A novel method for Q (quality factor Q) estimation is proposed based on crosscorrelation function and S-transform (CRST). We use the S-transform to analyse the time–frequency spectra of the crosscorrelation coefficients and extract the amplitude spectra corresponding to the maximum energy time in time–frequency spectra. The Q can be estimated using the spectra ratio based on the linear relationship between spectral ratio and frequency. Meanwhile, two time window factors are added to the Gaussian window function in S-transform to make the S-transform applicable for Q estimation. Firstly, through numerical tests and standard sample experiments, the feasibility and noise immunity of the CRST method are studied. Secondly, the applicability and stability of this method are studied using artificial samples with different Q. Finally, the stability and accuracy of the CRST method are analysed by comparing with the conventional spectrum ratio method (SR) through rock samples. The experimental results show that the Q of samples can be obtained by using the time–frequency spectrum information of the crosscorrelation coefficient. The proposed time window factors can effectively eliminate the errors caused by the conventional Gaussian window function, which the relative errors can reach about 40%. The CRST can reduce the effect of the frequency bandwidth for regression analysis. The new method can ensure that the maximum error of different Q factors (Q > 15) is about 5%. Compared with the conventional spectrum ratio method, the CRST method not only has better noise immunity, but also has higher stability and accuracy.

提出了一种基于互相关函数和 S 变换 (CRST) 的Q（品质因数Q）估计新方法。我们使用 S 变换来分析互相关系数的时频谱，并提取时频谱中最大能量时间对应的幅度谱。所述Q可使用基于分光比和频率之间的线性关系的光谱比率来估计。同时，在S变换的高斯窗函数中加入了两个时间窗因子，使S变换适用于Q估计。首先，通过数值试验和标准样品实验，研究了CRST方法的可行性和抗噪性。其次，利用不同Q的人工样本研究了该方法的适用性和稳定性。最后，通过岩石样品与常规光谱比法（SR）进行比较，分析CRST方法的稳定性和准确性。实验结果表明，Q利用互相关系数的时频频谱信息可以得到样本的数量。提出的时间窗因子可以有效消除传统高斯窗函数带来的误差，相对误差可达40%左右。CRST 可以减少频率带宽对回归分析的影响。新方法可以保证不同Q因子( Q  >15)的最大误差在5%左右。与传统的谱比法相比，CRST 法不仅具有更好的抗噪性，而且具有更高的稳定性和准确性。