APPLICATION OF OPTICAL REMOTE SENSING IMAGES TO MONITORING RICE–GROWING AREAS IN KIEN GIANG PROVINCE

Abstract

This study used multi–temporal Sentinel–2 images (10×10m) to map rice cultivation areas and cropping patterns in Kien Giang province for 2024. Sentinel–2 data were processed to generate maximum NDVI composites for each rice season, which were then used as input for a Decision Tree classification in ENVI. The results identified three major rice cropping systems: one rice crop (106,465ha, 28%), two rice crops (124,635ha, 33%), and three rice crops (148,362ha, 39%). Validation using 300 ground–truth samples achieved an overall accuracy of 86% and a Kappa coefficient of 0.82, demonstrating the classifier’s strong ability to distinguish among different rice cropping systems. These findings highlight the effectiveness of integrating Sentinel–2 multi–temporal data with Decision Tree classification for monitoring rice seasonality. Moreover, the study provides valuable scientific evidence to support agricultural management, land–use monitoring, and sustainable policy–making in the Mekong Delta.