A Learning Dempster-shafer Model for Automated Building Detection

This paper presents a learning Dempster-Shafer model for the detection of buildings in aerial image and range data. The process of evidence assignment in the Dempster-Shafer method is implemented through membership functions in an adaptivenetwork-based fuzzy inference system, where a back propagation learning rule is employed to tune the evidence assignment functions using training samples. The advantage of this method is that it incorporates our knowledge about various features that can be extracted from multisource aerial data, and the evidence that these features provide for buildings and other objects in urban and suburban areas. Experimental results show that the proposed learning model improves the performance of the Dempster-Shafer classifier in detecting buildings in multisource aerial data.