1 /*M///////////////////////////////////////////////////////////////////////////////////////
2 //
3 // IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
4 //
5 // By downloading, copying, installing or using the software you agree to this license.
6 // If you do not agree to this license, do not download, install,
7 // copy or use the software.
8 //
9 //
10 // License Agreement
11 // For Open Source Computer Vision Library
12 //
13 // Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
14 // Copyright (C) 2009, Willow Garage Inc., all rights reserved.
15 // Third party copyrights are property of their respective owners.
16 //
17 // Redistribution and use in source and binary forms, with or without modification,
18 // are permitted provided that the following conditions are met:
19 //
20 // * Redistribution's of source code must retain the above copyright notice,
21 // this list of conditions and the following disclaimer.
22 //
23 // * Redistribution's in binary form must reproduce the above copyright notice,
24 // this list of conditions and the following disclaimer in the documentation
25 // and/or other materials provided with the distribution.
26 //
27 // * The name of the copyright holders may not be used to endorse or promote products
28 // derived from this software without specific prior written permission.
29 //
30 // This software is provided by the copyright holders and contributors "as is" and
31 // any express or implied warranties, including, but not limited to, the implied
32 // warranties of merchantability and fitness for a particular purpose are disclaimed.
33 // In no event shall the Intel Corporation or contributors be liable for any direct,
34 // indirect, incidental, special, exemplary, or consequential damages
35 // (including, but not limited to, procurement of substitute goods or services;
36 // loss of use, data, or profits; or business interruption) however caused
37 // and on any theory of liability, whether in contract, strict liability,
38 // or tort (including negligence or otherwise) arising in any way out of
39 // the use of this software, even if advised of the possibility of such damage.
40 //
41 //M*/
42
43 #include "precomp.hpp"
44 #include "opencv2/photo.hpp"
45 #include "opencv2/imgproc.hpp"
46 //#include "opencv2/highgui.hpp"
47 #include "hdr_common.hpp"
48
49 namespace cv
50 {
51
52 class CalibrateDebevecImpl : public CalibrateDebevec
53 {
54 public:
CalibrateDebevecImpl(int _samples,float _lambda,bool _random)55 CalibrateDebevecImpl(int _samples, float _lambda, bool _random) :
56 name("CalibrateDebevec"),
57 samples(_samples),
58 lambda(_lambda),
59 random(_random),
60 w(tringleWeights())
61 {
62 }
63
process(InputArrayOfArrays src,OutputArray dst,InputArray _times)64 void process(InputArrayOfArrays src, OutputArray dst, InputArray _times)
65 {
66 std::vector<Mat> images;
67 src.getMatVector(images);
68 Mat times = _times.getMat();
69
70 CV_Assert(images.size() == times.total());
71 checkImageDimensions(images);
72 CV_Assert(images[0].depth() == CV_8U);
73
74 int channels = images[0].channels();
75 int CV_32FCC = CV_MAKETYPE(CV_32F, channels);
76
77 dst.create(LDR_SIZE, 1, CV_32FCC);
78 Mat result = dst.getMat();
79
80 std::vector<Point> sample_points;
81 if(random) {
82 for(int i = 0; i < samples; i++) {
83 sample_points.push_back(Point(rand() % images[0].cols, rand() % images[0].rows));
84 }
85 } else {
86 int x_points = static_cast<int>(sqrt(static_cast<double>(samples) * images[0].cols / images[0].rows));
87 int y_points = samples / x_points;
88 int step_x = images[0].cols / x_points;
89 int step_y = images[0].rows / y_points;
90
91 for(int i = 0, x = step_x / 2; i < x_points; i++, x += step_x) {
92 for(int j = 0, y = step_y / 2; j < y_points; j++, y += step_y) {
93 if( 0 <= x && x < images[0].cols &&
94 0 <= y && y < images[0].rows )
95 sample_points.push_back(Point(x, y));
96 }
97 }
98 }
99
100 std::vector<Mat> result_split(channels);
101 for(int channel = 0; channel < channels; channel++) {
102 Mat A = Mat::zeros((int)sample_points.size() * (int)images.size() + LDR_SIZE + 1, LDR_SIZE + (int)sample_points.size(), CV_32F);
103 Mat B = Mat::zeros(A.rows, 1, CV_32F);
104
105 int eq = 0;
106 for(size_t i = 0; i < sample_points.size(); i++) {
107 for(size_t j = 0; j < images.size(); j++) {
108
109 int val = images[j].ptr()[3*(sample_points[i].y * images[j].cols + sample_points[i].x) + channel];
110 A.at<float>(eq, val) = w.at<float>(val);
111 A.at<float>(eq, LDR_SIZE + (int)i) = -w.at<float>(val);
112 B.at<float>(eq, 0) = w.at<float>(val) * log(times.at<float>((int)j));
113 eq++;
114 }
115 }
116 A.at<float>(eq, LDR_SIZE / 2) = 1;
117 eq++;
118
119 for(int i = 0; i < 254; i++) {
120 A.at<float>(eq, i) = lambda * w.at<float>(i + 1);
121 A.at<float>(eq, i + 1) = -2 * lambda * w.at<float>(i + 1);
