OpenCV-Python -- Morphological Transformations

学习目标

• 学习不同类型的形态学操作，比如腐蚀（Erosion），膨胀（Dilation），开操作（Opening），闭操作（Closing）等
• 学习不同的函数，比如，cv2.erode()，cv2.dilate()，cv2.morphologyEx()等

理论知识

形态学变换是一系列基于图像形状的简单操作。通常处理二值图像，输入有2个参数，1.输入图像；2.结构化元素或核，形状决定该操作的性质。两个基本的形态学操作为腐蚀和膨胀。他们的变体，比如开操作，闭操作，梯度等。我们将使用下面的图像一一介绍形态学操作。

1. Erosion（腐蚀操作）

腐蚀的基本概念类似与土壤腐蚀，通常腐蚀前景的边界。操作过程：定义的核在图像上滑动，图像当前位置的像素为0或者1，只有核内所有的像素都为1，当前像素位置的值为1，否则就是0.
所以，根据核的大小，边界附近的像素为0。所以腐蚀操作相当于是将目标的边界变细。因此适合去除小的白色噪声。函数使用如下：

dst = cv2.erode(src, kernel, iterations) 

def erode(src, kernel, dst=None, anchor=None, iterations=None, borderType=None, borderValue=None): 
    """
       @brief Erodes an image by using a specific structuring 		 
       element(结构化元素，比如卷积核).
       The function supports the in-place mode. Erosion can be 
       applied several ( iterations ) times. In case of multi-
       channel images, each channel is processed independently.
       
       @param src input image; the number of channels can be 
       arbitrary, but the depth should be one of CV_8U, CV_16U, 
       CV_16S, CV_32F or CV_64F.
       
       @param dst output image of the same size and type as src.
	   
	   @param kernel structuring element used for erosion; if 
	   `element=Mat()`, a `3 x 3` rectangular structuring element is 
	   used. Kernel can be created using #getStructuringElement.
       
       @param anchor position of the anchor within the element; 
       default value (-1, -1) means that the anchor is at the 
       element center.
       
       @param iterations number of times erosion is applied.可以对图像	
       使用多次腐蚀操作.
       
       @param borderType pixel extrapolation method, see 
       #BorderTypes
       
       @param borderValue border value in case of a constant border
       @sa  dilate, morphologyEx, getStructuringElement
    """

下面的例子：

class Morphological:
    def __init__(self, image):
        self.img = cv2.imread(image, 0)
        self.kernel = np.ones((5, 5), np.uint8)
        self.kernel_tophat = np.ones((9, 9), np.uint8)


    def erosion(self):
        erosion = cv2.erode(self.img, self.kernel, iterations=1)
        return erosion
    
    def dilate(self):
        dilation = cv2.dilate(self.img, self.kernel, iterations=1)
        return dilation
    
    def opening(self):
        opening = cv2.morphologyEx(self.img, cv2.MORPH_OPEN, self.kernel)
        return opening
	
	def closing(self):
        closing = cv2.morphologyEx(self.img, cv2.MORPH_CLOSE, self.kernel)
        return closing
    
    def morphological_gradient(self):
        gradient = cv2.morphologyEx(self.img, cv2.MORPH_GRADIENT, self.kernel)
        return gradient
    
    def top_hat(self):
        tophat = cv2.morphologyEx(self.img, cv2.MORPH_TOPHAT, self.kernel_tophat)
        return tophat
    
    def black_hat(self):
        blackhat = cv2.morphologyEx(self.img, cv2.MORPH_BLACKHAT, self.kernel)
        return blackhat
    
    def show(self, name, result):
        cv2.imshow('origin', self.img)
        cv2.imshow(name, result)
        cv2.waitKey(0)


if __name__ == '__main__':
    image = 'j.png'  # j_1.png
    mor = Morphological(image)
    # result1 = mor.erosion()
    # result1 = mor.dilate()
    # result1 = mor.opening()
    # result1 = mor.closing()
    # result1 = mor.morphological_gradient()
    # result1 = mor.top_hat()
    result1 = mor.black_hat()
    mor.show('black_hat', result1)

结果如下：

2. Dilation(膨胀)

相当于腐蚀的反操作，只要图像的至少一个非零元素在核内，那么该位置的像素值置1。通常会增粗边界。一般情况下，在去除噪声时，会先腐蚀后膨胀：腐蚀将白色噪声去除，但是会边界变细，那么膨胀可以补边界。

dilation = cv2.dilate(img,kernel,iterations = 1)

def dilate(src, kernel, dst=None, anchor=None, iterations=None, borderType=None, borderValue=None):
    """
       @brief Dilates an image by using a specific structuring 	
       element.
       The function supports the in-place mode. Dilation can be 
       applied several ( iterations ) times. In case of multi-
       channel images, each channel is processed independently.
  
