Digital Rock Core Image Super-Resolution via Deep Convolutional Neural Networks
Yunfeng Bai, Vladimir V. Berezovsky, Jing Ming Guo, В. Н. Попов · 2020
Digital Rock is a research hotspot in the field of oil and gas. In practice, oil and gas departments hope to get clearer images of rock core which are helpful for them to understand characteristics of reservoir rock better. And based on that they can develop a better mining strategy. Through literature research, we found that in recent years, deep learning algorithms have achieved great success in processing images, thus we introduced them into the field of processing core images. At the same time, we noticed that core images generally have colourful blocks of different sizes. Based on this feature, we did not directly apply the classic algorithm, but made some improvements. Experiments show that the effect of our proposed method is equivalent to that of the classic algorithms, however, in the pre-processing stage, it saves time significantly.