-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathFE_Wrapper.m
179 lines (167 loc) · 7.16 KB
/
FE_Wrapper.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
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
function [PSF,GNyq] = FE_Wrapper(I_MS_LR, I_PAN, ratio, sensorInf, type)
if ~exist('type','var')
type = 'MBFE';
end
tap = 25;
lambda = 10^5;
mu = 10^5;
num_iter = 10;
th = 0;
filtername = 'other';
%%%%%%%%%%% LR有bug，目前只为DPSR考虑。貌似只能用FE和FE_MS
if size(I_MS_LR, 1) ~= size(I_PAN, 1)
I_MS = interpWrapper(I_MS_LR,ratio,sensorInf.upsampling);
else
I_MS = I_MS_LR;
end
switch type
case 'FE'
PSF = FE(I_MS,I_PAN,ratio,tap,lambda,mu,th,num_iter,filtername);
GNyq = MTF_GNyq_Est(PSF, ratio);
case 'FE_MS'
if size(I_PAN,3) == 1
I_PAN = repmat(I_PAN, [1,1,size(I_MS,3)]);
end
% Hist match?
for ii = 1:size(I_MS,3)
I_PAN(:,:,ii) = (I_PAN(:,:,ii) - mean2(LPfilterGauss(I_PAN(:,:,ii),ratio)))...
.*(std2(I_MS(:,:,ii))./std2(LPfilterGauss(I_PAN,ratio))) + mean2(I_MS(:,:,ii));
end
PSF = FE_MS(I_MS,I_PAN,tap,lambda,mu,th,filtername);
GNyq = MTF_GNyq_Est(PSF, ratio);
case 'MBFE_H'
%I_MS_hat = BDSD_PC(I_MS,I_PAN,ratio,sensorInf);
%I_MS_hat = GSA(I_MS,I_PAN,I_MS_LR,ratio);
I_MS_hat1 = MTF_GLP_HPM_Haze_min(I_MS,I_PAN,sensorInf.sensor,ratio,1);
I_MS_hat2 = BroveyRegHazeMin(I_MS,I_PAN,ratio);
I_MS_hat3 = MTF_GLP_AWLP_Haze(I_PAN,I_MS,ratio,1,sensorInf);
PSF1 = FE_MS(I_MS,I_MS_hat1,tap,lambda,mu,th,filtername);
PSF2 = FE_MS(I_MS,I_MS_hat2,tap,lambda,mu,th,filtername);
PSF3 = FE_MS(I_MS,I_MS_hat3,tap,lambda,mu,th,filtername);
GNyq1 = MTF_GNyq_Est(PSF1, ratio);
GNyq2 = MTF_GNyq_Est(PSF2, ratio);
GNyq3 = MTF_GNyq_Est(PSF3, ratio);
GNyq = (GNyq1 + GNyq2 + GNyq3)./3;
case 'MBFE'
%I_MS_hat = BDSD_PC(I_MS,I_PAN,ratio,sensorInf);
I_MS_hat = GSA(I_MS,I_PAN,I_MS_LR,ratio);
PSF = FE_MS(I_MS,I_MS_hat,tap,lambda,mu,th,filtername);
GNyq = MTF_GNyq_Est(PSF, ratio);
case 'bruse'
off = 20;
I_MS = I_MS(1+off:end-off, 1+off:end-off,:);
I_PAN = I_PAN(1+off:end-off, 1+off:end-off,:);
I_MS_LR = I_MS_LR(1+off/ratio:end-off/ratio, 1+off/ratio:end-off/ratio,:);
n_band = size(I_MS,3);
GNyq = zeros(1,n_band);
PSF = zeros(tap, tap, n_band);
min_err = ones(1,n_band)*Inf;
sensorInf_P = sensorInf;
Lap_MS_LR = zeros(size(I_MS_LR));
for i = 1:size(I_MS_LR,3)
Lap_MS_LR(:,:,i)= imfilter(I_MS_LR(:,:,i),fspecial('sobel'));
end
% 保留其中方差大的小块
%imageHR = GSA(I_MS,I_PAN,I_MS_LR,ratio);% spectral refine
imageHR = BroveyRegHazeMin(I_MS,I_PAN,ratio);
% if size(I_PAN,3) == 1
% imageHR = repmat(I_PAN, [1,1,size(I_MS,3)]);
% else
% imageHR = I_PAN;
% end
% % Hist match?
% imageHR0 = imageHR;
for iGNyq = 0.1:0.01:0.9
iPSF = MTF_GNyq2PSF(iGNyq*ones(1, size(I_MS_LR,3)), tap, ratio);
sensorInf_P.PSF_G = iPSF;
% I_ = MTF_GLP(I_PAN,I_MS,sensorInf_P,ratio);
imageHR_LR = MTF_conv_sample(imageHR, sensorInf_P, ratio, 1);
% for ii = 1:size(I_MS,3)
% imageHR0(:,:,ii) = (imageHR(:,:,ii) - mean2(imageHR_LR))...
