opencv处理图片批量添加噪声、以及光照、黑暗处理

本文整理了常见的通过opencv实现数据处理的函数：包括、添加噪声（高斯、椒盐、随机噪声）、灰暗化处理、光照处理。目的在于往数据集里增加噪声，略微增加数据集的挑战性。并通过函数批量实现数据的处理。

import os
import cv2
import numpy as np
import random
####变亮
def imgBrightness(img1, c, b):
    h, w, ch = img1.shape
    blank = np.zeros([h, w, ch], img1.dtype)
    rst = cv2.addWeighted(img1, c, blank, 1-c, b)
    return rst
### 变暗
def darker(image, percetage=0.9):
    image_copy = image.copy()
    w = image.shape[1]
    h = image.shape[0]
    # get darker
    for xi in range(0, w):
        for xj in range(0, h):
            image_copy[xj, xi, 0] = int(image[xj, xi, 0] * percetage)
            image_copy[xj, xi, 1] = int(image[xj, xi, 1] * percetage)
            image_copy[xj, xi, 2] = int(image[xj, xi, 2] * percetage)
    return image_copy
###添加椒盐噪声
def sp_noise(noise_img, proportion):
    '''
    添加椒盐噪声
    proportion的值表示加入噪声的量，可根据需要自行调整
    return: img_noise
    '''
    height, width = noise_img.shape[0], noise_img.shape[1]#获取高度宽度像素值
    num = int(height * width * proportion) #一个准备加入多少噪声小点
    for i in range(num):
        w = random.randint(0, width - 1)
        h = random.randint(0, height - 1)
        if random.randint(0, 1) == 0:
            noise_img[h, w] = 0
        else:
            noise_img[h, w] = 255
    return noise_img
######添加高斯 noise
def gaussian_noise(img, mean, sigma):
# 将图片灰度标准化
    img = img / 255
    # 产生高斯 noise
    noise = np.random.normal(mean, sigma, img.shape)
    # 将噪声和图片叠加
    gaussian_out = img + noise
    # 将超过 1 的置 1，低于 0 的置 0
    gaussian_out = np.clip(gaussian_out, 0, 1)
    # 将图片灰度范围的恢复为 0-255
    gaussian_out = np.uint8(gaussian_out*255)
    # 将噪声范围搞为 0-255
    # noise = np.uint8(noise*255)
    return gaussian_out# 这里也会返回噪声，注意返回值
###添加随机噪点
def random_noise(image,noise_num):
    '''
    添加随机噪点（实际上就是随机在图像上将像素点的灰度值变为255即白色）
    param image: 需要加噪的图片
    param noise_num: 添加的噪音点数目
    return: img_noise
    '''
    # 参数image：，noise_num：
    img_noise = image
    # cv2.imshow("src", img)
    rows, cols, chn = img_noise.shape
    # 加噪声
    for i in range(noise_num):
        x = np.random.randint(0, rows)#随机生成指定范围的整数
        y = np.random.randint(0, cols)
        img_noise[x, y, :] = 255
    return img_noise
######批量转换操作函数
def convert(input_dir, output_dir,Gaussian_rate,Darker_rate,Bright_rate,Fuse_rate):
    set_all_image = set(os.listdir(input_dir))
###总数  、高斯图片数.....
    image_number = len(set_all_image)
    Gaussian_number = int(image_number*Gaussian_rate)
    Darker_number = int(image_number*Darker_rate)
    Bright_number = int(image_number*Bright_rate)
    Fuse_number = int(image_number*Fuse_rate)
#######guassion_noise
    set_Gaussian = set(random.sample(set_all_image,Gaussian_number))
    for filename in set_Gaussian:
        print(output_dir)
        path = input_dir + "/" + filename # 获取文件路径
        print("Add gaussian noise doing... ", path)
        noise_img = cv2.imread(path)#读取图片
        img_noise = gaussian_noise(noise_img, 0, 0.4) # 高斯噪声
        img_noise = darker(img_noise, percetage=0.1)  #越小越暗
        #img_noise = sp_noise(noise_img,0.02)# 椒盐噪声
        #img_noise  = random_noise(noise_img,500)# 随机噪声
        cv2.imwrite(output_dir+'/'+filename,img_noise )
#######darker_noise        
    set_last2 =  set_all_image.difference(set_Gaussian)
    set_darker = set(random.sample(set_last2,Darker_number))
    for filename in set_darker:
        path = input_dir + "/" + filename # 获取文件路径
        print("darker doing... ", path)
        origin_img = cv2.imread(path)#读取图片
        img_darker = darker(origin_img, percetage=0.1)
        cv2.imwrite(output_dir+'/'+filename,img_darker )
#######brighter   
    set_last3 =  set_last2.difference(set_darker)
    set_bright = set(random.sample(set_last3,Bright_number))
    for filename in set_bright:
        path = input_dir + "/" + filename # 获取文件路径
        print("Add bright noise doing... ", path)
        noise_img = cv2.imread(path)#读取图片
        img_noise  = imgBrightness(noise_img, 2.5, 3)# 随机噪声
        cv2.imwrite(output_dir+'/'+filename,img_noise )
######Fuse_noise
    set_last4 =  set_last3.difference(set_bright)
    # = set(random.sample(set_last4,Fuse_number))
    for filename in set_last4:
        path = input_dir + "/" + filename # 获取文件路径
        print("Add fuseed noise doing... ", path)
        noise_img = cv2.imread(path)#读取图片
        noise_img = gaussian_noise(noise_img, 0, 0.40) # 高斯噪声
        noise_img = sp_noise(noise_img,0.1)# 椒盐噪声
        img_noise  = random_noise(noise_img,500)# 随机噪声
        cv2.imwrite(output_dir+'/'+filename,img_noise )
    print(image_number,Gaussian_number,Darker_number,Bright_number,len(set_last4))
if __name__ == '__main__':
    path = "/data2/houb/data/Birds_256_2022_lower/"
    path_output ="/data2/houb/data/Birds_256_2022_noise4.0/"
    for i in os.listdir(path):
        output_dir = os.path.join(path_output,i)
        os.makedirs(output_dir)
        input_dir = os.path.join(path,i)    # 输入数据文件夹
        convert(input_dir, output_dir,0.05,0.3,0.15,0.5)