Accurate quantification of multiphase fluid saturations in reservoir rocks remains a central challenge in petrophysics, with direct implications for hydrocarbon recovery optimization. Here, we present an integrated methodology that combines high-field nuclear magnetic resonance (NMR) imaging with relaxation-based contrast to quantify oil and water distributions in synthetic porous media. Controlled porous samples were fabricated by sintering soda–lime glass microspheres with defined grain sizes (75–125 \(\upmu {\text {m}}\) , 150–212 \(\upmu {\text {m}}\) , 250–300 \(\upmu {\text {m}}\) ) and saturated with known oil–water proportions. Chemical shift contrast was employed to distinguish fluid phases, while transverse relaxation ( \(T_2\) ) measurements revealed distinct signal decay behaviors that enabled robust grayscale differentiation in high-field NMR images. A custom image-analysis software was developed to extract intensity histograms from \(T_2\) -weighted images, allowing pixel-level quantification of each fluid phase. Across all tested samples, the methodology achieved an average quantification error below 10%, demonstrating its robustness and potential as a non-invasive approach for fluid characterization in reservoir analogs, with relevance to reservoir characterization, enhanced oil recovery, and environmental applications.
