Innovative Research Publication International Journal of Innovative Research in Computer Science and Technology 2347-5552 Volume 14 Issue 5 English Sep-Oct - February 2026 N N Y 2026 10 27 Computer Science Satellite Image Based Agricultural Monitoring Using Machine Learning and Explainable AI English Y 83 96 Nandhini B English Y Nithyanandh S English N Thara L English N Mohanapriya D English N https://doi.org/10.55524/ijircst.2026.14.5.10 Approaches for agriculture monitoring should consistently assess vegetation conditions throughout farm fields and limit the need for manual surveys. Satellite-derived information offers one source of spatially-explicit data that can be used to monitor crops using their spectral reflectance properties. Thresholding vegetation indices to classify vegetation conditions may be difficult to interpret across cropland with variable canopy structure, phenology, and reflectance. Here, we introduce a crop monitoring framework which integrates Sentinel-2 multitemporal data, vegetation indices, machine learning, outlier detection, and explainable AI. Sentinel-2 multispectral (MS) and RGB data streams are handled with separate processing components. Vegetation indices including Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), Soil-Adjusted Vegetation Index (SAVI), Normalized Difference Water Index (NDWI), and Excess Green Index (ExG) are computed from spectral bands when applicable. Crop condition class labels are predicted with a Random Forest classification model. Thresholding on crop-specific NDVI values is included for comparison as a rules-based method. An Isolation Forest algorithm detects anomalous vegetation conditions. Finally, SHapley Additive exPlanations (SHAP) are used to interpret ML model predictions. Six crop types are included in the framework: cotton, sugarcane, paddy, carrot, wheat, and barley. An initial test of the framework is done using a single Sentinel-2 MS image and five historical records yielded a NDVI value of 0.75 for the satellite image along with a classification of Healthy, 99% model certainty, and 0.0% of pixels identified as stressed vegetation. English Crop Health Monitoring, Explainable AI Remote Sensing, Precision Agriculture, NDVI, Sentinal-2, Random Forest, SHAP, Vegetation Indices https://ijircst.org/view_abstract.php?title=Satellite Image Based Agricultural Monitoring Using Machine Learning and Explainable AI&year=2026&vol=14&primary=QVJULTE0ODU=