Python+OpenCV：阈值分割

红太狼 2022-12-20 14:54 172阅读 0赞

## Python+OpenCV：阈值分割 ##

### 阈值分割 ###

####################################################################################################
    # 阈值分割
    def lmc_cv_thresholding():
        """
            函数功能: 阈值分割.
        """
    
        # 读取图像
        image = lmc_cv.imread('D:/99-Research/Python/Image/Gradient.png')
    
        # 阈值分割
        rows, cols, channel = image.shape
        ret, thresh1_image = lmc_cv.threshold(image, 127, 255, lmc_cv.THRESH_BINARY)
        ret, thresh2_image = lmc_cv.threshold(image, 127, 255, lmc_cv.THRESH_BINARY_INV)
        ret, thresh3_image = lmc_cv.threshold(image, 127, 255, lmc_cv.THRESH_TRUNC)
        ret, thresh4_image = lmc_cv.threshold(image, 127, 255, lmc_cv.THRESH_TOZERO)
        ret, thresh5_image = lmc_cv.threshold(image, 127, 255, lmc_cv.THRESH_TOZERO_INV)
    
        # 显示图像
        pyplot.figure('Image Display')
        titles = ['Original', 'BINARY', 'BINARY_INV', 'TRUNC', 'TOZERO', 'TOZERO_INV']
        images = [image, thresh1_image, thresh2_image, thresh3_image, thresh4_image, thresh5_image]
        for i in range(6):
            pyplot.subplot(2, 3, i + 1)
            pyplot.imshow(images[i], 'gray', vmin=0, vmax=255)
            pyplot.title(titles[i])
            pyplot.xticks([])
            pyplot.yticks([])
        pyplot.show()
    
        # 根据用户输入保存图像
        if ord("q") == (lmc_cv.waitKey(0) & 0xFF):
            # 销毁窗口
            pyplot.close()
        return

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2xpdWJpbmc4NjA5_size_16_color_FFFFFF_t_70][]

### 自适应阈值分割 ###

####################################################################################################
    # 自适应阈值分割
    def lmc_cv_adaptive_thresholding():
        """
            函数功能: 自适应阈值分割.
        """
    
        # 读取图像
        image = lmc_cv.imread('D:/99-Research/Python/Image/Lena.jpg')
        image = lmc_cv.cvtColor(image, lmc_cv.COLOR_BGR2GRAY)
    
        # 自适应阈值分割
        image = lmc_cv.medianBlur(image, 5)
        ret, thresh1_image = lmc_cv.threshold(image, 127, 255, lmc_cv.THRESH_BINARY)
        thresh2_image = lmc_cv.adaptiveThreshold(image, 255, lmc_cv.ADAPTIVE_THRESH_MEAN_C,
                                                 lmc_cv.THRESH_BINARY, 11, 2)
        thresh3_image = lmc_cv.adaptiveThreshold(image, 255, lmc_cv.ADAPTIVE_THRESH_GAUSSIAN_C,
                                                 lmc_cv.THRESH_BINARY, 11, 2)
    
        # 显示图像
        pyplot.figure('Image Display')
        titles = ['Original Image', 'Global Thresholding (v = 127)',
                  'Adaptive Mean Thresholding', 'Adaptive Gaussian Thresholding']
        images = [image, thresh1_image, thresh2_image, thresh3_image]
        for i in range(4):
            pyplot.subplot(2, 2, i + 1)
            pyplot.imshow(images[i], 'gray')
            pyplot.title(titles[i])
            pyplot.xticks([])
            pyplot.yticks([])
        pyplot.show()
    
        # 根据用户输入保存图像
        if ord("q") == (lmc_cv.waitKey(0) & 0xFF):
            # 销毁窗口
            pyplot.close()
        return

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2xpdWJpbmc4NjA5_size_16_color_FFFFFF_t_70 1][]

### 大津法(OTSU)阈值分割 ###

####################################################################################################
    # 大津法(OTSU)阈值分割
    def lmc_cv_otsu_thresholding():
        """
            函数功能: 大津法(OTSU)阈值分割.
        """
    
        # 读取图像
        image = lmc_cv.imread('D:/99-Research/Python/Image/Noise.png')
        image = lmc_cv.cvtColor(image, lmc_cv.COLOR_BGR2GRAY)
    
