Research on Environmental Target Image Recognition Method of Coal Mine Rescue Autonomous Robot

Abstract

Aiming at the problems of redundant computation and poor real-time performance in searching key points of visual image in standard SIFT algorithm, an Improved SIFT algorithm is proposed and applied to coal mine rescue autonomous robot to realize environment information perception, target recognition and matching in coal mine underground environment. In this method, Mahalanobis distance is used to replace the Euclidean distance and simplify the feature points in the traditional SIFT algorithm, so as to enhance the real-time performance of SIFT algorithm in recognizing coal mine environment targets and matching. The experimental results demonstrate that the improved algorithm greatly improves the real-time performance of visual recognition and the accuracy of target matching.