超声电视测井图像中异常点剔除算法  

A NEW ALGORITHM FOR EXTRACTING OUTLIERS IN THE ULTRASONIC TELEVIEWER LOGGING IMAGE

作　　者：涂继辉[1] 李长文 余厚全[1] 邹伟[1] 余春昊 李国军 

机构地区：[1]长江大学电信学院,湖北荆州434023 [2]中国石油集团测井公司技术中心,陕西西安710021

出　　处：《物探与化探》2012年第2期307-311,共5页Geophysical and Geochemical Exploration

基　　金：国家"863"计划重点科研项目(2006AA060103)资助

摘　　要：针对超声电视测井图像中经常会出现的异常点问题,提出了一种采用全变分的修复算法(TV算法)来剔除异常点。该方法先在选定的区域里根据阈值来确定要剔除的异常点,然后利用全变分的图像修复算法去修复这些异常点。实验结果证明,该方法能够很好地对异常点进行剔除,保证测井图像的完整性,明显地改善图像质量,对实际测井资料的处理具有很重要实际意义。The image of ultrasonic televiewer logging often has a lot of abnormal points,and hence this paper proposes a new method for reducing outliers based on the TV-inpainting to solve the problem of extracting these outliers.Firstly,it chooses reduced outliers according to the threshold in the selected regions,and then utilizes TV-inpainting to repair these outliers.The processing results show that the proposed method can effectively extract the outliers,ensure logging image integrity,and significantly improve image quality,which is of very important practical significance for processing the logging data.

关 键 词：超声电视图像 超声测井 全变分的图像修复 异常点 

分 类 号：P631.8[天文地球—地质矿产勘探]