Reservoir pore structure is an important factor to control reservoir physical properties and productivity, and also an important basis for reservoir characterization. Core slice image is an important method of pore structure analysis. At present, the analysis of core slice image mainly depends on the traditional image processing technology, and uses expert experience as an assistant. This method can only process a single image one by one. It has a long analysis period, a slow speed, and no effective use of the historical data that has been analyzed. In this paper, a large number of core slice images are collected, and the pore structure analysis model is constructed by Conditional GAN network. When a core slice image is input into the model, it can automatically give the corresponding pore structure distribution map. In this way, the intelligent analysis of reservoir pore structure is realized by the method of deep learning. At the same time, we put forward C-IoU, C-MSE evaluation indexes, which effectively solve the difficulty of model evaluation in the application of Conditional GAN.
Presentation Date: Monday, October 12, 2020
Session Start Time: 1:50 PM
Presentation Time: 2:15 PM
Location: Poster Station 1
Presentation Type: Poster
