Deforestation detection: Datasets & Models
Segment cleared, degraded, and intact forest areas in aerial and satellite imagery to track deforestation over time. Turns raw drone or geospatial captures into pixel-level maps of forest loss.
Deforestation detection
This workflow runs segmentation on aerial, drone, and satellite imagery to produce per-pixel masks of forest cover and forest loss. Trained on a deforestation dataset, it separates intact canopy from cleared land, degraded forest, logging roads, and burn scars so the results can be measured and compared across dates.
Try it out
Powered by general detection model
Run on custom image
Drop an image here or click to upload
Why automate deforestation detection
Manually reviewing thousands of aerial tiles or satellite scenes to flag forest clearing is slow, inconsistent between analysts, and impossible to keep current across large concessions or protected areas. In the drones and geospatial context, automating segmentation lets you process new captures as soon as they land and quantify canopy loss in hectares without hand-digitizing every polygon. A trained model applies the same criteria to every image, so change between two survey dates reflects real ground conditions rather than differences in who did the labeling. This makes it practical to run recurring monitoring over regions that would be cost-prohibitive to inspect by hand.
Key benefits
Instance segmentation gives you precise per-area masks that convert directly to measurable cleared hectares and support time-series change detection. The workflow plugs into the drone and satellite imagery you already collect and integrates with standard GIS and geospatial tooling, so you can go from raw capture to forest-loss report without manual tracing. As your image volume grows, the same model scales across new regions and survey dates, and you can retrain on your own deforestation dataset to match local forest types.
How to use it
- Upload an image: Drop any photo into the widget above to try deforestation detection for free, right on this page.
- Preview instant results: The model detects cleared land, forest canopy, logging road, burn scar, bare soil within seconds and returns an annotated image with structured predictions, so you can immediately see how well it works. Adjust the detection classes and re-run until the results fit your scenario.
- Like what you see?: With a free Roboflow account, you can deploy this workflow with the Serverless Cloud API, customize it in your workspace, or train RF-DETR (instance segmentation) on your own images for production accuracy.
- Get started for free: No code or training data required to begin: upload, review the results, and make it part of your stack in a single click.
Recommended model for deforestation detection
RF-DETR is state of the art for real-time object detection, instance segmentation, and keypoint detection, with higher accuracy and lower latency than comparable models, and it's Apache 2.0 licensed. When you're ready to build a production model for deforestation detection, fine-tune it on your own labeled images, then deploy to a Serverless Cloud API endpoint.
Not sure where to start?
Describe your Deforestation Detection project to Roboflow Agent.