Supervised land cover classification of Nueva Ecija using random forest in google earth

Bensi Papa Leonylyn *, Michael Enrile Bensi and Apple Grace Galang Oliveros

College of Information and Communications Technology, Nueva Ecija University of Science and Technology, Cabanatuan City, Nueva Ecija, Philippines.
 
Research Article
Global Journal of Engineering and Technology Advances, 2024, 21(01), 115–118.
Article DOI: 10.30574/gjeta.2024.21.1.0186
Publication history: 
Received on 02 September 2024; revised on 12 October 2024; accepted on 15 October 2024
 
Abstract: 
Land cover classification is essential for environmental monitoring, urban planning, and sustainable land use management. This study presents a supervised land cover classification in Nueva Ecija, Philippines, utilizing Google Earth Engine (GEE) and the Random Forest (RF) algorithm, applied to Landsat 8 imagery. A total of 1,523 samples were collected representing five land cover types: built-up areas, agricultural lands, water bodies, forests, and barren land. The classification achieved an overall accuracy of 87.69% with a Kappa coefficient of 0.841. Future work should explore the integration of seasonal imagery and topographic indices for improved performance. This methodology provides crucial insights for resource management and supports regional policy development in Nueva Ecija.
 
Keywords: 
Accuracy Assessment; Google Earth Engine; Land Cover Classification; Landsat 8; Nueva Ecija; Remote Sensing; Random Forest
 
Full text article in PDF: