Technique of optical target fusion recognition based on fuzzy evidence theory

Abstract

As a technique of uncertainty inferring, D-S evidence theory is widely used in the region of information fusion. The problem is how to create the basic probability assignment function (BPAF) of evidence. A technique based on the membership function that can generate BPA is given. It is used to recognize optical target by extracting the peak power, pulse-broaden and flicker variance of the echo of active laser detection. The preponderances of non-imaging laser detection are high-precision, high-flexibility, and anti-electromagnetic interference, and that the cat's eye optical target is easy to be detected. Consequently, this paper presents the active probing of cat's eye target as an example. Firstly, the BPA of evidence is got by fuzzy theory. Then the result is got through combining them together by evidence theory. The experiment indicates that the technique can discriminate the cat's eye target exactly and improve the efficiency of recognition compared to identification with single attribute.