opencv - 过滤以从文档的彩色图像创建Xerox/影印图像

Question

我有彩色的文档图像。样本输入

(取自https://www.indiamart.com/proddetail/pan-card-21687536812.html)
我希望以此创建逼真的xerox /影印图像。大多数复印/复印设置为低调。
这是一个示例输出(示例输出不是示例输入的施乐-我在网上都找不到相同的图像)
(取自http://shrikantmail7862.blogspot.com/2016/06/)
音符输出不是彩色图像的简单黑白。
我正在寻找一种工具/代码/算法来为我做这件事。我需要对超过0.1M的图像进行此操作
我确信图像处理中必须有这样的过滤器。寻找合适的指针

Answer 1

为了产生类似xerox的效果，一种方法是将图像转换为灰度，然后减少色彩空间。

def xeroxFilter(imgPath, colorSpaceReduction=8, rotate=False, fillColor=(255,255,255)):
  '''Takes image as input and returns a xerox like image of the input image.
  impPath: Name of Image(keep the images in the same folder as this script)
  colorSpaceReduction:  Reducing the color space. Higher the value, Higher the reduction. 
  rotate: Rotate to the image
  fillColor: Specify the RGB value of the background color after roatation
  '''
  print("Original Image")
  cv2_imshow(cv2.imread(imgPath))

  print('GreyScale')
  cv2_imshow(cv2.imread(imgPath, cv2.IMREAD_GRAYSCALE))
  
  print('Color Space Reduced')
  color_reduced_img = cv2.imread(imgPath, cv2.IMREAD_GRAYSCALE) // colorSpaceReduction * colorSpaceReduction + colorSpaceReduction // 2
  cv2_imshow(color_reduced_img)
  file_name = 'ColorSpaceReduced-'+imgPath.split('.')[0]+'.png'
  cv2.imwrite(file_name, color_reduced_img)
  
  if rotate:
    print("Rotated")
    img = Image.open(file_name)
    rgb_img = Image.new("RGBA", img.size)
    rgb_img.paste(img)
    rotated_img = rgb_img.rotate(random.randint(-25,25), expand = 1, fillcolor = fillColor)
    display(rotated_img)
    rotated_img.save('Rotated-'+file_name)
  print("Original Image")
  cv2_imshow(cv2.imread(imgPath))

xeroxFilter('MCRV7.jpg', rotate=True)

减少色彩空间代码(记数 @elizer)
将第一个图像传递到脚本时的输出可以在下面找到:
[ 1