python—图像增强

def img_chuli_plt2():
    import cv2
    import random
    # import imutils
    import numpy as np

    # 彩色图像每个像素值是[x,y,z], 灰度图像每个像素值便是一个np.uint8
    path = './img/0.jpg'  # 路径二维码
    # path = './img/sachin.jpg'  # 路径
    image = cv2.imread(path)
    gray_img = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)

    # 图像大小调整
    ori_h, ori_w = image.shape[:2]
    height, width = gray_img.shape[:2]
    image = cv2.resize(image, (int(ori_w / ori_h * 400), 400), interpolation=cv2.INTER_CUBIC)
    gray_img = cv2.resize(gray_img, (int(width / height * 400), 400), interpolation=cv2.INTER_CUBIC)

    # a<0 and b=0: 图像的亮区域变暗，暗区域变亮
    a, b = -0.5, 0
    new_img1 = np.ones((gray_img.shape[0], gray_img.shape[1]), dtype=np.uint8)
    for i in range(new_img1.shape[0]):
        for j in range(new_img1.shape[1]):
            new_img1[i][j] = gray_img[i][j] * a + b

    # a>1: 增强图像的对比度,图像看起来更加清晰
    a, b = 1.5, 20
    new_img2 = np.ones((gray_img.shape[0], gray_img.shape[1]), dtype=np.uint8)
    for i in range(new_img2.shape[0]):
        for j in range(new_img2.shape[1]):
            if gray_img[i][j] * a + b > 255:
                new_img2[i][j] = 255
            else:
                new_img2[i][j] = gray_img[i][j] * a + b

    # a<1: 减小了图像的对比度, 图像看起来变暗
    a, b = 0.5, 0
    new_img3 = np.ones((gray_img.shape[0], gray_img.shape[1]), dtype=np.uint8)
    for i in range(new_img3.shape[0]):
        for j in range(new_img3.shape[1]):
            new_img3[i][j] = gray_img[i][j] * a + b

    # a=1且b≠0, 图像整体的灰度值上移或者下移, 也就是图像整体变亮或者变暗, 不会改变图像的对比度
    a, b = 1, -50
    new_img4 = np.ones((gray_img.shape[0], gray_img.shape[1]), dtype=np.uint8)
    for i in range(new_img4.shape[0]):
        for j in range(new_img4.shape[1]):
            pix = gray_img[i][j] * a + b
            if pix > 255:
                new_img4[i][j] = 255
            elif pix < 0:
                new_img4[i][j] = 0
            else:
                new_img4[i][j] = pix

    # a=-1, b=255, 图像翻转
    new_img5 = 255 - gray_img

    cv2.imshow('origin', image)
    cv2.imshow('gray', gray_img)
    cv2.imshow('a<0 and b=0', new_img1)
    cv2.imshow('a>1 and b>=0',new_img2)
    cv2.imshow('a<1 and b>=0',new_img3)
    cv2.imshow('a=1 and b><0',new_img4)
    cv2.imshow('a=-1 and b=255',new_img5)
    if cv2.waitKey(0) == 27:
        cv2.destroyAllWindows()

    # 感谢：https: // blog.csdn.net / feilong_csdn / article / details / 82780664