Shape of object recognition based on information fusion for intelligent robot

When moving under unknown environment, the intelligent robot must have the capability of recognizing object. In this article, we focused on studying two aspects during object recognition, one was extraction of target shape feature, and the other was recognition algorithm. On studying feature extraction, we proposed the apothem sequence that is shape descriptor based on object border and took it as characteristic quantity of recognition. Experiment show that apothem sequence is a simple and effective. On studying recognition algorithm, the RS-ANN information fusion algorithm combined rough set theory with neural network was proposed. At first, we reduced the information table by Rough set, which was formed by training sample set, in order to unearth minimal decision-making regulations, and then the structure of BP network was confirmed, and the shape of object is recognized finally. Experimental results show that the algorithm solved the problem of redundant feature samples, so met real-time requirements of object recognition of the mobile robot in a dynamic environment.