-
Notifications
You must be signed in to change notification settings - Fork 7
/
CH7_color_image_processing.cpp
617 lines (511 loc) · 18.2 KB
/
CH7_color_image_processing.cpp
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
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
#include <opencv2/opencv.hpp>
#include <opencv2/core/types.hpp>
#include <opencv2/imgproc.hpp> //line
#include <opencv2/imgcodecs.hpp>
#include <opencv2/core/hal/interface.h> //CV_8UC3
#include <iostream>
#include <map>
#include <cmath> //M_PI
#include "CH7.h"
// #define MODE 1
using namespace std;
double eps = std::numeric_limits<double>::epsilon();
// double eps = 0.0;
bool RGB2CMY(cv::Mat& img){
//p.239
//it also serves as CMY2RGB
//Note that even if there is RGB in function name, the image's channel order is BGR
if(type2str(img.type()) != "8UC3") return false;
for(int row = 0; row < img.rows; row++){
for(int col = 0; col < img.cols; col++){
cv::Vec3b pixel = img.at<cv::Vec3b>(row, col);
pixel[0] = 255 - pixel[0]; //B
pixel[1] = 255 - pixel[1]; //G
pixel[2] = 255 - pixel[2]; //R
img.at<cv::Vec3b>(row, col) = pixel;
}
}
return true;
};
bool CMY2RGB(cv::Mat& img){
return RGB2CMY(img);
};
bool RGB2HSI(cv::Mat& img){ //?
//p.243
if(type2str(img.type()) != "8UC3") return false;
for(int row = 0; row < img.rows; row++){
for(int col = 0; col < img.cols; col++){
cv::Vec3b pixel = img.at<cv::Vec3b>(row, col);
double B = pixel[0]/255.0;
double G = pixel[1]/255.0;
double R = pixel[2]/255.0;
double minColor = min({B, G, R});
double maxColor = max({B, G, R});
double H;
double S;
double I = (R+G+B)/3.0; //[0,1]
// given in textbook??
// if(I < 0.078431){
// S = 0;
// }else if(I > 0.92){
// S = 0;
// }else{
// S = 1 - 3*minColor/(R+G+B);
// }
if(minColor == maxColor){
H = 0;
S = 0;
}else{
//its unit is radiance
double cosVal = (R-G+R-B)/2.0 / sqrt((R-G)*(R-G) + (R-B)*(G-B) + eps);
double radiance = (abs(cosVal-1) < eps) ? 0.0 : (abs(cosVal-(-1)) < eps) ? M_PI : acos(cosVal);
// double radiance = acos(radiance);
H = (B <= G) ? radiance : 2.0*M_PI - radiance; //[0,2.0*M_PI]
S = 1 - 3.0*minColor/(R+G+B + eps); //[0,1]
}
//normalize
H = H/(2.0*M_PI) * 255.0;
S = S * 255.0;
I = I * 255.0;
//put H in R channel, S in G channel, I in B channel
img.at<cv::Vec3b>(row, col) = cv::Vec3b((int)I, (int)S, (int)H);
}
}
return true;
};
bool HSI2RGB(cv::Mat& img){ //?
