opencv自动阈值分割

柔光的暖阳◎ 2023-01-10 13:25 270阅读 0赞

#自适应阈值与固定阈值对比
    import cv2
    import matplotlib.pyplot as plt
    img = cv2.imread('test.jpg',0)
    ret,th1 = cv2.threshold(img,127,255,cv2.THRESH_BINARY)
    th2 = cv2.adaptiveThreshold(img,255,cv2.ADAPTIVE_THRESH_MEAN_C,cv2.THRESH_BINARY,55,4)
    th3 = cv2.adaptiveThreshold(img,255,cv2.ADAPTIVE_THRESH_GAUSSIAN_C,cv2.THRESH_BINARY,17,6)
    titles =['Original','Global(v=127)','Adaptive Mean','Adaptive Gaussian']
    images = [img,th1,th2,th3]
    for i in range(4):
        plt.subplot(2,2,i+1),plt.imshow(images[i],'gray')
        plt.title(titles[i],fontsize=8)
        plt.xticks([]),plt.yticks([])
    plt.show()

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dpbmdneW4_size_16_color_FFFFFF_t_70][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dpbmdneW4_size_16_color_FFFFFF_t_70 1][]

迭代法阈值分割：

步骤：

1求出图象中的最大灰度值和最小灰度值，分别记为ZMAX和ZMIN，令初始阈值T0=(ZMAX+ZMIN)/2;

2根据阈值TK将图像分割为前景和背景，分别求出两者的平均灰度值ZO和ZB；

3求出新阈值TK+1=(Z0+ZB)/2;

4若TK==TK+1，则所得即为阈值；否则转2，迭代计算；

5使用计算后的阈值进行固定阈值分割。

大津阈值 也叫最大类间方差法，是一种基于全局阈值的自适应方法。图像分为前景和背景。当取最佳阈值时候，两部分之间的差别应该是最大的，衡量差别的标准为最大类间方差。直方图有两个峰值得图像，大津法求得的T近似于两个峰值之间的低谷。

OTSU大津法：

符号说明：

T:图像(x,y)前景和背景的分割阈值；

![\\omega \_\{1\}][omega _1]：前景像素点数占整幅图像的比例，其平均灰度![\\mu \_\{1\}][mu _1]；

![\\omega \_\{2\}][omega _2]：背景像素点数占整幅图像的比例，其平均灰度![\\mu \_\{2\}][mu _2]；

![\\mu][mu]：图像的总平均灰度；

g:类间方差；

![N\_\{1\}][N_1]：设图像的大小为![M\\times N][M_times N]，图像中像素的灰度值小于阈值T的像素个数；

![N\_\{2\}][N_2]：像素灰度大于阈值T 的像素个数。

![g=\\omega \_\{1\}\\times (\\mu -\\mu \_\{1\})^\{2\}+\\omega \_\{2\}\\times (\\mu -\\mu \_\{2\})^\{2\}][g_omega _1_times _mu -_mu _1_2_omega _2_times _mu -_mu _2_2]

等价公式：

![g=\\omega \_\{1\}\\times \\omega \_\{2\}\\times (\\mu \_\{1\}-\\mu \_\{2\})^\{2\}][g_omega _1_times _omega _2_times _mu _1_-_mu _2_2]

import cv2
    from matplotlib import pyplot as plt
    img = cv2.imread('test.jpg',0)
    #固定阈值法
    ret,th1 = cv2.threshold(img,100,255,cv2.THRESH_BINARY)
    #OTSU阈值法
    ret,th2 = cv2.threshold(img,0,255,cv2.THRESH_BINARY+cv2.THRESH_OTSU)
    #先进行高斯滤波，再使用Otsu阈值法
    blur = cv2.GaussianBlur(img,(5,5),0)
    ret3,th3 = cv2.threshold(blur,0,255,cv2.THRESH_BINARY+cv2.THRESH_OTSU)
    images = [img,0,th1,img,0,th2,blur,0,th3]
    titles = ['Original','Histogram','Global(v=100)','Original','Histogram',"Otsu's","Gaussian filtered Image","Histogram","Otsu's"]
    for i in range(3):
        #绘制原图
        plt.subplot(3,3,i*3+1)
        plt.imshow(images[i*3],'gray')
        plt.title(titles[i*3],fontsize=8)
        plt.xticks([]),plt.yticks([])
        #绘制直方图，plt,hist,ravel函数将数组降成一维
        plt.subplot(3,3,i*3+2)
        plt.hist(images[i*3].ravel(),256)
        plt.title(titles[i*3+1],fontsize=8)
        plt.xticks([]),plt.yticks([])
        #绘制阈值图
        plt.subplot(3,3,i*3+3)
        plt.imshow(images[i*3+2],'gray')
        plt.title(titles[i*3+2],fontsize=8)
        plt.xticks([]),plt.yticks([])
    plt.show()

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dpbmdneW4_size_16_color_FFFFFF_t_70 2][]

test.jpg:

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dpbmdneW4_size_16_color_FFFFFF_t_70 3][]

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dpbmdneW4_size_16_color_FFFFFF_t_70]: /images/20221119/9710b6dac5e644539d98bc6faf3bad6e.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dpbmdneW4_size_16_color_FFFFFF_t_70 1]: /images/20221119/995f5e48beb645739c174a4d6210a134.png
[omega _1]: https://private.codecogs.com/gif.latex?%5Cfn_cm%20%5Comega%20_%7B1%7D
[mu _1]: https://private.codecogs.com/gif.latex?%5Cfn_cm%20%5Cmu%20_%7B1%7D
[omega _2]: https://private.codecogs.com/gif.latex?%5Cfn_cm%20%5Comega%20_%7B2%7D
[mu _2]: https://private.codecogs.com/gif.latex?%5Cfn_cm%20%5Cmu%20_%7B2%7D
[mu]: https://private.codecogs.com/gif.latex?%5Cfn_cm%20%5Cmu
[N_1]: https://private.codecogs.com/gif.latex?%5Cfn_cm%20N_%7B1%7D
[M_times N]: https://private.codecogs.com/gif.latex?%5Cfn_cm%20M%5Ctimes%20N
[N_2]: https://private.codecogs.com/gif.latex?%5Cfn_cm%20N_%7B2%7D
[g_omega _1_times _mu -_mu _1_2_omega _2_times _mu -_mu _2_2]: https://private.codecogs.com/gif.latex?%5Cfn_cm%20g%3D%5Comega%20_%7B1%7D%5Ctimes%20%28%5Cmu%20-%5Cmu%20_%7B1%7D%29%5E%7B2%7D&plus;%5Comega%20_%7B2%7D%5Ctimes%20%28%5Cmu%20-%5Cmu%20_%7B2%7D%29%5E%7B2%7D
[g_omega _1_times _omega _2_times _mu _1_-_mu _2_2]: https://private.codecogs.com/gif.latex?%5Cfn_cm%20g%3D%5Comega%20_%7B1%7D%5Ctimes%20%5Comega%20_%7B2%7D%5Ctimes%20%28%5Cmu%20_%7B1%7D-%5Cmu%20_%7B2%7D%29%5E%7B2%7D
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dpbmdneW4_size_16_color_FFFFFF_t_70 2]: https://img-blog.csdnimg.cn/20210124172514267.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dpbmdneW4=,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dpbmdneW4_size_16_color_FFFFFF_t_70 3]: https://img-blog.csdnimg.cn/2021012417264413.jpg?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dpbmdneW4=,size_16,color_FFFFFF,t_70