-
Notifications
You must be signed in to change notification settings - Fork 1
/
main.cpp
4420 lines (3127 loc) · 132 KB
/
main.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
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
// Copyright (c) 2016 Guilherme POTJE.
// This Source Code Form is subject to the terms of the Mozilla Public
// License, v. 2.0. If a copy of the MPL was not distributed with this
// file, You can obtain one at http://mozilla.org/MPL/2.0/.
//Sites de referencia:
//http://docs.opencv.org/doc/tutorials/features2d/feature_homography/feature_homography.html
//https://code.ros.org/trac/opencv/browser/trunk/opencv/samples/c/detectors_sample.cpp?rev=3052
//---------------------------------------------------------------------------------------------
//#include <pcl/point_types.h>
//#include <pcl/filters/statistical_outlier_removal.h>
#include<utility>
#include<string.h>
#include<stdio.h>
#include<iostream>
#include<stdlib.h>
#include<queue>
#include<map>
#include<set>
#include "opencv2/core.hpp"
#include <opencv2/core/utility.hpp>
#include "opencv2/imgproc.hpp"
#include "opencv2/calib3d.hpp"
#include "opencv2/imgcodecs.hpp"
#include "opencv2/videoio.hpp"
#include "opencv2/highgui.hpp"
#include <opencv2/features2d.hpp>
#include <opencv2/opencv.hpp>
#include <unistd.h>
//#include <sys/time.h>
//#include "codigos/Utilitarios.h"
#include "codigos/ceres_BA.cpp"
#include "codigos/BundleAdjustment.h" //
#include "codigos/VocabTree.h"
#include "codigos/AC_Ransac.h"
//#include <pcl/visualization/cloud_viewer.h>
//reconstruction
/*#include <pcl/surface/mls.h>
#include <pcl/surface/poisson.h>
#include <pcl/common/common.h>
#include <pcl/io/pcd_io.h>
#include <pcl/features/normal_3d_omp.h>
#include <pcl/io/vtk_io.h>
#include <pcl/point_types.h>
#include <pcl/kdtree/kdtree_flann.h>
#include <pcl/features/normal_3d.h>
#include <pcl/surface/gp3.h>
#include <pcl/filters/passthrough.h>
*/
#define EPSILON 0.0001
#define GLOBAL_BA true //run full ba in the end
#define LOCAL_BA true //perform local BA in the reconstruction progress
#define ACONTRARIO true //use the a contrario scheme for RANSAC
#define GBA_FACTOR 2.0 //1.33, 1.5, 1.7
#define UPPER_BOUND 40.0
#define FILTER_POINTS true
using namespace cv;
using namespace std;
//
cv::Mat K; //intrinsic calib
cv::Mat D;
vector<cv::Mat> Kvec; //intrinsic for each cams
vector<cv::Mat> Dvec; // distortion for each cam
cv::Mat distortion; //distortion coeffs of lens
int c_img=0;
int rectf_count=0;
int pair_count=0;
int pnp_count=0;
long ba_counter = 1;
int next_ba;
pair<double,long> confidence;
struct timeval start, end;
long long mtime, seconds, useconds;
struct pairMatch
{
vector<long> idx_cam1, idx_cam2;
vector<Point2f> cam1, cam2;
Mat F_mat;
bool triangulated;
bool h_tested;
int homog_inliers;
};
#include "codigos/thread_Match.h"
#include "codigos/parallel_Features.h"
#include "codigos/input_output.h"
#include "codigos/graph_utils.h"
void remove_outliersfromcloud(vector<cloudPoint*> Cloud);
void reconstruct_surface(vector<cloudPoint*> Cloud);
cv::Mat find_ray(Matx34d P1, Point2f n_x1);
cv::Mat find_ray_3d(Matx34d P1, Mat X); //Normalized P_mat and 3D point position
