#include #include #include using namespace std; using namespace cv; static void help() { cout << "\n This program demonstrates how to detect compute and match ORB BRISK and AKAZE descriptors \n" "Usage: \n" " ./matchmethod_orb_akaze_brisk --image1= --image2=\n" "Press a key when image window is active to change algorithm or descriptor"; } int main(int argc, char *argv[]) { vector typeDesc; vector typeAlgoMatch; vector fileName; // This descriptor are going to be detect and compute typeDesc.push_back("AKAZE-DESCRIPTOR_KAZE_UPRIGHT"); // see http://docs.opencv.org/trunk/d8/d30/classcv_1_1AKAZE.html typeDesc.push_back("AKAZE"); // see http://docs.opencv.org/trunk/d8/d30/classcv_1_1AKAZE.html typeDesc.push_back("ORB"); // see http://docs.opencv.org/trunk/de/dbf/classcv_1_1BRISK.html typeDesc.push_back("BRISK"); // see http://docs.opencv.org/trunk/db/d95/classcv_1_1ORB.html // This algorithm would be used to match descriptors see http://docs.opencv.org/trunk/db/d39/classcv_1_1DescriptorMatcher.html#ab5dc5036569ecc8d47565007fa518257 typeAlgoMatch.push_back("BruteForce"); typeAlgoMatch.push_back("BruteForce-L1"); typeAlgoMatch.push_back("BruteForce-Hamming"); typeAlgoMatch.push_back("BruteForce-Hamming(2)"); cv::CommandLineParser parser(argc, argv, "{ @image1 | ../data/basketball1.png | }" "{ @image2 | ../data/basketball2.png | }" "{help h ||}"); if (parser.has("help")) { help(); return 0; } fileName.push_back(parser.get(0)); fileName.push_back(parser.get(1)); Mat img1 = imread(fileName[0], IMREAD_GRAYSCALE); Mat img2 = imread(fileName[1], IMREAD_GRAYSCALE); if (img1.rows*img1.cols <= 0) { cout << "Image " << fileName[0] << " is empty or cannot be found\n"; return(0); } if (img2.rows*img2.cols <= 0) { cout << "Image " << fileName[1] << " is empty or cannot be found\n"; return(0); } vector desMethCmp; Ptr b; // Descriptor loop vector::iterator itDesc; for (itDesc = typeDesc.begin(); itDesc != typeDesc.end(); ++itDesc) { Ptr descriptorMatcher; // Match between img1 and img2 vector matches; // keypoint for img1 and img2 vector keyImg1, keyImg2; // Descriptor for img1 and img2 Mat descImg1, descImg2; vector::iterator itMatcher = typeAlgoMatch.end(); if (*itDesc == "AKAZE-DESCRIPTOR_KAZE_UPRIGHT"){ b = AKAZE::create(AKAZE::DESCRIPTOR_KAZE_UPRIGHT); } if (*itDesc == "AKAZE"){ b = AKAZE::create(); } if (*itDesc == "ORB"){ b = ORB::create(); } else if (*itDesc == "BRISK"){ b = BRISK::create(); } try { // We can detect keypoint with detect method b->detect(img1, keyImg1, Mat()); // and compute their descriptors with method compute b->compute(img1, keyImg1, descImg1); // or detect and compute descriptors in one step b->detectAndCompute(img2, Mat(),keyImg2, descImg2,false); // Match method loop for (itMatcher = typeAlgoMatch.begin(); itMatcher != typeAlgoMatch.end(); ++itMatcher){ descriptorMatcher = DescriptorMatcher::create(*itMatcher); if ((*itMatcher == "BruteForce-Hamming" || *itMatcher == "BruteForce-Hamming(2)") && (b->descriptorType() == CV_32F || b->defaultNorm() <= NORM_L2SQR)) { cout << "**************************************************************************\n"; cout << "It's strange. You should use Hamming distance only for a binary descriptor\n"; cout << "**************************************************************************\n"; } if ((*itMatcher == "BruteForce" || *itMatcher == "BruteForce-L1") && (b->defaultNorm() >= NORM_HAMMING)) { cout << "**************************************************************************\n"; cout << "It's strange. You shouldn't use L1 or L2 distance for a binary descriptor\n"; cout << "**************************************************************************\n"; } try { descriptorMatcher->match(descImg1, descImg2, matches, Mat()); // Keep best matches only to have a nice drawing. // We sort distance between descriptor matches Mat index; int nbMatch=int(matches.size()); Mat tab(nbMatch, 1, CV_32F); for (int i = 0; i(i, 0) = matches[i].distance; } sortIdx(tab, index, SORT_EVERY_COLUMN + SORT_ASCENDING); vector bestMatches; for (int i = 0; i<30; i++) { bestMatches.push_back(matches[index.at(i, 0)]); } Mat result; drawMatches(img1, keyImg1, img2, keyImg2, bestMatches, result); namedWindow(*itDesc+": "+*itMatcher, WINDOW_AUTOSIZE); imshow(*itDesc + ": " + *itMatcher, result); // Saved result could be wrong due to bug 4308 FileStorage fs(*itDesc + "_" + *itMatcher + ".yml", FileStorage::WRITE); fs<<"Matches"<::iterator it; cout<<"**********Match results**********\n"; cout << "Index \tIndex \tdistance\n"; cout << "in img1\tin img2\n"; // Use to compute distance between keyPoint matches and to evaluate match algorithm double cumSumDist2=0; for (it = bestMatches.begin(); it != bestMatches.end(); ++it) { cout << it->queryIdx << "\t" << it->trainIdx << "\t" << it->distance << "\n"; Point2d p=keyImg1[it->queryIdx].pt-keyImg2[it->trainIdx].pt; cumSumDist2=p.x*p.x+p.y*p.y; } desMethCmp.push_back(cumSumDist2); waitKey(); } catch (Exception& e) { cout << e.msg << endl; cout << "Cumulative distance cannot be computed." << endl; desMethCmp.push_back(-1); } } } catch (Exception& e) { cout << "Feature : " << *itDesc << "\n"; if (itMatcher != typeAlgoMatch.end()) { cout << "Matcher : " << *itMatcher << "\n"; } cout << e.msg << endl; } } int i=0; cout << "Cumulative distance between keypoint match for different algorithm and feature detector \n\t"; cout << "We cannot say which is the best but we can say results are differents! \n\t"; for (vector::iterator itMatcher = typeAlgoMatch.begin(); itMatcher != typeAlgoMatch.end(); ++itMatcher) { cout<<*itMatcher<<"\t"; } cout << "\n"; for (itDesc = typeDesc.begin(); itDesc != typeDesc.end(); ++itDesc) { cout << *itDesc << "\t"; for (vector::iterator itMatcher = typeAlgoMatch.begin(); itMatcher != typeAlgoMatch.end(); ++itMatcher, ++i) { cout << desMethCmp[i]<<"\t"; } cout<<"\n"; } return 0; }