I'm hoping that some helpful can point me in the right direction for some guides on how to perform onboard image processing either with something like opencv or a deep learning approach that can be run onboard a Timer Camera F unit.

The main aim is to provide an image classifier hopefully via simple morphological operations and thresholding, but may need some added robustness via a very low compute cnn model.

I've used arduinos and various single board computers in the past, but this is my first time using the m5stack, so any experience/guidance would be extremely helpful - even recommendations as to whether to what IDE to use (ESP-IDF or arduino IDE for example).

Many thanks

Jim