Real-time Detection and Recognition of Landmarks for Mobile Robot in Indoor Scenes

This paper proposes a robust real-time artificial landmarks detection and recognition system for indoor mobile robot. Landmarks detection and recognition for indoor robots faces two major difficulties, one is the illumination changes and the other is processing speed. In this paper, first, histograms of oriented gradient (HOG) features are extracted to resolve the problem of illumination changes. Second, AdaBoost based algorithm is used in detection phase to increase the processing speed. Finally, RBF-SVM classifier is used for recognition. Experimental results show a high detection and recognition accuracy and the processing speed is about 10 frames per second.