Timber Survey & Tree Inventory

Species Breakdown

Species	Count

Data Source & Methodology

Individual trees were detected and classified by a vision model over four co-registered rasters derived from the drone flight (DJI Matrice 4TD): the orthophoto (crown/species), DSM (canopy surface), DTM (bare ground), and a canopy-height model (DSM−DTM) standing in for the 3D point cloud. Detections are georeferenced by converting pixel coordinates through the orthophoto geotransform to GPS.

• Detection — per-crown bounding boxes + species labels (AI vision)
• Height — sampled from DSM−DTM at each crown center
• Crown diameter — bounding-box width × ground sample distance
• Georeferencing — pixel → projected → WGS84 (GeoJSON/KML)

Outputs: GeoJSON, KML, segmentation/crown overlays, summary statistics.

← View interactive orthophoto map (toggle tree overlays)