Based on three-dimensional (3D) chaotic system, combined with maximal overlap discrete wavelet transform (MODWT), a novel underwater weak target detection method is proposed. This research focuses on the problem that the current detection systems are difficult to achieve weak target detection and frequency extraction under ultra-low signal-to-noise ratio (SNR). In this paper, we propose a new 3D non-autonomous chaotic system with complex attractors, and introduce a cosine function to achieve chaos control and enhancement. Its nonlinear properties have been detailed analyzed, such as Lyapunov exponent, bifurcation and entropy. Then, combined the idea of scale transformation and geometric sequence, a target detection array covering all frequencies is constructed. The detection array can achieve resonate with the external signal through the built-in cosine term, thereby turning the attractor into a periodic or intermittent chaotic state. The existence of the target can be determined by observing the attractor state. Considering the relationship between the generation of intermittent chaos and the target frequency, we design a novel frequency extraction method based on MODWT and Hilbert transform. This method can realize the noise reduction and envelope analysis of intermittent chaotic signal to estimate the real target frequency accurately. The experimental results demonstrate that the designed detection system can detect weak ship signals and extract their frequency information in complex ocean background. The detection SNR can reach – 43.4 dB, and the frequency extraction error is less than 0.2\(\%\). Importantly, the analog circuit of the detection system has been implemented to verify its realizability.

基于三维（3D）混沌系统，结合最大重叠离散小波变换（MODWT），提出了一种新颖的水下弱目标检测方法。本研究主要针对当前检测系统难以实现超低信噪比（SNR）下的弱目标检测和频率提取的问题。在本文中，我们提出了一种新的具有复杂吸引子的3D非自治混沌系统，并引入余弦函数来实现混沌控制和增强。对其非线性性质进行了详细分析，如Lyapunov指数、分岔和熵等。然后，结合尺度变换和几何序列的思想，构建了覆盖所有频率的目标检测阵列。检测阵列可以通过内置的余弦项实现与外部信号的共振，从而使吸引子进入周期性或间歇性的混沌状态。通过观察吸引子状态可以确定目标的存在。考虑到间歇性混沌的产生与目标频率之间的关系，我们设计了一种基于MODWT和Hilbert变换的新型频率提取方法。该方法可以实现间歇混沌信号的降噪和包络分析，从而准确估计真实目标频率。实验结果表明，所设计的检测系统能够在复杂的海洋背景下检测微弱的船舶信号并提取其频率信息。检测信噪比可达 – 43.4 dB，频率提取误差小于0.2 \(\%\)。重要的是，检测系统的模拟电路已经实现，验证了其可实现性。