基于小波包变换与RBF网络的大地目标辨识

利用正交小波包变换的高信号时频分辨特性引入小波包能量特征提取法，用于五类典型大地目标的特征提取和距离可分性测度计算。结果显示，小波包能量特征提取法提取的五类典型大地目标特征具有更好的可分性。将小波包能量特征提取法与径向基函数网络相结合，提出了一种大地目标辨识算法。仿真实验表明，对于分解层次m＝5，信噪比SNR≥0dB，五类典型大地目标能完全辨识。表明该大地目标辨识方法在一定条件下是可行的，具有强的抗噪声干扰能力。

Ground Target Identification Based on Wavelet Packet Transform and RBF Network

A conception and algorithm of wavelet packet energy feature is presented and applied to feature extraction of five types of typical ground target echoes. The computation results of distance measure show that the separability of wavelet packet energy feature is much better than that of Fourier transform feature. Combined with wavelet packet energy feature and radial basis function network, an algorithm for ground target identification is proposed. Simulation results show that for decomposition layer 5m= and signal-noise-ratio B0dSNR, five kinds of typical ground targets can be completely identified. It implies that under certain conditions, the proposed approach for ground target identification is of the better feasibility and strong anti-noise-interference ability.