Movement detection, calculation std between last 10 images

Hi together,

i want to detect if an object (like an eating animal) is moving at the moment or if it is sitting still. My plan was to record 10 frames and then calculate the standard deviation of each pixel per frame (threw 3rd dimension).

But I’m new to OpenMV, I’m not sure how to perform this matrix calculation on a OpenMV Cam H7 Plus.

Optional I could send the image to an PC and do the calculation there with numpy but I want to read analog and digital signals with the module as well and transferring all the data and the image to a PC does not seems to be trivial as well.

Does anyone has a suggestion how to solve this Problem?


Use a CNN trained from images using Edge Impulse.

1 Like

Thanks for the fast reply. CNN goes beyond my expertise, is there an example for a similar workflow?

I found the example snapshot_on_movement which almost does what I need.

Hi, I highly recommend going the CNN approach. It’s really easy: OpenMV - Edge Impulse Documentation

1 Like