% .*(std2(I_MS(:,:,ii))./std2(imageHR_LR)) + mean2(I_MS(:,:,ii));
% end
% imageHR_LR = MTF_conv_sample(imageHR0, sensorInf_P, ratio, 1);
% for ii = 1:size(I_MS,3)
% imageHR(:,:,ii) = (imageHR(:,:,ii) - mean2(imageHR_LR(:,:,ii)))...
% .*(std2(I_MS(:,:,ii))./std2(imageHR_LR(:,:,ii))) + mean2(I_MS(:,:,ii));
% end
% imageHR_LR = MTF_conv_sample(imageHR, sensorInf_P, ratio, 1);
Lap_imageHR_LR = imfilter(imageHR_LR,fspecial('sobel'));
Lap_ERR = Lap_MS_LR - Lap_imageHR_LR;
%Lap_ERR = I_MS_LR - imageHR_LR;
for ii = 1:n_band
t = norm(Lap_ERR(:,:,ii), 'fro');
if t < min_err(ii)
min_err(ii) = t;
GNyq(ii) = iGNyq;
PSF(:,:,ii) = iPSF(:,:,1);
end
end
end
case 'FE_LR'
PSF = FE_LR(I_MS, I_PAN,20,tap,lambda,mu,th,filtername);
GNyq = MTF_GNyq_Est(PSF, 1);
case 'bruse_LR'
n_band = size(I_MS,3);
GNyq = zeros(1,n_band);
PSF = zeros(tap, tap, n_band);
min_err = ones(1,n_band)*Inf;
sensorInf_P = sensorInf;
off = 10;
I_MS_LR = I_MS_LR(1+off:end-off, 1+off:end-off,:);
I_PAN = I_PAN(1+off:end-off, 1+off:end-off,:);
Lap_MS_LR = zeros(size(I_MS_LR));
for i = 1:size(I_MS_LR,3)
Lap_MS_LR(:,:,i)= imfilter(I_MS_LR(:,:,i),fspecial('sobel'));
end
for iGNyq = 0.1:0.01:0.9
iPSF = MTF_GNyq2PSF(iGNyq*ones(1, size(I_MS_LR,3)), tap, 1);
sensorInf_P.PSF_G = iPSF;
imageHR_LP = MTF_conv_sample(I_PAN, sensorInf_P, 1, 0);
Lap_imageHR_LP = imfilter(imageHR_LP,fspecial('sobel'));
Lap_ERR = Lap_MS_LR - Lap_imageHR_LP;
for ii = 1:n_band
t = norm(Lap_ERR(:,:,ii), 'fro');
if t < min_err(ii)
min_err(ii) = t;
GNyq(ii) = iGNyq;
PSF(:,:,ii) = iPSF(:,:,1);
end
end
end
case 'Hysure'
intersection = cell(1,1);
intersection{1} = 1:size(I_MS,3);
p = size(I_MS,3);
contiguous = intersection;
lambda_R = 1e1;
lambda_B = 1e1;
hsize_h = 2*ratio-1;
hsize_w = 2*ratio-1;
shift = sensorInf.downsampling.offset; % 'phase' parameter in MATLAB's 'upsample' function
blur_center = mod(ratio+1,2); % to center the blur kernel according to the simluated data
[~, ~, B_est] = sen_resp_est(I_MS_LR, I_PAN, ratio, intersection, contiguous, ...
p, lambda_R, lambda_B, hsize_h, hsize_w, shift, blur_center);
PSF = MatrixToKernel(B_est,hsize_h,hsize_w);
GNyq = MTF_GNyq_Est(PSF, ratio);
PSF = repmat(PSF, [1,1,size(I_MS_LR,3)]);
GNyq = repmat(GNyq, [1,1,size(I_MS_LR,3)]);
%TODO:多通道？
% for iGNyq = -0.05+min(GNyq(:)):0.01:max(GNyq(:))+0.05
% iPSF = MTF_GNyq2PSF(iGNyq*ones(1, size(I_MS_LR,3)), tap, ratio);
% sensorInf_P.PSF_G = iPSF;
% % I_ = MTF_GLP(I_PAN,I_MS,sensorInf_P,ratio);
%
% imageHR_LR = MTF_conv_sample(imageHR, sensorInf_P, ratio, 1);
% Lap_imageHR_LR = imfilter(imageHR_LR,fspecial('sobel'));
% for ii = 1:n_band
% Lap_ERR = Lap_MS_LR - Lap_imageHR_LR;
% t = norm(Lap_ERR(:,:,ii), 'fro');
% if t < min_err(ii)
% min_err(ii) = t;
% GNyq(ii) = iGNyq;
% PSF(:,:,ii) = iPSF(:,:,1);
% end
% end
% end
end