-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathslam.cpp
1151 lines (984 loc) · 35.8 KB
/
slam.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
#define ENABLE_EXTRACTOR_ORB 0
#define ENABLE_EXTRACTOR_GPU_SURF 0
#define ENABLE_EXTRACTOR_SURF 0
#define ESTIMATE_RANSAC 0
#define ESTIMATE_RANSAC_SVD 0
#define REFINE_GICP 0
#define REFINE_PCL 0
#include "slam.h"
//CV
#include <opencv2/nonfree/nonfree.hpp>
#include <opencv2/highgui/highgui.hpp>
#include <opencv2/core/core.hpp>
#include <opencv2/features2d/features2d.hpp>
#include <opencv2/imgproc/imgproc.hpp>
#include <opencv2/calib3d/calib3d.hpp>
#include <opencv2/core/eigen.hpp>
//Std
#include <fstream>
#include <sstream>
#include <vector>
#include <string>
#include <cstdlib>
#include <cstdio>
#include <ctime>
#include <algorithm>
//PCL
#include <pcl/ModelCoefficients.h>
#include <pcl/io/pcd_io.h>
#include <pcl/point_types.h>
#include <pcl/sample_consensus/method_types.h>
#include <pcl/sample_consensus/model_types.h>
#include <pcl/segmentation/sac_segmentation.h>
#include <pcl/filters/extract_indices.h>
#include <pcl/filters/voxel_grid.h>
#include <pcl/common/transforms.h>
#include <pcl/filters/passthrough.h>
//G2O
#include <g2o/types/slam3d/types_slam3d.h>
#include <g2o/core/sparse_optimizer.h>
#include <g2o/core/block_solver.h>
#include <g2o/core/factory.h>
#include <g2o/core/optimization_algorithm_factory.h>
#include <g2o/core/optimization_algorithm_gauss_newton.h>
#include <g2o/solvers/csparse/linear_solver_csparse.h>
#include <g2o/core/robust_kernel.h>
#include <g2o/core/robust_kernel_factory.h>
#if ENABLE_EXTRACTOR_ORB
#include "FeatureExtractor_ORB.h"
#elif ENABLE_EXTRACTOR_GPU_SURF
#include "FeatureExtractor_GPU_SURF.h"
#elif ENABLE_EXTRACTOR_SURF
#include "FeatureExtractor_SURF.h"
#endif
#include "FeatureMatcher.h"
#if ESTIMATE_RANSAC
#include "TransformEstimate_RANSAC.h"
#elif ESTIMATE_RANSAC_SVD
#include "TransformEstimate_RANSAC_SVD.h"
#endif
#if REFINE_GICP
#include "Refiner_GICP.h"
#elif REFINE_PCL
#include "Refiner_PCL.h"
#endif
//#include "test2/FeatureMatcher1.h"
using namespace std;
using namespace cv;
using namespace g2o;
using namespace Eigen;
//static const char *VOC_FILE = "/home/lc/workspace/paper_related/Appolo/test/surf64_k10L6.voc.gz";
typename Surf64LoopDetector::Parameters params;
GraphicEnd::GraphicEnd():
currCloud(new PointCloud),m_height(480),m_width(640),
m_vocPath("/home/lc/workspace/paper_related/Appolo/test/surf64_k10L6.voc.gz"),
bow_voc(m_vocPath),db(bow_voc,false,0),//bow_detector(bow_voc,params),
loop_fout("/home/lc/workspace/paper_related/Appolo/test/result/loop.txt")
{
//matcher=FlannMatcher();
use_voxel=false;
//bow initialize
/***********************bow init*******************************/
params.image_rows=m_height;
params.image_cols=m_width;
params.use_nss = true; // use normalized similarity score instead of raw score
params.alpha = 0.3; // nss threshold
params.k = 1; // a loop must be consistent with 1 previous matches
params.geom_check = GEOM_DI; // use direct index for geometrical checking
params.di_levels =2; // use two direct index levels
//cout<<"Initiate loop detector with the vocabulary !!!"<<endl;
