-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathwork4.m
162 lines (128 loc) · 4 KB
/
work4.m
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
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
%% 数字图像处理 作业4 (3/25)
% 阮泉源 201930033629
%% 1 平均去高斯白噪声
im_coser = imread('coser.jpg');
[w, h, d] = size(im_coser);
depth = [1,16,64,128];
figure;
for N=depth
im_coser_noise = zeros(w, h, d, N);
for i=1:N
im_coser_noise(:,:,:,i) = imnoise(im_coser, 'gaussian', 0, 1);
end
im_mean = zeros(w, h, d);
for i=1:N
im_mean = im_mean + im_coser_noise(:,:,:,i)/N;
end
im_mean = uint8(im_mean);
subplot(2,2,find(N==depth))
imshow(im_mean);
title(['Depth:',num2str(N)])
end
suptitle('Different Depth Average of Image with Gaussian Noise')
%%
% 结果分析：
%
% 原始图像加上均值为0，方差为1的高斯噪声后，整体画面变得模糊不清，如果能够利用多张叠加了噪声的图片进行均值化处理，能够显著消除噪声，且效果会随着叠加图片的张数的增多而提高。
%% 2 均值滤波去高斯白噪声
%%
im_coser_noise = imnoise(im_coser, 'gaussian', 0, 1);
kernel_size = [1,2,4,8];
figure;
for N=kernel_size
kernel = ones(N,N)/(N^2);
im_out = convn(im_coser_noise, kernel);
im_out = uint8(im_out);
subplot(2, 2, find(N==kernel_size));
imshow(im_out);
title(['Kernel Size:',num2str(N)]);
end
suptitle('Average Filter of Image with Gaussian Noise')
%%
% 结果分析：
%
% 采用均值化去高斯白噪声的方法有一定的效果，但效果并不好。在噪声较大时依旧存在着大量的模糊。
%% 3 中值滤波去脉冲噪声
%%
im_coser_noise = imnoise(im_coser, 'salt & pepper');
win_size = [1,2,3,4];
figure;
for N = win_size
im_max = zeros(w+floor(N/2)*2, h+floor(N/2)*2, d);
im_max(1+floor(N/2):w+floor(N/2), 1+floor(N/2):h+floor(N/2), :) = im_coser_noise;
if N ~= 1
im_out = zeros(w, h, d);
for j = 1:w
for k = 1:h
im_cut = im_max(j:j+N-1,k:k+N-1,:);
im_cut = reshape(im_cut,N^2,3);
im_cut_sort = sort(im_cut);
if mod(N^2/2,2) == 0
im_out(j,k,:) = [mean([im_cut_sort(floor(N^2/2),1),im_cut_sort(floor(N^2/2)+1,1)]);
mean([im_cut_sort(floor(N^2/2),2),im_cut_sort(floor(N^2/2)+1,2)]);
mean([im_cut_sort(floor(N^2/2),3),im_cut_sort(floor(N^2/2)+1,3)])];
else
im_out(j,k,:) = [im_cut_sort(floor(N^2/2)+1,1);
im_cut_sort(floor(N^2/2)+1,2);
im_cut_sort(floor(N^2/2)+1,3)];
end
end
end
else
im_out = im_coser_noise;
end
subplot(2,2,find(N==win_size));
im_out = uint8(im_out);
imshow(im_out);
title(['Kernel Size:', num2str(N)]);
end
suptitle('Median Filter of Image with Impulse Noise');
%%
% 结果分析：
%
% 采用中值滤波的方法处理脉冲噪声（椒盐噪声）非常有效，但随着选择窗口的增大图片会变得模糊不清。
%% 4 图像锐化增强
%%
im_moon = imread('moon.jpeg');
im_moon = imresize(im_moon,0.5);
Laplacian_4 = [0, -1, 0; -1, 5 -1; 0, -1, 0];
Laplacian_8 = [-1, -1, -1; -1, 9 -1; -1, -1, -1];
Sobel_x = [-1, 0, 1;-2, 1, 2;-1, 0, 1];
Sobel_y = [-1, -2, -1;0, 1, 0;1, 2, 1];
im_L4 = convn(im_moon,Laplacian_4);
im_L8 = convn(im_moon,Laplacian_8);
im_Sx= convn(im_moon,Sobel_x);
im_Sy = convn(im_moon,Sobel_y);
im_L4 = uint8(im_L4);
im_L8 = uint8(im_L8);
im_Sx = uint8(im_Sx);
im_Sy = uint8(im_Sy);
for i=1:1
figure;
imshow(im_moon);
suptitle('Original Image without process');
figure;
subplot(1,2,1);
imshow(im_L4);
title(' 4 neiberhood Laplacian Kernel')
subplot(1,2,2);
imshow(im_L8);
title(' 8 neiberhood Laplacian Kernel')
suptitle('Sharpen Enhancing of Image with Laplacian Kernel')
figure;
subplot(1,2,1);
imshow(im_Sx);
title('Sobel Kernel in X-axis')
subplot(1,2,2);
imshow(im_Sy);
title('Sobel Kernel in Y-axis')
suptitle('Sharpen Enhancing of Image with Sobel Kernel')
end
%%
% 结果分析：
%
% 相较于原始图片，锐化后的图片边缘变得更加明显。
%
% 八邻域拉普拉斯算子的效果比四邻域的效果更好，图片边缘变得更加突出。
%
% 相较于拉普拉斯算子，索贝尔算子的锐化效果更为显著。