Case Study: Global Forest Watch improves deforestation driver classification with Ellipsis Drive

Deep Learning for Deforestation Alerts

Global Forest Watch (GFW) needed to identify the specific drivers behind their near-real-time deforestation alerts, as the alerts only detected tree loss without indicating its cause. Partnering with Ellipsis Drive and another firm, GFW sought to develop a deep learning model to classify these drivers, a critical step in prioritizing alerts for conservation action. The challenge included working with satellite imagery often obscured by cloud cover.

Ellipsis Drive provided a platform for the partners to visualize spatial data, comment on specific locations, and facilitate communication throughout the model's iterative development. This collaboration resulted in a well-performing model capable of predicting drivers like selective logging and small-scale agriculture. Ellipsis Drive's solution was vital for improving the model and contributed to an initiative with the potential to help forest monitors better focus their resources.

Global Forest Watch (GFW)