为了在现有禁核试验条件下开展核爆电磁脉冲探测技术研究，对实测的非均匀采样核爆电磁脉冲数据进行阈值检波、直流剔除、滤波、插值和归一化处理后，采用小波分析提取信号第四尺度小波系数能谱熵、第一尺度和第四尺度小波系数计盒分形维数特征值，在统计分析的基础上分别建立其专用隶属度函数；利用模糊识别理论，对核爆电磁脉冲信号和作为主要探测干扰信号源的闪电电磁脉冲信号进行了模糊识别，获得了总识别率为99.41%的识别效果。

The paper investigates an NEMP(nuclear electromagnetic pulse) detection technique in the absence of nuclear testing. After threshold demodulation, DC elimination, filtering, interpolation and normalization, the method uses wavelet analysis to extract the forth segment of wavelet pack energy spectrum, the first and forth segments of wavelet coefficient box-counting dimension as eigenvalues, establishes their respective membership functions according to statistical analysis, and then uses the fuzzy identification theory to identify NEMP signals. The method has been used to process measured NEMP signals and LEMP(lightning electromagnetic pulse) signals, which are the main source of interference. The identification rate is 99.41%.