Try AIDM

You need Docker, or Podman, on Linux or on a Mac with Apple silicon. On Windows, use Docker Desktop with the WSL 2 backend; a native Windows package is on the roadmap.

Process a real survey

Download the small vineyard, 73 images of 20 megapixels, about 0.8 GiB:

hf download AIDroneMap/aidm-vineyard-small --repo-type dataset --include "images/*" --local-dir datasets/vineyard

The hf command comes with the Hugging Face client (pip install huggingface_hub). Any other way to fetch the images/ folder works too.

Process it:

docker run -ti --rm -v $(pwd)/datasets:/datasets ghcr.io/aidronemap/aidm:0.1.0 --project-path /datasets --dsm vineyard

The outputs land in datasets/vineyard: an orthophoto, elevation models, a point cloud and a textured model. On a modern 8-core machine this takes about 16 minutes.

Your own images go in a folder of the same shape: datasets/<project>/images/.

Run the task API

docker run -p 3000:3000 ghcr.io/aidronemap/aidm-api:2.5.0

Open http://localhost:3000 for the built-in page, or add http://<host>:3000 as a processing node in WebODM. The API is compatible with NodeODX, so existing clients work unchanged.

GPU

The GPU image uses CUDA when an NVIDIA GPU is present and falls back to the CPU when it is not:

docker run --gpus all -p 3000:3000 ghcr.io/aidronemap/aidm-api:2.5.0-gpu

It needs the NVIDIA Container Toolkit on the host.

Images

Each release lists the digest of every image it published, so you can pin exactly what was tested.

Flags, outputs, settings and the API reference are on aidronemap.dev.