//p.247
if(type2str(img.type()) != "8UC3") return false;
for(int row = 0; row < img.rows; row++){
for(int col = 0; col < img.cols; col++){
cv::Vec3b pixel = img.at<cv::Vec3b>(row, col);
double I = pixel[0]/255.0; //normalize to [0,1]
double S = pixel[1]/255.0; //normalize to [0,1]
double H = pixel[2]/255.0*2*M_PI; //normalize to [0,2*M_PI]
double B, G, R;
if(H >= 0 && H < M_PI*2.0/3){ //0 to 120 degrees
//RG sector
B = I * (1.0 - S);
R = I * (1.0 + S*cos(H)/cos(M_PI/3.0 - H));
G = 3.0*I - (R+B);
}else if(H >= M_PI*2.0/3 && H < M_PI*4.0/3){ //120 to 240 degrees
//GB sector
H = H - M_PI*2.0/3;
R = I * (1.0 - S);
G = I * (1.0 + S*cos(H)/cos(M_PI/3.0 - H));
B = 3.0*I - (R+G);
}else if(H >= M_PI*4.0/3){ //larger than 240 degrees
//BR sector
H = H - M_PI*4.0/3;
G = I * (1.0 - S);
B = I * (1.0 + S*cos(H)/cos(M_PI/3.0 - H));
R = 3.0*I - (G+B);
}
R *= 255.0;
G *= 255.0;
B *= 255.0;
img.at<cv::Vec3b>(row, col) = cv::Vec3b((int)B, (int)G, (int)R);
}
}
return true;
};
bool RGB2HSV(cv::Mat& img){
//p.250
if(type2str(img.type()) != "8UC3") return false;
for(int row = 0; row < img.rows; row++){
for(int col = 0; col < img.cols; col++){
cv::Vec3b pixel = img.at<cv::Vec3b>(row, col);
double B = pixel[0]/255.0;
double G = pixel[1]/255.0;
double R = pixel[2]/255.0;
double minColor = min({B, G, R});
double maxColor = max({B, G, R});
double H, S, V;
V = maxColor;
if(maxColor == 0){
S = 0;
H = 0; //undefined color
}else{
S = (maxColor - minColor)/maxColor;
if(R == maxColor){
H = (G-B)/(maxColor-minColor)*M_PI/3.0;
}else if(G == maxColor){
H = (2 + (B-R)/(maxColor-minColor))*M_PI/3.0;
}else if(B == maxColor){
H = (4 + (R-G)/(maxColor-minColor))*M_PI/3.0;
}
if(H < 0){
H += 2*M_PI;
}
}
//normalize
H = H/(2.0*M_PI) * 255.0;
S = S * 255.0;
V = V * 255.0;
//put H in R channel, S in G channel, V in B channel
img.at<cv::Vec3b>(row, col) = cv::Vec3b((int)V, (int)S, (int)H);
}
}
return true;
};
bool HSV2RGB(cv::Mat& img){
//p.250
if(type2str(img.type()) != "8UC3") return false;
for(int row = 0; row < img.rows; row++){
for(int col = 0; col < img.cols; col++){
cv::Vec3b pixel = img.at<cv::Vec3b>(row, col);
double V = pixel[0]/255.0; //normalize to [0,1]
double S = pixel[1]/255.0; //normalize to [0,1]
double H = pixel[2]/255.0*2*M_PI; //normalize to [0,2*M_PI]
double B, G, R;
int i = H/(M_PI/3.0); //falls in which sector
double f = H/(M_PI/3.0) - i;
double p = V * (1-S);
double q = V * (1-f*S);
double t = V * (1 - (1-f) * S);
switch(i){
case 0:
R = V;
G = t;
B = p;
break;
case 1:
R = q;
G = V;
B = p;
break;
case 2:
R = q;
G = V;
B = t;
break;
case 3:
R = p;
G = q;
B = V;
break;
case 4:
R = t;
G = p;
B = V;
break;
case 5:
R = V;
G = p;
B = q;
break;
}
R *= 255.0;
G *= 255.0;
B *= 255.0;
img.at<cv::Vec3b>(row, col) = cv::Vec3b((int)B, (int)G, (int)R);
}
}
return true;
};
bool RGB2YUV(cv::Mat& img){
//p.256
if(type2str(img.type()) != "8UC3") return false;
for(int row = 0; row < img.rows; row++){
for(int col = 0; col < img.cols; col++){
cv::Vec3b pixel = img.at<cv::Vec3b>(row, col);
double B = pixel[0];
double G = pixel[1];
double R = pixel[2];
double Y = 0.299*R + 0.587*G + 0.114*B;
double U = 0.567*(B-Y);
double V = 0.713*(R-Y);
//truncate value exceed the range, it seems we can use a better normalization method?
