A Visual Localization Method Based on Indoor Signs

A visual localization method is proposed to provide a way of localizing intelligent vehicles and mobile robots in indoor environment. The proposed method exploits various signs within indoor environment and uses the boosted efficient binary local image descriptor(BEBLID)algorithm. The proposed method enforces the ability to characterize the whole image by improving the classic BEBLID. The localization method consists of an offline and online component. For the offline component, a scene sign map is created. For the online component, the localization progress is divided into 3 parts. In the first part, the holistic BEBLID features are matched. The closet sign site and the closet image are located by using the KNN method. In the second part, the correspondences of key points are identified by local BEBLID features matching. In the third part, the current position is localized by metric calculation using coordinate information stored in the scene sign map. The experiment is conducted in three kinds of indoor scenes, including a teaching building, an office building, and an indoor parking lot. The results show that the recognition rate of signs in the scene reaches 90%, and the average localization error is less than 1 m. Compared with the traditional methods, the proposed method improves about 10% of relative recognition rate with the same test set, which verifies the effectiveness of the proposed method.