中文摘要：

为了抑制地震数据噪声,增强可视化效果,保证地质特征识别算法有效,提出一种基于地震数据局部结构属性与全局纹理属性的三边滤波算法.在局部结构属性中定义数据强度、空间距离和朝向权值函数,使得滤波算法能够综合考虑边结构特征及其类型;全局纹理属性分析则借助于鲁棒紊乱度地震属性估计与条件参数定义,采用参数自动调整办法自适应调整滤波效果.通过数据体噪声层的结构残留量以及层位识别结果比较,解决了评价地震数据滤波算法效果的难题.

英文摘要：

A trilateral filtering algorithm based on local structure and global texture attributes is presented.The proposed algorithm can highlight the interesting features of seismic image,enhance visualization efficiency and make the geological feature detection algorithms work well.Since data intensity,spatial distance and orientation weight functions are defined by local structure attributes,each edge feature and its type can be handled with filter algorithm.The proposed global texture attribute analysis is based on the robust chaos estimation and conditional parameters definition.The effect of filter algorithms can be adaptively adjusted by the chaos attribute.We solved the difficulty of seismic filter assessment by evaluating the structure residual of noise layer together with the horizon detection methods.