Friend wishes to venture in the agri drones biz soon, so they’ll be needing a team that consists of computer visionists and deep learning enggs for the software side, and what hardware and other misc setups would be ideal for the team?
I’ve seen two popular methods for drone use cases:
- Pull the video off the drone after it lands and run bulk inference in the cloud.
- Put an edge device (e.g., a Jetson) in each drone and run inference locally.
For a brand new startup, you’re going to want to go with the cloud option unless the drone will need model results to navigate. It’s going to be much cheaper to get started, and you can always switch to edge deployment later on if you find it necessary. That way you also don’t lock yourself into hardware choices you regret later.
As for the deep learning side of things, the whole point of Roboflow is you don’t need an entire ML team ready when you found the company. A solo psuedo-technical founder could get a decent model in an afternoon and a model sufficient for customer or investor demos after a week. Roboflow handles labeling, training, and model deployment out-of-the-box.
This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.