Over the past few decades, the amount of carbon dioxide has increased in the atmosphere. Carbon dioxide is the major greenhouse gas contributing to climate change. Because of the vital role that plants play in the carbon cycle, forests can act as a sink to sequester and store carbon. Due mainly to anthropogenic activities, much of the world's forest coverage has been lost. Over the past few decades, though, conservation efforts have been undertaken to restore what has been lost and to mitigate climate change. Therefore, monitoring, verifying, and reporting of carbon offsets are crucial for accountability and transparency. Current machine learning models often overpredict the carbon stock. The goal of this project is to explore the potential of tree species classification and carbon stock estimation using a state-of-the-art dataset, ReforesTree, that combines hand-measured field data and low-cost drone imagery of agro-forestry sites in Ecuador.