Investigation and Monitoring Deforestation by Evaluating the Satellite Images and Machine Learning

A woodland can be restored by planting new trees, caring for existing ones, and conserving natural resources. The effects of climate change could be mitigated with the help of new financing for woodland rehabilitation initiatives around the globe. Current poll methods for rehabilitation, while useful, could be improved so that such vast efforts can be tracked more effectively. Using satellite data, a machine-learning algorithm was developed to monitor the trees and plants in the forest area. The purpose of this paper is to investigate the viability of monitoring replanting initiatives using machine learning and high-resolution satellite images. The purpose of this research is to create an innovative but challenging collection using high-resolution satellite images. We discuss the training and testing of a random forest machine learning model on this one-of-a-kind dataset, which was collected from a location in northern Fiji where a native forest was being restored. Testing the proposed RF-Ranger model on a dataset, it was found to be highly accurate, with a categorization accuracy of 97.5% and a Macro-F1 score of 97.4% on unobserved pixels. Conservation International (CI), a nonprofit environmental group studied in cooperation for this study, may oversee the current replanting location assessed here. The proposed model, RF-Ranger, was able to correctly recognize plants in forest regeneration areas and therefore has the potential to be used for large-scale monitoring of such efforts.