一种改进的鱼雷电磁引信目标信号实时小波消噪算法

An Improved Algorithm of Real-time Wavelet Denoising for Target Signal of Torpedo Electromagnetic Fuze

针对Mallat小波变换在算法原理上不具备数据流动性,无法满足连续采样信号实时消噪处理要求的问题,介绍了一种基于Mallat算法改进的垒墙式小波变换算法,分析了该算法所具有的数据流动性,推导了数据流动性与小波分解层数的关系,并将该算法运用到被加性高斯白噪声污染的鱼雷电磁引信目标信号的实时消噪处理中,选取双正交样条小波作为小波元,并对目标信号做两层小波分解与重构。通过MATLAB环境下的仿真试验,验证了采用该算法实现连续采样信号实时消噪的可行性。消噪后的目标信号具有失真度较小,波形平滑的特点。

The wavelet transform based on Mallat algorithm cannot process flowing data, it is therefore unable to meet the requirement for real-time denoising of continuous sampling signals. As a result, this paper proposes a novel wavelet transform based on Mallat algorithm modification, called bricklaying algorithm. The ability of the proposed algorithm to process flowing data is analyzed, and the relationship between data flowing and level of wavelet decomposition is de- duced. The bricklaying algorithm is applied to real-time denoising of torpedo electromagnetic fuze signal in the back- ground of additive white Gaussian noise, where a biorthogonal spline wavelet is taken as the wavelet function, and 2-level wavelet decomposition and reconstruction are performed. The feasibility of the bricklaying algorithm in real-time denoising of continuous sampling signals is verified via MATLAB simulations. The results show that the de- noised target signal gains the feature of smooth waveform with smaller distortion.