I've been working on a project over the last year (time permitting) and would love some feedback.
We have a cat who is a very avid hunter and wanted to stop her from bringing her "wares" into the kitchen and beyond. I've seen other cat flap projects but they all seem to rely on sending data outside of the cat flap or building special boxes for cats to pass through.
I wanted to make something that was self contained, hence The Integrated Cat Flap.
The system is based around a stock Petsafe Staywell Deluxe Magnetic cat flap with a Pi Zero using Open CV 3.1.
A Pi Camera NoIR 2.1 points underneath the cats mouth alongside a dawn / dusk IR LED. When enough motion is detected, OpenCV takes a picture and if prey is detected in the cats mouth, a servo will lock the flap mechanism and fluffy can't come in.
I've had the flap in operation since September last year and now have over 500 photos of cat with prey.
The plan is to create an Open CV Haar Cascade from these images which Open CV will use to lock the flap.
Question is, is using a Haar Cascade the best option for recognition? Would LBP or HOG be better?