Tropical Landcover Classification with focus on palm oil plantations

This project presents land use classification maps obtained through satellite image analyses for 10 study areas in Mesoamerica. A classification algorithm was adapted specifically for the tropics and developed using the latest satellite images available. We applied Object Object-Based Image Analysis method & Random Forest algorithm on 10m Sentinel 1 (S1) & Sentinel 2 (S2) image collections of the years 2022/2023. The study area encompassed 10 landscapes spanning 8 countries across Mesoamerica. These areas cover a total of 256,971 km² and were previously defined by Osa Conservation as potential climate adaptation hubs.

Final results are available in Zenodo: https://zenodo.org/badge/DOI/10.5281/zenodo.13329296.svg

Paper:

Integrating high-resolution remote sensing and empirical wildlife detection data for climate-resilient corridors across tropical elevational gradients