double angle_between_vec(Mat v1, Mat v2);
Mat LinearLSTriangulation(Point3d u, //homogenous image point (u,v,1)
Matx34d P, //camera 1 matrix
Point3d u1, //homogenous image point in 2nd camera
Matx34d P1 //camera 2 matrix
);
bool DecomposeEtoRandT(
Mat_<double>& E,
Mat_<double>& R1,
Mat_<double>& R2,
Mat_<double>& t1,
Mat_<double>& t2);
Mat_<double> IterativeLinearLSTriangulation(Point3d u, //homogenous image point (u,v,1)
Matx34d P, //camera 1 matrix
Point3d u1, //homogenous image point in 2nd camera
Matx34d P1 //camera 2 matrix
);
bool CheckCoherentRotation(cv::Mat_<double>& R) {
if(fabsf(determinant(R))-1.0 > 1e-07) {
//cerr << "det(R) != +-1.0, this is not a rotation matrix" << endl;
return false;
}
return true;
}
bool TestTriangulation(vector<Point3d> pcloud_pt3d, Matx34d P)
{
vector<Point3d> pcloud_pt3d_projected(pcloud_pt3d.size());
vector<uchar> status;
Matx44d P4x4 = Matx44d::eye();
for(int i=0;i<12;i++)
P4x4.val[i] = P.val[i];
perspectiveTransform(pcloud_pt3d, pcloud_pt3d_projected, P4x4);
status.resize(pcloud_pt3d.size(),0);
for (int i=0; i<pcloud_pt3d.size(); i++) {
status[i] = (pcloud_pt3d_projected[i].z > 0) ? 1 : 0;
}
int count = countNonZero(status);
double percentage = ((double)count / (double)pcloud_pt3d.size());
//cout << count << "/" << pcloud_pt3d.size() << " = " << percentage*100.0 << "% are in front of camera" << endl;
if(percentage < 0.75) // || percentage > 0.985)
return false; //less than 75% of the points are in front of the camera
//check for coplanarity of points
if(false) //not
{
cv::Mat_<double> cldm(pcloud_pt3d.size(),3);
for(unsigned int i=0;i<pcloud_pt3d.size();i++) {
cldm.row(i)(0) = pcloud_pt3d[i].x;
cldm.row(i)(1) = pcloud_pt3d[i].y;
cldm.row(i)(2) = pcloud_pt3d[i].z;
}
cv::Mat_<double> mean;
cv::PCA pca(cldm,mean,CV_PCA_DATA_AS_ROW);
double p_to_plane_thresh = pca.eigenvalues.at<double>(2);
int num_inliers = 0;
cv::Vec3d nrm = pca.eigenvectors.row(2); nrm = nrm / norm(nrm);
cv::Vec3d x0 = pca.mean;
for (int i=0; i<pcloud_pt3d.size(); i++)
{
Vec3d w = Vec3d(pcloud_pt3d[i]) - x0;
double D = fabs(nrm.dot(w));
if(D < p_to_plane_thresh) num_inliers++;
}
cout << num_inliers << "/" << pcloud_pt3d.size() << " are coplanar" << endl;
if((double)num_inliers / (double)(pcloud_pt3d.size()) > 0.9)
return false;
}
return true;
}
double triangulatePoints_opencv(cv::Matx34d P1, cv::Matx34d P, vector<Point2f> cam1, vector<Point2f> cam2, vector<Point3d> &pcloud, vector<double> reproj_error)
{
vector<Point2d> cam_1, cam_2;
Mat triangulated_3d;
for (int i=0; i<cam1.size(); i++)
{
Point3f pt3d;
Point2f kp = cam1[i];
Point3d u(kp.x,kp.y,1.0);
Mat_<double> um = K.inv() * Mat_<double>(u);
u.x = um(0); u.y = um(1); u.z = um(2);
Point2f kp1 = cam2[i];
Point3d u1(kp1.x,kp1.y,1.0);
Mat_<double> um1 = K.inv() * Mat_<double>(u1);
u1.x = um1(0); u1.y = um1(1); u1.z = um1(2);
Point2d pc1 = Point2d(u.x/u.z,u.y/u.z);
Point2d pc2 = Point2d(u1.x/u1.z, u1.y/u1.z);
cam_1.push_back(pc1);
cam_2.push_back(pc2);
}
triangulatePoints(P,P1,cam_1,cam_2,triangulated_3d);
for(int i=0;i<triangulated_3d.cols;i++)
{
double h = triangulated_3d.at<double>(3,i);
