ios上opencv的图像特征点检测，核心代码。如果你做的更完善了，希望能反馈给我

//
//  simpleMatching.cpp
//  openCV_test
//
//  Created by mark on 14-6-3.
//  Copyright (c) 2014年 mark. All rights reserved.
//

#include "simpleMatching.h"

//#define DRAW_RICH_KEYPOINTS_MODE 1  //mark definition
//#define DRAW_OUTLIERS_MODE 1  //mark definition

const string winName = "correspondences";

enum { NONE_FILTER = 0, CROSS_CHECK_FILTER = 1 };

int getMatcherFilterType( const string& str ){
    if( str == "NoneFilter" )
        return NONE_FILTER;
    if( str == "CrossCheckFilter" )
        return CROSS_CHECK_FILTER;
    CV_Error(CV_StsBadArg, "Invalid filter name");
    return -1;
}

static void simpleMatching( Ptr<DescriptorMatcher>& descriptorMatcher,
                           const Mat& descriptors1, const Mat& descriptors2,
                           vector<DMatch>& matches12 ){
    vector<DMatch> matches;
    descriptorMatcher->match( descriptors1, descriptors2, matches12 );
}

static void crossCheckMatching( Ptr<DescriptorMatcher>& descriptorMatcher,
                               const Mat& descriptors1, const Mat& descriptors2,
                               vector<DMatch>& filteredMatches12, int knn=1 ){
    filteredMatches12.clear();
    vector<vector<DMatch> > matches12, matches21;
    descriptorMatcher->knnMatch( descriptors1, descriptors2, matches12, knn );
    descriptorMatcher->knnMatch( descriptors2, descriptors1, matches21, knn );
    for( size_t m = 0; m < matches12.size(); m++ ){
        bool findCrossCheck = false;
        for( size_t fk = 0; fk < matches12[m].size(); fk++ ){
            DMatch forward = matches12[m][fk];

            for( size_t bk = 0; bk < matches21[forward.trainIdx].size(); bk++ ){
                DMatch backward = matches21[forward.trainIdx][bk];
                if( backward.trainIdx == forward.queryIdx ){
                    filteredMatches12.push_back(forward);
                    findCrossCheck = true;
                    break;
                }
            }
            if( findCrossCheck ) break;
        }
    }
}

static void warpPerspectiveRand( const Mat& src, Mat& dst, Mat& H, RNG& rng ){
    H.create(3, 3, CV_32FC1);
    H.at<float>(0,0) = rng.uniform( 0.8f, 1.2f);
    H.at<float>(0,1) = rng.uniform(-0.1f, 0.1f);
    H.at<float>(0,2) = rng.uniform(-0.1f, 0.1f)*src.cols;
    H.at<float>(1,0) = rng.uniform(-0.1f, 0.1f);
    H.at<float>(1,1) = rng.uniform( 0.8f, 1.2f);
    H.at<float>(1,2) = rng.uniform(-0.1f, 0.1f)*src.rows;
    H.at<float>(2,0) = rng.uniform( -1e-4f, 1e-4f);
    H.at<float>(2,1) = rng.uniform( -1e-4f, 1e-4f);
    H.at<float>(2,2) = rng.uniform( 0.8f, 1.2f);

    warpPerspective( src, dst, H, src.size() );
}

UIImage* doIteration( const Mat& img1, Mat& img2, bool isWarpPerspective,
                     vector<KeyPoint>& keypoints1, const Mat& descriptors1,
                     Ptr<FeatureDetector>& detector, Ptr<DescriptorExtractor>& descriptorExtractor,
                     Ptr<DescriptorMatcher>& descriptorMatcher, int matcherFilter, bool eval,
                     double ransacReprojThreshold, RNG& rng ){
    assert( !img1.empty() );
    Mat H1to2;
    if( isWarpPerspective )
        warpPerspectiveRand(img1, img2, H1to2, rng );
    else
        assert( !img2.empty()/* && img2.cols==img1.cols && img2.rows==img1.rows*/ );

