小波脊线能较好地揭示信号瞬时频率变化情况。针对目前多小波脊线提取方法存在的问题，提出了一种多小波脊线提取的新方法。该法通过连续小波变换得到小波尺度谱后，利用重分配算法对其进行处理，再通过奇异值分解降噪，然后通过求小波系数的模极大值点来提取各分量的小波脊线。与其它方法相比，该法更加适合于某些分量具有较大载波频率的低信噪比多分量AM-FM信号的小波脊线提取。在齿轮故障特征提取中的应用结果也验证了所提方法的有效性和实用性。

Wavelet ridge can reveal the information about the instantaneous frequencies of signals. According to the problems of the current multi wavelet ridge extraction methods, a new method for multi wavelet ridge extraction is proposed. After the wavelet scalogram is obtained by continuous wavelet transform, the reassigned algorithm is applied to it, and the reassigned wavelet scalogram is denoised by the singular value decomposition further, then the wavelet ridge of each component in the signal is extracted via modulus maxima of wave-let coefficients. Compared with other methods, the proposed method is better suitable for extracting the wavelet ridges of the multicompo-nent AM-FM signal with low signal-to-noise ratio, in which some components have large carrier frequencies. The result of the application to gearbox fault characteristic extraction verifies the effectivity and practicability of the proposed method.