Point3d p = Point3d(triangulated_3d.at<double>(0,i)/h,triangulated_3d.at<double>(1,i)/h,triangulated_3d.at<double>(2,i)/h);
pcloud.push_back(p);
}
return 0;
}
double TriangulatePoints(cv::Matx34d P1, cv::Matx34d P, vector<Point2f> cam1, vector<Point2f> cam2, vector<Point3d> &pcloud, vector<double> &reproj_error)
{
cv::Mat Kinv = K.inv();
pcloud.clear();
reproj_error.clear();
//transforming to NCC
/* for(int i=0;i<cam1_.size();i++)
{
Mat p = (cv::Mat_<float>(1,3) << cam1_[i].x,cam1_[i].y,1);
Mat p2 = (cv::Mat_<float>(1,3) << cam2_[i].x,cam2_[i].y,1);
p = K.inv()*p;
p2= K.inv()*p2;
Point2f p_1 = Point2f(p.at<float>(0,0),p.at<float>(0,1));
Point2f p_2 = Point2f(p2.at<float>(0,0),p2.at<float>(0,1));
cam1.push_back(p_1);
cam2.push_back(p_2);
}*/
//vector<double> reproj_error;
Mat_<double> KP1 = K * Mat(P1);
Mat_<double> KP2 = K* Mat(P);
for (int i=0; i<cam1.size(); i++)
{
Point3f pt3d;
Point2f kp = cam1[i];
Point3d u(kp.x,kp.y,1.0);
Mat_<double> um = Kinv * Mat_<double>(u);
u.x = um(0); u.y = um(1); u.z = um(2);
Point2f kp1 = cam2[i];
Point3d u1(kp1.x,kp1.y,1.0);
Mat_<double> um1 = Kinv * Mat_<double>(u1);
u1.x = um1(0); u1.y = um1(1); u1.z = um1(2);
Mat_<double> X = IterativeLinearLSTriangulation(u,P,u1,P1);
pt3d.x = X(0);
pt3d.y = X(1);
pt3d.z = X(2);
pcloud.push_back(pt3d);
Mat_<double> xPt_img = KP1 * X; //reproject
Point2f xPt_img_(xPt_img(0)/xPt_img(2),xPt_img(1)/xPt_img(2));
double reprj_err = norm(xPt_img_-kp1);
Mat_<double> xPt_img2 = KP2 * X; //reproject
Point2f xPt_img_2(xPt_img2(0)/xPt_img2(2),xPt_img2(1)/xPt_img2(2));
double reprj_err2 = norm(xPt_img_2-kp);
reproj_error.push_back(max(reprj_err,reprj_err2));
}
//Scalar mse = mean(reproj_error);
std::sort(reproj_error.begin(),reproj_error.end());
//return mse[0];
return reproj_error[reproj_error.size()/2];
}
double TriangulatePointsRectified(cv::Matx34d P1, cv::Matx34d P, vector<Point2f> cam1, vector<Point2f> cam2, int idxc1, int idxc2, vector<Point3d> &pcloud, vector<double> &reproj_error, vector<bool> &inliers)
{
cv::Mat Kinv = K.inv();//
pcloud.clear();
reproj_error.clear();
inliers.clear();
Mat_<double> KP1 = Kvec[idxc2] * Mat(P1);
Mat_<double> KP2 = Kvec[idxc1]* Mat(P);
//undistort points
openMVG::sfm::undistort_points(cam1, Kvec[idxc1], Dvec[idxc1]);
openMVG::sfm::undistort_points(cam2, Kvec[idxc2], Dvec[idxc2]);
int cont_outs=0;
int cont_reprjs=0;
for (int i=0; i<cam1.size(); i++)
{
Point3f pt3d;
Point2f kp = cam1[i];
Point3d u(kp.x,kp.y,1.0);
Mat_<double> um = Kvec[idxc1].inv() * Mat_<double>(u); //normalized
u.x = um(0); u.y = um(1); u.z = um(2);
Point2f kp1 = cam2[i];
Point3d u1(kp1.x,kp1.y,1.0);
Mat_<double> um1 = Kvec[idxc2].inv() * Mat_<double>(u1);
u1.x = um1(0); u1.y = um1(1); u1.z = um1(2);
Point2f pt_n1 = Point2f(u.x,u.y);
Point2f pt_n2 = Point2f(u1.x,u1.y);
//cout<<"Angle = "<<angle_between_vec(find_ray(P,pt_n1), find_ray(P1,pt_n2))<<endl; getchar();
Mat_<double> X = IterativeLinearLSTriangulation(u,P,u1,P1);
pt3d.x = X(0);
pt3d.y = X(1);
pt3d.z = X(2);
pcloud.push_back(pt3d);
Mat_<double> xPt_img = KP1 * X; //reproject
Point2f xPt_img_(xPt_img(0)/xPt_img(2),xPt_img(1)/xPt_img(2));
double reprj_err = norm(xPt_img_-kp1);
Mat_<double> xPt_img2 = KP2 * X; //reproject
