Python 2.x 中两个图像的直方图匹配？
声明：本页面是StackOverFlow热门问题的中英对照翻译，遵循CC BY-SA 4.0协议，如果您需要使用它，必须同样遵循CC BY-SA许可，注明原文地址和作者信息，同时你必须将它归于原作者（不是我）：StackOverFlow
原文地址: http://stackoverflow.com/questions/32655686/
Warning: these are provided under cc-by-sa 4.0 license. You are free to use/share it, But you must attribute it to the original authors (not me):
StackOverFlow
Histogram matching of two images in Python 2.x?
提问by anon01
I'm trying to match the histograms of two images (in MATLAB this could be done using imhistmatch). Is there an equivalent function available from a standard Python library? I've looked at OpenCV, scipy, and numpy but don't see any similar functionality.
我正在尝试匹配两个图像的直方图（在 MATLAB 中，这可以使用 完成imhistmatch）。标准 Python 库中是否有可用的等效函数？我看过 OpenCV、scipy 和 numpy，但没有看到任何类似的功能。
采纳答案by ali_m
I previously wrote an answer hereexplaining how to do piecewise linear interpolation on an image histogram in order to enforce particular ratios of highlights/midtones/shadows.
我之前在这里写了一个答案，解释了如何对图像直方图进行分段线性插值，以强制执行特定的高光/中间调/阴影比率。
The same basic principles underlie histogram matchingbetween two images. Essentially you compute the cumulative histograms for your source and template images, then interpolate linearly to find the unique pixel values in the template image that most closely match the quantiles of the unique pixel values in the source image:
相同的基本原理是两幅图像之间直方图匹配的基础。本质上，您计算源图像和模板图像的累积直方图，然后线性插值以找到模板图像中与源图像中唯一像素值的分位数最匹配的唯一像素值：
import numpy as np
def hist_match(source, template):
"""
Adjust the pixel values of a grayscale image such that its histogram
matches that of a target image
Arguments:
-----------
source: np.ndarray
Image to transform; the histogram is computed over the flattened
array
template: np.ndarray
Template image; can have different dimensions to source
Returns:
-----------
matched: np.ndarray
The transformed output image
"""
oldshape = source.shape
source = source.ravel()
template = template.ravel()
# get the set of unique pixel values and their corresponding indices and
# counts
s_values, bin_idx, s_counts = np.unique(source, return_inverse=True,
return_counts=True)
t_values, t_counts = np.unique(template, return_counts=True)
# take the cumsum of the counts and normalize by the number of pixels to
# get the empirical cumulative distribution functions for the source and
# template images (maps pixel value --> quantile)
s_quantiles = np.cumsum(s_counts).astype(np.float64)
s_quantiles /= s_quantiles[-1]
t_quantiles = np.cumsum(t_counts).astype(np.float64)
t_quantiles /= t_quantiles[-1]
# interpolate linearly to find the pixel values in the template image
# that correspond most closely to the quantiles in the source image
interp_t_values = np.interp(s_quantiles, t_quantiles, t_values)
return interp_t_values[bin_idx].reshape(oldshape)
For example:
例如：
from matplotlib import pyplot as plt
from scipy.misc import lena, ascent
source = lena()
template = ascent()
matched = hist_match(source, template)
def ecdf(x):
"""convenience function for computing the empirical CDF"""
vals, counts = np.unique(x, return_counts=True)
ecdf = np.cumsum(counts).astype(np.float64)
ecdf /= ecdf[-1]
return vals, ecdf
x1, y1 = ecdf(source.ravel())
x2, y2 = ecdf(template.ravel())
x3, y3 = ecdf(matched.ravel())
fig = plt.figure()
gs = plt.GridSpec(2, 3)
ax1 = fig.add_subplot(gs[0, 0])
ax2 = fig.add_subplot(gs[0, 1], sharex=ax1, sharey=ax1)
ax3 = fig.add_subplot(gs[0, 2], sharex=ax1, sharey=ax1)
ax4 = fig.add_subplot(gs[1, :])
for aa in (ax1, ax2, ax3):
aa.set_axis_off()
ax1.imshow(source, cmap=plt.cm.gray)
ax1.set_title('Source')
ax2.imshow(template, cmap=plt.cm.gray)
ax2.set_title('template')
ax3.imshow(matched, cmap=plt.cm.gray)
ax3.set_title('Matched')
ax4.plot(x1, y1 * 100, '-r', lw=3, label='Source')
ax4.plot(x2, y2 * 100, '-k', lw=3, label='Template')
ax4.plot(x3, y3 * 100, '--r', lw=3, label='Matched')
ax4.set_xlim(x1[0], x1[-1])
ax4.set_xlabel('Pixel value')
ax4.set_ylabel('Cumulative %')
ax4.legend(loc=5)
For a pair of RGB images you could apply this function separately to each channel. Depending on the effect you are trying to achieve, you might want to transform the images into a different color space first. For example, you could transform into HSV spaceand then do matching just on the V channel if you wanted to match lightness, but not hue or saturation.
