ビルド手順例
cd /usr/local/share/OpenCV/samples/c/ g++ -o /tmp/find_obj find_obj.cpp -I/usr/local/include/opencv2 -I/usr/local/include/opencv -L/usr/local/lib -lopencv_nonfree -lopencv_legacy -lopencv_calib3d -lopencv_flann -lopencv_objdetect -lopencv_highgui -lopencv_imgproc -lopencv_core
実行手順例
cd /tmp wget http://133.5.18.161/rinkou/od/2013-05-23-sozai/DSC00227.JPG wget http://133.5.18.161/rinkou/od/2013-05-23-sozai/DSC00228.JPG # 11520 x 1080 -> 2880 x 270 convert -resize 2880x DSC00227.JPG /tmp/DSC00227s.JPG convert -resize 2880x DSC00228.JPG /tmp/DSC00228s.JPG /tmp/find_obj /tmp/DSC00227s.JPG /tmp/DSC00228s.JPG
ビルド手順例
cd /tmp g++ -o /tmp/opencv_surf opencv_surf.cpp -I/usr/local/include/opencv2 -I/usr/local/include/opencv -L/usr/local/lib -lopencv_nonfree -lopencv_legacy -lopencv_calib3d -lopencv_flann -lopencv_objdetect -lopencv_highgui -lopencv_imgproc -lopencv_core
実行手順例
cd /tmp wget http://133.5.18.161/rinkou/od/2013-05-23-sozai/DSC00227.JPG wget http://133.5.18.161/rinkou/od/2013-05-23-sozai/DSC00228.JPG # 11520 x 1080 -> 2880 x 270 convert -resize 2880x DSC00227.JPG /tmp/DSC00227s.JPG convert -resize 2880x DSC00228.JPG /tmp/DSC00228s.JPG /tmp/opencv_surf /tmp/DSC00227s.JPG /tmp/DSC00228s.JPG
ソースコード (ファイル名: opencv_surf.cpp)
/* * A Demo to OpenCV Implementation of SURF * Further Information Refer to "SURF: Speed-Up Robust Feature" * Author: Liu Liu * liuliu.1987+opencv@gmail.com */ #include "opencv2/objdetect/objdetect.hpp" #include "opencv2/features2d/features2d.hpp" #include "opencv2/highgui/highgui.hpp" #include "opencv2/calib3d/calib3d.hpp" #include "opencv2/nonfree/nonfree.hpp" #include "opencv2/imgproc/imgproc_c.h" #include "opencv2/legacy/legacy.hpp" #include "opencv2/legacy/compat.hpp" #include<iostream> #include<vector> #include<stdio.h> using namespace std; static void help() { printf( "This program demonstrated the use of the SURF Detector and Descriptor using\n" "either FLANN (fast approx nearst neighbor classification) or brute force matching\n" "on planar objects.\n" "Usage:\n" "./find_obj <object_filename> <scene_filename>, default is box.png and box_in_scene.png\n\n"); return; } // define whether to use approximate nearest-neighbor search #define USE_FLANN #ifdef USE_FLANN static void flannFindPairs( const CvSeq*, const CvSeq* objectDescriptors, const CvSeq*, const CvSeq* imageDescriptors, vector<int>& ptpairs ) { int length = (int)(objectDescriptors->elem_size/sizeof(float)); cv::Mat m_object(objectDescriptors->total, length, CV_32F); cv::Mat m_image(imageDescriptors->total, length, CV_32F); // copy descriptors CvSeqReader obj_reader; float* obj_ptr = m_object.ptr<float>(0); cvStartReadSeq( objectDescriptors, &obj_reader ); for(int i = 0; i < objectDescriptors->total; i++ ) { const float* descriptor = (const float*)obj_reader.ptr; CV_NEXT_SEQ_ELEM( obj_reader.seq->elem_size, obj_reader ); memcpy(obj_ptr, descriptor, length*sizeof(float)); obj_ptr += length; } CvSeqReader img_reader; float* img_ptr = m_image.ptr<float>(0); cvStartReadSeq( imageDescriptors, &img_reader ); for(int i = 0; i < imageDescriptors->total; i++ ) { const float* descriptor = (const float*)img_reader.ptr; CV_NEXT_SEQ_ELEM( img_reader.seq->elem_size, img_reader ); memcpy(img_ptr, descriptor, length*sizeof(float)); img_ptr += length; } // find nearest neighbors using FLANN cv::Mat m_indices(objectDescriptors->total, 2, CV_32S); cv::Mat m_dists(objectDescriptors->total, 2, CV_32F); cv::flann::Index flann_index(m_image, cv::flann::KDTreeIndexParams(4)); // using 4 randomized kdtrees flann_index.knnSearch(m_object, m_indices, m_dists, 2, cv::flann::SearchParams(64) ); // maximum number of leafs checked int* indices_ptr = m_indices.ptr<int>(0); float* dists_ptr = m_dists.ptr<float>(0); // kaneko fprintf( stderr, "pairnum, objkeypoint, imagekeypoint, dist\n" ); int j=0; for (int i=0;i<m_indices.rows;++i) { if (dists_ptr[2*i]<0.6*dists_ptr[2*i+1]) { ptpairs.push_back(i); ptpairs.push_back(indices_ptr[2*i]); // kaneko fprintf( stderr, "%d, %d, %d, %f\n", j, i, indices_ptr[2*i], dists_ptr[2*i] ); j++; } } } #else static double compareSURFDescriptors( const float* d1, const float* d2, double best, int length ) { double total_cost = 0; assert( length % 4 == 0 ); for( int i = 0; i < length; i += 4 ) { double t0 = d1[i ] - d2[i ]; double t1 = d1[i+1] - d2[i+1]; double t2 = d1[i+2] - d2[i+2]; double t3 = d1[i+3] - d2[i+3]; total_cost += t0*t0 + t1*t1 + t2*t2 + t3*t3; if( total_cost > best ) break; } return total_cost; } static int // kaneko naiveNearestNeighbor( const float* vec, int laplacian, const CvSeq* model_keypoints, const CvSeq* model_descriptors, const int i ) { int length = (int)(model_descriptors->elem_size/sizeof(float)); int i, neighbor = -1; double d, dist1 = 1e6, dist2 = 1e6; CvSeqReader reader, kreader; cvStartReadSeq( model_keypoints, &kreader, 0 ); cvStartReadSeq( model_descriptors, &reader, 0 ); for( i = 0; i < model_descriptors->total; i++ ) { const CvSURFPoint* kp = (const CvSURFPoint*)kreader.ptr; const float* mvec = (const float*)reader.ptr; CV_NEXT_SEQ_ELEM( kreader.seq->elem_size, kreader ); CV_NEXT_SEQ_ELEM( reader.seq->elem_size, reader ); if( laplacian != kp->laplacian ) continue; d = compareSURFDescriptors( vec, mvec, dist2, length ); if( d < dist1 ) { dist2 = dist1; dist1 = d; neighbor = i; } else if ( d < dist2 ) dist2 = d; } if ( dist1 < 0.6*dist2 ) { // kaneko fprintf( stderr, "%d, %d, %f \n", i, neighbor, dist1 ); return neighbor; } return -1; } static void findPairs( const CvSeq* objectKeypoints, const CvSeq* objectDescriptors, const CvSeq* imageKeypoints, const CvSeq* imageDescriptors, vector<int>& ptpairs ) { int i; CvSeqReader reader, kreader; cvStartReadSeq( objectKeypoints, &kreader ); cvStartReadSeq( objectDescriptors, &reader ); ptpairs.clear(); for( i = 0; i < objectDescriptors->total; i++ ) { const CvSURFPoint* kp = (const CvSURFPoint*)kreader.ptr; const float* descriptor = (const float*)reader.ptr; CV_NEXT_SEQ_ELEM( kreader.seq->elem_size, kreader ); CV_NEXT_SEQ_ELEM( reader.seq->elem_size, reader ); // kaneko int nearest_neighbor = naiveNearestNeighbor( descriptor, kp->laplacian, imageKeypoints, imageDescriptors, i ); if( nearest_neighbor >= 0 ) { ptpairs.push_back(i); ptpairs.push_back(nearest_neighbor); } } } #endif /* a rough implementation for object location */ static int locatePlanarObject( const CvSeq* objectKeypoints, const CvSeq* objectDescriptors, const CvSeq* imageKeypoints, const CvSeq* imageDescriptors, const CvPoint src_corners[4], CvPoint dst_corners[4] ) { double h[9]; CvMat _h = cvMat(3, 3, CV_64F, h); vector<int> ptpairs; vector<CvPoint2D32f> pt1, pt2; CvMat _pt1, _pt2; int i, n; #ifdef USE_FLANN flannFindPairs( objectKeypoints, objectDescriptors, imageKeypoints, imageDescriptors, ptpairs ); #else findPairs( objectKeypoints, objectDescriptors, imageKeypoints, imageDescriptors, ptpairs ); #endif n = (int)(ptpairs.size()/2); if( n < 4 ) return 0; pt1.resize(n); pt2.resize(n); for( i = 0; i < n; i++ ) { pt1[i] = ((CvSURFPoint*)cvGetSeqElem(objectKeypoints,ptpairs[i*2]))->pt; pt2[i] = ((CvSURFPoint*)cvGetSeqElem(imageKeypoints,ptpairs[i*2+1]))->pt; } _pt1 = cvMat(1, n, CV_32FC2, &pt1[0] ); _pt2 = cvMat(1, n, CV_32FC2, &pt2[0] ); if( !cvFindHomography( &_pt1, &_pt2, &_h, CV_RANSAC, 5 )) return 0; for( i = 0; i < 4; i++ ) { double x = src_corners[i].x, y = src_corners[i].y; double Z = 1./(h[6]*x + h[7]*y + h[8]); double X = (h[0]*x + h[1]*y + h[2])*Z; double Y = (h[3]*x + h[4]*y + h[5])*Z; dst_corners[i] = cvPoint(cvRound(X), cvRound(Y)); } return 1; } int main(int argc, char** argv) { const char* object_filename = argc == 3 ? argv[1] : "box.png"; const char* scene_filename = argc == 3 ? argv[2] : "box_in_scene.png"; cv::initModule_nonfree(); help(); IplImage* object = cvLoadImage( object_filename, CV_LOAD_IMAGE_GRAYSCALE ); IplImage* image = cvLoadImage( scene_filename, CV_LOAD_IMAGE_GRAYSCALE ); if( !object || !image ) { fprintf( stderr, "Can not load %s and/or %s\n", object_filename, scene_filename ); exit(-1); } CvMemStorage* storage = cvCreateMemStorage(0); cvNamedWindow("Object Correspond", 1); static CvScalar colors[] = { {{0,0,255}}, {{0,128,255}}, {{0,255,255}}, {{0,255,0}}, {{255,128,0}}, {{255,255,0}}, {{255,0,0}}, {{255,0,255}}, {{255,255,255}} }; IplImage* object_color = cvCreateImage(cvGetSize(object), 8, 3); cvCvtColor( object, object_color, CV_GRAY2BGR ); CvSeq* objectKeypoints = 0, *objectDescriptors = 0; CvSeq* imageKeypoints = 0, *imageDescriptors = 0; int i; CvSURFParams params = cvSURFParams(500, 1); double tt = (double)cvGetTickCount(); cvExtractSURF( object, 0, &objectKeypoints, &objectDescriptors, storage, params ); printf("Object Descriptors: %d\n", objectDescriptors->total); cvExtractSURF( image, 0, &imageKeypoints, &imageDescriptors, storage, params ); printf("Image Descriptors: %d\n", imageDescriptors->total); tt = (double)cvGetTickCount() - tt; // printf( "Extraction time = %gms\n", tt/(cvGetTickFrequency()*1000.)); CvPoint src_corners[4] = {{0,0}, {object->width,0}, {object->width, object->height}, {0, object->height}}; CvPoint dst_corners[4]; IplImage* correspond = cvCreateImage( cvSize(image->width, object->height+image->height), 8, 1 ); cvSetImageROI( correspond, cvRect( 0, 0, object->width, object->height ) ); cvCopy( object, correspond ); cvSetImageROI( correspond, cvRect( 0, object->height, correspond->width, correspond->height ) ); cvCopy( image, correspond ); cvResetImageROI( correspond ); #ifdef USE_FLANN printf("Using approximate nearest neighbor search\n"); #endif if( locatePlanarObject( objectKeypoints, objectDescriptors, imageKeypoints, imageDescriptors, src_corners, dst_corners )) { for( i = 0; i < 4; i++ ) { CvPoint r1 = dst_corners[i%4]; CvPoint r2 = dst_corners[(i+1)%4]; cvLine( correspond, cvPoint(r1.x, r1.y+object->height ), cvPoint(r2.x, r2.y+object->height ), colors[8] ); } } vector<int> ptpairs; #ifdef USE_FLANN flannFindPairs( objectKeypoints, objectDescriptors, imageKeypoints, imageDescriptors, ptpairs ); #else findPairs( objectKeypoints, objectDescriptors, imageKeypoints, imageDescriptors, ptpairs ); #endif // kaneko fprintf( stderr, "id, objkeypoint, imagekeypoint, p1x, p1y, radius1, p2x, p2y, radius2\n" ); for( i = 0; i < (int)ptpairs.size(); i += 2 ) { CvSURFPoint* r1 = (CvSURFPoint*)cvGetSeqElem( objectKeypoints, ptpairs[i] ); CvSURFPoint* r2 = (CvSURFPoint*)cvGetSeqElem( imageKeypoints, ptpairs[i+1] ); cvLine( correspond, cvPointFrom32f(r1->pt), cvPoint(cvRound(r2->pt.x), cvRound(r2->pt.y+object->height)), colors[8] ); // int radius; radius = cvRound(r1->size*1.2/9.*2); cvCircle( correspond, cvPointFrom32f(r1->pt), radius, colors[0], 1, 8, 0 ); // kaneko int radius1 = cvRound(r1->size*1.2/9.*2);; int radius2 = cvRound(r2->size*1.2/9.*2);; fprintf( stderr, "%d, %d, %d, %f, %f, %d, %f, %f, %d\n", i/2, ptpairs[i], ptpairs[i+1], r1->pt.x, r1->pt.y, radius1, r2->pt.x, r2->pt.y, radius2 ); } cvShowImage( "Object Correspond", correspond ); fprintf( stderr, "done\n" ); cvWaitKey(0); cvDestroyWindow("Object Correspond"); return 0; }
ソースコード (ファイル名: opencv_surf.cpp)
今度は、 ウインドウを開かない、 「fprintf( stderr, "%d, %d, %d, %f, %f, %d, %f, %f, %d\n", i/2, ptpairs[i], ptpairs[i+1], r1->pt.x, r1->pt.y, radius1, r2->pt.x, r2->pt.y, radius2 );」で表示のみ、 というように書き換え
/* * A Demo to OpenCV Implementation of SURF * Further Information Refer to "SURF: Speed-Up Robust Feature" * Author: Liu Liu * liuliu.1987+opencv@gmail.com */ #include "opencv2/objdetect/objdetect.hpp" #include "opencv2/features2d/features2d.hpp" #include "opencv2/highgui/highgui.hpp" #include "opencv2/calib3d/calib3d.hpp" #include "opencv2/nonfree/nonfree.hpp" #include "opencv2/imgproc/imgproc_c.h" #include "opencv2/legacy/legacy.hpp" #include "opencv2/legacy/compat.hpp" #include#include #include using namespace std; static void help() { printf( "This program demonstrated the use of the SURF Detector and Descriptor using\n" "either FLANN (fast approx nearst neighbor classification) or brute force matching\n" "on planar objects.