/***********************************************************/
}
GraphicEnd::~GraphicEnd()
{
}
void GraphicEnd::init(SLAMEnd* pSLAMEnd)
{
cout<<"*****************************"<<endl;
cout<<"slam init....!!!"<<endl;
cout<<"*****************************"<<endl;
cout << "Loading " <<"vocabulary..." << endl;
_pSLAMEnd=pSLAMEnd;
index=1;
rgbPath="/media/新加卷/dataset/dataset1/rgb_index/";
depPath="/media/新加卷/dataset/dataset1/dep_index/";
pcdPath="/media/新加卷/dataset/dataset1/pcd/";
//read first image
readimage();
cout<<"index is: "<<index<<endl;
currKF.id=0;
currKF.frame_index=index;
currKF.fullpose=Eigen::Isometry3d::Identity();
currKF.frame.rgb=currRGB.clone();
currKF.frame.dep=currDep.clone();
//pcl::copyPointCloud(*currCloud,*(currKF.frame.cloud));
//currKF.frame.cloud=currCloud;
previous=currKF;
//set first frame as keyframe
keyframes.push_back(currKF);
/******************************************************/
//set first estimate
double initQua[7]={1.2764 ,-0.9763, 0.6837, 0.8187, 0.3639, -0.1804, -0.4060};
Eigen::Quaterniond q(initQua[6],initQua[3],initQua[4],initQua[5]);
Eigen::Matrix3d r=q.toRotationMatrix();
Eigen::AngleAxisd angle(r);
Eigen::Isometry3d T=Eigen::Isometry3d::Identity();
T=angle;
T(0,3)=initQua[0];
T(1,3)=initQua[1];
T(2,3)=initQua[2];
/*****************************************************/
//put first image into slamEnd
SparseOptimizer& opt=_pSLAMEnd->globalOptimizer;
VertexSE3* v=new VertexSE3();
v->setId(currKF.id);
v->setEstimate(Eigen::Isometry3d::Identity());
//v->setEstimate(T);
v->setFixed(true);
opt.addVertex(v);
index++;
/*
//add the first image into database
vector<cv::KeyPoint> keys;
vector<FSurf64::TDescriptor> descrip;
DetectionResult result;
bow_extractor(currKF.frame.rgb,keys,descrip);
bow_detector.detectLoop(keys,descrip,result);
*/
vector<cv::KeyPoint> keys;
vector<vector<float> > descrip;
loadFeatures(currKF.frame.rgb,descrip,keys);
db.add(descrip);
cout<<"*****************************"<<endl;
}
void GraphicEnd::run()
{
readimage();
cout<<"index is: "<<index<<endl;
present.frame.rgb=currRGB.clone();
present.frame.dep=currDep.clone();
//present.frame.cloud=currCloud;
//pcl::copyPointCloud(*currCloud,*(present.frame.cloud));
Eigen::Matrix4f H;
//bool validTrans=calcTrans(keyframes.back(),present,H,true);
bool validTrans=calcTrans2(keyframes.back(),present,H);
FlannMatcher matcher;
bool isKeyframe=matcher.isKeyframe(H);
//bool isKeyframe=isKeyframe2(H);
Eigen::Isometry3d T(H.cast<double>());
T=T.inverse();
cout<<"T is:"<<endl<<T.matrix()<<endl;
/*
//calc present fullpose
if(!validTrans)
{
cout<<"this frame is lost!!!"<<endl;
lost++;
}
else
{
present.fullpose=previous.fullpose*T;
FlannMatcher matcher;
Eigen::Matrix4f H1=present.fullpose.matrix().cast<float>();
bool isKeyframe=matcher.isKeyframe(H1);
if(isKeyframe)
{
//generated keyframes
generateKeyframe(T);
//check the loop
loopClosure();
lost=0;
}
else
{
cout<<"update robot pose!!!"<<endl;
lost=0;
}
previous=present;
}
*/
if(!validTrans)
{
cout<<"this frame is lost!!!"<<endl;
lost++;
}
else if(validTrans && isKeyframe)
{
//generated keyframes
generateKeyframe(T);
//check the loop
//loopClosure();
