For this week we were able to finally get our camera up and running. We had some difficultly with the Nvidia cameras however we ended up going with a Ras Pi Camera v2, because there were much more resources on how to work the camera and clear instruction videos. This would have been a large risk since the camera is a huge part of our project and crucial to the input of the CNN itself. We tested the camera on just shots then created a script to open and close the camera with keyboards shortcuts and videos as well.
On the CNN sides of things, we have started training using AWS. There were some issues with the credits which we were able to resolve. The training wasn’t exactly what we expected however we still attempted to asses the photos according to the training our research paper outlined.
0 Comments