基于小波分解的模极大值算法在钛合金超声检测信号去噪中的应用

本文针对钛合金超声检测信号中普遍存在的噪声过大、信噪比过低的现象，结合钛合金超声检测信号与材料散射噪声信号的产生特点，构造了一组数学模型，并提出了基于小波分解的一种改进的模极大值去噪算法。该算法利用小波分解中的低频系数来确定信号位置，在高频各尺度系数的对应位置寻找相应的频率信息。以达到提取信号去除噪声的目的。实例分析表明：该模型能基本反映超声检测信号的特点，而且此算法对钛合金超声检测信号去噪效果比较理想。

Study of modulus maximum algorithm based on wavelet analysis in ultrasonic testing signal denoising for titanium alloy

The ultrasonic testing signal of titanium alloy is always immingled wi th the scattering noise. According to the phenomena, this paper presents a kind of signal model of ultrasonic testing for titanium alloy and a wavelet transform denoising algorithm based on modulus maximum. The characters of producing mode and wavelet transform of noise and signal are discussed. The processing of the a lgorithm is followed: firstly, ascertain the location of signal in the low frequ ency coefficient of wavelet transform; secondly, find corresponding frequency in formation in the high frequency coefficient; thirdly, reconstruct the signal thr ough inverse wavelet transform. Examples prove that the model can reflect the ch aracter of the ultrasonic signal and noise correctly, and the result of the deno ising using this algorithm is satisfactory.