation in unknown environment using geometrical information of mobile camera images

Abstract

In this paper, a novel and robust algorithm for avoiding obstacles in an unknown environment is proposed that utilises the geometrical analysis of the distances present in the images. A mobile camera with pan feature is used for increasing the field of view of the robot. Low level image properties (red green blue colour; hue saturation value colour and brightness gradient) are used for obstacle detection. Once the positions of the obstacles are determined in the image these are transformed into actual distances by the geometrical analysis and a safe passage is determined. The algorithm has been successfully implemented to avoid obstacle(s) during navigation of a mobile robot. It has been observed that the proposed algorithm increases the robustness of obstacle avoidance and helps in the effective path planning of an autonomous mobile robot.