Dataset for Different Methods of Permeability Calculation in Thin-Pore Tight Sandstones
Description
The micro x-ray computed tomography images of five clastic tight sandstone samples from one Russian oilfield. General Electric v|tome|x L240 CT system was used to obtain images with spatial resolution 1.2 μm/vox. CT images have size 1400 vox. in each direction. The Manual DIP, Cross-Laboratory Control DIP, and Automated DIP as three different methodologies have been applied to create 3D digital rock models. The number of voxels in the final model, consistency, and consequence of image processing and binarization techniques are the main differences.
Files
Steps to reproduce
This research is concerned with the influence of various Digital Image Processing (DIP) procedures on the calculations of favorite parameters. The effects of taking a local or global threshold for segmentation, the cubic size of the digital rock model, and implemented types of filters to increase the quality of images have been taken into account to investigate how far they bring about changes of permeability. 1) The 3D models constructed with the help of Manual DIP on the personal computers are computationally less expensive and much faster. The small size of the models (cropping of homogenous part of the images), simple mean filter for denoising and global thresholding OTSU for binarization were used to build the models. 2) Automated DIP allows forming more precise 3D models although it is time-consuming, and its implementation requires having access to the architecture of a High-Performance Computing (HPC) unit. The largest size of the models, the combination of "heavy" band-pass (Fourier transformation) and bilateral filters for denoising and local thresholding Random walker algorithm for binarization were used to build the models. 3) Down to the details of the Cross-Laboratory Control (Cross-lab) DIP, the implemented algorithm utilizes the spatial covariance of the image in conjunction with indicator kriging to determine object edges. The use of indicator kriging makes the thresholding local and guarantees smoothness in the threshold surface.