Y = min(max((int)Y, 0), 255);
U = min(max((int)U, 0), 255);
V = min(max((int)V, 0), 255);
img.at<cv::Vec3b>(row, col) = cv::Vec3b((int)V, (int)U, (int)Y);
}
}
return true;
};
bool YUV2RGB(cv::Mat& img){
//p.259
if(type2str(img.type()) != "8UC3") return false;
for(int row = 0; row < img.rows; row++){
for(int col = 0; col < img.cols; col++){
cv::Vec3b pixel = img.at<cv::Vec3b>(row, col);
double V = pixel[0];
double U = pixel[1];
double Y = pixel[2];
double R = Y + 1.402 * V;
double G = Y - 0.344 * U - 0.714 * V;
double B = Y + 1.772 * U;
R = min(max((int)R, 0), 255);
G = min(max((int)G, 0), 255);
B = min(max((int)B, 0), 255);
img.at<cv::Vec3b>(row, col) = cv::Vec3b((int)B, (int)G, (int)R);
}
}
return true;
};
bool RGB2YIQ(cv::Mat& img){
//p.261
if(type2str(img.type()) != "8UC3") return false;
for(int row = 0; row < img.rows; row++){
for(int col = 0; col < img.cols; col++){
cv::Vec3b pixel = img.at<cv::Vec3b>(row, col);
double B = pixel[0];
double G = pixel[1];
double R = pixel[2];
double Y = 0.299*R + 0.587*G + 0.114*B;
double I = 0.596*R - 0.274*G - 0.322*B;
double Q = 0.211*R - 0.523*G + 0.312*B;
//truncate value exceed the range, it seems we can use a better normalization method?
Y = min(max((int)Y, 0), 255);
I = min(max((int)I, 0), 255);
Q = min(max((int)Q, 0), 255);
img.at<cv::Vec3b>(row, col) = cv::Vec3b((int)Q, (int)I, (int)Y);
}
}
return true;
};
bool YIQ2RGB(cv::Mat& img){
//p.263
if(type2str(img.type()) != "8UC3") return false;
for(int row = 0; row < img.rows; row++){
for(int col = 0; col < img.cols; col++){
cv::Vec3b pixel = img.at<cv::Vec3b>(row, col);
double Q = pixel[0];
double I = pixel[1];
double Y = pixel[2];
double R = Y + 0.956*I + 0.114*Q;
double G = Y - 0.272*I - 0.647*Q;
double B = Y - 1.106*I + 1.703*Q;
R = min(max((int)R, 0), 255);
G = min(max((int)G, 0), 255);
B = min(max((int)B, 0), 255);
img.at<cv::Vec3b>(row, col) = cv::Vec3b((int)B, (int)G, (int)R);
}
}
return true;
};
bool compensate(cv::Mat& img){
//p.266
/*
select most red, green, blue points int the image and
learn a linear transformation to convert them to true r,g,b
*/
if(type2str(img.type()) != "8UC3") return false;
int rMaxR = INT_MIN;
int cMaxR = INT_MIN;
int rMaxG = INT_MIN;
int cMaxG = INT_MIN;
int rMaxB = INT_MIN;
int cMaxB = INT_MIN;
double maxR = std::numeric_limits<double>::lowest();
double maxG = std::numeric_limits<double>::lowest();
double maxB = std::numeric_limits<double>::lowest();
double minR = std::numeric_limits<double>::max();
double minG = std::numeric_limits<double>::max();
double minB = std::numeric_limits<double>::max();
for(int row = 0; row < img.rows; row++){
for(int col = 0; col < img.cols; col++){
cv::Vec3b pixel = img.at<cv::Vec3b>(row, col);
double B = pixel[0];
double G = pixel[1];
double R = pixel[2];
#if MODE == 1
if(R-(B+G) > maxR){
#else
if(255-R+B+G < minR){
#endif
rMaxR = row;
cMaxR = col;
#if MODE == 1
maxR = R-(B+G);
#else
minR = 255-R+B+G;
#endif
}
#if MODE == 1
if(G-(B+R) > maxG){
