Start WebODM and Process Images¶

Spin up the WebODM photogrammetry server and process geotagged aerial images into orthophotos, elevation models, 3D meshes, and point clouds.

Prerequisites

Docker and Docker Compose installed
A folder of geotagged JPEG images (see Transfer Images)

Start WebODM¶

cd ground/odm
docker compose up -d

This launches three services:

Service	Description	Port
`webapp`	WebODM web interface	8000
`db`	PostgreSQL database	--
`worker`	NodeODM processing engine	3000

Once the containers are running, open http://localhost:8000 in your browser.

Process via Web UI¶

Log in to WebODM (create an account on first launch)
Click Add Project and give it a name
Click Select Images and upload your geotagged JPEGs
Click Start Task to begin processing
When complete, download results (orthophoto, 3D model, point cloud)

The web UI shows real-time progress and lets you preview results in a built-in map viewer.

Process via CLI¶

For scripted or headless workflows, use the CLI wrapper:

./ground/odm/process.sh <image_folder> [profile]

Examples¶

Process with the default survey_standard profile:

./ground/odm/process.sh ./captures_20260308_143022

Process with a custom profile:

./ground/odm/process.sh ./captures_20260308_143022 high_detail

Profiles are JSON files in ground/odm/profiles/ that map to ODM CLI flags.

Output Files¶

Results are written to <image_folder>/odm_output/:

Directory	Output	Format
`odm_orthophoto/`	Orthophoto (georeferenced mosaic)	`.tif`
`odm_dem/`	Digital Surface Model	`.tif`
`odm_texturing/`	Textured 3D mesh	`.obj`
`odm_georeferencing/`	Georeferenced point cloud	`.laz`

Large Datasets

For large datasets (hundreds of images), the CLI approach is faster than the web UI because it avoids the upload step -- images are mounted directly from disk into the Docker container.

Stop WebODM¶

cd ground/odm
docker compose down

Data is persisted in Docker volumes (webodm_data, webodm_db), so your projects and results survive restarts.