Scene recognition based on fusion of color and corner features

Abstract

The advance of science and technology has motivated to face new and more complex engineering applications. These new challenges must involve not only the design of adaptive and dynamic systems but also the use of correct information. Everyday, it is more evident that good multicriteria decision making systems require different types of information; therefore data fusion is becoming a paramount point in the design of multicriteria decision making systems. This chapter presents a scenery recognition system for robot navigation using a neural network hierarchical approach. The system is based on information fusion in indoor scenarios. The neural systems consist on two levels. The first level is built with one neural network and the second level with two. The system extracts relevant information with respect to color and landmarks. Color information is related mainly to localization of doors. Landmarks are related to corner detection. The hierarchical neural system, based on feedforward architectures, presents 90% of correct recognition in the first level in training, and 95% in validation. The first ANN in the second level shows 90.90% of correct recognition during training, and 87.5% in validation. The second ANN has a performance of 93.75% and 91.66% during training and validation, respectively. The total performance of the systems was 86.6% during training, and 90% in validation. © 2010 Springer-Verlag Berlin Heidelberg.