Research on Service Robot Vision Alignment Algorithm Based on the SIFT Characteristic

Abstract

For service robot vision localization requirements, the localization technology research based on the image SIFT characteristic was introduced. The space examination, the precise positions of characteristic points, the direction parameters of the assigned operator and the description of the characteristic point were analyzed. At the same time, the stability under the condition of image zoom, rotation and affine transformation was analyzed according to experiments. Experiments result shows that the SIFT characteristic has the proportion zoom invariability, the revolving invariability, the part affine invariability and a high recognition rate at complex environments. On the basis of above work, the vision localization method based on SIFT characteristic turns out to be an applicable technology in in-building complex environment.