How can I change the hue of a certain area with OpenCV-Python

Question

i want to edit the hsv value of a certain area The picture I want to change color of

I to change all the purple parts of the image to green. Here is the result I made in an image editing software just to show a example.The Result Yeah i made it darker for no reason but i just wanna make the blade green but not the handle

Answer 1

Here is a way to do that in Python/OpenCV by shifting hues.

Input:

import cv2
import numpy as np

# load image with alpha channel
img = cv2.imread('sword_purple.png', cv2.IMREAD_UNCHANGED)

# extract alpha channel
alpha = img[:,:,3]

# extract bgr channels
bgr = img[:,:,0:3]

# convert to HSV
hsv = cv2.cvtColor(bgr, cv2.COLOR_BGR2HSV)
#h = hsv[:,:,0]
#s = hsv[:,:,1]
#v = hsv[:,:,2]
h,s,v = cv2.split(hsv)

# purple is 276 in range 0 to 360; so half in OpenCV
# green is 120 in range 0 to 360; so half in OpenCV
purple = 138
green = 60

# diff color (green - hue)
diff_color = green - purple

# modify hue channel by adding difference and modulo 180
hnew = np.mod(h + diff_color, 180).astype(np.uint8)

# recombine channels
hsv_new = cv2.merge([hnew,s,v])

# convert back to bgr
bgr_new = cv2.cvtColor(hsv_new, cv2.COLOR_HSV2BGR)

# put alpha back into bgr_new
bgra = cv2.cvtColor(bgr_new, cv2.COLOR_BGR2BGRA)
bgra[:,:,3] = alpha

# save output
cv2.imwrite('sword_alpha.png', alpha)
cv2.imwrite('sword_bgr.png', bgr)
cv2.imwrite('sword_bgr_new.png', bgr_new)
cv2.imwrite('sword_green.png', bgra)

# Display various images to see the steps
cv2.imshow('alpha',alpha)
cv2.imshow('bgr',bgr)
cv2.imshow('bgr_new',bgr_new)
cv2.imshow('bgra',bgra)
cv2.waitKey(0)
cv2.destroyAllWindows()

Result:

Answer 2

The simple way in Python/OpenCV is to create a mask where the image is purple using inRange() and then change the image to green where the mask is white using Numpy.

Input:

import cv2
import numpy as np

# load image with alpha channel
img = cv2.imread('sword_purple.png', cv2.IMREAD_UNCHANGED)

# extract alpha channel
alpha = img[:,:,3]

# extract bgr channels
bgr = img[:,:,0:3]

# select purple
lower_purple = (140,40,110)
upper_purple = (170,60,130)
mask = cv2.inRange(bgr, lower_purple, upper_purple)

# change the image to make it green where the mask is white
bgr_new = bgr.copy()
bgr_new[mask==255] = (0,255,0)

# put alpha back into rgb_new
bgra = cv2.cvtColor(bgr_new, cv2.COLOR_BGR2BGRA)
bgra[:,:,3] = alpha

# save output
cv2.imwrite('sword_alpha.png', alpha)
cv2.imwrite('sword_bgr.png', bgr)
cv2.imwrite('sword_mask.png', mask)
cv2.imwrite('sword_masked_green.png', bgr_new)
cv2.imwrite('sword_green.png', bgra)

# Display various images to see the steps
cv2.imshow('alpha',alpha)
cv2.imshow('mask',mask)
cv2.imshow('bgr_new',bgr_new)
cv2.imshow('bgra',bgra)
cv2.waitKey(0)
cv2.destroyAllWindows()

Alpha channel:

BGR channels:

Mask:

Recolored BGR channels:

Result with alpha channel put back:

Alternately, you can do the same by converting BGR to HSV and changing the purple to green the same way. Then convert back to BGR and put the alpha channel back.