A Managerial Approach of Visual Features in Mobile Robot SLAM

A managerial approach of features for mobile robot Simultaneous Localization and Mapping with visual sensor is proposed in this paper. The regular pattern of features matching in image sequence which is obtained in the moving process of robot was researched. Based on the regular pattern, the features of good quality which could be matched continuously were reserved, and the bad features which could be matched only once were eliminated to reduce the number of features. So the computation complexity was reduced by the number reducing of features, meanwhile the efficiency of SLAM was increased. It is demonstrated by mobile robot SLAM experiments with omnidirectional vision sensor that the computation consuming time of mobile robot SLAM is reduced obviously with a good accuracy kept.