122 A.at<float>(eq, i + 2) = lambda * w.at<float>(i + 1);
123 eq++;
124 }
125 Mat solution;
126 solve(A, B, solution, DECOMP_SVD);
127 solution.rowRange(0, LDR_SIZE).copyTo(result_split[channel]);
128 }
129 merge(result_split, result);
130 exp(result, result);
131 }
132
getSamples() const133 int getSamples() const { return samples; }
setSamples(int val)134 void setSamples(int val) { samples = val; }
135
getLambda() const136 float getLambda() const { return lambda; }
setLambda(float val)137 void setLambda(float val) { lambda = val; }
138
getRandom() const139 bool getRandom() const { return random; }
setRandom(bool val)140 void setRandom(bool val) { random = val; }
141
write(FileStorage & fs) const142 void write(FileStorage& fs) const
143 {
144 fs << "name" << name
145 << "samples" << samples
146 << "lambda" << lambda
147 << "random" << static_cast<int>(random);
148 }
149
read(const FileNode & fn)150 void read(const FileNode& fn)
151 {
152 FileNode n = fn["name"];
153 CV_Assert(n.isString() && String(n) == name);
154 samples = fn["samples"];
155 lambda = fn["lambda"];
156 int random_val = fn["random"];
157 random = (random_val != 0);
158 }
159
160 protected:
161 String name;
162 int samples;
163 float lambda;
164 bool random;
165 Mat w;
166 };
167
createCalibrateDebevec(int samples,float lambda,bool random)168 Ptr<CalibrateDebevec> createCalibrateDebevec(int samples, float lambda, bool random)
169 {
170 return makePtr<CalibrateDebevecImpl>(samples, lambda, random);
171 }
172
173 class CalibrateRobertsonImpl : public CalibrateRobertson
174 {
175 public:
CalibrateRobertsonImpl(int _max_iter,float _threshold)176 CalibrateRobertsonImpl(int _max_iter, float _threshold) :
177 name("CalibrateRobertson"),
178 max_iter(_max_iter),
179 threshold(_threshold),
180 weight(RobertsonWeights())
181 {
182 }
183
process(InputArrayOfArrays src,OutputArray dst,InputArray _times)184 void process(InputArrayOfArrays src, OutputArray dst, InputArray _times)
185 {
186 std::vector<Mat> images;
187 src.getMatVector(images);
188 Mat times = _times.getMat();
189
190 CV_Assert(images.size() == times.total());
191 checkImageDimensions(images);
192 CV_Assert(images[0].depth() == CV_8U);
193
194 int channels = images[0].channels();
195 int CV_32FCC = CV_MAKETYPE(CV_32F, channels);
196
197 dst.create(LDR_SIZE, 1, CV_32FCC);
198 Mat response = dst.getMat();
199 response = linearResponse(3) / (LDR_SIZE / 2.0f);
200
201 Mat card = Mat::zeros(LDR_SIZE, 1, CV_32FCC);
202 for(size_t i = 0; i < images.size(); i++) {
203 uchar *ptr = images[i].ptr();
204 for(size_t pos = 0; pos < images[i].total(); pos++) {
205 for(int c = 0; c < channels; c++, ptr++) {
206 card.at<Vec3f>(*ptr)[c] += 1;
207 }
208 }
209 }
210 card = 1.0 / card;
211
212 Ptr<MergeRobertson> merge = createMergeRobertson();
213 for(int iter = 0; iter < max_iter; iter++) {
214
215 radiance = Mat::zeros(images[0].size(), CV_32FCC);
216 merge->process(images, radiance, times, response);
217
218 Mat new_response = Mat::zeros(LDR_SIZE, 1, CV_32FC3);
219 for(size_t i = 0; i < images.size(); i++) {
220 uchar *ptr = images[i].ptr();
221 float* rad_ptr = radiance.ptr<float>();
222 for(size_t pos = 0; pos < images[i].total(); pos++) {
223 for(int c = 0; c < channels; c++, ptr++, rad_ptr++) {
224 new_response.at<Vec3f>(*ptr)[c] += times.at<float>((int)i) * *rad_ptr;
225 }
226 }
227 }
228 new_response = new_response.mul(card);
229 for(int c = 0; c < 3; c++) {
230 float middle = new_response.at<Vec3f>(LDR_SIZE / 2)[c];
231 for(int i = 0; i < LDR_SIZE; i++) {
232 new_response.at<Vec3f>(i)[c] /= middle;
233 }
234 }
235 float diff = static_cast<float>(sum(sum(abs(new_response - response)))[0] / channels);
236 new_response.copyTo(response);
237 if(diff < threshold) {
238 break;
239 }
240 }
241 }
242
getMaxIter() const243 int getMaxIter() const { return max_iter; }
setMaxIter(int val)244 void setMaxIter(int val) { max_iter = val; }
245
getThreshold() const246 float getThreshold() const { return threshold; }
setThreshold(float val)247 void setThreshold(float val) { threshold = val; }
248
getRadiance() const249 Mat getRadiance() const { return radiance; }
250
write(FileStorage & fs) const251 void write(FileStorage& fs) const
252 {
253 fs << "name" << name
254 << "max_iter" << max_iter
255 << "threshold" << threshold;
256 }
257
read(const FileNode & fn)258 void read(const FileNode& fn)
259 {
260 FileNode n = fn["name"];
261 CV_Assert(n.isString() && String(n) == name);
262 max_iter = fn["max_iter"];
263 threshold = fn["threshold"];
264 }
265
266 protected:
267 String name;
268 int max_iter;
269 float threshold;
270 Mat weight, radiance;
271 };
272
createCalibrateRobertson(int max_iter,float threshold)273 Ptr<CalibrateRobertson> createCalibrateRobertson(int max_iter, float threshold)
274 {
275 return makePtr<CalibrateRobertsonImpl>(max_iter, threshold);
276 }
277
278 }
279