基于改进UNet++的铸体薄片图像喉道分割算法

岳枫云; 何小海; 滕奇志; 龚剑

doi:10.3969/j.issn.2096-6091.2023.23.003

您当前的位置：

首页 >

文章列表页 >

基于改进UNet++的铸体薄片图像喉道分割算法

研究论文 | 更新时间：2024-10-08

- 基于改进UNet++的铸体薄片图像喉道分割算法
- An Improved UNet++ Based Algorithm for Throat Segmentation in Cast Slice Images
- 新一代信息技术 2023年6卷第23期页码：15-22
- 作者机构：
  
  1.四川大学电子信息学院，四川成都 610065
  2.成都西图科技有限公司，四川成都 610065
- 作者简介：
  
  岳枫云（1999—），男，现为四川大学硕士研究生。研究方向：数字图像处理。E-mail：1218688274@qq.com
  何小海（1964—），男，教授，现为四川大学博士生导师，研究方向：图像处理与网络通信。E-mail：hxh@scu.edu.cn
- 基金信息：
- DOI：10.3969/j.issn.2096-6091.2023.23.003
  中图分类号： TP391.41
- 纸质出版日期：2023-12-15，
- 稿件说明：
移动端阅览
岳枫云, 何小海, 滕奇志, 等. 基于改进UNet++的铸体薄片图像喉道分割算法[J]. 新一代信息技术, 2023, 6(23): 15-22

YUE Feng-yun, HE Xiao-hai, TENG Qi-zhi, et al. An Improved UNet++ Based Algorithm for Throat Segmentation in Cast Slice Images[J]. New Generation of Information Technology, 2023, 6(23): 15-22
岳枫云, 何小海, 滕奇志, 等. 基于改进UNet++的铸体薄片图像喉道分割算法[J]. 新一代信息技术, 2023, 6(23): 15-22 DOI： 10.3969/j.issn.2096-6091.2023.23.003.

YUE Feng-yun, HE Xiao-hai, TENG Qi-zhi, et al. An Improved UNet++ Based Algorithm for Throat Segmentation in Cast Slice Images[J]. New Generation of Information Technology, 2023, 6(23): 15-22 DOI： 10.3969/j.issn.2096-6091.2023.23.003.

摘要

分析岩石中的喉道特征对岩石储层质量的衡量有着重要意义。传统的岩石铸体薄片图像喉道分割方法在提取喉道时主要依赖于孔隙的二值化图像，当分析复杂的孔隙结构时，这些方法容易使提取的喉道出现过分割和分割精度不高的问题，从而导致算法泛化性能差。针对上述问题，本文提出了一种基于深度学习的喉道分割算法。该算法采用UNet++作为特征提取的主干网络，在下采样过程中结合CBAM（Convolutional Block Attention Module）注意力模块以增强模型捕获不同维度特征之间相关性的能力，减小跳跃连接中的语义差异，从而提高喉道特征提取的准确性。本文算法与其他现有喉道分割算法分别就正确分割率CR（Correctness Rate）、欠分割率UR（Under-segmentation Rate）、过分割OR（Over-segmentation Rate）和MIoU（Mean Intersection over Union）四项指标进行了实验对比分析，结果表明本文方法能够有效抑制喉道过分割，在分割喉道时具有更高的准确度。

Abstract

Analyzing the characteristics of channels in rocks is important for assessing reservoir quality. Traditional methods for segmenting channels in rock thin section images mainly rely on binary images of pores for channel extraction. However

these methods tend to produce over-segmentation and low segmentation accuracy when analyzing complex pore structures

leading to poor generalization performance of the algorithm. To address these issues

this paper proposes a deep learning-based channel segmentation algorithm. The algorithm uses UNet++ as the backbone network for feature extraction. It incorporates CBAM (Convolutional Block Attention Module) attention modules during downsampling to enhance the model's ability to capture correlations between features of different dimensions and reduce semantic differences in skip connections. This improves the accuracy of channel feature extraction. Experimental comparisons of the proposed algorithm and other existing channel segmentation algorithms are conducted based on four metrics: CR (Correctness Rate)

UR (Under-segmentation Rate)

OR (Over-segmentation Rate)

and MIoU (Mean Intersection over Union). The results show that the proposed method can effectively suppress channel over-segmentation and achieve higher accuracy in channel segmentation.

关键词

铸体薄片图像喉道分割UNet++CBAM跳跃连接

Keywords

cast thin section imagesthroat segmentationUNet++CBAMskip connections

references

肖渊甫, 郑荣才, 邓江红. 岩石学简明教程[M]. 第3版. 北京: 地质出版社, 2009.

代贺. 岩矿全薄片图像自动分割算法研究[D]. 成都: 四川大学, 2017.

蔡宇恒. 铸体薄片图像孔隙及粒度分析相关算法研究[D]. 成都: 四川大学, 2021.

SILIN D, PATZEK T. Pore space morphology analysis using maximal inscribed spheres[J]. Physica A Statistical Mechanics and Its Applications, 2006, 371(2): 336-360.

刘凯文. 基于深度学习的岩石铸体薄片图像分析[D]. 成都: 四川大学, 2022.

LIU Z, LIN Y T, CAO Y, et al. Swin Transformer: Hierarchical vision Transformer using shifted windows[C]//2021 IEEE/CVF International Conference on Computer Vision (ICCV). Piscataway: IEEE, 2021: 10012-10022.

RONNEBERGER O, FISCHER P, BROX T. U-Net: Convolutional networks for biomedical image segmentation[M]//Lecture Notes in Computer Science. Cham: Springer International Publishing, 2015: 234-241.

ZHOU Z W, RAHMAN SIDDIQUEE M M, TAJBAKHSH N, et al. UNet++: A nested U-Net architecture for medical image segmentation[M]//Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support. Cham: Springer International Publishing, 2018: 3-11.

VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. New York: ACM, 2017: 6000-6010.

WU H P, XIAO B, CODELLA N, et al. CvT: Introducing convolutions to vision Transformers[C]//2021 IEEE/CVF International Conference on Computer Vision (ICCV). Piscataway: IEEE, 2021: 22-31.

YUAN K, GUO S P, LIU Z W, et al. Incorporating convolution designs into visual Transformers[C]//2021 IEEE/CVF International Conference on Computer Vision (ICCV). Piscataway: IEEE, 2021: 579-588.

WOO S, PARK J, LEE J Y, et al. CBAM: Convolutional block attention module[C]//Computer Vision – ECCV 2018. Cham: Springer International Publishing, 2018: 3-19.

WEI J, WANG S H, HUANG Q M. F³Net: Fusion, feedback and focus for salient object detection[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2020, 34(7): 12321-12328.

邓亮. 微观大模型驱替图像分析与处理研究[D]. 成都: 四川大学, 2023.

ZHAO H S, SHI J P, QI X J, et al. Pyramid scene parsing network[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway: IEEE, 2017: 2881-2890.

WANG S H, LI L, ZHUANG X H. AttU-NET: Attention U-Net for brain tumor segmentation[M]//Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries. Cham: Springer International Publishing, 2022: 302-311.

WANG H N, CAO P, WANG J Q, et al. UCTransNet: Rethinking the skip connections in U-Net from a channel-wise perspective with Transformer[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2022, 36(3): 2441-2449.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于深度学习的微观驱替图像分类