#else
if(255-G+B+R < minG){
#endif
rMaxG = row;
cMaxG = col;
#if MODE == 1
maxG = G-(B+R);
#else
minG = 255-G+B+R;
#endif
}
#if MODE == 1
if(B-(R+G) > maxB){
#else
if(255-B+R+G < minB){
#endif
rMaxB = row;
cMaxB = col;
#if MODE == 1
maxB = B-(R+G);
#else
minB = 255-B+R+G;
#endif
}
}
}
// cout << "Red: " << rMaxR << " " << cMaxR << " " << maxR << " " << minR << endl;
// cout << "Green: " << rMaxG << " " << cMaxG << " " << maxG << " " << minG << endl;
// cout << "Blue: " << rMaxB << " " << cMaxB << " " << maxB << " " << minB << endl;
//old three point's rgb value
double r1 = img.at<cv::Vec3b>(rMaxR, cMaxR)[2], g1 = img.at<cv::Vec3b>(rMaxR, cMaxR)[1], b1 = img.at<cv::Vec3b>(rMaxR, cMaxR)[0];
double r2 = img.at<cv::Vec3b>(rMaxG, cMaxG)[2], g2 = img.at<cv::Vec3b>(rMaxG, cMaxG)[1], b2 = img.at<cv::Vec3b>(rMaxG, cMaxG)[0];
double r3 = img.at<cv::Vec3b>(rMaxB, cMaxB)[2], g3 = img.at<cv::Vec3b>(rMaxB, cMaxB)[1], b3 = img.at<cv::Vec3b>(rMaxB, cMaxB)[0];
vector<vector<double>> A1 = {
{r1, r2, r3},
{g1, g2, g3},
{b1, b2, b3}
};
//which does the best?
//given by textbook, brightness will be lower?
double R = 0.3*r1 + 0.59*g1 + 0.11*b1;
double G = 0.3*r2 + 0.59*g2 + 0.11*b2;
double B = 0.3*r3 + 0.59*g3 + 0.11*b3;
//*3 to restore its brightness?
// double R = (0.3*r1 + 0.59*g1 + 0.11*b1)*3;
// double G = (0.3*r2 + 0.59*g2 + 0.11*b2)*3;
// double B = (0.3*r3 + 0.59*g3 + 0.11*b3)*3;
//assume each channel contribute same brightness?
// double R = (r1+b1+g1)/3.0;
// double G = (r2+b2+g2)/3.0;
// double B = (r3+b3+g3)/3.0;
//*3 to restore its brightness?
// double R = (r1+b1+g1);
// double G = (r2+b2+g2);
// double B = (r3+b3+g3);
vector<vector<double>> A2 = {
{R, 0, 0},
{0, G, 0},
{0, 0, B}
};
vector<vector<double>> invA1 = A1;
InvMat(invA1);
vector<vector<double>> invA2 = A2;
InvMat(invA2);
vector<vector<double>> C;
ProdMat(A1, invA2, C);
vector<vector<double>> invC = C;
InvMat(invC);
// cout << "A1" << endl;
// PrintMatrix(A1);
// cout << "A2" << endl;
// PrintMatrix(A2);
// cout << "invA2" << endl;
// PrintMatrix(invA2);
// cout << "C" << endl;
// PrintMatrix(C);
// cout << "invC" << endl;
// PrintMatrix(invC);
for(int row = 0; row < img.rows; row++){
for(int col = 0; col < img.cols; col++){
cv::Vec3b pixel = img.at<cv::Vec3b>(row, col);
double B = pixel[0];
double G = pixel[1];
double R = pixel[2];
vector<vector<double>> oldColor = {{R}, {G}, {B}}; // 3 x 1 matrix
vector<vector<double>> newColor; //also a 3 x 1 matrix
ProdMat(invC, oldColor, newColor);
// ProdMat(invA1, oldColor, newColor);
// oldColor = newColor;
// ProdMat(A2, oldColor, newColor);
for(int i = 0; i < 3; i++){
newColor[i][0] = min(max((int)newColor[i][0], 0), 255); ;
}
// if(row == 0 && col % (img.cols/10) == 0){
// cout << "old" << endl;
// PrintMatrix(oldColor);
// cout << "new" << endl;
// PrintMatrix(newColor);
// }
img.at<cv::Vec3b>(row, col) = cv::Vec3b((int)newColor[2][0], (int)newColor[1][0], (int)newColor[0][0]);
}
}
return true;
};
