I would like to separate the red colour, however faint, from the rest of the image. I can't find a satisfactory method. I have used otsu threshold and adjusted the contrast, but so far I can't find anything satisfactory.
I receive different images, so the colour quantity can vary more or less. It would be necessary for the program to adapt to the image received using a threshoding method.
Do you have an idea?
This is the code I wrote. I get the green channel to be able to separate the colours and then I amplify the different colours with contrast_adjuster to get a better threshold at the end. But I can't get a very good threshold.
green = img[..., 1]
adj_green = contrast_adjuster(green)
ret = cv2.threshold(adj_green, 0, 255, cv2.THRESH_BINARY + cv2.THRESH_OTSU)[0]
mask = cv2.threshold(green, ret, 255, cv2.THRESH_BINARY)[1]
Orange = cv2.bitwise_and(img, img, mask=mask)
Red = img - Orange
Red = cv2.cvtColor(Red, cv2.COLOR_BGR2GRAY)
def contrast_adjuster(img):
equ = cv2.equalizeHist(img)
return equ