对于一对 RGB 图像，您可以将此功能单独应用于每个通道。根据您尝试实现的效果，您可能希望首先将图像转换为不同的色彩空间。例如，如果你想匹配亮度，而不是色调或饱和度，你可以转换到HSV 空间，然后只在 V 通道上进行匹配。
回答by Sandipan Dey
Here is yet another implementation based on thisand scikit-image exposure's cumulative_distributionfunction which uses np.interpsimilar to ali_m's implementation. The input and template image is assumed to be gray-scale and with pixel values as integers in [0,255].
这是另一种实现基于此和scikit-image exposure的cumulative_distribution功能，它采用np.interp类似于ali_m的实现。假设输入和模板图像是灰度的，像素值为 [0,255] 中的整数。
from skimage.exposure import cumulative_distribution
import matplotlib.pylab as plt
import numpy as np
def cdf(im):
'''
computes the CDF of an image im as 2D numpy ndarray
'''
c, b = cumulative_distribution(im)
# pad the beginning and ending pixels and their CDF values
c = np.insert(c, 0, [0]*b[0])
c = np.append(c, [1]*(255-b[-1]))
return c
def hist_matching(c, c_t, im):
'''
c: CDF of input image computed with the function cdf()
c_t: CDF of template image computed with the function cdf()
im: input image as 2D numpy ndarray
returns the modified pixel values
'''
pixels = np.arange(256)
# find closest pixel-matches corresponding to the CDF of the input image, given the value of the CDF H of
# the template image at the corresponding pixels, s.t. c_t = H(pixels) <=> pixels = H-1(c_t)
new_pixels = np.interp(c, c_t, pixels)
im = (np.reshape(new_pixels[im.ravel()], im.shape)).astype(np.uint8)
return im
The output is shown below:
输出如下所示：
回答by abk
I would like to add a minor addition to both the solution written above. If somebody plans to make this as a global function (such as for grayscale images), it would be a good idea to convert the final matched array into its corresponding format (numpy.uint8). This might help in future image conversions without creating conflicts.
我想对上面写的两个解决方案添加一个小补充。如果有人计划将其作为全局函数（例如用于灰度图像），最好将最终匹配的数组转换为其相应的格式 (numpy.uint8)。这可能有助于未来的图像转换而不会产生冲突。
def hist_norm(source, template):
olddtype = source.dtype
oldshape = source.shape
source = source.ravel()
template = template.ravel()
s_values, bin_idx, s_counts = np.unique(source, return_inverse=True,
return_counts=True)
t_values, t_counts = np.unique(template, return_counts=True)
s_quantiles = np.cumsum(s_counts).astype(np.float64)
s_quantiles /= s_quantiles[-1]
t_quantiles = np.cumsum(t_counts).astype(np.float64)
t_quantiles /= t_quantiles[-1]
interp_t_values = np.interp(s_quantiles, t_quantiles, t_values)
interp_t_values = interp_t_values.astype(olddtype)
return interp_t_values[bin_idx].reshape(oldshape)
