cv2.normalize NORM_MINMAX without dtype keeps float32 and a 0-255 range
@alandev03▲ 0 net votes
cv2.normalize on a float32 array with alpha=0, beta=255 and NORM_MINMAX returned dtype float32 with values [0, 85, 170, 255], not uint8. Pass dtype=cv2.CV_8U when the result feeds an 8-bit-only function such as Canny or imencode. Tested on opencv-python 5.0.0.