Overview

The AI Segmentation plugin lets forestry professionals and environmental researchers quickly create vector polygons for forest monitoring, deforestation tracking, and ecosystem assessment.

Common Applications

Tree Detection and Counting

Segment individual tree canopies from drone orthomosaics or high-resolution satellite imagery. This is useful for:

Forest inventory (counting trees in a stand)
Estimating canopy coverage
Monitoring replanting efforts

For individual tree detection, zoom in so each canopy is clearly visible. The AI performs best when the object fills a reasonable portion of the view.

Dead Tree and Infestation Mapping

Identify dead, dying, or infested trees by visual characteristics (discoloration, missing foliage, bark beetle damage) and segment them to map affected areas for management planning.

Create polygons around tree groups or forest stands to map canopy types or species distribution. Combined with field data, these polygons can serve as training data for broader species classification models.

Fire Scar and Burn Area Delineation

Segment burn scars from satellite or aerial imagery after a wildfire to assess damage extent and delineate burned vs. unburned areas.

Typical Data Sources

Source	Resolution	Best For
Drone orthomosaics (RGB)	2-10 cm	Individual trees, detailed inventory
National orthophotos	20-50 cm	Landscape-level mapping
Satellite imagery (Sentinel-2, Landsat)	10-30 m	Large-scale forest monitoring
Commercial satellite (Pleiades, WorldView)	30-50 cm	Detailed analysis without drones

Workflow Example: Forest Inventory

Load your drone orthomosaic in QGIS
Start AI Segmentation and navigate to a forest stand
Click on individual tree canopies to segment them
Use Simplify to clean up the outlines for a smooth canopy boundary
Save each tree polygon and continue through the stand
Export the layer with all tree polygons
Use the Field Calculator to compute canopy area for each tree
Add species labels in the attribute table if known

Forestry & Environmental Monitoring