opencv+python用直方图进行相似度判断、对比

1、直方图作为一种常用的方法，经常用在数据分析和图片处理过程，

采用直方图对比图片相似性，简单明了直观。

根据官网函数说明：

# compareHist(H1, H2, method) -> retval# @param H1 First compared histogram.# @param H2 Second compared histogram of the same size as H1.# @param method Comparison method, see cv::HistCompMethods

HistCompMethods：

HISTCMP_BHATTACHARYYA

Correlation ( HISTCMP_CORREL)相关性，

Chi-Square ( HISTCMP_CHISQR) 卡方，

Intersection ( HISTCMP_INTERSECT )交集法，

Bhattacharyya distance ( HISTCMP_BHATTACHARYYA)常态分布比对的Bhattacharyya距离法。

2、import cv2  as cvimport  numpy as npimport copyfrom matplotlib import pyplot as pltimage = cv.imread('c:\\meiping1.png')cv.imshow("image", image)pic = cv.imread('c:\\meiping4.png')cv.imshow("pic", pic)

两幅图有一定相似，但大有不同。

另外注意 必须大小一样，否则不行！

3、# 计算图1的直方图 然后分别归一化

histGrayImage = cv.calcHist([image], [1], None, [256], [0, 256])

cv.normalize(histGrayImage, histGrayImage,0,255*0.9,cv.NORM_MINMAX)

# 计算图2的直方图然后分别归一化

histGrayPic = cv.calcHist([pic], [1], None, [256], [0, 256])

cv.normalize(histGrayPic, histGrayPic,0,255*0.9,cv.NORM_MINMAX)

4、显示直方图

plt.subplot(2, 1, 1)

plt.plot(histGrayImage)

plt.subplot(2, 1, 2)

plt.plot(histGrayPic)

plt.show()

5、retval1  = cv.compareHist(histGrayImage, histGrayPic, cv.HISTCMP_BHATTACHARYYA)

print("retval1: {}".format(retval1))

retval2  = cv.compareHist(histGrayImage, histGrayPic, cv.HISTCMP_CORREL)

print("retval2: {}".format(retval2))

retval3  = cv.compareHist(histGrayImage, histGrayPic, cv.HISTCMP_CHISQR)

print("retval3: {}".format(retval3))

cv.waitKey(0)

6、从计算可以看出 相似度还是有一些的!

比想象的相似度要高.