Stable Physical Property Estimation Across Resolutions for Digital Rock Utilizing Texture and Intensity Features

Daren Zhang,Gang Ma,Jia Zhang,Zhuang Cheng,Xiaolin Chang,Wei Zhou
DOI: https://doi.org/10.1007/s00603-024-04304-8
IF: 6.518
2024-12-12
Rock Mechanics and Rock Engineering
Abstract:The physical properties of rock can be obtained through various methods, such as the non-destructive X-ray computed tomography (X-ray CT). The urgency for a stable property estimation method using X-ray CT images becomes increasingly important with the development of large-scale engineering applications and cross-resolution image fusion technologies. This study introduces an innovative image processing method utilizing texture and intensity features for digital sandstone samples, involving an improved particle swarm optimization algorithm for image enhancement and a random forest method for segmentation. The proposed method allows for stable physical property estimation across resolutions from 2.25 to 21.6 μm by enhancing the key structural areas to reduce the randomness inherent in the random forest method. The efficacy of this method is demonstrated through its application to three different sandstone samples, showing consistent results in both macroscopic and microscopic property characterization. Cross-resolution studies reveal that changes in the number of isolated and circle-shape pores due to resolution variations critically affect the porosity and permeability of sandstones. This research advances resolution-robust analysis in digital rock physics and offers a valuable tool for future explorations.
geosciences, multidisciplinary,engineering, geological
What problem does this paper attempt to address?