Unmanned Visual Localization Based on Satellite and Image Fusion

Abstract

Unmanned driving is an important means for future human to achieve locomotion, and it will have broad application prospects. The unmanned vehicle still has the following difficulties in its long distance displacement: Single source sensor cannot meet the requirements of the positioning accuracy of the changeable and complex scenes, and the image analysis accuracy of visual processing in complex interference needs to be improved. In order to solve these problems, an unmanned vision localization algorithm based on multi-sensor fusion is proposed in this paper, by analyzing the positioning perception accuracy of unmanned vehicle, the precision and range of different perception methods at large, medium and small scales are obtained. A vision localization algorithm of multi-source fusion based on pseudo-range equivalence is designed in this paper. In order to reduce the influence of image distortion on localization accuracy, a visual localization algorithm based on image feature matching is proposed. The localization accuracy in complex environment is effectively improved by the multi-source fusion localization algorithm of pseudo-range equivalence. The MATLAB simulation shows that the positioning accuracy of the unmanned driving is improved to a certain extent at different scales.