为了从一类非高斯噪声——多模噪声背景中有效提取目标信号,把高阶谱理论与小波包网络相融合,根据高斯分布的斜度为零,以及结合小波包网络具有改善信号检测性能和提高检测手段自适应性的特性,采用频带分割检测技术,提出一种基于斜度-小波包网络的弱信号检测方法.大量计算和MATLAB仿真表明,该方法在多模噪声且信噪比较低的情况下具有良好的消噪能力,并比小波包变换的检测性能优越. In order to effectively extract the target signal from a class of non-Gaussian noise-multimode noise background, the high-order spectral theory is integrated with the wavelet packet network, the slope is zero according to the Gaussian distribution, and combined with the wavelet packet network, the signal detection performance is improved And to improve the adaptability of the detection means, a band-division detection technique is proposed to detect the weak signal based on the gradient-wavelet packet network.Many calculations and MATLAB simulation show that the proposed method is effective in multi-mode noise and signal-to- It has good de-noising ability under low conditions and superior performance than wavelet packet transform.