数字图像处理：图像彩色通道均衡化

珀伊桦哆

已于 2022-03-17 19:07:19 修改

阅读量1k

点赞数 1

文章标签：计算机视觉 python opencv

于 2022-03-17 19:03:09 首次发布

本文链接：https://blog.csdn.net/CrisSss1/article/details/123557886

版权

import cv2
import numpy as np


image_src = cv2.imread('.\picture.bmp')     # 读入原图像

#图像彩色通道均衡化
def normalizePicture(image):
    rows, cols, BGR = image.shape  # 获取原图像像素点的宽度和高度和三通道
    BGR_Hist = np.zeros((BGR,256))
    normalize_image = image.copy()
    for i in range(rows):  # 统计各个通道值的个数
        for j in range(cols):
            for k in range(BGR):
                BGR_Hist[k,image[i,j,0]] = BGR_Hist[k,image[i,j,0]]+1

    p_BGR_hist = BGR_Hist/(rows *cols)
    p_BGR_add_hist = np.zeros(BGR_Hist.shape)
    p_BGR_add = p_BGR_hist[:,0]

    for i in range(1, 256):  # 令p_add_hist每个元素都为p_hist同下标的前面元素的累加值
        p_BGR_add +=p_BGR_hist[:,i]
        p_BGR_add_hist[:,i] = p_BGR_add

    for i in range(rows):
        for j in range(cols):
            for k in range(BGR):
                normalize_image[i,j,k] = p_BGR_add_hist[k,image[i,j,k]] * 255

    return normalize_image

normalize_image=normalizePicture(image_src)
cv2.imshow('image', normalize_image)
cv2.waitKey(0)
cv2.destroyAllWindows()

关于网上灰度图像的均衡化挺多了，一般彩色通道均衡化的结果也没什么用，

但是因为要写作业，网上又没有搜到，就基于灰度均衡化的写法照着写了一个，

区别也就在于通道数*3，写的时候在张量的形状上稍微注意下就好