决策模板法在决策层融合目标识别中的应用和改进

Modification and application of decision template method in decision-level fusion for target recognition

决策模板法是一种简单直观的决策层融合识别算法,但是经典的决策模板法没有充分利用各传感器对于不同类目标鉴别能力的先验信息。本文提出利用传感器平均度量熵对决策模板法进行修正,合理度量各个传感器对不同类目标的分类鉴别能力,仿真结果表明改进的决策模板法能提高目标正确识别率。

Decision template method is a simple and intuitionistic decision fusion algorithm, but the prior information of different sensors＇ discrimination capability towards different targets is not used sufficiently. This paper represents a modified decision template method based on the entropy of sensors＇ uncertain measurement, which weights the classifying capacity of each sensor toward each class of target. Simulation result shows that target recognition ratio can be improved by the modified algorithm.