Point2f xPt_img_2(xPt_img2(0)/xPt_img2(2),xPt_img2(1)/xPt_img2(2));
double reprj_err2 = norm(xPt_img_2-kp);
double max_rerr = max(reprj_err,reprj_err2);
Mat_<double> pt3d_n1 = Mat(P1)*X;
Mat_<double> pt3d_n2 = Mat(P)*X;
reproj_error.push_back(max_rerr);
if(max_rerr > 2.0)
cont_reprjs++;
if(angle_between_vec(find_ray_3d(P,X), find_ray_3d(P1,X)) < 2.0 || pt3d_n1(2) <= 0.0 || pt3d_n2(2) <= 0.0) // validate by triangulation angle and verify if they are in the back of cameras
{inliers.push_back(false); cont_outs++;}
else
inliers.push_back(true);
}
/*cv::Mat_<double> ray1(3,1), ray2(3,1);
ray1(0) = 413; ray1(2) =585; ray1(3) = 0;
ray2(0) = 721; ray2(2) = 333; ray2(3) = 0;
cout<<angle_between_vec(ray1,ray2)<<endl;getchar();
*/
//Scalar mse = mean(reproj_error);
std::sort(reproj_error.begin(),reproj_error.end());
cout<<"Count triangulation outliers = "<<cont_outs <<" in a total of "<<reproj_error.size()<<endl;//". % Points w/ large error "<<100.0*(cont_reprjs/(double)reproj_error.size())<<endl;
//return mse[0];
// cout<<"triangulation median: "<<reproj_error[reproj_error.size()*0.9]<<endl; getchar();
return reproj_error[reproj_error.size()*0.5];
}
bool find_P1(cv::Matx34d &P1, cv::Mat F, vector<Point2f> cam1, vector<Point2f> cam2, vector<double> &reproj_error, int idx_c1, int idx_c2, vector<bool> &inliers)
{
//Essential matrix: compute then extract cameras [R|t]
Mat_<double> E = F;//Kvec[idx_c1].t() * F * Kvec[idx_c2]; //according to HZ (9.12)
vector<Point3d> pcloud;
vector<Point3d> pcloud2;
//vector<bool> inliers;
cv::Matx34d P(1,0,0,0,
0,1,0,0,
0,0,1,0);
if(fabsf(determinant(E)) > 1e-07)
{
//cout << "det(E) != 0 : " << determinant(E) << "\n";
// P1 = 0;
//return false;
}
Mat_<double> R1(3,3);
Mat_<double> R2(3,3);
Mat_<double> t1(1,3);
Mat_<double> t2(1,3);
//decompose E to P' , HZ (9.19)
if (!DecomposeEtoRandT(E,R1,R2,t1,t2)) return false;
if(determinant(R1)+1.0 < 1e-09) {
//according to http://en.wikipedia.org/wiki/Essential_matrix#Showing_that_it_is_valid
//cout << "det(R) == -1 ["<<determinant(R1)<<"]: flip E's sign" << endl;
E = -E;
DecomposeEtoRandT(E,R1,R2,t1,t2);
}
if (!CheckCoherentRotation(R1)) {
cout << "resulting rotation is not coherent\n";
// P1 = 0;
return false;
}
P1 = Matx34d(R1(0,0), R1(0,1), R1(0,2), t1(0),
R1(1,0), R1(1,1), R1(1,2), t1(1),
R1(2,0), R1(2,1), R1(2,2), t1(2));
//cout << "Testing 1 " << endl << Mat(P1) << endl;
double reproj_error1 = TriangulatePointsRectified(P1,P,cam1,cam2,idx_c1,idx_c2,pcloud,reproj_error,inliers);
double reproj_error2 = TriangulatePointsRectified(P,P1,cam2,cam1,idx_c2,idx_c1,pcloud2,reproj_error,inliers);
//cout<<"Reproj error: "<<reproj_error1<<endl;
//cout<<" reproj error 1:"<<reproj_error1<<endl;
//check if pointa are triangulated --in front-- of cameras for all 4 ambiguations
if (!TestTriangulation(pcloud,P1) || !TestTriangulation(pcloud2,P)|| reproj_error1 > 100.0 || reproj_error2>100.0)
{
P1 = Matx34d(R1(0,0), R1(0,1), R1(0,2), t2(0),
R1(1,0), R1(1,1), R1(1,2), t2(1),
R1(2,0), R1(2,1), R1(2,2), t2(2));
//cout << "Testing 2 "<< endl << Mat(P1) << endl;
reproj_error1 = TriangulatePointsRectified(P1,P,cam1,cam2,idx_c1,idx_c2,pcloud,reproj_error,inliers);
