聚类分析方法识别水下目标

水下目标的固有振动频率与其本身的材料参数密切相关。如能从测得的目标共振频率反演目标材料参数，结合某些先验知识，便可对目标的某些物理属性作出判断，实现或有助于实现水下目标识别。Ｂａｔａｒｄ等人研究了柱形目标中声速与其固有振动频率间的关系。本文分析并简化了这种关系，然后应用模态识别中的动态聚类方法，由目标共振频率反演目标材料参数，所得结果与实际材料参数符合较好。

RECOGNIZING UNDERWATER TARGET BY CLUSTERING ANALYSIS METHOD

The resonance frequencies of the underwater target closely relate to its material parameters. If we can obtain the material parameters from the resonance frequencies of a target and use some priori knowledge, then we can estimate physical characters of the target, and it is helpful to recognize the underwater target-H. Batard obtained an approximate relationship between resonance frequencies and sound velocities in the target- This relationship is analysed and simplified in this paper. The material parameters of the target are inversely deducted by a clustering analysis method in pattern recognition. The obtained results are in agree with the real parameters of the target quite well.