【摘要】 本文基于Liu等的改进算法,以Morlet小波作为基本匹配子波,在控制参数的确定过程中,充分利用地震信号的瞬时特征确定动态扫描范围,并根据地震信号本身的特点创建动态子波库;通过一次迭代同时确定多个匹配子波,再利用匹配子波控制参数之间的关系,提出了双参数快速匹配追踪算法,进一步减少了扫描参数的个数,使计算效率得到了质的提高。为了使算法更快收敛,利用残差信号的能量对双参数快速匹配追踪算法的迭代次数进行有效控制,不仅避免了毫无意义的计算,而且可在一定程度上去除随机噪声。最后,本文利用各匹配子波的独立性,提出了一种变窗口短时傅里叶变换方法,得到了高精度时频分布图。更多还原

【Abstract】 Based on improved algorithm by liu et al.,thking Morlet wavelet as basic matching wavelet,making fully using of the instantaneous characteristics of seismic signal to determine the scanning range dynamically and creating a dynamic wavelet library according to characteristics of seismic signal itself in determination of controlling parameters;and then determning the several matchng wavelet simultaneously through one iteration,using the matching wavelets to control the relationship among the parameters,the dual-parameter fast matching tracing algorithm was proposed in this paper,which further reduced the number of scanning parameters and greatly raised the calculation efficiency.In order to gain faster convergence of the algorithm,the energy of the residual signal was utilized to effectively control the iteration times for the fast matching tracing algorithm,which not only avoid the meaningless calculation,but also could remove random noise. Finaly,by utilizing independence of each matching wavelet,a variable window short-time Fourier transform method was proposed and used,through which the time-frequency distribution map was obtained.更多还原