remove the black line on the edge

Question

I have a program that detects floors, so the floors I detect are removed and become a transparent png, but there are still black lines around the edges

before adding threshold

enter image description here

enter image description here

src = cv2.imread(file_name)
    
tmp = cv2.cvtColor(src, cv2.COLOR_BGR2GRAY)
_,alpha = cv2.threshold(tmp,0, 255, cv2.THRESH_BINARY)
 b, g, r = cv2.split(src)
 rgba = [b,g,r, alpha]
 dst = cv2.merge(rgba,1)
    

Answer 1

You can mitigate the effect of the black transition line in Python/OpenCV/Skimage by antialiasing the alpha channel as follows:

• Read the image with alpha
• Extract the bgr base image
• Extract the alpha channel
• Gaussian blur the alpha channel (choose the blur amount so as to match the black transition)
• Stretch the dynamic range of the alpha channel so that 255 -> 255 and mid-gray goes to 0 and save as mask. (Choose the mid-gray level to mitigate further)
• Put the resulting mask image into the alpha channel of the bgr image
• Save the results

Input:

import cv2
import numpy as np
import skimage.exposure

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

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

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

# apply Gaussian blur to alpha
blur = cv2.GaussianBlur(alpha, (0,0), sigmaX=5, sigmaY=5, borderType = cv2.BORDER_DEFAULT)

# stretch so that 255 -> 255 and 192 -> 0
mask = skimage.exposure.rescale_intensity(blur, in_range=(192,255), out_range=(0,255)).astype(np.uint8)

# put mask into alpha channel
result = np.dstack((bgr, mask))

# save result
cv2.imwrite('room_new_alpha.png', result)

# display result, though it won't show transparency
cv2.imshow("bgr", bgr)
cv2.imshow("alpha", alpha)
cv2.imshow("mask", mask)
cv2.imshow("result", result)
cv2.waitKey(0)
cv2.destroyAllWindows()

Result: