Forest mapping in Peninsular Malaysia using Random Forest and Support Vector Machine Classifiers on Google Earth Engine

Abstract

Forests play a crucial role in maintaining the balance of the global ecosystem by sustaining the interactions between living and non-living entities. Changes in forest areas encompass both growth and loss, often driven by development activities. Assessing forest cover and its changes is also a pivotal issue in forest management. Therefore, this study aims to investigate the performance of machine learning algorithms, namely Random Forest (RF) and Support Vector Machine (SVM), in mapping forest cover in Peninsular Malaysia. Landsat 5 TM and Landsat 8 OLI images were utilized to derive forest cover information. The classification process was automated using the remote sensing data management platform, Google Earth Engine (GEE). The accuracy assessment test using the Kappa coefficient resulted in a value of 0.7893 for the RF algorithm and 0.6328 for the SVM algorithm for the year 2010. Whereas, for the year 2020 the Kappa coefficient yielded 0.7475 for RF and 0.5893 for SVM. However, forest cover returned highest RF Kappa coefficient values of 0.875 (2010) and 0.8793 (2020), and SVM Kappa coefficient values of 0.8116 (2010) and 0.7313 (2020). The results implied that RF performed better in the land use classification compared to SVM. It is evident that this study can aid various stakeholders in assessing future plans and developments without compromising the environment.