A robust method of mark recognition for robot autonomous navigation

Abstract

The vision system of an autonomous robot can obtain and analyze the necessary environmental information used for autonomous decisions. The processing and recognition of the images captured through the autonomous robot visual system is discussed in this paper. The discussion mainly includes three key issues, namely the extraction of image region of interest, the canceling of image inverse projection error, and pattern recognition and matching. For these three key issues, the paper proposes respective solution and verifies the validity of the above solutions through simulation experiments. Finally, the paper analyzes the existing problems in the scheme.