批量去雾的代码

最新推荐文章于 2024-08-08 10:19:28 发布

小龙呮

最新推荐文章于 2024-08-08 10:19:28 发布

阅读量306

点赞数

本文链接：https://blog.youkuaiyun.com/weixin_44472033/article/details/114990231

版权

本文介绍了如何使用Python实现一种批量图像去雾算法，通过应用先进的图像处理技术，有效去除图像中的雾气，提升图像清晰度。内容包括预处理、去雾算法的详细步骤以及后期处理，适合对图像处理感兴趣的读者学习。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

import os
from PIL import Image
import numpy as np

def haze_removal(image, windowSize=24, w0=0.6, t0=0.1):
    darkImage = image.min(axis=2)
    maxDarkChannel = darkImage.max()
    darkImage = darkImage.astype(np.double)

    t = 1 - w0 * (darkImage / maxDarkChannel)
    T = t * 255
    T.dtype = 'uint8'

    t[t < t0] = t0

    J = image
    J[:, :, 0] = (image[:, :, 0] - (1 - t) * maxDarkChannel) / t
    J[:, :, 1] = (image[:, :, 1] - (1 - t) * maxDarkChannel) / t
    J[:, :, 2] = (image[:, :, 2] - (1 - t) * maxDarkChannel) / t
    result = Image.fromarray(J)

    return result

if __name__ == '__main__':
    # 源图片文件夹路径
    dirname_read = "F:/images1/"
    # 新图片文件夹路径
    dirname_write = "F:/images2/"

    names = os.listdir(dirname_read)
    count = 1
    
    for name in names:
        image = np.array(Image.open(dirname_read + name))
        imageSize = image.shape
        result = haze_removal(image)

        to_save_path = dirname_write + name
        result.save(to_save_path)
        print(to_save_path, "------conut：", count)
        count += 1