Color transformation from Bayer to RGB
Published at February 17, 2020 · Last Modified at November 19, 2021 · 4 min read · Tags: bayer rgb computer-vision
A Bayer filter mosaic is a color filter array (CFA) for arranging RGB color filters on a square grid of photosensors. Its particular arrangement of color filters is used in most single-chip digital image sensors used in digital cameras, camcorders, and scanners to create a color image. The filter pattern is 50% green, 25% red and 25% blue - see here.
Color conversion algorithm
The conversion algorithm is based on simple linear interpolation of pixels to find the missing value. Since Bayer filter dived into 4 pixels (red, green, green, blue), there are four types of equations for each pixel in the square.
The following is RGB triple for the pixels on red channel: $$ RGB_{} = ( I_{n,m} , \frac{I_{n-1,m} +I_{n+1,m}+I_{n,m-1} + I_{n,m+1}}{4} , \frac{I_{n-1,m-1} +I_{n+1,m+1}+I_{n-1,m+1} + I_{n+1,m-1}}{4} ) $$
The following is RGB triple for the pixels on green1 channel:
The following is RGB triple for the pixels on green2 channel:
The following is RGB triple for the pixels on blue channel:
where $$ I_{n,m} $$ is raw bayer pixel
Implemantation
Implementation using python including that modules: matplotlib, numpy,cv2. To install these modules with pip:
pip install matplotlib --user
pip install numpy --user
#pip install cv2 --user - not sure this will work , I did it with the gentoo package manager (in debian/ubuntu you can try apt-cache search)
Some images to work with can be download from here 1 2 3 4
- read raw image at a size of 540x600 in Bayer format
import matplotlib.pyplot as plt
import numpy as np
import cv2
file_path = 'path-to-bayer.raw'
imrows = 540
imcols = 600
imsize = imrows*imcols
with open(file_path, "rb") as rawimage:
bayer_img = np.fromfile(rawimage, np.dtype('u1'), imsize).reshape((imrows, imcols))
- define a conversion function
def pixel (img):
img = img.astype(np.float64)
pixel = lambda x,y : {
0: [ img[x][y] , (img[x][y-1] + img[x-1][y] + img[x+1][y] + img[x][y+1]) / 4 , (img[x-1][y-1] + img[x+1][y-1] + img[x-1][y+1] + img[x+1][y+1]) / 4 ] ,
1: [ (img[x-1][y] + img[x+1][y]) / 2,img[x][y] , (img[x][y-1] + img[x][y+1]) / 2 ],
2: [(img[x][y-1] + img[x][y+1]) / 2 ,img[x][y], (img[x-1][y] + img[x+1][y]) / 2],
3: [(img[x-1][y-1] + img[x+1][y-1] + img[x-1][y+1] + img[x+1][y+1]) / 4 , (img[x][y-1] + img[x-1][y] + img[x+1][y] + img[x][y+1]) / 4 ,img[x][y] ]
} [ x % 2 + (y % 2)*2]
res = np.zeros ( [ np.size(img,0) , np.size(img,1) , 3] )
for x in range (1,np.size(img,0)-2):
for y in range (1,np.size(img,1)-2):
p = pixel(x,y)
p.reverse();
res[x][y] = p
res = res.astype(np.uint8)
return res
- break image to to 3 channels : RGB
def channel_break (img):
img = img.astype(np.float64)
red=np.copy (img);red [1::2,:]=0;red[:,1::2]=0
blue=np.copy (img);blue [0::2,:]=0;blue[:,0::2]=0
green=np.copy (img);green [0::2,0::2]=0;green [1::2,1::2]=0;
red = red.astype(np.float64)
blue = blue.astype(np.float64)
green = green.astype(np.float64)
return (red,green,blue)
- conver RGB to gray
def rgb2gray(img):
res = np.zeros ( [ np.size(img,0) , np.size(img,1) , 3] )
res = res.astype(np.float64)
for x in range (1,np.size(img,0)-1):
for y in range (1,np.size(img,1)-1):
res[x][y]=img[x][y][0]*0 + img[x][y][1]*0.5 + img[x][y][2]*0.5;
res = res.astype(np.uint8)
return res
- save results to png files
# plot bayer imager
plt.imshow(bayer_img)
plt.title ('bayer img')
plt.imsave('bayer_img.png', bayer_img)
#plt.show()
# this algorithm conversion
rgb_res = pixel (bayer_img)
plt.imshow(rgb_res)
plt.title ('the article conversion')
plt.imsave('the_article_conversion.png', rgb_res)
#plt.show()
# open cv conversion
colour = cv2.cvtColor(bayer_img, cv2.COLOR_BAYER_BG2BGR)
plt.imshow(colour)
plt.title ('color image by open cv')
plt.imsave('color_image_by_opencv.png', colour)
#plt.show()
# convert to gray level
gray = rgb2gray(rgb_res)
plt.imshow(gray)
plt.title ('gray conversion')
plt.imsave('gray_level.png', gray)
#plt.show()
# break to RGB channels
RGB = channel_break(bayer_img)
blue_only = pixel (RGB[0])
plt.imshow(blue_only)
plt.title ('blue only')
plt.imsave('blue_only.png',blue_only)
plt.show()
green_only = pixel (RGB[1])
plt.imshow(green_only)
plt.title ('green only')
plt.imsave('green_only.png',green_only)
plt.show()
red_only = pixel (RGB[2])
plt.imshow(red_only)
plt.title ('red only')
plt.imsave('red_only.png',red_only)
plt.show()
#plt.show()
Results
Raw Bayer Image
Conversion to RGB
Conversion by OpenCV
Conversion to Gray
Red Only
Green Only
Blue Only
References
Here are some reference sources that used to create this post
[1] Bayer to RGB
[2] RGB color conversions matrixes
[3] Bayer to RGB algorithm
[4] Bayer Patterns in Digicam CCDs
[5] RGB to gray