基于纹理的超声成像测井图像“城墙效应”修复研究  

Ultrasonic Logging Image Restoration Based on Texture

作　　者：倪路桥[1] 余厚全[1] 李长文 李国军 余春昊 

机构地区：[1]长江大学电子信息学院,湖北荆州434000 [2]中国石油集团测井有限公司,陕西西安710077

出　　处：《测井技术》2010年第5期428-431,共4页Well Logging Technology

基　　金：国家"863"计划重点科研项目"先进测井技术与装备"中"方位声波成像测井技术研究"(编号:2006AA060103)

摘　　要：根据测井图像中病态图像的特点,基于现代图像修复技术,提出了一种基于纹理的图像修复算法。介绍了该算法的基本思想,给出了利用基于纹理的图像修复算法处理超声测井图像的具体流程。该算法按照优先修复含有已知信息最多、中心点处等照度线方向与待修复区域边界垂直方向一致性最好的待修复块的原则,通过最佳匹配,在已知信息区域中寻找出最佳匹配块,对"城墙效应"进行修复。处理结果表明,该算法能够有效修复图像中的"城墙效应",保持结构的完整性,明显改善图像质量。该算法简单、实用、实时性强、无需人工干预,对超声成像测井资料的后期处理具有较高实用价值。Proposed is a modified inpainting algorithm based on texture for logging image restora tion according to the characteristics of ill image. Introduced are the principle of this algorithm and specific processing procedure for ultrasonic logging image restoration by using this algorithm. The inpainting algorithm gives priority to blocks which have the most information and those central pointffs isophotes direction have best consistency with the vertical direction of the boundary in the area. And then found out is the best matching block by matching in the neighborhood area to repair the ＂wall effect＂. The processing results show this algorithm can effectively repair the ＂wall effect＂ image, maintain structural integrity, and improve significantly image quality. The algorithm is simple, real-time, and no more human interface, and has practical value for post-pro- cessing of the BHTV logging imaging data.

关 键 词：超声成像测井 纹理修复 井下电视 图像处理 

分 类 号：P631.83[天文地球—地质矿产勘探]