<![CDATA[Rice data is essential for policymakers in designing food security strategies in Indonesia. Currently, harvested area estimates are produced monthly using the Area Sampling Frame (ASF), although this method faces operational and cost-related limitations. Satellite Imagery Time Series (SITS) data, particularly from Sentinel-1, offers an alternative for identifying rice growth stages through machine learning-based modelling. This study applies that approach in South Sulawesi Province to estimate harvested rice area. The workflow includes regional clustering, satellite data integration, preprocessing, growth stage modelling using XGBoost, and phase area estimation. The results show that most clusters achieved high classification accuracy. Moreover, the predicted harvest area patterns closely match those from the ASF method. These findings demonstrate that using SITS data combined with machine learning offers an effective and practical alternative, especially in modernizing agricultural statistics systems in major rice-producing regions like South Sulawesi.]]>
Copyrights © 2025
