A Fast Indoor Positioning Method for Mobile Robot

摘要

This paper mainly studies the indoor positioning system of mobile robot based on visual technology. In this positioning system, the detection target is mobile robots with special label and the scene information of robot is obtained through a visual sensor. Then the target will be segmented from the image by using image processing technology. In order to reduce the interference of background light source with high brightness such as reflected light, lamplight and other color light source, a filtering algorithm based on color vector features is designed in this paper. For other objects that are similar to the target color, a simple but effective screening technique based on shape feature is proposed to improve the anti-interference ability of the system. Finally, the pixel coordinates of each robot can be calculated easily with previously processed image and different robots are classified by using the Bayesian decision. The experimental results show that the method proposed in this paper can get a highly accurate pixel coordinate of the robot, the fluctuation is less than 0.5 pixels, and achieves high classification accuracy with extremely low computational costs.