An evidence fusion approach for characterization of heterogeneous images under complex environment

Characterization, recognition under complex environment is a challenging task. The measured signal will be submerged by noise in complex environment, which makes it difficult to characterize targets, especially when the targets share the similar characteristics. Multi-sensor information fusion will improve characterization significantly and DS evidence theory is an effective approach in heterogeneous information fusion. However, evidence from multi-sensor information is always affected by subjective factors in the process of evidence fusion. In this paper, a new evidence fusion approach for improving characterization under complex environment is proposed. To characterize the heterogeneous images better, a concept of comprehensive credibility is introduced into the proposed approach and a new update rule of evidence is designed. Some experimental results show the efficiency and effectiveness of the proposed approach.