Abstract:Digital Rock Imaging is constrained by detector hardware, and a trade-off between the image field of view (FOV) and the image resolution must be made. This can be compensated for with super resolution (SR) techniques that take a wide FOV, low resolution (LR) image, and super resolve a high resolution (HR), high FOV image. The Enhanced Deep Super Resolution Generative Adversarial Network (EDSRGAN) is trained on the Deep Learning Digital Rock Super Resolution Dataset, a diverse compilation 12000 of raw and processed uCT images. The network shows comparable performance of 50% to 70% reduction in relative error over bicubic interpolation. GAN performance in recovering texture shows superior visual similarity compared to SRCNN and other methods. Difference maps indicate that the SRCNN section of the SRGAN network recovers large scale edge (grain boundaries) features while the GAN network regenerates perceptually indistinguishable high frequency texture. Network performance is generalised with augmentation, showing high adaptability to noise and blur. HR images are fed into the network, generating HR-SR images to extrapolate network performance to sub-resolution features present in the HR images themselves. Results show that under-resolution features such as dissolved minerals and thin fractures are regenerated despite the network operating outside of trained specifications. Comparison with Scanning Electron Microscope images shows details are consistent with the underlying geometry of the sample. Recovery of textures benefits the characterisation of digital rocks with a high proportion of under-resolution micro-porous features, such as carbonate and coal samples. Images that are normally constrained by the mineralogy of the rock (coal), by fast transient imaging (waterflooding), or by the energy of the source (microporosity), can be super resolved accurately for further analysis downstream.

Enhancing digital core image resolution using optimal upscaling algorithm: with application to paired SEM images

3D resoution enhancement in saturated competition microscopy

Sub-60-nm 3D super-resolution imaging via saturated I5S

Counter‐Propagating Evanescent Illumination Super‐Resolution Chip

Resolution Enhancement of Scanning Electron Micrographs using Artificial Intelligence

Digital Rock Core Images Resolution Enhancement with Improved Super Resolution Convolutional Neural Networks

A comparative study of paired versus unpaired deep learning methods for physically enhancing digital rock image resolution

Super Resolution Convolutional Neural Network Models for Enhancing Resolution of Rock Micro-CT Images

Boosting Resolution and Recovering Texture of micro-CT Images with Deep Learning

Remote Sensing Single-Image Resolution Improvement Using A Deep Gradient-Aware Network with Image-Specific Enhancement

SRFS:Super Resolution in Fast-Slow Scanning Mode of SEM Based on Denoising Diffusion Probability Model

Applicability of 2D algorithms for 3D characterization in digital rocks physics: an example of a machine learning-based super resolution image generation

Digital Rock Core Images Super Resolution via SRCNN Based on Accelerated Bicubic Interpolation

Image Super Resolution for Scanning Tunneling Microscopy and Atomic Force Microscopy

SRFS: Super Resolution in Fast-Slow Scanning Mode of SEM Based on Denoising Diffusion Probability Model

Multiple-image super-resolution of cryo-electron micrographs based on deep internal learning

Resolution Enhancement of Polar Ice Core Micro CT Scans via Deep Learning: A Comparative Study

Deep Learning Enables Large Depth-of-Field Images for Sub-Diffraction-Limit Scanning Superlens Microscopy

Resolution enhancement in scanning electron microscopy using deep learning

Automated high-resolution backscattered-electron imaging at macroscopic scale

Cryo-ZSSR: multiple-image super-resolution based on deep internal learning