| Sumario: | "Accurate land use classification plays a critical role in agricultural monitoring, resource management, and policy planning. Remote sensing, particularly the use of high-resolution multispectral imagery, has emerged as a powerful tool for mapping and assessing agricultural production systems with enhanced precision. In Cambodia, where rice farming dominates the landscape, understanding spatial variations in land use and cropping patterns is essential for improving agricultural productivity and sustainability.
This study aims to classify land use and assess agricultural production systems in selected sites in Takeo and Prey Veng provinces, Cambodia, using high-resolution satellite imagery from Pleiades (0.5 m) and SPOT 7 (1.5 m). By integrating satellite-derived data with field-based validation techniques, this study seeks to improve classification accuracy and enhance our understanding of land use dynamics in these regions.
The study employs Object-Based Image Analysis (OBIA) and a Support Vector Machine (SVM) classifier within the Orfeo Toolbox (OTB) in QGIS. This approach leverages spectral, textural, and spatial attributes to enhance classification accuracy while minimizing misclassification errors commonly associated with pixel-based methods. The classification results are further validated using ground truth data collected through field surveys and supplementary sources such as Google Earth and the RIICE project’s rice area maps.
The findings provide insights into the spatial distribution of key land cover types, including rice fields, fallow croplands, built-up areas, and tree cover. Additionally, the study highlights challenges in differentiating specific land use classes due to spectral similarities and seasonal variations. The results contribute to improved land use planning and decision-making for agricultural development in Cambodia."
|
|---|