DST与DSmT自适应融合算法

A Adaptive Integration Algorithms with DST and DSmT

与DST相比,DSmT可以很好的解决证据矛盾时的证据组合问题,但是DSmT计算量过大,且在低冲突情况下融合结果次于DST。文章在DST和DSmT的基础上提出一种将两种融合方法结合使用的融合算法――DST与DSmT自适应融合算法。以冲突率作为判决依据,在冲突率较低情况下采取DST融合算法,当冲突率高于一定阈值时采用DSmT融合算法,并给出了DST和DSmT之间转化的方法。在进行三维飞机序列图像的目标类型识别中,采用本文提出的自适应算法进行迭代运算,可以快速准确地进行目标识别。

Compared with DST, DSmT can be good evidence of contradictions to resolve the issue of evidence portfolios, But DSmT computation quantity oversized, also fuses the result under the low conflict situation time to DST. This paper proposes one kind of adaptive integration algorithms on basis of DST and DSmt. The method use conflict rates as a basis for judgement in conflict situations, and adopt DST integration algorithms when conflict rates is lower, and adopt DSmt integration algorithms when conflicts rate is higher. During the three-dimensional images of aircraft sequence types identified objectives, the purpose of this algorithm using iterative computation can rapidly and accurately identify a target.