【摘要】 经典的地震波形反演采用梯度算法,该方法计算量小,但收敛速度慢,不容易重建速度场的低波数成分。我们在文中讨论了雅可比矩阵的计算,实现了高斯—牛顿法反演。数值例子表明,与梯度法相比,高斯—牛顿法收敛迅速,反演精度高,是一种更加有效可行的物性反演策略。更多还原

【Abstract】 The classical seismic waveform inversion adopts gradient method,it doesn’t require much computing consumption.While it’s convergence rate is slow and the reconstruction for the low wave num- ber component of the velocity filed is difficult.In this article,we discussed the calculation of Jacobian ma- trix,accomplishing Gauss-Newton waveform inversion.The numerical example shows that compared with gradient method,Gauss-Newton method converges faster,and the inversion accuracy is higher,hence,it’s a more effective and feasible inversion strategy for physical properties.更多还原