提出了基于小波包分解特征的神经网络导弹目标识别算法。在该算法中,首先通过小波包分解获取能够反映时间 序列时频信息的稳定特征,然后利用训练过的神经网络提取特征对导弹目标进行识别。文中运用一组仿真数据和一组试 验数据对该算法进行测试,结果表明该算法具有较高的识别概率。

Missile target recognition algorithm of neural networks based on wavelet packet decomposition is presented. Firstly, stable feature is obtained by wavelet packet decomposition which can reflect the time-frequency information of time serials. Secondly, the feature which represents some target is identified by the trained neural networks. This algorithm is tested using artificial data and real data. The result indicates upper recognition probability.