//Bowloopclosure();
checkLoopClosure();
lost=0;
}
else if(validTrans && !isKeyframe)
{
cout<<"update robot pose!!!"<<endl;
lost=0;
}
if(lost>5)
{
cout<<"robot is lost, excuse lost recovery!!!!"<<endl;
lostRecovery();
}
cout<<"current keyframe size is: "<<keyframes.size()<<endl;
index++;
}
bool GraphicEnd::run2()
{
readimage();
cout<<"index is: "<<index<<endl;
present.frame.rgb=currRGB.clone();
present.frame.dep=currDep.clone();
//present.frame.cloud=currCloud;
pcl::copyPointCloud(*currCloud,*(present.frame.cloud));
Eigen::Matrix4f H;
bool validTrans=calcTrans(previous,present,H,true);
Eigen::Isometry3d T(H.cast<double>());
T=T.inverse();
if(!validTrans)
{
cerr<<"invalid sequences!!!"<<endl;
return false;
}
present.fullpose=previous.fullpose*T;
FlannMatcher matcher;
Eigen::Matrix4f H1=present.fullpose.matrix().cast<float>();
bool isKeyframe=matcher.isKeyframe(H1);
if(isKeyframe)
{
//generated keyframes
generateKeyframe(T);
//check the loop
loopClosure();
}
previous.frame.rgb=present.frame.rgb.clone();
previous.frame.dep=present.frame.dep.clone();
pcl::copyPointCloud(*(present.frame.cloud),*(previous.frame.cloud));
cout<<"current keyframe size is: "<<keyframes.size()<<endl;
index++;
return true;
}
void GraphicEnd::readimage()
{
cout<<"loading data from dataset...."<<endl;
ss<<rgbPath<<index<<".png";
currRGB=cv::imread(ss.str());
ss.str("");
ss.clear();
ss<<depPath<<index<<".png";
currDep=cv::imread(ss.str(),-1);
ss.str("");
ss.clear();
ss<<pcdPath<<index<<".pcd";
PointCloud::Ptr currCloudTmp(new PointCloud);
pcl::io::loadPCDFile(ss.str(),*currCloudTmp);
//in order to release memery stress,
//we use filter to make pointcloud into
//no order pointcloud(h=1)
//downsamplePointCloud(currCloudTmp,currCloud);
ss.str("");
ss.clear();
}
void GraphicEnd::generateKeyframe(Eigen::Isometry3d& T)
{
cout<<"generating keyframes..."<<endl;
currKF.id++;
currKF.frame_index=index;
currKF.frame.rgb=present.frame.rgb.clone();
currKF.frame.dep=present.frame.dep.clone();
//currKF.frame.cloud=present.frame.cloud;
pcl::copyPointCloud(*(present.frame.cloud),*(currKF.frame.cloud));
currKF.fullpose=present.fullpose;
//keyframes.reserve(100);
keyframes.push_back(currKF);
//cout<<"keyframes capacity is: "<<keyframes.capacity()<<endl;
SparseOptimizer& opt = _pSLAMEnd->globalOptimizer;
VertexSE3* v = new VertexSE3();
v->setId( currKF.id );
v->setEstimate( Eigen::Isometry3d::Identity() );
opt.addVertex( v );
cout<<"add vertex success!!!"<<endl;
//edge
EdgeSE3* edge = new EdgeSE3();
edge->vertices()[0] = opt.vertex( currKF.id - 1 );
edge->vertices()[1] = opt.vertex( currKF.id );
Eigen::Matrix<double, 6,6> information = Eigen::Matrix<double, 6, 6>::Identity();
information(0, 0) = information(2,2) = 100;
information(1, 1) = 100;
information(3,3) = information(4,4) = information(5,5) = 100;
edge->setInformation( information );
edge->setMeasurement( T );
opt.addEdge( edge );
cout<<"add edge success!!!"<<endl;
}
//use featurematcher2
bool GraphicEnd::calcTrans2(KEYFRAME& frame1,KEYFRAME& frame2,Eigen::Matrix4f& H)
{
bool result;
FlannMatcher2 matcher;
vector<cv::DMatch> bestMatches;
Eigen::Matrix4f H1;
