研究了小波网络的构造问题,提出了一种根据小波的局部时频特性,删除不包含任何目标识别样本数据的小波,初步确定网络隐层节点数的方法;采用提出的算法构造小波网络,可以减小网络的结构,加快网络的收敛速度。将构造好的小波网络作为雷达目标识别的分类器,获得了较好的识别效果。 The problem of constructing the wavelet network is studied. A method is proposed based on the local time-frequency characteristics of the wavelet to delete the wavelet that does not contain any target recognition sample data and to initially determine the number of nodes in the hidden layer. Using the proposed algorithm to construct a wavelet network, Reduce the structure of the network and speed up the convergence of the network. The constructed wavelet network as a radar target recognition classifier, get a better recognition effect.