from PIL import Image    img = Image.open('01.jpg')  imgGrey = img.convert('L')    img.show()  imgGrey.show()    img.save('img_copy.jpg')  imgGrey.save('img_gray.jpg')

from PIL import Image  import numpy as np    img = Image.open('01.jpg')    width, height = img.size  channel_mode = img.mode  mean_value = np.mean(img)    print(width)  print(height)  print(channel_mode)  print(mean_value)

from PIL import Image    width = 200  height = 100    img_white = Image.new('RGB', (width,height), (255,255,255))  img_black = Image.new('RGB', (width,height), (0,0,0))  img_L = Image.new('L', (width, height), (255))    img_white.show()  img_black.show()  img_L.show()

from PIL import Image    img = Image.open('01.jpg')    width, height = img.size    # 获取指定坐标位置像素值  pixel_value = img.getpixel((width/2, height/2))  print(pixel_value)    # 或者使用load方法  pim = img.load()  pixel_value1 = pim[width/2, height/2]  print(pixel_value1)    # 设置指定坐标位置像素的值  pim[width/2, height/2] = (0, 0, 0)    # 或使用putpixel方法  img.putpixel((w//2, h//2), (255,255,255))    # 设置指定区域像素的值  for w in range(int(width/2) - 40, int(width/2) + 40):  for h in range(int(height/2) - 20, int(height/2) + 20):  pim[w, h] = (255, 0, 0)  # img.putpixel((w, h), (255,255,255))  img.show()

from PIL import Image    img = Image.open('01.jpg')    # 通道分离  R, G, B = img.split()    R.show)  G.show()  B.show()    # 通道合并  img_RGB = Image.merge('RGB', (R, G, B))  img_BGR = Image.merge('RGB', (B, G, R))  img_RGB.show()  img_BGR.show()

from PIL import Image, ImageDraw, ImageFont    img = Image.open('01.jpg')    # 创建Draw对象:  draw = ImageDraw.Draw(img)  # 字体颜色  fillColor = (255, 0, 0)    text = 'print text on PIL Image'  position = (200,100)    draw.text(position, text, fill=fillColor)  img.show()

from PIL import Image    img = Image.open('01.jpg')    width, height = img.size    img_NEARESET = img.resize((width//2, height//2)) # 缩放默认模式是NEARESET(最近邻插值)  img_BILINEAR = img.resize((width//2, height//2), Image.BILINEAR) # BILINEAR 2x2区域的双线性插值  img_BICUBIC = img.resize((width//2, height//2), Image.BICUBIC) # BICUBIC 4x4区域的双三次插值  img_ANTIALIAS = img.resize((width//2, height//2), Image.ANTIALIAS) # ANTIALIAS 高质量下采样滤波

from PIL import Image    img = Image.open('01.jpg').convert('L')    width, height = img.size    pim = img.load()    for w in range(width):  for h in range(height):  if pim[w, h] > 100:  img.putpixel((w, h), 255)  # pim[w, h] = 255  else:  img.putpixel((w, h), 0)  # pim[w, h] = 0    img.show()

from PIL import Image    img = Image.open('01.jpg').convert('L')    width, height = img.size    threshold = 125    for w in range(width):  for h in range(height):  if img.getpixel((w, h)) > threshold:  img.putpixel((w, h), 255)  else:  img.putpixel((w, h), 0)    img.save('binary.jpg')

from PIL import Image    img = Image.open('01.jpg')    width, height = img.size    # 前两个坐标点是左上角坐标  # 后两个坐标点是右下角坐标  # width在前， height在后  box = (100, 100, 550, 350)    region = img.crop(box)    region.save('crop.jpg')

# 边界扩展  from PIL import Image    img = Image.open('test.png')    width, height = img.size  channel_mode = img.mode    img_makeBorder_full = Image.new(channel_mode, (2*width, height))  img_makeBorder_part = Image.new(channel_mode, (width+200, height))    # 图像水平扩展整个图像  img_makeBorder_full.paste(img, (0, 0, width, height))  img_makeBorder_full.paste(img, (width, 0, 2*width, height))    # 前两个坐标点是左上角坐标  # 后两个坐标点是右下角坐标  # width在前， height在后  box = (width-200, 0, width, height)  region = img.crop(box)    # 图像水平右侧扩展一个ROI  img_makeBorder_part.paste(img, (0, 0, width, height))  img_makeBorder_part.paste(region, (width, 0, width+200, height))  img_makeBorder_part.show()  img_makeBorder_full.show()

from PIL import Image  import numpy as np    img = Image.open('01.jpg')    array = np.array(img) # PIL.Image 转 numpy    img1 = Image.fromarray(array) # numpy转 PIL.Image  img1 = Image.fromarray(array.astype('uint8'))    img1.save('from_array.jpg')