Multiscale modeling of shale samples based on low- and high-resolution images

Accurate modeling of shale samples is very crucial for evaluating and predicting the physical properties. However, obtaining large-scale and high-resolution images of shales using a single imaging tool such as X-ray computed tomography or scanning electron microscopy machine might not be plausible as there is always a trade-off between the resolution and field-of-view. On the other hand, relying on small-scale images may jeopardize the accuracy of porous media modeling. One solution scheme for addressing these issues is to fuse the spatial information from the multiscale and multiresolution images into a multiscale high-resolution image. As such, a novel multiscale modeling algorithm based on template matching for complex rocks is proposed in this study. This algorithm can fuse multiscale and multiresolution structural patterns from the images with different scales, resolution, and phases into one single image. The accuracy of the presented algorithm is tested using two heterogeneous shale samples and the results are compared with the actual large and high-resolution images. We found that the algorithm can generate an accurate image after comparing the geometrical properties of pore systems, correlation functions, and image similarity indicators of the generated image and reference image.