Mahotas – Otsu 的方法
在本文中,我们将了解如何在 mahotas 中实现 otsu 的方法。在计算机视觉和图像处理中,以 Nobuyuki Otsu 命名的 Otsu 方法用于执行自动图像阈值处理。在最简单的形式中,该算法返回一个单一的强度阈值,将像素分为前景和背景两类。
在本教程中,我们将使用“luispedro”图像,下面是加载它的命令。
mahotas.demos.load('luispedro')
下面是 luispedro 图像
为此,我们将使用 mahotas.otsu 方法
Syntax : mahotas.otsu(image)
Argument : It takes image object as argument
Return : It returns integer
注意:输入图像应被过滤或应加载为灰色
为了过滤图像,我们将获取图像对象 numpy.ndarray 并在索引的帮助下对其进行过滤,下面是执行此操作的命令
image = image[:, :, 0]
示例 1:
Python3
# importing required libraries
import mahotas
import mahotas.demos
import numpy as np
from pylab import imshow, gray, show
from os import path
# loading the image
photo = mahotas.demos.load('luispedro')
# showing original image
print("Original Image")
imshow(photo)
show()
# loading image as grey
photo = mahotas.demos.load('luispedro', as_grey = True)
# converting image type to unit8
# because as_grey returns floating values
photo = photo.astype(np.uint8)
# otsu method
T_otsu = mahotas.otsu(photo)
# printing otsu value
print("Otsu Method value : " + str(T_otsu))
print("Image threshold using Otsu Method")
# showing image
# image values should be greater than otsu value
imshow(photo > T_otsu)
show()
Python3
# importing required libraries
import mahotas
import numpy as np
from pylab import imshow, show
import os
# loading image
img = mahotas.imread('dog_image.png')
# setting filter to the image
img = img[:, :, 0]
imshow(img)
show()
# otsu method
T_otsu = mahotas.otsu(img)
# printing otsu value
print("Otsu Method value : " + str(T_otsu))
print("Image threshold using Otsu Method")
# showing image
# image values should be greater than otsu value
imshow(img > T_otsu)
show()
输出 :
示例 2:
Python3
# importing required libraries
import mahotas
import numpy as np
from pylab import imshow, show
import os
# loading image
img = mahotas.imread('dog_image.png')
# setting filter to the image
img = img[:, :, 0]
imshow(img)
show()
# otsu method
T_otsu = mahotas.otsu(img)
# printing otsu value
print("Otsu Method value : " + str(T_otsu))
print("Image threshold using Otsu Method")
# showing image
# image values should be greater than otsu value
imshow(img > T_otsu)
show()
输出 :