图像直方图处理技巧-优快云博客

本文链接：https://blog.youkuaiyun.com/l_Flipped_/article/details/116585260

文章目录

前言
一、直方图是什么？
二、使用步骤
- 1.引入库
- 2.读入数据
- 3.代码解释分析
- 4.代码展示效果
总结

前言

直方图是图像处理过程中的一种非常重要的分析工具。直方图从图像内部灰度级的角度对图像进行表述，包含十分丰富而重要的信息。从直方图的角度对图像进行处理，可以达到增强图像显示效果的目的。

一、直方图是什么？

从统计的角度讲，直方图是图像内灰度值的统计特性与图像灰度值之间的函数，直方图统计图像内各个灰度级出现的次数。从直方图的图形上观察，横坐标是图像中各像素点的灰度级，纵坐标是具有灰度级（像素值）的像素个数。

二、使用步骤

1.引入库

代码如下：

import cv2
import numpy as np

2.读入数据

代码如下：

import cv2
import numpy as np

def calcGrayHist(grayimage):

    rows,cols = grayimage.shape

    print(grayimage.shape)

    grayHist = np.zeros([256],np.uint64)

    for r in range(rows):

        for c in range(cols):

            grayHist[grayimage[r][c]] += 1

    return grayHist


def threshTwoPeaks(image):

    if len(image.shape) == 2:

        gray = image

    else:

        gray = cv2.cvtColor(image,cv2.COLOR_BGR2GRAY)

        print(666666)

    histogram = calcGrayHist(gray)

    maxLoc = np.where(histogram==np.max(histogram))

    firstPeak = maxLoc[0][0]

    measureDists = np.zeros([256],np.float32)

    for k in range(256):

        measureDists[k] = pow(k-firstPeak,2)*histogram[k]

    maxLoc2 = np.where(measureDists==np.max(measureDists))

    secondPeak = maxLoc2[0][0]

    thresh = 0

    if firstPeak > secondPeak:

        temp = histogram[int(secondPeak):int(firstPeak)]

        minloc = np.where(temp == np.min(temp))

        thresh = secondPeak + minloc[0][0] + 1

    else:

        temp = histogram[int(firstPeak):int(secondPeak)]

        minloc = np.where(temp == np.min(temp))

        thresh = firstPeak + minloc[0][0] + 1

    threshImage_out = gray.copy()

    threshImage_out[threshImage_out > thresh] = 255

    threshImage_out[threshImage_out <= thresh] = 0

    return thresh, threshImage_out


if __name__ == "__main__":

    img = cv2.imread("mao.jpg")

    thresh,threshImage_out = threshTwoPeaks(img)

    print(thresh)

    cv2.imshow("threshImage_out",threshImage_out)

    cv2.waitKey(0)

    cv2.destroyAllWindows()