bool colorBalance(cv::Mat& img){
//p.268
//fix green color, do linear transform on red and blue to make the two chosen points "gray"
if(type2str(img.type()) != "8UC3") return false;
cv::Vec3b F1 = img.at<cv::Vec3b>(0, 0);
//select farther point to avoid "Floating point exception"
// cv::Vec3b F2 = img.at<cv::Vec3b>(50, 50);
cv::Vec3b F2 = img.at<cv::Vec3b>(0, 1);
cv::Vec3b newF1(F1[1], F1[1], F1[1]);
cv::Vec3b newF2(F2[1], F2[1], F2[1]);
//red
double K1 = (double)(newF1[2]-newF2[2])/(double)(F1[2]-F2[2]);
double K2 = newF1[2] - K1 * F1[2];
//blue
double L1 = (double)(newF1[0]-newF2[0])/(double)(F1[0]-F2[0]);
double L2 = newF1[0] - L1 * F1[0];
for(int row = 0; row < img.rows; row++){
for(int col = 0; col < img.cols; col++){
cv::Vec3b pixel = img.at<cv::Vec3b>(row, col);
double B = pixel[0];
double G = pixel[1];
double R = pixel[2];
double newR = min(max((int)(K1 * R + K2), 0), 255);
double newG = G;
double newB = min(max((int)(L1 * B + K2), 0), 255);
img.at<cv::Vec3b>(row, col) = cv::Vec3b((int)newB, (int)newG, (int)newR);
}
}
return true;
};
#ifdef CH7
int main(){
cv::Mat img_lenna = cv::imread("images/Lenna.png");
cv::Mat img_plane = cv::imread("images/plane.bmp");
cv::Mat img_rgb = cv::imread("images/rgb.tif");
bool isSave = false;
cv::Mat cmy = img_lenna.clone();
RGB2CMY(cmy);
cv::Mat cmy_back_color = cmy.clone();
CMY2RGB(cmy_back_color);
vector<cv::Mat> cmys = {img_lenna, cmy, cmy_back_color};
ShowHorizontal(cmys, "CMY vs CMY BACK RGB", isSave);
cv::Mat hsi = img_lenna.clone();
RGB2HSI(hsi);
cv::Mat hsi_back_color = hsi.clone();
HSI2RGB(hsi_back_color);
vector<cv::Mat> hsis = {img_lenna, hsi, hsi_back_color};
ShowHorizontal(hsis, "HSI vs HSI BACK RGB", isSave);
cv::Mat hsv = img_lenna.clone();
RGB2HSV(hsv);
cv::Mat hsv_back_color = hsv.clone();
HSV2RGB(hsv_back_color);
vector<cv::Mat> hsvs = {img_lenna, hsv, hsv_back_color};
ShowHorizontal(hsvs, "HSV vs HSV BACK RGB", isSave);
cv::Mat yuv = img_lenna.clone();
RGB2YUV(yuv);
cv::Mat yuv_back_color = yuv.clone();
YUV2RGB(yuv_back_color);
vector<cv::Mat> yuvs = {img_lenna, yuv, yuv_back_color};
ShowHorizontal(yuvs, "YUV vs YUV BACK RGB", isSave);
cv::Mat yiq = img_lenna.clone();
RGB2YIQ(yiq);
cv::Mat yiq_back_color = yiq.clone();
YIQ2RGB(yiq_back_color);
vector<cv::Mat> yiqs = {img_lenna, yiq, yiq_back_color};
ShowHorizontal(yiqs, "YIQ vs YIQ BACK RGB", isSave);
cv::Mat compensated = img_plane.clone();
compensate(compensated);
vector<cv::Mat> compensateds = {img_plane, compensated};
ShowHorizontal(compensateds, "plane before and after compensating", isSave);
cv::Mat rgbCompensated = img_rgb.clone();
compensate(rgbCompensated);
vector<cv::Mat> rgbCompensateds = {img_rgb, rgbCompensated};
ShowHorizontal(rgbCompensateds, "rgb before and after compensating", isSave);
cv::Mat balanced = img_plane.clone();
colorBalance(balanced);
vector<cv::Mat> balanceds = {img_plane, balanced};
ShowHorizontal(balanceds, "Before and after color balancing", isSave);
}
#endif