0
votes

I am developing an app which detects major colour in the frame. I used python code and HSV colour ranges to do this. My sample code for detecting blue is as below.

import cv2    
def detectBlue(frame):
        # Convert BGR to HSV
        hsv = cv2.cvtColor(frame, cv2.COLOR_BGR2HSV)
        # define range of green color in HSV
        lower_blue = np.array([101, 39, 64])
        upper_blue = np.array([140, 255, 255])
        mask = cv2.inRange(hsv, lower_blue, upper_blue)
        kernel = np.ones((5,5),'int')
        dilated = cv2.dilate(mask,kernel)
        # Bitwise-AND mask and original image
        res = cv2.bitwise_and(frame,frame, mask=mask)
        ret,thrshed = cv2.threshold(cv2.cvtColor(res,cv2.COLOR_BGR2GRAY),3,255,cv2.THRESH_BINARY)
        #img,contours,hier = cv2.findContours(thrshed,cv2.RETR_LIST,cv2.CHAIN_APPROX_SIMPLE)
        contours,hier = cv2.findContours(thrshed,cv2.RETR_LIST,cv2.CHAIN_APPROX_SIMPLE)
        area = [0]
        for cnt in contours:
            #Contour area is taken
            area.append(cv2.contourArea(cnt))
        return max(area)

I created similar functions for Red and Green with HSV color ranges as below.

Green:
lower_green = [65,60,60])
upper_green = [80,255,255]

Red:
lower_red = [170,120,70] 
upper_red = [180,255,255]

Now I am trying to do the same for all VIBGYOR colors, but I am not getting proper color ranges for all these. I tried converting RGB code to HSV using this python code.

import colorsys

def rgb_hsv_converter(rgb):
    (r,g,b) = rgb_normalizer(rgb)
    hsv = colorsys.rgb_to_hsv(r,g,b)
    (h,s,v) = hsv_normalizer(hsv)
    upper_band = [h+10, s+40, v+40]
    lower_band = [h-10, s-40, v-40]
    return {
        'upper_band': upper_band,
        'lower_band': lower_band
    }

def rgb_normalizer(rgb):
    (r,g,b) = rgb
    return (r/255, g/255, b/255)

def hsv_normalizer(hsv):
    (h,s,v) = hsv
    return (h*360, s*255, v*255)

rgb_hsv_converter((255,255,255))

But, these ranges did not work properly when tested. I searched a lot to find out best HSV ranges for all these colors, but no use. Can anybody suggest any approach to find best and more accurate color ranges for many colors.?

1

1 Answers

1
votes

You may take a pure red color, RGB( 255, 0, 0 ), convert it to HSV and the resulting hue (something around 0 -- zero) will be the center value for red. Do the same for the rest of the colors you want to define in HSV space. Green would be somewhere around 55, blue -- 120 and so on...

enter image description here