       @param src input image; the number of channels can be 
       arbitrary, but the depth should be one of CV_8U, CV_16U, 
       CV_16S, CV_32F or CV_64F.
       @param dst output image of the same size and type as src.
       @param kernel structuring element used for dilation; if 
       elemenat=Mat(), a 3 x 3 rectangular structuring element is 
       used. Kernel can be created using #getStructuringElement
       @param anchor position of the anchor within the element; 
       default value (-1, -1) means that the anchor is at the 
       element center.
       @param iterations number of times dilation is applied.
       @param borderType pixel extrapolation method, see 
       #BorderTypes
       @param borderValue border value in case of a constant border
       @sa  erode, morphologyEx, getStructuringElement
    """

示例代码见上述的使用，运行结果如下：

3. Opening

开操作：腐蚀+膨胀的连续处理。常用在去除噪声，下面介绍函数的使用：

cv2.morphologyEx(src, op, kernel)

上面的op主要包括下面的类型：

"""
	enum MorphTypes{
	    MORPH_ERODE    = 0, //腐蚀
	    MORPH_DILATE   = 1, //膨胀
	    MORPH_OPEN     = 2, //开操作
	    MORPH_CLOSE    = 3, //闭操作
	    MORPH_GRADIENT = 4, //梯度操作
	    MORPH_TOPHAT   = 5, //顶帽操作
	    MORPH_BLACKHAT = 6, //黑帽操作
	    MORPH_HITMISS  = 7  
	};
"""

def morphologyEx(src, op, kernel, dst=None, anchor=None, iterations=None, borderType=None, borderValue=None): 
    """
       @brief Performs advanced morphological transformations.
      
       @param src Source image. The number of channels can be 
       arbitrary. The depth should be one of CV_8U, CV_16U, CV_16S, 
       CV_32F or CV_64F.
       @param dst Destination image of the same size and type as 
       source image.
       @param op Type of a morphological operation, see #MorphTypes
       @param kernel Structuring element. It can be created using 
       #getStructuringElement.
       @param anchor Anchor position with the kernel. Negative 
       values mean that the anchor is at the kernel center.
       @param iterations Number of times erosion and dilation are 
       applied.
       @param borderType Pixel extrapolation method, see 
       #BorderTypes
       @param borderValue Border value in case of a constant 
       border. The default value has a special meaning.
       
       当iteration大于1时，注意下面的操作顺序：
       @note The number of iterations is the number of times 
       erosion or dilatation operation will be applied.
       
       For instance, an opening operation (#MORPH_OPEN) with two 
       iterations is equivalent to apply
       successively: erode -> erode -> dilate -> dilate (and not 
       erode -> dilate -> erode -> dilate).
    """

示例代码见上述的使用，运行结果如下：

4. Closing

闭操作是开操作的逆过程：膨胀+腐蚀。常用于去除前景目标里面的小洞，或者目标里面小的黑色点。
代码在上面，运行结果如下：

5. Morphological Gradient（形态学梯度）

具体操作如下：梯度 = 膨胀-腐蚀，结果类似于得到目标的外轮廓。
运行结果如下：

6. Top Hat（顶帽操作）

顶帽 = Img - Opening，输入图像与开操作的差。
代码如上，核的大小为9x9，运行结果如下：

7. Black Hat（黑帽）

黑帽 = 闭运算-原始输入。

Structuring Element（结构化元素）

在上面的例子中，我们可以通过Numpy手动创建结构化元素，它是矩形。但是在一些情况下，我们需要特定形状的核。鉴于此，OpenCV提供了函数，具体的定义和使用如下：

cv2.getStructuringElement(shape, ksize, anchor=None)
"""
	@brief: 生成特定形状的核或着结构化元素
	@shape：结构化元素的形状
		矩形：MORPH_RECT = 0
		交叉形:：MORPH_CROSS = 1
		椭圆形:：MORPH_ELLIPSE = 2
	@ksize：核的大小
"""

    # Rectangular Kernel
    a = cv2.getStructuringElement(cv2.MORPH_RECT, (5, 5))

    # Elliptical Kernel
    b = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (5, 5))
    
    # Cross-shaped Kernel
    c = cv2.getStructuringElement(cv2.MORPH_CROSS, (5, 5))
    print(a)
    print('=============')
    print(b)
    print('=============')
    print(c)
    print('=============')

输出如下：

[[1 1 1 1 1]
 [1 1 1 1 1]
 [1 1 1 1 1]
 [1 1 1 1 1]
 [1 1 1 1 1]]
=============
[[0 0 1 0 0]
 [1 1 1 1 1]
 [1 1 1 1 1]
 [1 1 1 1 1]
 [0 0 1 0 0]]
=============
[[0 0 1 0 0]
 [0 0 1 0 0]
 [1 1 1 1 1]
 [0 0 1 0 0]
 [0 0 1 0 0]]
=============