AHE算法python实现

AHE（自适应直方图均衡化）算法是一种用于增强图像对比度的方法。以下是Python实现AHE算法的示例代码：

import cv2
import numpy as np

# 自适应直方图均衡化函数
def adaptive_histogram_equalization(img, clip_limit=2.0, tile_size=8):
    # 将图像转为灰度图
    img_gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    img_height, img_width = img_gray.shape

    # 计算每个tile的大小
    tile_rows = int(np.ceil(img_height / tile_size))
    tile_cols = int(np.ceil(img_width / tile_size))

    # 计算clip_limit
    map_size = tile_rows * tile_cols
    clip_limit = int(clip_limit * map_size / 256)

    # 对每个tile进行直方图均衡化
    map_array = np.zeros((tile_rows, tile_cols, 256), dtype=np.uint32)
    for i in range(tile_rows):
        for j in range(tile_cols):
            x1, y1 = i * tile_size, j * tile_size
            x2, y2 = min(x1 + tile_size, img_height), min(y1 + tile_size, img_width)
            tile = img_gray[x1:x2, y1:y2]
            hist, _ = np.histogram(tile.flatten(), bins=256, range=(0, 255))
            cdf = np.cumsum(hist)
            cdf = (cdf * 255) // cdf[-1]
            map_array[i, j, :] = cdf

    # 对整张图像进行映射
    img_out = np.zeros_like(img_gray)
    for i in range(tile_rows):
        for j in range(tile_cols):
            x1, y1 = i * tile_size, j * tile_size
            x2, y2 = min(x1 + tile_size, img_height), min(y1 + tile_size, img_width)
            tile = img_gray[x1:x2, y1:y2]
            cdf = map_array[i, j, :]
            cdf_min = np.min(cdf)
            cdf[cdf < cdf_min + clip_limit] = cdf_min + clip_limit
            cdf[cdf > 255 - clip_limit] = 255 - clip_limit
            cdf = ((cdf - cdf_min) * 255) // (256 - 2 * clip_limit)
            img_out[x1:x2, y1:y2] = cdf[tile]

    return img_out

使用方法：

img = cv2.imread('input_image.jpg')
img_out = adaptive_histogram_equalization(img, clip_limit=3.0, tile_size=16)
cv2.imshow('Input Image', img)
cv2.imshow('Output Image', img_out)
cv2.waitKey(0)

其中，`clip_limit`是限制像素值范围的参数，`tile_size`是每个tile的大小。