Method of Implementation of Fusing Correlated Attribute Informationin Target Recognition
Yongcheng Wang · Acta Armamentarii · 2005
Target recognition is an important branch of application in multi-source information fusion. The attribute informations measured by each sensor are often correlated and dependent and therefore it is frequently a hard task to fuse data from different sensors. A novel weighted combination method based on the MADM and D-S evidence theory was put forward and the methods of establishing weight coefficients and analyzing their sensitivity were given in this paper. Moreover, the weighted conditional probability is also introduced into target recognition.