Raspberry Pi camera script using grayscale

Report bugs
#1

Question/Issue:
If you use the RGB version then you have to parse a RGB image, which works fine. But if you want to use grayscale you have to parse in a BGR-Image. I’m not sure if this is intended.

Line 115, 116 in EdgeImpulseRunner
if self.isGrayscale:
cropped = cv2.cvtColor(cropped, cv2.COLOR_BGR2GRAY)

#2

Hi @Merten,

Thanks for the heads up! My guess is that it should use COLOR_RGB2GRAY. I’ve made a note for our engineering team to look at it. The difference from RGB and BGR to grayscale should be fairly slight, though.

Raspberry Pi and FOMO