result=matcher.getFinalTransform(frame1.frame.rgb,frame2.frame.rgb,frame1.frame.dep,
frame2.frame.dep,frame1.frame.cloud,frame2.frame.cloud,
bestMatches,H);
/*
PointCloud::Ptr aligned_cloud(new PointCloud);
pcl::IterativeClosestPoint<PointT,PointT> *icp;
icp=new pcl::GeneralizedIterativeClosestPoint<PointT,PointT>();
//icp=new pcl::IterativeClosestPoint<PointT,PointT>();
//set icp parameters
icp->setMaximumIterations(16);
icp->setMaxCorrespondenceDistance(0.1);
icp->setRANSACOutlierRejectionThreshold(0.05);
icp->setTransformationEpsilon(1e-5);
icp->setInputCloud(frame1.frame.cloud);
icp->setInputTarget(frame2.frame.cloud);
if(result)
{
icp->align(*aligned_cloud,H1);
if(icp->hasConverged())
{
H=icp->getFinalTransformation();
cout<<"GICP Refine H is:"<<endl<<H<<endl;
}
else
{
cout<<"GICP has not converged!!"<<endl;
H=H1;
}
}
*/
return result;
}
//use featurematcher1
bool GraphicEnd::calcTrans(KEYFRAME& frame1,KEYFRAME& frame2,Eigen::Matrix4f& H,bool gicpRefine)
{
bool result;
bool validTrans;
vector<cv::DMatch> bestMatches;
FlannMatcher matcher;
Eigen::Matrix4f H1;
//here we use depth image instead of cloud
validTrans=matcher.getFinalTransform(frame1.frame.rgb,frame2.frame.rgb,frame1.frame.dep,
frame2.frame.dep,bestMatches,H1);
//validTrans=matcher.getFinalTransform(frame1.frame.rgb,frame2.frame.rgb,frame1.frame.cloud,
// frame2.frame.cloud,bestMatches,H1);
//validTrans=matcher.getFinalTransform(frame2.frame.rgb,frame1.frame.rgb,frame2.frame.cloud,
// frame1.frame.cloud,bestMatches,H);
//if(!validTrans)
// H1=Eigen::Matrix4f::Identity();
if(gicpRefine)
{
//use GICP,if it do not converge return false
//1.first downsample the pointcloud
//PointCloud::Ptr cloud1(new PointCloud);
//PointCloud::Ptr cloud2(new PointCloud);
PointCloud::Ptr aligned_cloud(new PointCloud);
//downsamplePointCloud(frame1.frame.cloud,cloud1);
//downsamplePointCloud(frame2.frame.cloud,cloud2);
//initial gicp
pcl::IterativeClosestPoint<PointT,PointT> *icp;
icp=new pcl::GeneralizedIterativeClosestPoint<PointT,PointT>();
//icp=new pcl::IterativeClosestPoint<PointT,PointT>();
//set icp parameters
icp->setMaximumIterations(16);
icp->setMaxCorrespondenceDistance(0.1);
icp->setRANSACOutlierRejectionThreshold(0.05);
icp->setTransformationEpsilon(1e-5);
icp->setInputCloud(/*cloud1*/frame1.frame.cloud);
icp->setInputTarget(/*cloud2*/frame2.frame.cloud);
//icp->align(*aligned_cloud);
if(validTrans)
icp->align(*aligned_cloud,H1);
else
icp->align(*aligned_cloud);
H=icp->getFinalTransformation();
if(icp->hasConverged())
{
//H=icp->getFinalTransformation();
cout<<"GICP Refine H is:"<<endl<<H<<endl;
}
else
{
cout<<"GICP has not converged!!"<<endl;
}
result= icp->hasConverged();
}
else
{
H=H1;
result= validTrans;
}
//return validTrans;
return result;
}
bool GraphicEnd::isKeyframe2(Eigen::Matrix4f& T)
{
bool isKey;
Eigen::Matrix3f rotate;
rotate=T.block<3,3>(0,0);
cv::Mat R,rvec;
cv::eigen2cv(rotate,R);
cv::Rodrigues(R,rvec);
float trans[1][3]={T(0,3),T(1,3),T(2,3)};
cv::Mat t(1,3,CV_64F,trans);
float norm_transformation=cv::min(cv::norm(rvec),2*M_PI-cv::norm(rvec))+cv::norm(t);
if(norm_transformation>=0.2 && norm_transformation<0.8)
isKey=true;
else
isKey=false;
