【摘要】 蒙特卡洛速度谱的自动拾取采用一种非线性优化的快速方法 ,所得叠加速度的精度可以满足常规地震资料处理中叠加的要求 ,转换为层速度后可用于叠前深度偏移成象 ,在给定初始速度模型前提下建立优化目标函数。采用蒙特卡洛非线性优化算法 ,以最大相似度量准则在考虑实际地质条件前提下给出速度约束条件 ,对初始速度模型加以扰动 ,自动寻找速度谱中叠加能量的全局最优解 ,从而获得合理的速度模型。实际地震资料处理的应用结果表明 ,自动拾取速度谱比常规的人工速度谱解释工作的效率大幅度提高。在SUNUltra6 0机上 ,仅需 1h即可完成 2 0 0 0km的地震资料的速度谱拾取任务 ,而且所获得的速度模型不受人为因素影响 ,转换所得的层速度模型用于叠前深度偏移后 ,获得了满意的成象效果。更多还原

【Abstract】 Automatic Monte Carlo Velocity Picking (AMCVP), developed a non-linear optimization algorithm is a effective stack velocity spectrum interpretation method to produce reasonable stack velocity method that can be used in pre-stack seismic processing and transformed to interval velocity for pre-stack depth migration. Based on given initial velocity model a optimal goal function is constructed. Constrained with real geological information non-linear Monte Carlo optimization searches the global optimal solution within stack velocity spectrum on the light of maximum semblance criterion, and the reasonable stack velocity methods can be obtained effectively.The result of seismic data has shown that AMCVP has move efficient than manual velocity spectrum interpretation. It only spent one hour to finish 2 000 km seismic data AMCVP on a sun Ultra60 computer, The precision of velocity method produced by AMCVP is better than manual velocity interpretation. The interval velocity method converted form picking stack velocity method was developed for pre-stack depth migration, and images are high quality.更多还原