Super-resolution Reconstruction of Rock Thin-Section Image based on SinGAN

Abstract

The rock thin section images are of great significance to the study of petroleum geological characteristics and oil and gas exploration. Due to the limitations of various factors, the obtained images of rock thin-sections often have low resolution, which limits the researchers’ grasp of their detailed information to some extent. The traditional super-resolution algorithm of neural network will need a large amount of data as a training set, in order to improve rock thin-section image super-resolution reconstruction algorithm texture detail information reduction ability, the paper using single image generative adversarial network for rock thin-section image super-resolution reconstruction, does not need to input a large number of data sets, of single image super-resolution reconstruction image. Rock cast thin section images from an oil field area in Ordos basin were used for training, and peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) evaluation indexes were used for model evaluation. The experimental results show that the super-resolution image processing based on this method has good visual effects and evaluation indexes.