return isKey;
}
void GraphicEnd::downsamplePointCloud(PointCloud::Ptr& pc_in,PointCloud::Ptr& pc_downsampled)
{
if(use_voxel)
{
pcl::VoxelGrid<pcl::PointXYZRGB> grid;
grid.setLeafSize(0.05,0.05,0.05);
grid.setFilterFieldName ("z");
grid.setFilterLimits (0.0,5.0);
grid.setInputCloud(pc_in);
grid.filter(*pc_downsampled);
}
else
{
int downsamplingStep=8;
static int j;j=0;
std::vector<double> xV;
std::vector<double> yV;
std::vector<double> zV;
std::vector<double> rV;
std::vector<double> gV;
std::vector<double> bV;
pc_downsampled.reset(new pcl::PointCloud<pcl::PointXYZRGB> );
pc_downsampled->points.resize(640*480/downsamplingStep*downsamplingStep);
for(int r=0;r<480;r=r+downsamplingStep)
{
for(int c=0;c<640;c=c+downsamplingStep)
{
int nPoints=0;
xV.resize(downsamplingStep*downsamplingStep);
yV.resize(downsamplingStep*downsamplingStep);
zV.resize(downsamplingStep*downsamplingStep);
rV.resize(downsamplingStep*downsamplingStep);
gV.resize(downsamplingStep*downsamplingStep);
bV.resize(downsamplingStep*downsamplingStep);
for(int r2=r;r2<r+downsamplingStep;r2++)
{
for(int c2=c;c2<c+downsamplingStep;c2++)
{
//Check if the point has valid data
if(pcl_isfinite (pc_in->points[r2*640+c2].x) &&
pcl_isfinite (pc_in->points[r2*640+c2].y) &&
pcl_isfinite (pc_in->points[r2*640+c2].z) &&
0.3<pc_in->points[r2*640+c2].z &&
pc_in->points[r2*640+c2].z<5)
{
//Create a vector with the x, y and z coordinates of the square region and RGB info
xV[nPoints]=pc_in->points[r2*640+c2].x;
yV[nPoints]=pc_in->points[r2*640+c2].y;
zV[nPoints]=pc_in->points[r2*640+c2].z;
rV[nPoints]=pc_in->points[r2*640+c2].r;
gV[nPoints]=pc_in->points[r2*640+c2].g;
bV[nPoints]=pc_in->points[r2*640+c2].b;
nPoints++;
}
}
}
if(nPoints>0)
{
xV.resize(nPoints);
yV.resize(nPoints);
zV.resize(nPoints);
rV.resize(nPoints);
gV.resize(nPoints);
bV.resize(nPoints);
//Compute the mean 3D point and mean RGB value
std::sort(xV.begin(),xV.end());
std::sort(yV.begin(),yV.end());
std::sort(zV.begin(),zV.end());
std::sort(rV.begin(),rV.end());
std::sort(gV.begin(),gV.end());
std::sort(bV.begin(),bV.end());
pcl::PointXYZRGB point;
point.x=xV[nPoints/2];
point.y=yV[nPoints/2];
point.z=zV[nPoints/2];
point.r=rV[nPoints/2];
point.g=gV[nPoints/2];
point.b=bV[nPoints/2];
//Set the mean point as the representative point of the region
pc_downsampled->points[j]=point;
j++;
}
}
}
pc_downsampled->points.resize(j);
pc_downsampled->width=pc_downsampled->size();
pc_downsampled->height=1;
}
}
/*************************************************************************************/
//DBOW2
void GraphicEnd::loadFeatures(cv::Mat& image,
vector<vector<float> >& features,
vector<cv::KeyPoint>& keypoints)
{
cv::SURF surf(400, 4, 2, false); //do not extend surf
cv::Mat mask;
vector<float> descriptors;
surf(image,mask,keypoints,descriptors);
changeStructure(descriptors,features,surf.descriptorSize());
}
void GraphicEnd::changeStructure(const vector<float> &plain,
vector<vector<float> > &out,
int L)
{
out.resize(plain.size() / L);
unsigned int j = 0;
for(unsigned int i = 0; i < plain.size(); i += L, ++j)
{
out[j].resize(L);
std::copy(plain.begin() + i, plain.begin() + i + L, out[j].begin());
