当前位置:   article > 正文

【youcans 的 OpenCV 例程200篇】161. OTSU 阈值处理算法的实现_从头实现otsu算法

从头实现otsu算法

OpenCV 例程200篇 总目录-202205更新
欢迎关注 『youcans 的 OpenCV 例程 200 篇』 系列,持续更新中


【youcans 的 OpenCV 例程200篇】161. OTSU 阈值处理算法的实现


3.3 全局阈值处理 Otsu 方法

OTSU 方法又称大津算法,使用最大化类间方差(intra-class variance)作为评价准则,基于对图像直方图的计算,可以给出类间最优分离的最优阈值。

任取一个灰度值 T,可以将图像分割为两个集合 F 和 B,集合 F、B 的像素数的占比分别为 pF、pB,集合 F、B 的灰度值均值分别为 mF、mB,图像灰度值为 m,定义类间方差为:
I C V = p F ∗ ( m F − m ) 2 + p B ∗ ( m B − m ) 2 ICV = p_F * (m_F - m)^2 + p_B * (m_B - m)^2 ICV=pF(mFm)2+pB(mBm)2
使类间方差 ICV 最大化的灰度值 T 就是最优阈值。

因此,只要遍历所有的灰度值,就可以得到使 ICV 最大的最优阈值 T。


例程 11.18:OTSU 阈值处理算法的实现

    # 11.18 OTSU 阈值处理算法的实现
    img = cv2.imread("../images/Fig1039a.tif", flags=0)

    histCV = cv2.calcHist([img], [0], None, [256], [0, 256])  # 灰度直方图
    scale = range(256)  # 灰度级 [0,255]
    totalPixels = img.shape[0] * img.shape[1]  # 像素总数
    totalGray = np.dot(histCV[:,0], scale)  # 内积, 总和灰度值
    mG = totalGray / totalPixels  # 平均灰度

    icv = np.zeros(256)
    numFt, sumFt = 0, 0
    for t in range(256):  # 遍历灰度值
        numFt += histCV[t,0]   # F(t) 像素数量
        sumFt += histCV[t,0] * t  # F(t) 灰度值总和
        pF = numFt / totalPixels  # F(t) 像素数占比
        mF = (sumFt/numFt) if numFt>0 else 0  # F(t) 平均灰度
        numBt = totalPixels-numFt  # B(t) 像素数量
        sumBt = totalGray - sumFt  # B(t) 灰度值总和
        pB = numBt / totalPixels  # B(t) 像素数占比
        mB = (sumBt/numBt) if numBt>0 else 0  # B(t) 平均灰度
        icv[t] = pF * (mF-mG)**2 + pB * (mB-mG)**2  # 灰度 t 的类间方差
    maxIcv = max(icv)  # ICV 的最大值
    maxIndex = np.argmax(icv)  # 最大值的索引

    # 阈值处理
    ret, imgBin = cv2.threshold(img, maxIndex, 255, cv2.THRESH_BINARY)  # 以 maxIndex 作为最优阈值
    ret, imgOtsu = cv2.threshold(img, mG, 255, cv2.THRESH_OTSU)  # 阈值分割, OTSU
    print("t(maxICV)={}, retOtsu={}".format(maxIndex, round(ret)))

    plt.figure(figsize=(7,7))
    plt.subplot(221), plt.axis('off'), plt.title("Origin"), plt.imshow(img, 'gray')
    plt.subplot(222, yticks=[]), plt.title("Gray Hist")  # 直方图
    plt.plot(scale, histCV[:,0]/max(histCV))  # 灰度直方图
    plt.plot(scale, icv/maxIcv)  # 类间方差图
    plt.subplot(223), plt.title("global binary(T={})".format(maxIndex)), plt.axis('off')
    plt.imshow(imgBin, 'gray')
    plt.subplot(224), plt.title("OTSU binary(T={})".format(round(ret))), plt.axis('off')
    plt.imshow(imgOtsu, 'gray')
    plt.tight_layout()
    plt.show()
  • 1
  • 2
  • 3
  • 4
  • 5
  • 6
  • 7
  • 8
  • 9
  • 10
  • 11
  • 12
  • 13
  • 14
  • 15
  • 16
  • 17
  • 18
  • 19
  • 20
  • 21
  • 22
  • 23
  • 24
  • 25
  • 26
  • 27
  • 28
  • 29
  • 30
  • 31
  • 32
  • 33
  • 34
  • 35
  • 36
  • 37
  • 38
  • 39
  • 40

在这里插入图片描述



(本节完)


版权声明:

youcans@xupt 原创作品,转载必须标注原文链接:(https://blog.csdn.net/youcans/article/details/124281291)

Copyright 2022 youcans, XUPT
Crated:2022-4-18


欢迎关注 『youcans 的 OpenCV 例程 200 篇』 系列,持续更新中
欢迎关注 『youcans 的 OpenCV学习课』 系列,持续更新中

【youcans 的 OpenCV 例程200篇】158. 阈值处理之固定阈值法
【youcans 的 OpenCV 例程200篇】159. 图像分割之全局阈值处理
【youcans 的 OpenCV 例程200篇】160. 图像处理之OTSU 方法
【youcans 的 OpenCV 例程200篇】161. OTSU 阈值处理算法的实现
【youcans 的 OpenCV 例程200篇】162. 全局阈值处理改进方法
【youcans 的 OpenCV 例程200篇】163. 基于边缘信息改进全局阈值处理
【youcans 的 OpenCV 例程200篇】164.使用 Laplace 边缘信息改进全局阈值处理
【youcans 的 OpenCV 例程200篇】165.多阈值 OTSU 处理方法
【youcans 的 OpenCV 例程200篇】166.自适应阈值处理
【youcans 的 OpenCV 例程200篇】167.基于移动平均的可变阈值处理
更多内容,请见:
【OpenCV 例程200篇 总目录-202206更新】

声明:本文内容由网友自发贡献,不代表【wpsshop博客】立场,版权归原作者所有,本站不承担相应法律责任。如您发现有侵权的内容,请联系我们。转载请注明出处:https://www.wpsshop.cn/w/小小林熬夜学编程/article/detail/481139
推荐阅读
相关标签
  

闽ICP备14008679号