Target Identification of Correlation Information Based on BP Network D-S Evidence Theory

摘要

The characteristics of the D-S evidence theory is analyzed and based on the requirement that the evidences are independent mutually and probabilities are known, this paper presents a new target identification method of correlation information. From the detected correlation evidences, first the correlativity theory is used to complete decoupling of these evidences and BP network is applied to calculate confidence of these evidences, then the D-S evidence theory is used to fuse the information in order to get the high accurate and reliable results. The experiment shows that the method is practicable and it can effectively enhance the efficiency and reliability of the target identification of the correlation information.