}
}
cv::Mat GraphicEnd::changeVecToMat(vector<vector<float> >& desp)
{
cv::Mat despMat(desp.size(),64,CV_32F);
for(int i=0;i<desp.size();++i)
{
for(int j=0;j<64;++j)
{
despMat.at<float>(j,i)=desp[ i ][ j ];
}
}
return despMat;
}
bool GraphicEnd::checkFundumental(vector<cv::KeyPoint>& kp1,
vector<vector<float> > desp1,
vector<cv::KeyPoint>& kp2,
vector<vector<float> > desp2)
{
bool results=false;
vector<cv::DMatch> matches;
//convert descriptors into Mat
cv::Mat descriptors1=changeVecToMat(desp1);
cv::Mat descriptors2=changeVecToMat(desp2);
//flann match
//cv::FlannBasedMatcher matcher;
cv::BruteForceMatcher<cv::L2<float> > matcher;
//std::vector< cv::DMatch > matches1;
matcher.match( descriptors2, descriptors1, matches );
/*
double max_dist = 0; double min_dist = 100;
//-- Quick calculation of max and min distances between keypoints
for( int i = 0; i < descriptors1.rows; i++ )
{
double dist = matches1[i].distance;
if( dist < min_dist ) min_dist = dist;
if( dist > max_dist ) max_dist = dist;
}
//-- Draw only "good" matches (i.e. whose distance is less than 2*min_dist )
//-- PS.- radiusMatch can also be used here.
for( int i = 0; i < descriptors1.rows; i++ )
{
if( matches1[i].distance < 2*min_dist )
{
matches.push_back( matches1[i]);
}
}
//flann kdtree match
cv::Mat m_indices(descriptors1.rows,2,CV_32S);
cv::Mat m_dists(descriptors1.rows,2,CV_32S);
cv::flann::Index flann_index(descriptors2,cv::flann::KDTreeIndexParams(4));
flann_index.knnSearch(descriptors1,m_indices,m_dists,2,cv::flann::SearchParams(64));
int* indices_ptr=m_indices.ptr<int>(0);
float* dists_ptr=m_dists.ptr<float>(0);
cv::DMatch match;
//vector<cv::DMatch> matches;
for (int i=0;i<m_indices.rows;++i) {
if (dists_ptr[2*i]<0.6*dists_ptr[2*i+1]) {
match.queryIdx=i;
match.trainIdx=indices_ptr[ 2*i ];
match.distance=dists_ptr[ 2*i ];
matches.push_back(match);
}
}
*/
if(matches.size()<12)
return false;
//calc fundamental mat
vector<cv::Point2f> points1,points2;
for(vector<cv::DMatch>::iterator itr=matches.begin();itr!=matches.end();++itr)
{
float x1=kp1[ itr->queryIdx ].pt.x;
float y1=kp2[ itr->queryIdx ].pt.y;
float x2=kp2[ itr->trainIdx ].pt.x;
float y2=kp2[ itr->trainIdx ].pt.y;
points1.push_back(cv::Point2f(x1,y1));
points2.push_back(cv::Point2f(x2,y2));
}
vector<uchar> inliers(points1.size(),0);
cv::Mat fundamental=cv::findFundamentalMat(cv::Mat(points1),
cv::Mat(points2),
inliers,
CV_FM_RANSAC,
3.0,
0.99);
int inlierNum=0;
for(size_t i=0;i<inliers.size();++i)
{
if(inliers[ i ])
inlierNum++;
}
if(inlierNum>=12)
results=true;
cout<<"num inliers: "<<inlierNum<<endl;
return results;
}
void GraphicEnd::checkLoopClosure()
{
cout<<"**************************************************"<<endl;
cout<<"checking loop closure!!!"<<endl;
cout<<"**************************************************"<<endl;
SparseOptimizer& opt = _pSLAMEnd->globalOptimizer;
//first extract currKF desp and keys
vector<cv::KeyPoint> keys;
vector<vector<float> > descrip;
loadFeatures(currKF.frame.rgb,descrip,keys);
//add currKF into database
db.add(descrip);
//query, first is the image itself
QueryResults ret;
db.query(descrip,ret,10);//only get 3 highest score frame;
