基于EMD和奇异值分解的搜救目标判别

Target Classification for Life-detection Based on Empirical Mode Decomposition and Singular Value Decomposition

为解决穿墙信号研究中目标分辨问题,提出基于经验模态分解和奇异值分解的分类方法。首先利用经验模态分解方法自适应地将回波信号分解为一组固有模态函数(IMF)和残余分量,利用固有模态函数构造初始特征向量矩阵,分解得到矩阵的奇异值,再由奇异值计算奇异熵,依据奇异熵和马氏距离判别函数对搜索目标类型进行分类。仿真结果表明本方法能方便有效地对目标类型进行识别判断,可用于相关领域。

Based on empirical mode decomposition and singular value decomposition,a classification method for targets behind barriers recognition is introduced in this paper.Firstly,by empirical mode decomposition（EMD） the echo signal can be decomposed into a group of intrinsic mode function（IMF） and a residual component.The initial feature vector matrix can be formed by rearranging the IMFs.Then,by singular value decomposition the feature vector matrix is decomposed into the singular values and the singular entropy of original feature vector matrix can be calculated.Finally,in virtue of the singular entropy values and mahalanobis distance criterion function,compared with standard feature values of training samples the experimental data of echo can be classified into different types.Simulation indicates that the new approach can identify the type of targets easily and effectively.