Geospatial Artificial Intelligence (GeoAI) and Satellite Imagery Fusion for Soil Physical Property Predicting

This study aims to predict vital soil physical properties, including clay, sand, and silt, which are essential for agricultural management environmental protection. Precision distribution of texture is crucial effective land resource precision agriculture. To achieve this, we propose an innovative approach that combines Geospatial Artificial Intelligence (GeoAI) with the fusion satellite imagery properties. We collected 317 samples from Iran’s Golestan province dependent data. The independent dataset encompasses 14 parameters Landsat-8 images, seven topographic Shuttle Radar Topography Mission (SRTM) DEM, two meteorological parameters. Using Random Forest (RF) algorithm, conducted feature importance analysis. employed a Convolutional Neural Network (CNN), RF, our hybrid CNN-RF model comparing their performance various metrics. network strengths CNN networks RF algorithm improved prediction. demonstrated superior across metrics, excelling in predicting sand (MSE: 0.00003%, RMSE: 0.006%), silt 0.00004%, clay 0.00005%, 0.007%). Moreover, exhibited (R2: 0.995), 0.992), 0.987), as indicated by R2 index. identified MRVBF, LST, B7 most influential prediction, respectively, underscoring significance remote sensing, topography, climate. Our integrated GeoAI-satellite provides valuable tools monitoring degradation, optimizing irrigation, assessing quality. methodology has significant potential advance agriculture practices.