opencv阈值分割之直方图（分割）技术法和OTSU

import cv2 as cv
import numpy as np
from matplotlib import pyplot as plt


#计算灰度直方图
def calcGrayHist(grayimage):
    #灰度图像矩阵的高，宽
    rows, cols = grayimage.shape
    print(grayimage.shape)
    #存储灰度直方图
    grayHist = np.zeros([256],np.uint64)
    for r in range(rows):
        for c in range(cols):
            grayHist[grayimage[r][c]] += 1
    return grayHist

#OTSU自动阈值分割
def OTSU(image):
    
    if len(image.shape) == 2:
        gray = image
    else:
        gray = cv.cvtColor(image, cv.COLOR_BGR2GRAY)
    rows,cols = gray.shape
    #1.计算灰度直方图
    grayHist = calcGrayHist(gray)
    #2.灰度直方图归一化
    uniformGrayHist = grayHist/float(rows*cols)
    #3.计算零阶累计矩何一阶累计矩
    zeroCumuMoment = np.zeros([256],np.float32)
    oneCumuMoment = np.zeros([256],np.float32)
    for k in range(256):
        if k == 0:
            zeroCumuMoment[k] = uniformGrayHist[0]
            oneCumuMoment[k] = (k)*uniformGrayHist[0]
        else:
            zeroCumuMoment[k] = zeroCumuMoment[k-1] + uniformGrayHist[k]
            oneCumuMoment[k] = oneCumuMoment[k-1] + k*uniformGrayHist[k]
    #计算类间方差
    variance = np.zeros([256],np.float32)
    for k in range(255):
        if zeroCumuMoment[k] == 0 or zeroCumuMoment[k] == 1:
            variance[k] = 0
        else:
            variance[k] = math.pow(oneCumuMoment[255]*zeroCumuMoment[k] -
            oneCumuMoment[k],2)/(zeroCumuMoment[k]*(1.0-zeroCumuMoment[k]))
    #找到阈值、
    threshLoc = np.where(variance[0:255] == np.max(variance[0:255]))
    thresh = threshLoc[0][0]
    #阈值处理
    threshold = np.copy(gray)
    threshold[threshold > thresh] = 255
    threshold[threshold <= thresh] = 0
    return threshold, thresh


#阈值分割：直方图技术法 
def threshTwoPeaks(image):
    if len(image.shape) == 2:
        gray = image
    else:
        gray = cv.cvtColor(image, cv.COLOR_BGR2GRAY)
        print(666666)
    #计算灰度直方图
    histogram = calcGrayHist(gray)
    #寻找灰度直方图的最大峰值对应的灰度值
    maxLoc = np.where(histogram==np.max(histogram))
    firstPeak = maxLoc[0][0]
    #寻找灰度直方图的第二个峰值对应的灰度值
    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]

    #找到两个峰值之间的最小值对应的灰度值，作为阈值
    thresh = 0
    if firstPeak > secondPeak:#第一个峰值再第二个峰值的右侧
        temp = histogram[int(secondPeak):int(firstPeak)]
        minloc = np.where(temp == np.min(temp))
        thresh = secondPeak + minloc[0][0] + 1
    else:#第一个峰值再第二个峰值的左侧
        temp = histogram[int(firstPeak):int(secondPeak)]
        minloc = np.where(temp == np.min(temp))
        thresh =firstPeak + minloc[0][0] + 1

    #找到阈值之后进行阈值处理，得到二值图
    threshImage_out = gray.copy()
    #大于阈值的都设置为255
    threshImage_out[threshImage_out > thresh] = 255
    threshImage_out[threshImage_out <= thresh] = 0
    return thresh, threshImage_out
    
def threshTwoPeaks(image):
    if len(image.shape) == 2:
        gray = image
    else:
        gray = cv.cvtColor(image, cv.COLOR_BGR2GRAY)
        print(666666)
    #计算灰度直方图
    histogram = calcGrayHist(gray)
    #寻找灰度直方图的最大峰值对应的灰度值
    maxLoc = np.where(histogram==np.max(histogram))
    firstPeak = maxLoc[0][0]
    #寻找灰度直方图的第二个峰值对应的灰度值
    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]

    #找到两个峰值之间的最小值对应的灰度值，作为阈值
    thresh = 0
    if firstPeak > secondPeak:#第一个峰值再第二个峰值的右侧
        temp = histogram[int(secondPeak):int(firstPeak)]
        minloc = np.where(temp == np.min(temp))
        thresh = secondPeak + minloc[0][0] + 1
    else:#第一个峰值再第二个峰值的左侧
        temp = histogram[int(firstPeak):int(secondPeak)]
        minloc = np.where(temp == np.min(temp))
        thresh =firstPeak + minloc[0][0] + 1

    #找到阈值之后进行阈值处理，得到二值图
    threshImage_out = gray.copy()
    #大于阈值的都设置为255
    threshImage_out[threshImage_out > thresh] = 255
    #小于阈值的都设置为0
    threshImage_out[threshImage_out <= thresh] = 0
    return thresh, threshImage_out
if __name__ == "__main__":
    img = cv.imread('./123.png')
    kkk,kkkk = threshTwoPeaks(img)
    print(kkk)
    cv.imshow('66',kkkk)
    cv.waitKey(0)

结果如下：
原图：

自适应阈值分割后的二值图：

可见，直方图阈值分割计数法能够较为有效的将背景何前景区分开来，比较完整的分割出图片中的目标物体。值得一提的是，对于任何一张图像，它的直方图中如果存在较为明显的双峰，用直方图分割技术法可以达到很好的效果，否则，达到的效果会很不理想.