【图像处理opencv】_阈值分割

目录

0 程序环境及所学函数

1 直方图阈值

2 三角法阈值

3 迭代法阈值及实现

4 大津法（otsu）阈值及实现

5 自适应阈值及实现

---

0 程序环境及所学函数

本章程序运行需要导入下面三个库，并定义了一个显示图像的函数

import cv2 as cv
import numpy as np
import matplotlib.pyplot as plt
 
def show(img):
    if img.ndim == 2:
        plt.imshow(img, cmap='gray')
    else:
        plt.imshow(cv.cvtColor(img, cv.COLOR_BGR2RGB))
    plt.show()

所学函数

直方图阈值

plt.hist(img.ravel(), 256, [0,256])
cv.threshold(img, th, 255, cv.THRESH_BINARY)

三角法阈值

th, img_bin = cv.threshold(img, th, 255, cv.THRESH_TRIANGLE)

大津法阈值

th, img_bin = cv.threshold(img, th, 255, cv.THRESH_OTSU)

自适应阈值分割

img_adapt = cv.adaptiveThreshold(img, 255, cv.ADAPTIVE_THRESH_MEAN_C,
                                 cv.THRESH_BINARY, 77, 0)

1 直方图阈值

原理

 threshold函数使

 读入图像

img = cv.imread('pic/eagle500x500.jpg', 0)
show(img)

显示

 程序实现

plt.hist(img.ravel(),256,[0,256]) #ravel拉平矩阵，数组降维
plt.show()

显示

plt.hist(img.flatten(),np.arange(-0.5,256,1),color='g')
plt.show()

显示

_,img_bin = cv.threshold(img,125,255,cv.THRESH_BINARY)#小于125置为0，大于125置为255
show(img_bin)

 显示

2 三角法阈值

原理

 程序实现

# img = cv.imread('pic/blossom500x500.jpg', 0)
img = cv.imread('pic/eagle500x500.jpg', 0)
show(img)

plt.hist(img.flatten(), np.arange(-0.5, 256, 1), color='g')
plt.show()

显示

th, img_bin = cv.threshold(img, 0, 255, cv.THRESH_TRIANGLE)
print(th)
show(np.hstack([img, img_bin]))

显示

3 迭代法阈值及实现

原理

 程序实现

img = cv.imread('pic/eagle500x500.jpg', 0)

T = img.mean()

while True:
    t0 = img[img < T].mean()
    t1 = img[img >= T].mean()
    t  = (t0 + t1) / 2
    if abs(T - t) < 1:
        break
    T = t
T = int(T)

print(f"Best threshold = {T}")

显示

4 大津法（otsu）阈值及实现

原理

程序实现

img = cv.imread('pic/eagle500x500.jpg', 0)

th, img_bin = cv.threshold(img, -1, 255, cv.THRESH_OTSU)
print(th)
show(img_bin)

显示

根据算法编程实现

img = cv.imread('pic/eagle500x500.jpg', 0)

Sigma = -1
T = 0

for t in range(0, 256):
    bg  = img[img <= t]
    obj = img[img > t]
    
    p0 = bg.size / img.size
    p1 = obj.size / img.size
    
    m0 = 0 if bg.size == 0 else bg.mean()
    m1 = 0 if obj.size == 0 else obj.mean()
    
    sigma = p0 * p1 * (m0 - m1)**2
    
    if sigma > Sigma:
        Sigma = sigma
        T = t
        
print(f"Best threshold = {T}")

5 自适应阈值及实现

原理

编程实现

img = cv.imread('pic/page760x900.jpg', 0)
show(img)

 显示

img_bin = cv.adaptiveThreshold(img, 255, cv.ADAPTIVE_THRESH_GAUSSIAN_C, 
                                                cv.THRESH_BINARY, 21, 6)
show(img_bin)

显示

 自适应阈值分割实现

img = cv.imread('pic/page760x900.jpg', 0)

C = 6
winSize = 21

img_blur = cv.blur(img, (winSize, winSize))
img_bin = np.uint8(img > img_blur.astype(np.int) - C) * 255

show(img_bin)

显示

img = cv.imread('pic/page760x900.jpg', 0)

alpha = 0.05
winSize = 21

img_blur = cv.GaussianBlur(img, (winSize, winSize), 5)
img_bin = np.uint8(img > (1 - alpha) * img_blur) * 255

show(img_bin)

 显示