A iterative learning method for indoor robots visual perception based on multi-feature fusion

For improving the accurate and the real-time requirement of indoor robots in environmental visual perception, a new model of visual perception based on image multi-feature fusion with iterative learning control is proposed. The hierarchical match mode is used to match real-time collected images of indoor robot with various multi-directional and multi-state images in a database. After establishing a database of 1000 images, average accuracy, average recall ratio and average time are used to evaluate the algorithm. Experimental results show that the algorithm can accurately and efficiently apperceive target images. Relative to single feature visual perception, the algorithm can not only achieve higher matching accuracy, but also meet the real-time requirement of robots.