The advent of hyperspectral remote sensing represented a breakthrough in the accurate, fast, and non-invasive estimation of important soil fertility parameters. The organic content in the soil acts as a strong indicator of soil fertility, which influences agricultural production and the global carbon cycle. The present study utilises non-imaging hyperspectral data in the spectral range of 350-2500nm collected proximally using an ASD FieldSpec spectroradiometer for estimating the soil organic carbon (SOC) content of a research farm in ICAR-Indian Agricultural Research Institute, New Delhi, India. The partial least squares regression (PLSR) scores were used as the independent variables for evaluating three multivariate regression models, such as support vector machine (SVM), random forest, and partial least squares regression, to estimate SOC. After pre-processing, the proximal spectral values were spatially interpolated using the ordinary kriging technique to construct a synthetic hyperspectral image of the experimental fields. The SVM outperformed other models, achieving an R² value of 0.83, which suggests an accurate prediction of SOC. On applying the regression model to this synthetic hyperspectral imagery, a high-resolution SOC map was generated. Our study indicated the potential of non-imaging proximal hyperspectral data for generating a high-resolution map showing the variability of organic content in the soil.
Previous Article in event
Previous Article in session
Next Article in event
Generation of Synthetic Hyperspectral Image Cube for Mapping Soil Organic Carbon using Proximal Remote Sensing
Published:
02 September 2025
by MDPI
in The 2nd International Electronic Conference on Land
session Soil Carbon Sequestration and Climate Change Mitigation
Abstract:
Keywords: Soil Organic Carbon; proximal; hyperspectral remote sensing; machine learning regression
