Detecting Soil pH from Open-Source Remote Sensing Data: A Case Study of Angul and Balangir Districts, Odisha State

Soil sampling, collection, and analysis are a costly labor-intensive activity that cannot cover the entire farmlands; hence, it was conceived to use high-speed open-source platforms like Google Earth Engine in this research estimate soil characteristics remotely using high-resolution satellite data. The objective of pH from Sentinel-1, Sentinel-2, Landsat-8 satellite-derived indices; data missions were used generate indices as proxies statistical model pH. Step-wise multiple regression (SWMR), artificial neural networks (ANN), random forest (RF) develop predictive models for pH, SWMR, ANN, RF models. SWMR greedy method variable selection select appropriate independent variables highly correlated with Variables retained B2, B11, Brightness index, Salinity index 2, 5 Sentinel-2 data; VH/VV Sentinel 1 TIR1 (thermal infrared band1) p-value < 0.05. Among four developed, class-wise performed better than other cumulative correlation coefficient 0.87 RMSE 0.35. performance can be attributed different spectral groups. More 70% soils Angul Balangir districts acidic soils, therefore, training dataset affected by leading misclassification neutral alkaline hindering single class Our results showed bands individual classes acidic, neutral, soils. This study has shown potential big analytics predict accurate mapping help decision support.