reproj_error2 = TriangulatePointsRectified(P,P1,cam2,cam1,idx_c2,idx_c1,pcloud2,reproj_error,inliers);
//cout<<"Reproj error: "<<reproj_error1<<endl;
if (!TestTriangulation(pcloud,P1) || !TestTriangulation(pcloud2,P)|| reproj_error1 > 100.0 || reproj_error2>100.0){
if (!CheckCoherentRotation(R2)) {
cout << "resulting rotation is not coherent\n";
P1 = 0;
return false;
}
P1 = Matx34d(R2(0,0), R2(0,1), R2(0,2), t1(0),
R2(1,0), R2(1,1), R2(1,2), t1(1),
R2(2,0), R2(2,1), R2(2,2), t1(2));
//cout << "Testing 3 "<< endl << Mat(P1) << endl;
reproj_error1 = TriangulatePointsRectified(P1,P,cam1,cam2,idx_c1,idx_c2,pcloud,reproj_error,inliers);
reproj_error2 = TriangulatePointsRectified(P,P1,cam2,cam1,idx_c2,idx_c1,pcloud2,reproj_error,inliers);
//cout<<"Reproj error: "<<reproj_error1<<endl;
if (!TestTriangulation(pcloud,P1) || !TestTriangulation(pcloud2,P)|| reproj_error1 > 100.0 || reproj_error2>100.0){
P1 = Matx34d(R2(0,0), R2(0,1), R2(0,2), t2(0),
R2(1,0), R2(1,1), R2(1,2), t2(1),
R2(2,0), R2(2,1), R2(2,2), t2(2));
//cout << "Testing 4 "<< endl << Mat(P1) << endl;
reproj_error1 = TriangulatePointsRectified(P1,P,cam1,cam2,idx_c1,idx_c2,pcloud,reproj_error,inliers);
reproj_error2 = TriangulatePointsRectified(P,P1,cam2,cam1,idx_c2,idx_c1,pcloud2,reproj_error,inliers);
//cout<<"Reproj error: "<<reproj_error1<<endl;
if (!TestTriangulation(pcloud,P1) || !TestTriangulation(pcloud2,P)|| reproj_error1 > 100.0 || reproj_error2>100.0) {
//cout << "nope" << endl;
return false;
}
}
}
}
return true;
}
double euclidean_distance_latlong(double lat1, double long1, double lat2, double long2)
{
Point2d p1, p2;
p1.x = lat1; p1.y = long1;
p2.x = lat2; p2.y = long2;
return norm(p1-p2);
}
double toRadians(double v)
{
return v*(M_PI/180.0);
}
double distFrom(double lat1, double lng1, double lat2, double lng2)
{
double earthRadius = 6378137; //meters
double dLat =toRadians(lat2-lat1);
double dLng =toRadians(lng2-lng1);
double a = sin(dLat/2) * sin(dLat/2) +
cos(toRadians(lat1)) * cos(toRadians(lat2)) *
sin(dLng/2) * sin(dLng/2);
double c = 2 * atan2(sqrt(a), sqrt(1-a));
double dist = (double) (earthRadius * c);
return abs(dist);
}
Mat getK(double focal_length, double sensor_size_x, double sensor_size_y, int img_width, int img_height)
{
cv::Mat estK;
double w_certo, h_certo;
if(img_width>img_height)
{
w_certo = img_width; h_certo = img_height;
}
else
{ w_certo = img_height; h_certo = img_width; }
if(focal_length==0)
{
focal_length = 4.3; // assuming a generic FoV
sensor_size_x = 6.16;
estK = (cv::Mat_<double>(3,3) << w_certo*1.0,0.0, (img_width/2.0),
0.0, w_certo*1.0,(img_height/2.0),
0.0, 0.0, 1.0 );
cout<<"Using est. calibration matrix: " <<endl << estK<<endl <<"Press key..."; getchar();
}
else{
estK = (cv::Mat_<double>(3,3) << (focal_length*(double)w_certo)/sensor_size_x,0.0, img_width/2.0 - 1.0,
0.0, (focal_length*(double)h_certo)/sensor_size_y, img_height/2.0 - 1.0,
0.0, 0.0, 1.0 );
}
return estK;
}
void generate_I_Graph(Mat &G, vector<imgMatches*> imgs_features)
{
if(true)
{
vector<vector<int> > all_queries;
all_queries = vocabulary_tree_prune(imgs_features); //returns the highest k scores from the vocabulary tree query for each image in imgs_features.
