Open-source photogrammetry

Turn photos into measurable 3D.

OpenReco reconstructs drone and camera imagery into dense point clouds, textured meshes, orthomosaics and elevation models — georeferenced, measurable, and fully reproducible. One desktop app, the entire pipeline.

GPU-accelerated dense MVS EPSG georeferencing Reproducible cache-keyed pipeline MIT licensed
Dense cloud · 483,902 pts
Measure volume & area
Georeferenced · EPSG:32632
OpenReco — Chunk 1 · 483,902 pts
The full OpenReco application — workspace tree, 3D viewport with a textured dense point cloud, properties panel and photo filmstrip

Built for people who measure the real world

Surveying & mapping Agriculture & forestry VFX & heritage Construction & earthworks Robotics & research
drone photo
36
Drone photos in
483,902
3D points solved
1 model
Measurable & georeferenced

Drop in a folder of overlapping images. OpenReco aligns the cameras, builds a dense cloud, meshes and textures it — then hands you a model you can measure, classify and export.

01The workflow

One pipeline, every stage.

From raw images to georeferenced deliverables — every step lives in a single, node-based workspace. Run the whole graph or any stage on its own; results are cached and reproducible.

01

Add Photos

Add a folder of images to a chunk (ingest + quality check). The first step of any project.

02

Align Photos

Detect features, match, and solve camera poses to produce a sparse point cloud.

03

Georeference

Place the model in a CRS from EXIF GPS or ground control points using the marker tool.

04

Build Dense Cloud

Dense multi-view stereo depth maps fused into a high-density point cloud.

05

Build Model

Reconstruct a polygonal mesh surface from the dense cloud.

06

Build Texture

Bake a UV texture atlas onto the model from the source images.

07

Build DEM

Rasterize a digital elevation model (DSM) from the dense cloud or mesh.

08

Build Orthomosaic

Orthorectified, georeferenced image mosaic ready for GIS.

09

Vegetation Indices

NDVI / ExG / VARI maps from the orthomosaic for agriculture & health analysis.

10

Classify Points

Classify the dense cloud into ground, building and vegetation.

11

Build Tiled Model

Split the model into streamable 3D Tiles for the web.

12

Build Contours

Generate contour lines at any interval from the DEM.

13

Clean Point Cloud

Remove noise and stray points with statistical outlier removal.

14

Clean Mesh

Remove small floating mesh islands and keep the main surface.

15

Merge Chunks

Align separate chunks with ICP and merge their clouds into one.

02Reconstruct

From sparse tie-points to a dense, textured surface.

A GPU multi-view stereo engine (COLMAP CUDA) fuses depth maps into clouds of hundreds of thousands of points, then meshes and textures them — all from the same source photos.

Dense MVS with geometric consistency and configurable plane-sweep depth.
Poisson meshing into a watertight polygonal model with vertex counts in the millions.
UV texture baking straight from the registered imagery.
Textured dense point cloud of buildings
Dense cloud · mvs1 483,902 pts
03Measure

Volumes, areas, profiles — straight off the model.

Pick a tool, click on the 3D model, and read real-world quantities instantly. Every measurement is georeferenced and exportable.

Volume with cut / fill breakdown for stockpiles and earthworks.
Profiles with elevation graph, length, Δ-height and slope %.
Areas & distances annotated directly in the viewport.
Export to GeoJSON, DXF or CSV in a click.
Measurement panel with volume, area and profile annotations on the model
Measurements · volume · profile · area GeoJSON · DXF · CSV
04Map & analyze

Orthomosaics, DEMs and vegetation maps.

Produce survey-grade 2D deliverables from the same reconstruction — orthorectified imagery, elevation rasters, contours and crop-health indices.

Orthomosaic — georeferenced, seam-aware image mosaic at your chosen GSD.
DSM + contours — elevation model with contour lines at any interval.
NDVI / ExG / VARI — vegetation indices for agriculture and forestry.
Top-down orthomosaic
Orthomosaic EPSG:32632
Digital elevation model
DSM
Vegetation index · ExG
BARECANOPY
05Visualize

A viewport built for point clouds.

Inspect hundreds of thousands of points at a smooth frame rate, with the controls photogrammetrists actually reach for.

Point cloud colored by elevation
Color by elevation · viridis EDL on

RGB & Elevation

Color points by true RGB or by height to read terrain at a glance.

Scientific colormaps

viridis and friends for perceptually-uniform elevation shading.

Eye-Dome Lighting

EDL shading brings out depth and structure in dense clouds.

Model · Ortho · Map · Photo

Switch views, adjust point size, and orbit with the nav cube.

06Precise & reproducible

Engineered like a measuring instrument.

Every stage is content-addressed by a cache key derived from its inputs and resolved parameters. Re-run a project and identical stages are reused — byte-for-byte reproducible, every time.

GPU dense MVS

COLMAP CUDA backend with an automatic CPU fallback.

CRS aware

EPSG georeferencing from EXIF GPS or ground control points.

Cache-keyed

Resolved parameters hashed per stage for instant re-runs.

Chunks & presets

Quality presets and multi-chunk projects that merge cleanly.

07Export

Open formats, no lock-in.

Every artifact leaves OpenReco in a standard format your GIS, CAD or web stack already speaks.

LAS / LAZ PLY OBJ GeoTIFF GeoJSON DXF CSV 3D Tiles Contours (SHP)
Free & open source

Start reconstructing in minutes.

Download OpenReco for your platform, drop in a folder of photos, and run the pipeline. No account, no subscription.

MIT licensed · build from source on GitHub