Visual Segmentation and Localization of Mobile Robots Based on the Feedback Regulation

Abstract

In order to realize the rapid and accurate localization of multiple mobile robots under the global vision, a new visual segmentation algorithm based on the feedback adjustment of threshold is presented in paper. First, the position information of the mobile robots is described by the color label plates which are designed by ourselves; then, the improved grid method is used to search the color labels; finally, the segmentation threshold of the next frame image is adjusted based on the segmentation results of the mobile robots and the amount of noise in the current frame, and the visual localization of multiple mobile robots is realized through the self-adjusting thresh segmentation of color label plates. The experiment results show that the proposed method is characterized by small calculation amount, good real-time and strong noise suppression capability, performance and has strong noise suppression abilities, and the accurate visual localization of multiple mobile robots under the global vision can be realized.