for(int i=0;i<all_queries.size();i++)
for(int j=0;j<all_queries[i].size();j++)
G.at<double>(i,all_queries[i][j]) = 1;
}
else
{
cout<<"Warning: Generating full-connected graph."<<endl;
for(int i=0;i<G.rows;i++)
for(int j=i+1;j<G.cols;j++)
G.at<double>(i,j) = 1;
}
//cout<<"Initial vocab graph:"<<G<<endl;
}
void generate_D_Graph(Mat &G, vector<Point2d> gps_data) //calculates distance in meters between all imgs
{
long cont_m=0;
//initializing
//cout<<"Before prune: "<<G<<endl;
for(int i=0;i<G.rows && i < G.cols;i++) //zeroing the main diagonal to avoid corresponding img to itself
G.at<double>(i,i)= 0;
for(int i=0;i<G.rows;i++)
for(int j=i+1;j<G.cols;j++)
{
// if(gps_data[i].x == 0 || gps_data[i].y == 0 || gps_data[j].x == 0 || gps_data[j].y == 0)
// G.at<double>(i,j) = -1;
// else
if(G.at<double>(i,j)==0) //check if vocab_tree considered this edge
{
// G.at<double>(i,j) = 1;
// G.at<double>(j,i) = 0;
if(!(gps_data[i].x == 0 || gps_data[i].y == 0 || gps_data[j].x == 0 || gps_data[j].y == 0))
{
G.at<double>(i,j) = distFrom(gps_data[i].x,gps_data[i].y,gps_data[j].x,gps_data[j].y);
}
}
//cout<<G.at<double>(i,j)<<endl; getchar();
// if(G.at<double>(i,j)==-1)
// cont_m++;
}
cout<<cont_m<<endl; //getchar();
}
void cut_D_Graph(Mat &G, double meters_threshold) // cut edges based on a threshold in meters
{
for(int i=0;i<G.rows;i++)
for(int j=i+1;j<G.cols;j++)
{
if(G.at<double>(i,j)!=0)
{
if( G.at<double>(i,j) > meters_threshold)
G.at<double>(i,j) = 0;
else
{
pair_count++;
G.at<double>(i,j) = 1;
}
}
}
}
bool sort_matches(DMatch p1, DMatch p2)
{
return (p1.distance < p2.distance);
}
void get_best_matches(vector<DMatch> matches, vector<DMatch> &good_matches, double percentile)
{
good_matches.clear();
vector<DMatch> ordered_matches;
for(unsigned int i=0;i<matches.size(); i++)
ordered_matches.push_back(matches[i]);
std::sort(ordered_matches.begin(),ordered_matches.end(),sort_matches);
for(unsigned int i=0;i< (unsigned int) ( (double)matches.size() * percentile); i++)
good_matches.push_back(ordered_matches[i]);
}
double match_pair(imgMatches *im1, imgMatches *im2, Mat &F_mat, int qualDescritor, vector<Point2f> &cam_1, vector<Point2f> &cam_2, vector<long> &idx_cam1, vector<long> &idx_cam2, vector<cloudPoint*> &Cloud, Mat img1, Mat img2, int &homog_inliers,bool use_flann, double good_matches_percentile, double inliers_threshold, int id_cam1, int id_cam2) //returns qtd de inliers
{
//calculating matching
//
int qualnorm;
bool write_img= true;
bool use_fast_matching = false;
int MAX_KEYPOINTS = 500;
if(qualDescritor<2)
qualnorm = NORM_L2;
else
qualnorm = NORM_HAMMING;
BFMatcher matcher(qualnorm,true);
FlannBasedMatcher fnn_matcher;
vector<DMatch> matches;
/////////////////////////////////////////////////
if(use_fast_matching)
{
int inlier_c=0;
cv::Mat im2k, im1k;
vector<Point2f> cam1t,cam2t;
/*im2k.create(MAX_KEYPOINTS,im2->descriptors_img.cols,CV_32F);
im1k.create(MAX_KEYPOINTS,im2->descriptors_img.cols,CV_32F);
for(int i=0;i<MAX_KEYPOINTS;i++)
{
for(int j=0;j<im2->descriptors_img.cols;j++)
{
im2k.at<float>(i,j)=im2->descriptors_img.at<float>(i,j);
