咸鱼哥学python OpenCV之边缘检测

女爷i 2023-02-23 12:06 9阅读 0赞

**1、Canny边缘检测**

**基本步骤：**  
1）使用高斯滤波，以平滑图像，滤除噪声  
2）计算每个像素点的梯度和方向  
3）应用费极大值抑制MNS，消除边缘检测带来的杂散相应  
4）应用双阈值检测来确定真实和潜在的边缘  
5）通过抑制孤立的弱边缘最终完成边缘检测

**1.1 高斯滤波**  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NTg4NTIzMg_size_16_color_FFFFFF_t_70]  
**1.2梯度和方向**  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NTg4NTIzMg_size_16_color_FFFFFF_t_70 1]![在这里插入图片描述][2020070711180230.png]  
**1.3非极大抑制**  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NTg4NTIzMg_size_16_color_FFFFFF_t_70 2]  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NTg4NTIzMg_size_16_color_FFFFFF_t_70 3]  
**1.4 双阈值检测**  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NTg4NTIzMg_size_16_color_FFFFFF_t_70 4]

在OpenCV中实现：

import numpy as np
    import cv2 as cv
    
    path = r'F:\Python_study\python_image\opencv\image\lena.png'
    img = cv.imread(path, cv.IMREAD_GRAYSCALE)
    
    
    def cv_show(img, name):
        cv.imshow(name, img)
        cv.waitKey(0)
        cv.destroyAllWindows()
    
    v1 = cv.Canny(img, 50, 100)
    v2 = cv.Canny(img, 80, 150)
    res = np.hstack((v1, v2))
    cv_show(res, 'canny')

结果如下：  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NTg4NTIzMg_size_16_color_FFFFFF_t_70 5]

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NTg4NTIzMg_size_16_color_FFFFFF_t_70]: https://img-blog.csdnimg.cn/20200707111702435.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NTg4NTIzMg==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NTg4NTIzMg_size_16_color_FFFFFF_t_70 1]: https://img-blog.csdnimg.cn/20200707111735835.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NTg4NTIzMg==,size_16,color_FFFFFF,t_70
[2020070711180230.png]: https://img-blog.csdnimg.cn/2020070711180230.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NTg4NTIzMg_size_16_color_FFFFFF_t_70 2]: https://img-blog.csdnimg.cn/20200707123715170.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NTg4NTIzMg==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NTg4NTIzMg_size_16_color_FFFFFF_t_70 3]: https://img-blog.csdnimg.cn/20200707144322309.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NTg4NTIzMg==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NTg4NTIzMg_size_16_color_FFFFFF_t_70 4]: https://img-blog.csdnimg.cn/20200707144548999.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NTg4NTIzMg==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NTg4NTIzMg_size_16_color_FFFFFF_t_70 5]: https://img-blog.csdnimg.cn/20200707145522590.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80NTg4NTIzMg==,size_16,color_FFFFFF,t_70