Looking for most effective way to leverage DigitalOcean

I’m brand new to drones and mapping, but experienced with cloud stuff. I have generated a couple of ~200-image datasets, and so far my process is: spin up a nice DigitalOcean droplet, install WebODM and friends, start it up, use the UI to upload my files and launch the processing, download the products and snapshot the droplet before deleting it to save money.

I got to thinking, if I could populate a directory in Spaces (DO’s S3-like product) with a JSON file containing my proposed configuration and all my images, then I could spin up a prebuilt droplet that would mount that directory, perform the processing, and store the results in that directory.

Is anyone else working on this?

1 Like

Not for DigitalOcean (the software is tied to Amazon AWS), but there’s GitHub - hobu/codm which performs something similar (you place images in a “magic bucket” and that kicks off the processing, storing the results back into the bucket).

Would be cool to write something similar that is cloud agnostic.

1 Like

Hmm! Deploying Lambda on DigitalOcean is a thing and I already have a managed Kubernetes cluster there. Hmmmmm.

Thank you for the pointer, it was inspirational. :slight_smile:

3 Likes