        # 大津法(OTSU)阈值分割
        sigma = 1
        ksize = (sigma * 3) | 1
        image = lmc_cv.GaussianBlur(image, (ksize, ksize), sigma)
        # global thresholding
        ret1, thresh1_image = lmc_cv.threshold(image, 127, 255, lmc_cv.THRESH_BINARY)
        # Otsu's thresholding
        ret2, thresh2_image = lmc_cv.threshold(image, 0, 255, lmc_cv.THRESH_BINARY + lmc_cv.THRESH_OTSU)
        # Otsu's thresholding after Gaussian filtering
        blur = lmc_cv.GaussianBlur(image, (5, 5), 0)
        ret3, thresh3_image = lmc_cv.threshold(blur, 0, 255, lmc_cv.THRESH_BINARY + lmc_cv.THRESH_OTSU)
    
        # 显示图像
        pyplot.figure('Image Display')
        images = [image, 0, thresh1_image,
                  image, 0, thresh2_image,
                  blur, 0, thresh3_image]
        titles = ['Original Noisy Image', 'Histogram', 'Global Thresholding (v=127)',
                  'Original Noisy Image', 'Histogram', "Otsu's Thresholding",
                  'Gaussian filtered Image', 'Histogram', "Otsu's Thresholding"]
        for i in range(3):
            pyplot.subplot(3, 3, i * 3 + 1)
            pyplot.imshow(images[i * 3], 'gray')
            pyplot.title(titles[i * 3])
            pyplot.xticks([]), pyplot.yticks([])
            pyplot.subplot(3, 3, i * 3 + 2)
            pyplot.hist(images[i * 3].ravel(), 256)
            pyplot.title(titles[i * 3 + 1])
            pyplot.xticks([]), pyplot.yticks([])
            pyplot.subplot(3, 3, i * 3 + 3)
            pyplot.imshow(images[i * 3 + 2], 'gray')
            pyplot.title(titles[i * 3 + 2])
            pyplot.xticks([]), pyplot.yticks([])
        pyplot.show()
    
        # 根据用户输入保存图像
        if ord("q") == (lmc_cv.waitKey(0) & 0xFF):
            # 销毁窗口
            pyplot.close()
        return

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2xpdWJpbmc4NjA5_size_16_color_FFFFFF_t_70 2][]

### 大津法(OTSU)阈值分割的具体实现 ###

本节将演示大津法(OTSU)阈值分割的Python实现，以展示它的实际工作原理。

由于要处理的是双峰图像，大津法(OTSU)试图找到一个阈值(t)，该阈值使类内加权方差最小:

![20201112223048945.png][]

![20201112223204716.png][]

它实际上找到了一个t的值，它位于两个峰之间，使得两个类内的方差都最小。它可以在Python中简单地实现如下:

img = cv.imread('noisy2.png',0)
    blur = cv.GaussianBlur(img,(5,5),0)
    # find normalized_histogram, and its cumulative distribution function
    hist = cv.calcHist([blur],[0],None,[256],[0,256])
    hist_norm = hist.ravel()/hist.sum()
    Q = hist_norm.cumsum()
    bins = np.arange(256)
    fn_min = np.inf
    thresh = -1
    for i in xrange(1,256):
        p1,p2 = np.hsplit(hist_norm,[i]) # probabilities
        q1,q2 = Q[i],Q[255]-Q[i] # cum sum of classes
        if q1 < 1.e-6 or q2 < 1.e-6:
            continue
        b1,b2 = np.hsplit(bins,[i]) # weights
        # finding means and variances
        m1,m2 = np.sum(p1*b1)/q1, np.sum(p2*b2)/q2
        v1,v2 = np.sum(((b1-m1)**2)*p1)/q1,np.sum(((b2-m2)**2)*p2)/q2
        # calculates the minimization function
        fn = v1*q1 + v2*q2
        if fn < fn_min:
            fn_min = fn
            thresh = i
    # find otsu's threshold value with OpenCV function
    ret, otsu = cv.threshold(blur,0,255,cv.THRESH_BINARY+cv.THRESH_OTSU)
    print( "{} {}".format(thresh,ret) )

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2xpdWJpbmc4NjA5_size_16_color_FFFFFF_t_70]: /images/20221120/36a968e51520477cae200c0e221d3b28.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2xpdWJpbmc4NjA5_size_16_color_FFFFFF_t_70 1]: /images/20221120/5e5339bb2dfd48b581d11a4d308fc647.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2xpdWJpbmc4NjA5_size_16_color_FFFFFF_t_70 2]: /images/20221120/96c54f5f4b9d4294a0f262ffbe4d1e74.png
[20201112223048945.png]: /images/20221120/fbea53f972544886beb12d6cb53e066c.png
[20201112223204716.png]: /images/20221120/d16ed2b5c0474a6bb7d166953d954ce1.png