基于GAN的岩石CT图像超分辨算法  

Super Resolution Algorithm of Rock CT image Based on GAN

作　　者：朱联祥 仝文东 裴圣武 ZHU Lianxiang;TONG Wendong;PEI Shengwu(College of Computer Science,Xi'an Shiyou University,Xi'an Shaanxi 710065,China)

机构地区：[1]西安石油大学计算机学院,陕西西安710065

出　　处：《信息与电脑》2022年第22期112-114,共3页Information & Computer

摘　　要：岩石电子计算机断层扫描(Computed Tomography,CT)图像可使岩石内部结构可视化,故可以作为评估油气储藏量的依据。但是,受采集条件的约束,岩石CT图像往往分辨率低、细节较模糊。因此,本文基于生成对抗网络(Generative Adversarial Network,GAN)方法引入岩石孔隙度损失,在原有损失函数约束的基础上增加新的约束项,并进行4倍超分辨重构实验,通过峰值信噪比(Peak Signal to Noise Ratio,PSNR)和结构相似性(Structural Similarity Index,SSIM)的指标比较重建结果。结果表明,本文方法在指标和视觉上均取得了良好效果。The Computed Tomography(CT)image of rock can make the internal structure of rock visible,so it can be used as a basis for evaluating oil and gas reserves.However,due to the constraints of acquisition conditions,rock CT images often have low resolution and fuzzy details.Therefore,based on the Generative Adversarial Network(GAN)method,this paper introduces the loss of rock porosity and adds new constraints on the basis of the original loss function constraints.The 4-fold super-resolution reconstruction experiment was carried out,and the reconstruction results were compared by the indexes of Peak Signal to Noise Ratio(PSNR)and Structural Similarity Index(SSIM).The results show that the method in this paper has achieved good results in terms of indicators and vision.

关 键 词：深度学习 生成对抗网络(GAN) 超分辨率重构 岩石电子计算机断层扫描(CT)图像 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]