Build reproducible, Python-driven photogrammetry pipelines for drone mapping. Automate image alignment, ground control point optimization, DEM/DSM generation, and export workflows with battle-tested open-source tooling.
This site is for UAV operators, surveying technicians, and Python GIS developers who need to move past black-box mapping software. Every guide treats photogrammetry as a deterministic software-engineering discipline — strict data schemas, explicit geospatial alignment, and automated resource controls that scale from a single mission to enterprise fleet operations.
You'll find production-grade solutions to the failures that actually break mapping jobs: RAM exhaustion during dense matching, silent CRS drift, alignment convergence failures, and batch scaling bottlenecks — each with copy-ready Python you can drop into your own pipeline.
83 pages across four sections now cover the whole chain — ingestion and flight geometry, feature matching and bundle adjustment, ground control and datum handling, and surface generation through to web delivery — including unattended job orchestration, GPU-accelerated dense matching, RTK/PPK geotagging, contour and hillshade derivation, and serving orthomosaics as web tiles.
Four focused areas, each a complete workflow from ingestion through reconstruction to survey-grade export and web delivery.
Treat photogrammetry as deterministic software: schema-enforced ingestion, geometric validation, strict CRS handling, and unattended job orchestration across a fleet.
Feature detection, bundle adjustment, parallel and GPU-accelerated matching, and memory-safe point-cloud workflows that survive large UAV datasets.
Auto-tag GCPs, geotag from RTK/PPK trajectories, distribute residual errors, transform coordinates with pyproj, and lock survey-grade accuracy thresholds.
Rasterize dense point clouds into DSM and DTM surfaces with PDAL, export validated Cloud-Optimized GeoTIFFs with rasterio, derive contours and hillshade with GDAL, and serve the result as web tiles.
Each opens onto its own in-depth guides, wired into the sections they extend.
Battle-tested, copy-ready solutions to the errors that stop a mapping job outright.
The full catalogue, grouped by the topic each guide sits under.