«上一篇/Previous Article|本期目录/Table of Contents|下一篇/Next Article»

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:

期数:

2025年03期

页码:

18-25

栏目:

媒体计算

出版日期:

2025-03-10

文章信息/Info

Title:

Research on Imaging Logging Image Restoration Method Based on VAEAC

文章编号:

1673-629X(2025)03-0018-08

作者:

曹茂俊; 张鹏杰

东北石油大学 计算机与信息技术学院,黑龙江 大庆 163318

Author(s):

CAO Mao-jun; ZHANG Peng-jie

School of Computer and Information Technology,Northeast Petroleum University,Daqing 163318,China

关键词:

电成像测井; 图像复原; VAEAC; 转置卷积; 卷积块注意力机制

Keywords:

electrical imaging logging; image restoration; VAEAC; transposed convolution; convolutional block attention module

分类号:

TP391

DOI:

10.20165/j.cnki.ISSN1673-629X.2024.0358

摘要:

成像测井图像能清晰展示地下的地质结构和特征,助力地质分析和资源勘探。 但由于电成像测井仪器的设计限制,电成像测井图像上会出现空白条带,这严重影响了图像的完整性和可用性。 现有的图像复原方法在处理空白条带的填充问题时效果不佳。为此,提出一种基于改进任意条件变分自编码器(Variational Autoencoder with Arbitrary Conditioning,VAEAC)的成像测井图像复原方法。 该方法在解码器部分引入卷积块注意力机制,以增强模型在通道和空间两个维度上对特征图重要性的自适应学习能力。 此外,采用转置卷积替代传统的上采样技术,提高了上采样过程中对细节信息的捕捉能力。 实验结果表明,测试集中五组具有不同缺失区域的成像测井图像的平均结构相似性度量为 0. 94,与其他同类方法比较提升了 0. 25 左右。 改进后的 VAEAC 模型在处理电成像测井图像复原任务时表现更为出色,不仅有效复原了成像测井图像的纹理特征,还保留了其语义结构,为后续的成像测井图像解释提供了更为准确的图像信息。

Abstract:

Imaging logging images can clearly show the geological structure and characteristics of the underground,which helps geological analysis and resource exploration. However, due to the design limitations of electrical imaging logging instruments, blank strips will appear on electrical imaging logging images,which seriously affects the integrity and usability of the image. Existing image restoration methods are not effective in dealing with the problem of filling blank strips. To this end,an imaging logging image restoration method based on improved Variational Autoencoder with Arbitrary Conditioning (VAEAC) is proposed. This method introduces a convolutional block attention module in the decoder part to enhance the model’s adaptive learning ability of the importance of feature maps in both channel and spatial dimensions. In addition,transposed convolution is used instead of traditional upsampling technology to improve the ability to capture detailed information during upsampling. Experimental results show that the average structural similarity measure of five groups of imaging logging images with different missing areas in the test set is 0.94,which is about 0. 25 higher than that of other similar methods. The improved VAEAC model performs better in processing electrical imaging logging image restoration tasks. It not only effectively restores the texture features of the imaging logging images,but also retains their semantic structure,providing more accurate image information for subsequent imaging logging image interpretation.

常用功能

更新日期/Last Update: 2025-03-10