Hyperspectral data as a proxy for porosity estimation of carbonate rocks

Rock porosity is one of the most significant parameters in fluid-flow simulation context carbonate reservoirs. The hydrocarbon industry uses to assess production potential oil and gas environments. Traditional methods determine are limited discrete measurements generally demand extra resources associated with careful analysis logs, rock sampling laboratory analysis. This paper investigates an alternative estimate rocks using pointwise hyperspectral data machine learning. method contiguous, does not require was validated various plug samples collected from two distinct outcrops. were analysed ground-truth values for effective reflectance visible infrared regions spectra. supervised regression applied able a robust relationship between spectral behaviour characteristic features carbonate, hydroxyl, molecular water Fe/Mn. results obtained here suggest soundness indirect approach models trained achieving coefficient determination above 0.8 mean absolute deviation less than 2%.KEY POINTSHyperspectral can be used as proxy estimation rocks.All tested learners achieved R2 greater 0.7.Regularised linear porosity.Support vector achieves error 1.0249 our dataset.