\n" "Usage:\n" "./find_obj , default is box.png and box_in_scene.png\n\n"); return; } // define whether to use approximate nearest-neighbor search #define USE_FLANN #ifdef USE_FLANN static void flannFindPairs( const CvSeq*, const CvSeq* objectDescriptors, const CvSeq*, const CvSeq* imageDescriptors, vector & ptpairs ) { int length = (int)(objectDescriptors->elem_size/sizeof(float)); cv::Mat m_object(objectDescriptors->total, length, CV_32F); cv::Mat m_image(imageDescriptors->total, length, CV_32F); // copy descriptors CvSeqReader obj_reader; float* obj_ptr = m_object.ptr (0); cvStartReadSeq( objectDescriptors, &obj_reader ); for(int i = 0; i < objectDescriptors->total; i++ ) { const float* descriptor = (const float*)obj_reader.ptr; CV_NEXT_SEQ_ELEM( obj_reader.seq->elem_size, obj_reader ); memcpy(obj_ptr, descriptor, length*sizeof(float)); obj_ptr += length; } CvSeqReader img_reader; float* img_ptr = m_image.ptr (0); cvStartReadSeq( imageDescriptors, &img_reader ); for(int i = 0; i < imageDescriptors->total; i++ ) { const float* descriptor = (const float*)img_reader.ptr; CV_NEXT_SEQ_ELEM( img_reader.seq->elem_size, img_reader ); memcpy(img_ptr, descriptor, length*sizeof(float)); img_ptr += length; } // find nearest neighbors using FLANN cv::Mat m_indices(objectDescriptors->total, 2, CV_32S); cv::Mat m_dists(objectDescriptors->total, 2, CV_32F); cv::flann::Index flann_index(m_image, cv::flann::KDTreeIndexParams(4)); // using 4 randomized kdtrees flann_index.knnSearch(m_object, m_indices, m_dists, 2, cv::flann::SearchParams(64) ); // maximum number of leafs checked int* indices_ptr = m_indices.ptr (0); float* dists_ptr = m_dists.ptr (0); // kaneko // fprintf( stderr, "objkeypoint, imagekeypoint, dist\n" ); int j=0; for (int i=0;i best ) break; } return total_cost; } static int // kaneko naiveNearestNeighbor( const float* vec, int laplacian, const CvSeq* model_keypoints, const CvSeq* model_descriptors, const int i ) { int length = (int)(model_descriptors->elem_size/sizeof(float)); int i, neighbor = -1; double d, dist1 = 1e6, dist2 = 1e6; CvSeqReader reader, kreader; cvStartReadSeq( model_keypoints, &kreader, 0 ); cvStartReadSeq( model_descriptors, &reader, 0 ); for( i = 0; i < model_descriptors->total; i++ ) { const CvSURFPoint* kp = (const CvSURFPoint*)kreader.ptr; const float* mvec = (const float*)reader.ptr; CV_NEXT_SEQ_ELEM( kreader.seq->elem_size, kreader ); CV_NEXT_SEQ_ELEM( reader.seq->elem_size, reader ); if( laplacian != kp->laplacian ) continue; d = compareSURFDescriptors( vec, mvec, dist2, length ); if( d < dist1 ) { dist2 = dist1; dist1 = d; neighbor = i; } else if ( d < dist2 ) dist2 = d; } if ( dist1 < 0.6*dist2 ) { // kaneko // fprintf( stderr, "%d, %d, %f \n", i, neighbor, dist1 ); return neighbor; } return -1; } static void findPairs( const CvSeq* objectKeypoints, const CvSeq* objectDescriptors, const CvSeq* imageKeypoints, const CvSeq* imageDescriptors, vector & ptpairs ) { int i; CvSeqReader reader, kreader; cvStartReadSeq( objectKeypoints, &kreader ); cvStartReadSeq( objectDescriptors, &reader ); ptpairs.clear(); for( i = 0; i < objectDescriptors->total; i++ ) { const CvSURFPoint* kp = (const