//ransac check
for(int j=1;j<ret.size();++j)
{
//extract chose image feature
vector<cv::KeyPoint> keypoint;
vector<vector<float> > descriptor;
loadFeatures(keyframes[ret[ j ].Id].frame.rgb,descriptor,keypoint);
bool ransacResult=checkFundumental(keys,descrip,keypoint,descriptor);
//if pass the ransac check
if(ransacResult)
{
Eigen::Matrix4f H;
bool validTrans=calcTrans2(keyframes[ ret[ j ].Id ],currKF,H);
loop_fout<<currKF.frame_index<<" and "<<keyframes[ret[ j ].Id].frame_index<<" : "<<validTrans<<endl;
if(!validTrans || H==Eigen::Matrix4f::Identity())
continue;
Eigen::Isometry3d T(H.cast<double>());
T = T.inverse();
EdgeSE3* edge = new EdgeSE3();
edge->vertices() [0] = opt.vertex( keyframes[ret[ j ].Id].id );
edge->vertices() [1] = opt.vertex( currKF.id );
Matrix<double, 6,6> information = Matrix<double, 6, 6>::Identity();
information(0, 0) = information(2,2) = 100;
information(1,1) = 100;
information(3,3) = information(4,4) = information(5,5) = 100;
edge->setInformation( information );
edge->setMeasurement( T );
edge->setRobustKernel( _pSLAMEnd->robustKernel );
opt.addEdge( edge );
}
}
}
/*************************************************************************************/
//bow loopclosure
void GraphicEnd::Bowloopclosure()
{
SparseOptimizer& opt = _pSLAMEnd->globalOptimizer;
vector<cv::KeyPoint> keys;
vector<FSurf64::TDescriptor> descrip;
DetectionResult result;
bow_extractor(currKF.frame.rgb,keys,descrip);
bow_detector.detectLoop(keys,descrip,result);
if(result.detection())
{
cout<<"Loop found with image "<<result.match<<"!"<<endl;
/******write loop into file*******/
//ofstream loop_fout("/home/lc/workspace/paper_related/Appolo/test/result/loop.txt");
/********************************/
Eigen::Matrix4f H;
bool validTrans=calcTrans(keyframes[ result.match ],currKF,H,true);
loop_fout<<currKF.frame_index<<" and "<<keyframes[result.match].frame_index<<" : "<<validTrans<<endl;
if(!validTrans)
return;
Eigen::Isometry3d T(H.cast<double>());
//T = T.inverse();
EdgeSE3* edge = new EdgeSE3();
edge->vertices() [0] = opt.vertex( keyframes[result.match].id );
edge->vertices() [1] = opt.vertex( currKF.id );
Matrix<double, 6,6> information = Matrix<double, 6, 6>::Identity();
information(0, 0) = information(2,2) = 100;
information(1,1) = 100;
information(3,3) = information(4,4) = information(5,5) = 100;
edge->setInformation( information );
edge->setMeasurement( T );
edge->setRobustKernel( _pSLAMEnd->robustKernel );
opt.addEdge( edge );
}
}
void GraphicEnd::loopClosure()
{
if (keyframes.size() >3 )
{
cout<<"Checking loop closure."<<endl;
waitKey(10);
vector<int> checked;
SparseOptimizer& opt = _pSLAMEnd->globalOptimizer;
for (int i=-2; i>-5; i--)
{
int n = keyframes.size() + i;
if (n>=0)
{
KEYFRAME& p1 = keyframes[n];
Eigen::Matrix4f H;
//bool validTrans=calcTrans(p1,currKF,H);
bool validTrans=calcTrans2(p1,currKF,H);
if(!validTrans)
continue;
Eigen::Isometry3d T(H.cast<double>());
T = T.inverse();
//
EdgeSE3* edge = new EdgeSE3();
edge->vertices() [0] = opt.vertex( keyframes[n].id );
edge->vertices() [1] = opt.vertex( currKF.id );
Matrix<double, 6,6> information = Matrix<double, 6, 6>::Identity();
