dronnix-io/drone-building-extraction
Preview • Updated
AI & ML interests
Geospatial AI, semantic segmentation, remote sensing, drone mapping, deep learning
Recent Activity
View all activity
Organization Card
🌍 Dronnix
Drone Mapping & Geospatial AI
🏢 About
Dronnix specializes in drone mapping, 3D modeling, digital twins, and geospatial AI for:
| Domain | Application |
|---|---|
| ☀️ Energy | Solar panel inspection & fault detection |
| 🌾 Agriculture | Crop monitoring, yield estimation & disease detection |
| 🏙️ Urban | Growth monitoring, land cover classification & change detection |
| 🏗️ Construction | Progress tracking, volume measurement & site monitoring |
| 🗺️ Mapping | Large-scale 2D orthomosaic & topographic mapping |
| 🏛️ Digital Twins | 3D city models, infrastructure twins & facility management |
| 🌲 Environment | Forest inventory, flood mapping & erosion monitoring |
| ⛏️ Mining | Stockpile volume, pit mapping & earthwork estimation |
| 🚧 Infrastructure | Road condition, powerline inspection & bridge assessment |
| 🏘️ Real Estate | 3D property models, site visualization & area measurement |
🚀 Open Resources
🔌 GeoSeg Studio
Open-source QGIS plugin for deep learning semantic segmentation of satellite and aerial imagery
- Full pipeline: Prepare → Train → Evaluate → Predict
- 7 model architectures (UNet, SegFormer, SwinUNet, and more)
- GPU-accelerated with PyTorch — no coding required
🗂️ Drone Building Extraction Dataset — Calgary
High-resolution drone imagery with manually annotated building polygons for binary semantic segmentation
- 4-band GeoTIFF orthomosaics (RGB + Alpha) captured with DJI Matrice 4E
- Manually annotated building footprints (QGIS) — Calgary, Alberta, Canada
- Ready to use directly with GeoSeg Studio — no pre-processing required
- License: CC BY 4.0
📬 Contact
For inquiries and support: salar@dronnix.com
Built with ❤️ by Dronnix
models 0
None public yet