Aimed at the problem of not taking into account both precision and efficiency, a novel ridges extraction algorithm based on “image” segmentation is proposed. For an asymptotic signal, its continuous wavelet coefficients with large absolute values usually distribute in some dispersed regions in the time-scale plane. Taking the absolute wavelet coefficients matrix as an “image”, the wavelet ridge can be detected by segmenting the “image” and locating the maximum value from each region. Experimental results show that the proposed algorithm not only outperforms general algorithms with respect to precision and efficiency, but also works well in signal denoising and signal separation.