Is there a way to draw the rectangle for each detected blob onto the original image if it was modified between snapshot and calling find_blobs? Better said, is there a way of storing the sensor.snapshot() to reuse it later in this way?
And is it possible to draw colored boxes even though the img is in grayscale mode?
Thx in advance 
Yes, now that there’s a lot of heap RAM you can just copy() an image.
image — machine vision — MicroPython 1.23 documentation (openmv.io)
And is it possible to draw colored boxes even though the img is in grayscale mode?
image — machine vision — MicroPython 1.23 documentation (openmv.io)
I recommend just doing processing in RGB565 if you need to draw color and setting the color thresholds to everything to ignore color in find_blobs().