CvSURFPoint*)kreader.ptr; const float* descriptor = (const float*)reader.ptr; CV_NEXT_SEQ_ELEM( kreader.seq->elem_size, kreader ); CV_NEXT_SEQ_ELEM( reader.seq->elem_size, reader ); // kaneko int nearest_neighbor = naiveNearestNeighbor( descriptor, kp->laplacian, imageKeypoints, imageDescriptors, i ); if( nearest_neighbor >= 0 ) { ptpairs.push_back(i); ptpairs.push_back(nearest_neighbor); } } } #endif /* a rough implementation for object location */ static int locatePlanarObject( const CvSeq* objectKeypoints, const CvSeq* objectDescriptors, const CvSeq* imageKeypoints, const CvSeq* imageDescriptors, const CvPoint src_corners[4], CvPoint dst_corners[4] ) { double h[9]; CvMat _h = cvMat(3, 3, CV_64F, h); vector ptpairs; vector pt1, pt2; CvMat _pt1, _pt2; int i, n; #ifdef USE_FLANN flannFindPairs( objectKeypoints, objectDescriptors, imageKeypoints, imageDescriptors, ptpairs ); #else findPairs( objectKeypoints, objectDescriptors, imageKeypoints, imageDescriptors, ptpairs ); #endif n = (int)(ptpairs.size()/2); if( n < 4 ) return 0; pt1.resize(n); pt2.resize(n); for( i = 0; i < n; i++ ) { pt1[i] = ((CvSURFPoint*)cvGetSeqElem(objectKeypoints,ptpairs[i*2]))->pt; pt2[i] = ((CvSURFPoint*)cvGetSeqElem(imageKeypoints,ptpairs[i*2+1]))->pt; } _pt1 = cvMat(1, n, CV_32FC2, &pt1[0] ); _pt2 = cvMat(1, n, CV_32FC2, &pt2[0] ); if( !cvFindHomography( &_pt1, &_pt2, &_h, CV_RANSAC, 5 )) return 0; for( i = 0; i < 4; i++ ) { double x = src_corners[i].x, y = src_corners[i].y; double Z = 1./(h[6]*x + h[7]*y + h[8]); double X = (h[0]*x + h[1]*y + h[2])*Z; double Y = (h[3]*x + h[4]*y + h[5])*Z; dst_corners[i] = cvPoint(cvRound(X), cvRound(Y)); } return 1; } int main(int argc, char** argv) { const char* object_filename = argc == 3 ? argv[1] : "box.png"; const char* scene_filename = argc == 3 ? argv[2] : "box_in_scene.png"; cv::initModule_nonfree(); help(); IplImage* object = cvLoadImage( object_filename, CV_LOAD_IMAGE_GRAYSCALE ); IplImage* image = cvLoadImage( scene_filename, CV_LOAD_IMAGE_GRAYSCALE ); if( !object || !image ) { fprintf( stderr, "Can not load %s and/or %s\n", object_filename, scene_filename ); exit(-1); } CvMemStorage* storage = cvCreateMemStorage(0); cvNamedWindow("Object Correspond", 1); static CvScalar colors[] = { {{0,0,255}}, {{0,128,255}}, {{0,255,255}}, {{0,255,0}}, {{255,128,0}}, {{255,255,0}}, {{255,0,0}}, {{255,0,255}}, {{255,255,255}} }; IplImage* object_color = cvCreateImage(cvGetSize(object), 8, 3); cvCvtColor( object, object_color, CV_GRAY2BGR ); CvSeq* objectKeypoints = 0, *objectDescriptors = 0; CvSeq* imageKeypoints = 0, *imageDescriptors = 0; int i; CvSURFParams params = cvSURFParams(500, 1); double tt = (double)cvGetTickCount(); cvExtractSURF( object, 0, &objectKeypoints, &objectDescriptors, storage, params ); // printf("Object Descriptors: %d\n", objectDescriptors->total); cvExtractSURF( image, 0, &imageKeypoints, &imageDescriptors, storage, params ); // printf("Image Descriptors: %d\n", imageDescriptors->total); tt = (double)cvGetTickCount() - tt; // printf( "Extraction time = %gms\n", tt/(cvGetTickFrequency()*1000.)); CvPoint src_corners[4] = {{0,0}, {object->width,0}, {object->width, object->height}, {0, object->height}}; CvPoint dst_corners[4]; // IplImage* correspond = cvCreateImage( cvSize(image->width, object->height+image->height), 8, 1 ); // cvSetImageROI( correspond, cvRect( 0, 0, object->width, object->height ) ); // cvCopy( object, correspond ); // cvSetImageROI( correspond, cvRect( 0, object->height, correspond->width, correspond->height ) ); // cvCopy( image, correspond ); // cvResetImageROI( correspond ); #ifdef USE_FLANN // printf("Using approximate nearest neighbor search\n"); #endif if( locatePlanarObject( objectKeypoints, objectDescriptors, imageKeypoints, imageDescriptors, src_corners, dst_corners )) { for( i = 0; i < 4; i++ ) { CvPoint r1 = dst_corners[i%4]; CvPoint r2 = dst_corners[(i+1)%4]; // cvLine( correspond, cvPoint(r1.x, r1.y+object->height ), // cvPoint(r2.x, r2.y+object->height ), colors[8] ); } } vector ptpairs; #ifdef USE_FLANN flannFindPairs( objectKeypoints, objectDescriptors, imageKeypoints, imageDescriptors, ptpairs ); #else findPairs( objectKeypoints, objectDescriptors, imageKeypoints, imageDescriptors, ptpairs ); #endif // kaneko fprintf( stderr, "id, objkeypoint, imagekeypoint, p1x, p1y, radius1, p2x, p2y, radius2\n" ); for( i = 0; i < (int)ptpairs.size(); i += 2 ) { CvSURFPoint* r1 = (CvSURFPoint*)cvGetSeqElem( objectKeypoints, ptpairs[i] ); CvSURFPoint* r2 = (CvSURFPoint*)cvGetSeqElem( imageKeypoints, ptpairs[i+1] ); // cvLine( correspond, cvPointFrom32f(r1->pt), cvPoint(cvRound(r2->pt.x), cvRound(r2->pt.y+object->height)), colors[8] ); // int radius; radius = cvRound(r1->size*1.2/9.*2); // cvCircle( correspond, cvPointFrom32f(r1->pt), radius, colors[0], 1, 8, 0 ); // kaneko int radius1 = cvRound(r1->size*1.2/9.*2);; int radius2 = cvRound(r2->size*1.2/9.*2);; fprintf( stderr, "%d, %d, %d, %f, %f, %d, %f, %f, %d\n", i/2, ptpairs[i], ptpairs[i+1], r1->pt.x, r1->pt.y, radius1, r2->pt.x, r2->pt.y, radius2 ); } // cvShowImage( "Object Correspond", correspond ); // fprintf( stderr, "done\n" ); // cvWaitKey(0); // cvDestroyWindow("Object Correspond"); return 0; }
surf 実験. まず、エンコード.5フレームごとに1フレームを /tmp/w に保存.
cd /tmp mkdir /tmp/w cp /home/kkaneko/rinkou/od/data/recorder2013/10540510/*.AVI /tmp ffmpeg -i 03430001.AVI 03430001-%07d.png cp *0.png *5.png /tmp/w rm /tmp/*.png ffmpeg -i 03430002.AVI 03430002-%07d.png cp *0.png *5.png /tmp/w rm /tmp/*.png ffmpeg -i 03430003.AVI 03430003-%07d.png cp *0.png *5.png /tmp/w rm /tmp/*.png ffmpeg -i 03430004.AVI 03430004-%07d.png cp *0.png *5.png /tmp/w rm /tmp/*.png ffmpeg -i 03430005.AVI 03430005-%07d.png cp *0.png *5.png /tmp/w rm /tmp/*.png ffmpeg -i 03430006.AVI 03430006-%07d.png cp *0.png *5.png /tmp/w rm /tmp/*.png ffmpeg -i 03430007.AVI 03430007-%07d.png cp *0.png *5.png /tmp/w rm /tmp/*.png ffmpeg -i 03430008.AVI 03430008-%07d.png cp *0.png *5.png /tmp/w rm /tmp/*.png ffmpeg -i 03430009.AVI 03430009-%07d.png cp *0.png *5.png /tmp/w rm /tmp/*.png # 総フレーム数 1810, fps は多分 10 [fps]
