Final Blog - Wildlife Monitor Project

RE: Final Blog - Wildlife Monitor Project

ralphjy — Tue, 25 Jun 2024 16:28:17 GMT

Just an epilogue with tribute to Extreme Environments. We've been having some hotter weather (up to 32 C) and the CPU started throttling when running continuous inferencing.

After a week of running with the enclosure high temperature getting to 60 C and the CPU to 84 C, I switched to running motion triggered inferencing and the lower duty cycle kept the CPU temp below throttling temp (80 C). However, I noticed that there was blurriness in the images.

A quick check of the setup revealed that my PLA Front Panel had deformed due to the heat. The enclosure had only been to 60 C so I guess that's something I'll have to watch for with PLA prints.

Good news is that I've had a lot of birds coming by... and the enclosure is doing great [emoticon:0e469978eb8947d18b9c41123db3725a].

RE: Final Blog - Wildlife Monitor Project

genebren — Sun, 16 Jun 2024 16:42:33 GMT

Looks like everything performed well. I really liked your bloopers, but it all honesty I can see how some of the images might have been easy to have been misinterpreted without using a size classifier of some sort. The horse and cow were somewhat understandable given the leg proportions of the two dogs. The descending dog could also easily be misclassified as there likely no images of a dog moving down the stairs in the learning data.

RE: Final Blog - Wildlife Monitor Project

javagoza — Sun, 16 Jun 2024 14:37:48 GMT

Congratulations, great summary and project!