The Earth Observation-based Anomaly Detection (EOAD) system: A simple, scalable approach to mapping in-field and farm-scale anomalies using widely available satellite imagery

To feed the world increasing population, expansion in area under arable cultivation is expected, with majority projected to occur Sub-Sahara Africa and Latin American countries. However, many existing Precision Agriculture (PA) techniques are difficult transfer agricultural systems these regions as they rely on prohibitively expensive crop monitoring systems. Satellite Earth Observation (EO) has ability provide affordable solutions, particularly identify yield-limiting conditions within site-specific management zones (MZs). This paper presents Observation-based Anomaly Detection (EOAD) approach, a novel system for detection of in-field anomalies through automatic thresholding optical Vegetation Index data, based their deviation from normal distribution. The EOAD sets dynamic thresholds pixel values parcel by removing atypical increments tails towards median until distribution normal. normality assessed upon measures skewness kurtosis each iteration. anomaly approach demonstrated strong agreement, 80% overall accuracy, identified when applied rice plots Ibague Plateau, Colombia, using both Sentinel-2 PlanetScope imagery. Areas anomalous during booting stage were shown be significantly (p ⩽0.005) associated decrease final yield. Additionally, percentage detected improved underperforming early growth stages. Using freely available data software, this automated demonstrates an exciting potential use improving practices low-resource regions.