im1k.at<float>(i,j)=im1->descriptors_img.at<float>(i,j);
}
}
*/
if(im2->descriptors_img.rows<10 || im1->descriptors_img.rows<10)
return 0;
for(int i=0;i< im2->descriptors_img.rows && i< im1->descriptors_img.rows && i < MAX_KEYPOINTS;i++)
{
im2k.push_back(im2->descriptors_img.row(i));
im1k.push_back(im1->descriptors_img.row(i));
}
matcher.match(im2k, im1k, matches);
for(unsigned int i = 0; i < matches.size(); i++ )
{
cam1t.push_back( im1->keypoints_img[ matches[i].trainIdx ].pt );
cam2t.push_back( im2->keypoints_img[ matches[i].queryIdx ].pt );
}
double threshold;
std::vector<uchar> inlierst(cam1t.size(),0);
Mat test_mat = findFundamentalMat(cam1t, cam2t, FM_RANSAC, 0.00320*(double)im1->cols, 0.999,inlierst);//0.00120
// Mat test_mat = openMVG::sfm::FindFundamentalMat_AC_RANSAC(cam1t, cam2t,Kvec[id_cam1], Kvec[id_cam2], im1->cols, im1->rows, im2->cols, im2->rows, threshold, inlierst);
std::vector<uchar>::const_iterator
itInt= inlierst.begin();
std::vector<cv::DMatch>::const_iterator
itMt= matches.begin();
// for all matches
for ( ;itInt!= inlierst.end(); ++itInt, ++itMt)
{
if (*itInt) // if inlier
{ // it is a valid match
inlier_c++;
}
}
if(inlier_c<0.15*matches.size())
{
F_mat = (Mat_<double>(3,3)<<0,0,0,0,0,0,0,0,0);
// cout<<"skipping match...."<<c_img <<" "<<inlier_c <<endl;
c_img++;
return 0;
}
cout<<"inlier pair ratio = "<<(inlier_c/(double)matches.size())*100<<"% "<<endl;
matches.clear();
}
//////////////////////////////////////////////////
//FlannBasedMatcher fnn_matcher2(new flann::LshIndexParams(20,10,2));
vector<vector<DMatch> > k_matches;
if(use_flann)
{ if(qualnorm == NORM_L2)
fnn_matcher.knnMatch(im2->descriptors_img, im1->descriptors_img,k_matches,2,Mat(),false);
else
;//fnn_matcher2.match(im2->descriptors_img, im1->descriptors_img, matches);
}
else
matcher.match(im2->descriptors_img, im1->descriptors_img, matches);
//cout<<matches.size()<<endl;
//desenha matches
vector<DMatch> good_matches;
double min_dist=9999, max_dist=0;
/*for( int i = 0; i < matches.size(); i++ )
{
double dist = matches[i].distance;
if( dist < min_dist ) min_dist = dist;
if( dist > max_dist ) max_dist = dist;
}
*/
//for(unsigned int i=0;i<matches.size();i++)
// good_matches.push_back(matches[i]);
//getting the best % keypoint matches
//get_best_matches(matches,good_matches,good_matches_percentile); //20
//Criterion ratio test
for(int i=0;i<k_matches.size();i++)
{
if(k_matches[i][0].distance < 0.8*k_matches[i][1].distance)
good_matches.push_back(k_matches[i][0]);
}
//waitKey(0);
Mat img_matches;
if(write_img)
{
drawMatches(img2, im2->keypoints_img, img1, im1->keypoints_img,
good_matches, img_matches, Scalar::all(-1), Scalar::all(-1),
vector<char>(), DrawMatchesFlags::NOT_DRAW_SINGLE_POINTS);
}
//waitKey(0);
//imshow("Good Matches", img_matches);
string path = "img_mina/result/matches_ransac";
char buffer [33];
sprintf(buffer," %d-%d ",c_img+1,pair_count);
path+=buffer;
path+=".jpg";
cout<<"PATH "<<path<<endl;
//imwrite( path, img_matches );
vector<Point2f> cam1;
vector<Point2f> cam2;
vector<Point2f> n_cam1;
vector<Point2f> n_cam2;
// Mat points3D;
