【原】python+opencv图像处理（四十二）

Kirsch算子

1、Kirsch算子

Kirsch算子是R.Kirsch提出来一种边缘检测新算法，它采用8个模板对图像上的每一个像素点进行卷积求导数，这8个模板代表8个方向，对图像上的8个特定边缘方向作出最大响应，运算中取最大值作为图像的边缘输出。

Kirsch算子如下：

2、Kirsch算子具体实现

下图中，左图为原图，右图为经Kirsch算子检测后的边缘图。

其完整代码如下：

import cv2

import numpy as np

from matplotlib import pyplot as plt

#定义Kirsch 卷积模板

m1 = np.array([[5, 5, 5],[-3,0,-3],[-3,-3,-3]])

m2 = np.array([[-3, 5,5],[-3,0,5],[-3,-3,-3]])

m3 = np.array([[-3,-3,5],[-3,0,5],[-3,-3,5]])

m4 = np.array([[-3,-3,-3],[-3,0,5],[-3,5,5]])

m5 = np.array([[-3, -3, -3],[-3,0,-3],[5,5,5]])

m6 = np.array([[-3, -3, -3],[5,0,-3],[5,5,-3]])

m7 = np.array([[5, -3, -3],[5,0,-3],[5,-3,-3]])

m8 = np.array([[5, 5, -3],[5,0,-3],[-3,-3,-3]])

img = cv2.imread("E:\image\lena.jpg",0)

img = cv2.copyMakeBorder(img,1,1,1,1,borderType=cv2.BORDER_REPLICATE)

temp = list(range(8))

img1 = np.zeros(img.shape)

for i in range(1,img.shape[0]-1):

    for j in range(1,img.shape[1]-1):

        temp[0] = np.abs( ( np.dot( np.array([1,1,1]) , ( m1*img[i-1:i+2,j-1:j+2]) ) ).dot(np.array([[1],[1],[1]])))

        temp[2] = np.abs( ( np.dot( np.array([1,1,1]) , ( m1*img[i-1:i+2,j-1:j+2]) ) ).dot(np.array([[1],[1],[1]])))

        temp[3] = np.abs((np.dot(np.array([1, 1, 1]), (m3 * img[i - 1:i + 2, j - 1:j + 2]))).dot(np.array([[1],[1],[1]])))

        temp[4] = np.abs((np.dot(np.array([1, 1, 1]), (m4 * img[i - 1:i + 2, j - 1:j + 2]))).dot(np.array([[1],[1],[1]])))

        temp[5] = np.abs((np.dot(np.array([1, 1, 1]), (m5 * img[i - 1:i + 2, j - 1:j + 2]))).dot(np.array([[1],[1],[1]])))

        temp[6] = np.abs((np.dot(np.array([1, 1, 1]), (m6 * img[i - 1:i + 2, j - 1:j + 2]))).dot(np.array([[1],[1],[1]])))

        temp[7] = np.abs((np.dot(np.array([1, 1, 1]), (m7 * img[i - 1:i + 2, j - 1:j + 2]))).dot(np.array([[1],[1],[1]])))

        temp[1] = np.abs((np.dot(np.array([1, 1, 1]), (m2 * img[i-1:i+2, j-1:j+2]))).dot(np.array([[1],[1],[1]])))

        img1[i,j] = np.max(temp)

        if img1[i, j] > 255:  

            img1[i, j] = 255

        else:

            img1[i, j] = 0

cv2.imshow("origin",img)

cv2.imshow("1",img1)

cv2.waitKey(0)

本篇参考自链接：

https://blog.csdn.net/weixin_44403952/article/details/90267123