摘要
全井眼微电阻率成像(Fullbore Formation Micro Imager,FMI)技术广泛应用于强非均质性的碳酸盐岩储层测井精细解释.然而,由于各极板间间隙的影响,FMI无法测量360°全井壁地层电阻率信息,这使得FMI图像出现空白条带,严重影响了图像视觉和测井解释进程.为获取全井壁的电阻率信息,必须有效的填充空白条带.本文在探讨现有的电成像图像空白条带填充方法基础上,提出了一种基于卷积神经网络的电成像图像空白条带填充方法.与常用的Criminisi插值法对比填充效果显示:该方法完全适用于各种特征的电成像图像空白条带填充,延续了所有待填充边缘的暗色条带和亮色背景,填充结果与原始图像的像素值频率占比相关性高达0.9581.Criminisi插值法在复杂区域电成像图像的填充结果连通性较差且存在干扰区域,填充结果与原始图像的像素值频率占比相关性仅为0.5924.本文提出的方法具有处理速度快、干扰区域少、连通性强、相关性高的特点,能够满足电成像空白条带成规模填充的实际应用需求,还可为测量极板间的测量缺失带信息处理提供参考手段.
Fullbore Formation Micro Imager(FMI) technology is widely used in elaborate logging interpretation of carbonate reservoirs with strong heterogeneity. However, due to the influence of the gap between the plates, the formation resistivity information of 360° whole borehole can not be measured, which makes the blank strip and seriously affects the image vision and logging interpretation process. In order to obtain the resistivity information of the whole borehole wall, it is necessary to fill the blank strip effectively. In this paper, a new method based on convolution neural network is proposed, based on the discussion of the existing blank strip filling methods for resistivity imaging images. Compared with the commonly used criminisi interpolation method, the filling effect shows that this method is completely suiTable for filling blank strips of resistivity imaging images with various features, and it continues all the dark stripes and bright background of the edges. The frequency ratio correlation between the filling results and the pixel value of the original image is as high as 0.9581. The results of criminisi interpolation method show that the connectivity of the filling results is poor and there are interference areas in the complex area. The frequency ratio correlation between criminisi’s filling results and the original image is only 0.5924. The method proposed in this paper has the characteristics of fast processing speed, less interference area, strong connectivity and high correlation. It can meet the practical application requirements of filling the blank strips of resistivity imaging in large scale, and can also provide a reference for processing the missing band information between the measuring plates.
作者
张浩
司马立强
王亮
车国琼
郭宇豪
杨琴琴
ZHANG Hao;SIMA LiQiang;WANG Liang;Che GuoQiong;GUO YuHao;YANG QinQin(School of Earth Science and technology,Southwest Petroleum University,Chengdu 610500,China;State Key Laboratory of Oil and Gas Reservoir Geology and Development Engineering,Chengdu 610500,China;School of Energy,Chengdu University of technology,Chengdu 610059,China;Chuanzhong Oil and Gas Mine of PetroChina Southwest Oil and Gas Field Company,Suining 629000,China)
出处
《地球物理学进展》
CSCD
北大核心
2021年第5期2136-2142,共7页
Progress in Geophysics
基金
国家科技重大专项“四川盆地大型碳酸盐岩气田开发示范工程”(2016ZX05052)资助。
关键词
全井眼微电阻率成像
测井精细解释
图像视觉
卷积神经网络
空白条带填充
测量缺失带信息处理
Fullbore Formation Micro Imager(FMI)
Elaborate logging interpretation
Visual effects
Convolution Neural Network
Blank strip filling
Information processing of measurement missing zone