std::vector<uchar> inliers(cam1.size(),0);
std::vector<uchar> inliers_homography(cam1.size(),0);
for(unsigned int i = 0; i < good_matches.size(); i++ )
{
cam1.push_back( im1->keypoints_img[ good_matches[i].trainIdx ].pt );
cam2.push_back( im2->keypoints_img[ good_matches[i].queryIdx ].pt );
}
// Mat fmat1 = findFundamentalMat(cam1, cam1, FM_RANSAC, 2, 0.99);
//Mat fmat1 = (cv::Mat_<double>(3,3) << 1, 0 , 0,
// 0, 1, 0,
// 0, 0, 1); // Fundamental mat image 1
double minVal, maxVal;
minMaxIdx(cam1,&minVal,&maxVal);
double threshold;
Mat fmat2 = cv::findEssentialMat(cam1, cam2, Kvec[id_cam1].at<double>(0,0),Point2d(Kvec[id_cam1].at<double>(0,2),Kvec[id_cam1].at<double>(1,2)),cv::RANSAC, 0.999,(0.00220*(double)im1->cols),inliers);
// Mat fmat2 = findFundamentalMat(cam1, cam2, FM_RANSAC, (0.00220*(double)im1->cols), 0.999,inliers); //0.00120
// Mat fmat2 = openMVG::sfm::FindFundamentalMat_AC_RANSAC(cam1, cam2,Kvec[id_cam1], Kvec[id_cam2], im1->cols, im1->rows, im2->cols, im2->rows, threshold, inliers);
cout<<"Found AC threshold: "<<threshold<<endl<<endl;
Mat mat_homog = findHomography(cam1, cam2, FM_RANSAC,(0.00350*(double)im1->cols),inliers_homography);
homog_inliers = 0;
std::vector<uchar>::const_iterator itIn_h= inliers_homography.begin();
// number of inliers homography
for ( ;itIn_h!= inliers_homography.end(); ++itIn_h)
{
if (*itIn_h) // if inlier
{ // it is a valid match
homog_inliers++;
}
}
//cout<<"eh "<<countNonZero(inliers)<<endl;
//good_matches.clear();
vector<DMatch> good_matches_RANSAC;
std::vector<uchar>::const_iterator
itIn= inliers.begin();
std::vector<cv::DMatch>::const_iterator
itM= good_matches.begin();
// for all matches
for ( ;itIn!= inliers.end(); ++itIn, ++itM)
{
if (*itIn) // if inlier
{ // it is a valid match
good_matches_RANSAC.push_back(*itM);
}
}
if(write_img)
{
drawMatches(img2, im2->keypoints_img, img1, im1->keypoints_img,
good_matches_RANSAC, img_matches, Scalar::all(-1), Scalar::all(-1),
vector<char>(), DrawMatchesFlags::NOT_DRAW_SINGLE_POINTS);
// waitKey(0);
}
// imshow("Good Matches 2", img_matches);
string path2 = "/local/Datasets/icex_small/images/result/matches/good_matches_ransac_";
char buffer2 [33];
sprintf(buffer2,"%d",c_img++);
path2+=buffer2;
path2+=".jpg";
//cout<<"PATH "<<path<<endl;
cv::resize(img_matches,img_matches,Size(),0.5,0.5);
if(write_img)
imwrite(path2, img_matches);
cam1.clear();
cam2.clear();
if(good_matches_RANSAC.size() > inliers_threshold) // dont create a point cloud for outlier pair, cuz it doesn't makes sense.
{
for( int i = 0; i < good_matches_RANSAC.size(); i++ )
{
cam1.push_back( im1->keypoints_img[ good_matches_RANSAC[i].trainIdx ].pt );
cam2.push_back( im2->keypoints_img[ good_matches_RANSAC[i].queryIdx ].pt );
idx_cam1.push_back(good_matches_RANSAC[i].trainIdx);
idx_cam2.push_back(good_matches_RANSAC[i].queryIdx);
//adding corresponding points into a common point in cloud
if(im1->pkeypoints[good_matches_RANSAC[i].trainIdx]->cloud_p ==NULL && im2->pkeypoints[good_matches_RANSAC[i].queryIdx]->cloud_p ==NULL ) //not set yet
{
cloudPoint *cp = new cloudPoint;
cp->is_in_cloud = false;