Python|图像处理中的形态学运算（开幕） |第一组

形态学操作用于提取对区域形状的表示和描述有用的图像分量。形态学操作是一些依赖于图片形状的基本任务。它通常在二进制图像上执行。它需要两个数据源，一个是输入图像，另一个是结构化组件。形态运算符将输入图像和结构组件作为输入，然后使用集合运算符组合这些元素。输入图像中的对象根据图像形状的属性进行处理，这些属性在结构化组件中进行编码。
开放类似于侵蚀，因为它倾向于从前景像素区域的边缘移除明亮的前景像素。运算符的作用是保护与结构化成分相似的前景区域，或者可以完全包含结构化成分的前景区域，同时取出每个其他区域的前景像素。开运算用于去除图像中的内部噪声。
开运算是腐蚀运算，然后是膨胀运算。

Syntax: cv2.morphologyEx(image, cv2.MORPH_OPEN, kernel)
Parameters: 
-> image: Input Image array. 
-> cv2.MORPH_OPEN: Applying the Morphological Opening operation. 
-> kernel: Structuring element. 
 

下面是解释打开形态学运算的Python代码——

Python3

# Python program to illustrate
# Opening morphological operation
# on an image
 
# organizing imports 
import cv2 
import numpy as np 
 
# return video from the first webcam on your computer. 
screenRead = cv2.VideoCapture(0)
 
# loop runs if capturing has been initialized.
while(1):
    # reads frames from a camera
    _, image = screenRead.read()
     
    # Converts to HSV color space, OCV reads colors as BGR
    # frame is converted to hsv
    hsv = cv2.cvtColor(image, cv2.COLOR_BGR2HSV)
     
    # defining the range of masking
    blue1 = np.array([110, 50, 50])
    blue2 = np.array([130, 255, 255])
     
    # initializing the mask to be
    # convoluted over input image
    mask = cv2.inRange(hsv, blue1, blue2)
 
    # passing the bitwise_and over
    # each pixel convoluted
    res = cv2.bitwise_and(image, image, mask = mask)
     
    # defining the kernel i.e. Structuring element
    kernel = np.ones((5, 5), np.uint8)
     
    # defining the opening function
    # over the image and structuring element
    opening = cv2.morphologyEx(mask, cv2.MORPH_OPEN, kernel)
    
    # The mask and opening operation
    # is shown in the window
    cv2.imshow('Mask', mask)
    cv2.imshow('Opening', opening)
     
    # Wait for 'a' key to stop the program
    if cv2.waitKey(1) & 0xFF == ord('a'):
        break
 
# De-allocate any associated memory usage 
cv2.destroyAllWindows()
 
# Close the window / Release webcam
screenRead.release()

输入帧：

面具：

输出帧：

系统将定义的蓝皮书识别为输入，在Opening函数的帮助下消除并简化了感兴趣区域的内部噪声。