-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathcanny.py
More file actions
103 lines (85 loc) · 3.23 KB
/
canny.py
File metadata and controls
103 lines (85 loc) · 3.23 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Wed Jun 19 21:53:00 2019
@author: vaeahc
"""
import cv2
import numpy as np
import matplotlib.pyplot as plt
im = cv2.imread('lena512color.tiff', 0)
print('原图')
plt.imshow(im, 'gray')
plt.show()
Gaussian = np.zeros((3, 3))
sigma = 0.5
for i in range(3):
for j in range(3):
Gaussian[i][j] = 1 / (2 * np.pi * sigma ** 2) * np.exp(-((i - 1) ** 2 + (j - 1) ** 2) / (2 * sigma ** 2))
Gaussian /= np.sum(Gaussian)
#padding处理
row = np.zeros((1, im.shape[1]))
im = np.row_stack((row, im, row))
col = np.zeros((im.shape[0], 1))
im = np.column_stack((col, im, col))
#高斯低通滤波
for i in range(1, im.shape[0] - 1):
for j in range(1, im.shape[1] - 1):
im[i][j] = np.sum(Gaussian * im[i - 1: i + 2, j - 1: j + 2])
print('高斯低通滤波')
plt.imshow(im[1:-1, 1:-1], 'gray')
plt.show()
#提取水平和竖直方向梯度，幅度和夹角
sobel_x = np.array([[-1, 0, 1], [-2, 0, 2], [-1, 0, 1]])
sobel_y = np.array([[1, 2, 1], [0, 0, 0], [-1, -2, -1]])
grad_x = np.zeros(np.shape(im))
grad_y = np.zeros(np.shape(im))
Amp = np.zeros(np.shape(im))
dire = np.zeros(np.shape(im))
for i in range(1, im.shape[0] - 1):
for j in range(1, im.shape[1] - 1):
grad_x[i][j] = np.sum(sobel_x * im[i - 1: i + 2, j - 1: j + 2])
grad_y[i][j] = np.sum(sobel_y * im[i - 1: i + 2, j - 1: j + 2])
Amp = np.sqrt(grad_x ** 2 + grad_y ** 2)
dire = np.arctan(grad_y / grad_x) + np.pi / 2
print('梯度图像')
plt.imshow(Amp[1:-1, 1:-1], 'gray')
plt.show()
#NMS(非极大值抑制), 非插值方法
for i in range(1, Amp.shape[0] - 1):
for j in range(1, Amp.shape[1] - 1):
angel = np.abs(dire[i][j])
if angel < np.pi / 8 or angel >= np.pi * 7 / 8:
if Amp[i][j] < Amp[i][j - 1] or Amp[i][j] < Amp[i][j + 1]:
Amp[i][j] = 0
elif angel >= np.pi / 8 and angel < np.pi * 3 / 8:
if Amp[i][j] < Amp[i + 1][j - 1] or Amp[i][j] < Amp[i - 1][j + 1]:
Amp[i][j] = 0
elif angel >= np.pi * 3 /8 and angel < np.pi * 5 /8:
if Amp[i][j] < Amp[i - 1][j] or Amp[i][j] < Amp[i + 1][j]:
Amp[i][j] = 0
elif angel >= np.pi * 5 / 8 and angel < np.pi * 7 / 8:
if Amp[i][j] < Amp[i - 1][j - 1] or Amp[i][j] < Amp[i + 1][j + 1]:
Amp[i][j] = 0
print('非极大值抑制后梯度图像')
plt.imshow(Amp[1:-1, 1:-1], 'gray')
plt.show()
#双阈值处理，取 B=100, A = 60
A = 100
B = 150
for i in range(1, Amp.shape[0] - 1):
for j in range(1, Amp.shape[1] - 1):
if Amp[i][j] >= B:
Amp[i][j] = 255
elif Amp[i][j] <= A:
Amp[i][j] = 0
for i in range(1, Amp.shape[0] - 1):
for j in range(1, Amp.shape[1] - 1):
if Amp[i][j] > A and Amp[i][j] < B:
if Amp[i - 1][j - 1] == 255 or Amp[i - 1][j] == 255 or Amp[i - 1][j + 1] == 255 or Amp[i][j - 1] == 255 or Amp[i][j + 1] == 255 or Amp[i + 1][j - 1] == 255 or Amp[i + 1][j] == 255 or Amp[i + 1][j + 1] == 255:
Amp[i][j] == 255
else:
Amp[i][j] == 0
print('双阈值后图像')
plt.imshow(Amp[1:-1, 1:-1], 'gray')
plt.show()