information(0, 0) = information(2,2) = 100;
information(1,1) = 100;
information(3,3) = information(4,4) = information(5,5) = 100;
edge->setInformation( information );
edge->setMeasurement( T );
edge->setRobustKernel( _pSLAMEnd->robustKernel );
opt.addEdge( edge );
}
else
break;
}
//
cout<<"checking seeds, seed.size()"<<seed.size()<<endl;
vector<int> newseed;
for (size_t i=0; i<seed.size(); i++)
{
KEYFRAME& p1 = keyframes[seed[i]];
//Eigen::Isometry3d T = calcTrans( p1, currKF.frame ).T;
Eigen::Matrix4f H;
//bool validTrans=calcTrans(p1,currKF,H);
bool validTrans=calcTrans2(p1,currKF,H);
if(!validTrans)
continue;
Eigen::Isometry3d T(H.cast<double>());
T = T.inverse();
//
checked.push_back( seed[i] );
newseed.push_back( seed[i] );
EdgeSE3* edge = new EdgeSE3();
edge->vertices() [0] = opt.vertex( keyframes[seed[i]].id );
edge->vertices() [1] = opt.vertex( currKF.id );
Matrix<double, 6,6> information = Matrix<double, 6, 6>::Identity();
information(0, 0) = information(2,2) = 100;
information(1,1) = 100;
information(3,3) = information(4,4) = information(5,5) = 100;
edge->setInformation( information );
edge->setMeasurement( T );
edge->setRobustKernel( _pSLAMEnd->robustKernel );
opt.addEdge( edge );
}
//
cout<<"checking random frames"<<endl;
for (int i=0; i<10; i++)
{
int frame = rand() % (keyframes.size() -3 ); //
if ( find(checked.begin(), checked.end(), frame) != checked.end() ) //
continue;
checked.push_back( frame );
KEYFRAME& p1 = keyframes[frame];
//Eigen::Isometry3d T = calcTrans( p1, currKF.frame ).T;
Eigen::Matrix4f H;
//bool validTrans=calcTrans(p1,currKF,H);
bool validTrans=calcTrans2(p1,currKF,H);
if(!validTrans)
continue;
Eigen::Isometry3d T(H.cast<double>());
T = T.inverse();
newseed.push_back( frame );
//
cout<<"find a loop closure between kf "<<currKF.id<<" and kf "<<frame<<endl;
EdgeSE3* edge = new EdgeSE3();
edge->vertices() [0] = opt.vertex(keyframes[frame].id );
edge->vertices() [1] = opt.vertex( currKF.id );
Matrix<double, 6,6> information = Matrix<double, 6, 6>::Identity();
information(0, 0) = information(2,2) = 100;
information(1,1) = 100;
information(3,3) = information(4,4) = information(5,5) = 100;
edge->setInformation( information );
edge->setMeasurement( T );
edge->setRobustKernel( _pSLAMEnd->robustKernel );
opt.addEdge( edge );
}
waitKey(10);
seed = newseed;
}
}
void GraphicEnd::lostRecovery()
{
cout<<"Lost Recovery..."<<endl;
//
currKF.id++;
currKF.frame.rgb=present.frame.rgb.clone();
currKF.frame.dep=present.frame.dep.clone();
//currKF.frame.cloud=present.frame.cloud;
pcl::copyPointCloud(*(present.frame.cloud),*(currKF.frame.cloud));
currKF.frame_index = index;
//waitKey(0);
keyframes.push_back( currKF );
//
SparseOptimizer& opt = _pSLAMEnd->globalOptimizer;
//
VertexSE3* v = new VertexSE3();
v->setId( currKF.id );
v->setEstimate( Eigen::Isometry3d::Identity() );
opt.addVertex( v );
/****************************************
//check bow loop
//first extract currKF desp and keys
vector<cv::KeyPoint> keys;
vector<vector<float> > descrip;
loadFeatures(currKF.frame.rgb,descrip,keys);
//add currKF into database
db.add(descrip);
//query, first is the image itself
QueryResults ret;