图像阈值处理-OpenCV

简单的阈值处理：

cv.threshold(img,threshold_value,max_value,model)
img：输入图像
threshold_value：设置阈值
max：设置的最大图像值，当原图像某点的值超高阈值既被赋予这个值
model：阈值的方式：

threshold_type=cv.THRESH_BINARY:
dst(x,y) = max_value, if src(x,y)>threshold 0, otherwise.
既大于阈值被赋予max,小于阈值被赋予0

threshold_type=cv.THRESH_BINARY_INV:
dst(x,y) = 0, if src(x,y)>threshold; dst(x,y) = max_value, otherwise.
既大于阈值被赋予0,小于阈值被赋予max

threshold_type=cv.THRESH_TRUNC:
dst(x,y) = threshold, if src(x,y)>threshold;   dst(x,y) = src(x,y), otherwise.
大于阈值则被赋予阈值，其他不变

threshold_type=cv.THRESH_TOZERO:
dst(x,y) = src(x,y), if (x,y)>threshold ;  dst(x,y) = 0, otherwise.
小于阈值被赋予0，其他不变

threshold_type=cv.THRESH_TOZERO_INV:
dst(x,y) = 0, if src(x,y)>threshold ;  dst(x,y) = src(x,y), otherwise
大于阈值被赋予0，其他不变

实例：

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

img = cv2.imread(r'XXXXX.jpg')
img = cv2.resize(img, (0, 0), fx=0.1, fy=0.1, interpolation=cv2.INTER_NEAREST)
img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
rows = img.shape[0]
cols = img.shape[1]

ret,thresh1 = cv.threshold(img,127,255,cv.THRESH_BINARY)
ret,thresh2 = cv.threshold(img,127,255,cv.THRESH_BINARY_INV)
ret,thresh3 = cv.threshold(img,127,255,cv.THRESH_TRUNC)
ret,thresh4 = cv.threshold(img,127,255,cv.THRESH_TOZERO)
ret,thresh5 = cv.threshold(img,127,255,cv.THRESH_TOZERO_INV)
titles = ['Original Image','BINARY','BINARY_INV','TRUNC','TOZERO','TOZERO_INV']
images = [img, thresh1, thresh2, thresh3, thresh4, thresh5]
for i in range(6):
    plt.subplot(2,3,i+1),plt.imshow(images[i],'gray',vmin=0,vmax=255)
    plt.title(titles[i])
    plt.xticks([]),plt.yticks([])
plt.show()

 自适应阈值处理：

自适应阈值处理用到了函数：cv.adaptiveThreshold（），有两种方法：cv.ADAPTIVE_THRESH_MEAN_C（平均）和cv.ADAPTIVE_THRESH_GAUSSIAN_C（高斯）。

blocksize为一次处理的面积，C为偏置，以平均方法为例，每个点的值会通过以这点为中心的blocksize大小的平均值减去C来确定。

如果使用高斯的方法，则（x,y）周围的像素的权值则根据其到中心点的距离通过高斯方程得到。

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

img = cv2.imread(r'\chess.jpg')
img = cv2.resize(img, (0, 0), fx=2, fy=2, interpolation=cv2.INTER_NEAREST)
img_gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
rows = img.shape[0]
cols = img.shape[1]

ret,thresh1 = cv.threshold(img_gray,127,255,cv.THRESH_BINARY)

maxVal = 255;
blockSize = 201 ;
C = 0;
th2 = cv.adaptiveThreshold(img_gray, maxVal, cv.ADAPTIVE_THRESH_MEAN_C, cv.THRESH_BINARY, blockSize, C);
th3 = cv.adaptiveThreshold(img_gray,maxVal,cv.ADAPTIVE_THRESH_GAUSSIAN_C,cv.THRESH_BINARY,blockSize,C)

titles = ['Original Image', 'Global Thresholding (v = 127)',
            'Adaptive Mean Thresholding', 'Adaptive Gaussian Thresholding']
images = [img_gray, thresh1, th2, th3]
for i in range(4):
    plt.subplot(2,2,i+1),plt.imshow(images[i],'gray')
    plt.title(titles[i])
    plt.xticks([]),plt.yticks([])
plt.show()

 Otsu的二值化

ret2,th2 = cv.threshold(img_gray,0,255,cv.THRESH_BINARY+cv.THRESH_OTSU)

Otsu的算法试图找到一个阈值（t），使加权的类内方差最小。

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

img = cv2.imread(r'XXXX.jpg')
img = cv2.resize(img, (0, 0), fx=1, fy=1, interpolation=cv2.INTER_NEAREST)
img_gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
rows = img.shape[0]
cols = img.shape[1]

ret,thresh1 = cv.threshold(img_gray,127,255,cv.THRESH_BINARY)

maxVal = 255;
blockSize = 201 ;
C = 0;
ret2,th2 = cv.threshold(img_gray,0,255,cv.THRESH_BINARY+cv.THRESH_OTSU)

th3 = cv.adaptiveThreshold(img_gray,maxVal,cv.ADAPTIVE_THRESH_MEAN_C,cv.THRESH_BINARY,blockSize,C)
titles = ['Original Image', 'Global Thresholding (v = 127)',
            'OTSU', 'Adaptive Mean Thresholding']
images = [img_gray, thresh1, th2, th3]
for i in range(4):
    plt.subplot(2,2,i+1),plt.imshow(images[i],'gray')
    plt.title(titles[i])
    plt.xticks([]),plt.yticks([])
plt.show()