たくさんできる。1秒あたり2フレームに間引きされていることに注意
surf
export f=`echo *.png` export f2='' for i in $f; do if [ ! f2 = '' ]; then /tmp/opencv_surf $f2 $i > $f2.csv export f2=$i fi done # number of keypoints rm -f lines.csv touch lines.csv for i in *.csv; do wc $i >> lines.csv done R require(data.table) x <- read.table("lines.csv", as.is=TRUE) # number of keypoint pairs plot( x[,1] - 7 ) lines( lowess( x[,1] - 7, f=0.01 ) )
inlier and outlier example
R x <- read.csv("03430001-0000405.png.csv", as.is=TRUE, skip=7, header=FALSE) d <- rbind(x$V4 - x$V7, x$V5 - x$V7) require(mvoutlier) corr.plot( x=d[1,], y=d[2,] )
#!/bin/bash rm -f lines.csv rm -f /tmp/1 touch /tmp/1 for i in *.csv; do echo $i rm -f /tmp/a.r cat > /tmp/a.r <<-RCOMMAND x <- read.csv("$i", as.is=TRUE, skip=7, header=FALSE) d <- rbind(x\$V4 - x\$V7, x\$V5 - x\$V7) require(mvoutlier) # corr.plot( x=d[1,], y=d[2,] ) # hist( sign2(t(d))\$x.dist ) a <- c( sum( sign2(t(d), qcrit=0.975)\$wfinal01 ), sum( sign2(t(d), qcrit=0.95)\$wfinal01 ), sum( sign2(t(d), qcrit=0.9)\$wfinal01 ), sum( sign2(t(d), qcrit=0.8)\$wfinal01 )) print(a) RCOMMAND cat /tmp/a.r | R --vanilla >> /tmp/1 done fgrep '[1]' /tmp/1 > /tmp/2 R x <- read.table("/tmp/2", header=FALSE, as.is=TRUE) plot( lowess( x[,2], f=0.01 ), type="l", col=1, xlim=c(0,362), ylim=c(0, 1400) ) par(new=T) plot( lowess( x[,3], f=0.01 ), type="l", col=2, xlim=c(0,362), ylim=c(0, 1400) ) par(new=T)Fast algorithm for identifying multivariate outliers in high-dimensional and/or large datasets, using spatial signs, see Filzmoser, Maronna, and Werner (CSDA, 2007)
#!/bin/bash rm -f lines.csv rm -f /tmp/3 touch /tmp/3 for i in *.csv; do echo $i rm -f /tmp/a.r cat > /tmp/a.r <<-RCOMMAND x <- read.csv("$i", as.is=TRUE, skip=7, header=FALSE) d <- rbind(x\$V4 - x\$V7, x\$V5 - x\$V7) require(mvoutlier) # corr.plot( x=d[1,], y=d[2,] ) # hist( sign2(t(d))\$x.dist ) a <- sign2(t(d), qcrit=0.975)\$wfinal01 d2 <- rbind(x[a==1,]\$V4 - x[a==1,]\$V7, x[a==1,]\$V5 - x[a==1,]\$V7) print( c( mean(d2[1,]), mean(d2[2,]) ) ) RCOMMAND cat /tmp/a.r | R --vanilla >> /tmp/3 done fgrep '[1]' /tmp/3 > /tmp/4 R x2 <- read.table("/tmp/4", header=FALSE, as.is=TRUE) plot( x2[,2], type="l", col=2, xlim=c(0,362) ) plot( x2[,3], type="l", col=4, xlim=c(0,362) )time change of mean value of x
time change of mean value of y
x <- read.table("/tmp/2", header=FALSE, as.is=TRUE) x2 <- read.table("/tmp/4", header=FALSE, as.is=TRUE) d <- rbind(x[,2], x2[,2], x2[,3]) # 状態 # 1 左に曲がる、2 右に曲がる, 3 止まっている y <- rep(0, 722) y[18:28] <- 1 y[58:66] <- 1 y[67:76] <- 1 y[94:104] <- 1 y[104:116] <- 1 y[246:254] <- 1 y[340:350] <- 1 y[46:56] <- 2 y[128:138] <- 2 y[164:172] <- 2 y[194:208] <- 2 y[492:502] <- 2 y[1:8] <- 3 y[272:334] <- 3 y[406:432] <- 3 plot(x=d[2,], y=d[3,], type="p", col=y+1) par(new=T) plot(x=d[2,], y=d[3,], type="l", lwd=0.2)
引き続き
# 正規化 d[1,] <- scale( d[1,] ) d[2,] <- scale( d[2,] ) d[3,] <- scale( d[3,] ) p1 <- cmdscale(dist(t(d))) plot(x=p1[,1], y=p1[,2], type="p", col=y+1) par(new=T) plot(x=p1[,1], y=p1[,2], type="l", lwd=0.2)