    cout << endl << "< Extracting keypoints from second image

";
    vector<KeyPoint> keypoints2;
#define mark 1
#if mark
    int cols = img2.cols;
    int rows = img2.rows;

    int delta = 0;
    img2 = Mat(img2, Range(delta, rows-delta), Range(delta, cols-delta));
    detector->detect(img2, keypoints2);
    /*
     1 2 3 4
     5 6 7 8
     2 3 4 5
     0 9 8 7
     */
//
//    for(int i= 0; i < 2; i++){
//        for(int j = 0; j < 2; j++){
//            Mat tmpMat = Mat(img2, Range(rows*i/2, rows*(i+1)/2), Range(cols*j/2, cols*(j+1)/2));
//            vector<KeyPoint> keypoints;
//            detector->detect( tmpMat, keypoints );
//            for(auto itr = keypoints.begin(); itr != keypoints.end(); ++itr){
//                keypoints2.push_back(*itr);
//            }
//        }
//    }

//    /*左上角
//     1 2
//     5 6
//     */
//    Mat tmpMat = Mat(img2, Range(0, rows/2), Range(0, cols/2));
//    vector<KeyPoint> keypoints;
//    detector->detect( tmpMat, keypoints );
//    for(auto itr = keypoints.begin(); itr != keypoints.end(); ++itr){
//        keypoints2.push_back(*itr);
//    }
//    /*右下脚
//     4 5
//     8 7
//     */
//    tmpMat = Mat(img2, Range(rows/2, rows), Range(cols/2, cols));
//    keypoints.clear();
//    detector->detect( tmpMat, keypoints );
//    for(auto itr = keypoints.begin(); itr != keypoints.end(); ++itr){
//        keypoints2.push_back(*itr);
//    }
//    /*右上角
//     3 4
//     7 8
//     */
//    tmpMat = Mat(img2, Range(rows/2, rows), Range(0, cols/2));
//    keypoints.clear();
//    detector->detect( tmpMat, keypoints );
//    for(auto itr = keypoints.begin(); itr != keypoints.end(); ++itr){
//        keypoints2.push_back(*itr);
//    }
//    /*左下角
//     2 3
//     0 9
//     */
//    tmpMat = Mat(img2, Range(0, rows/2), Range(cols/2, cols/2));
//    keypoints.clear();
//    detector->detect( tmpMat, keypoints );
//    for(auto itr = keypoints.begin(); itr != keypoints.end(); ++itr){
//        keypoints2.push_back(*itr);
//    }
#else
    detector->detect( img2, keypoints2 );
#endif
    cout << keypoints2.size() << " points" << ">" << endl;

    if( !H1to2.empty() && eval ){
        cout << "< Evaluate feature detector

" << endl;
        float repeatability;
        int correspCount;
        evaluateFeatureDetector( img1, img2, H1to2, &keypoints1, &keypoints2,
                                repeatability, correspCount );
        cout << "repeatability = " << repeatability << endl;
        cout << "correspCount = " << correspCount << endl << ">" << endl;
    }

    cout << "< Computing descriptors for keypoints from second image

";
    Mat descriptors2;
    descriptorExtractor->compute( img2, keypoints2, descriptors2 );
    cout << ">" << endl;

    cout << "< Matching descriptors

";
    vector<DMatch> filteredMatches;
    switch( matcherFilter ){
        case CROSS_CHECK_FILTER :
            crossCheckMatching( descriptorMatcher, descriptors1, descriptors2, filteredMatches, 1 );
            break;
        default :
            simpleMatching( descriptorMatcher, descriptors1, descriptors2, filteredMatches );
    }
    cout << ">" << endl;

    if( !H1to2.empty() && eval ){
        cout << "< Evaluate descriptor matcher

" << endl;
        vector<Point2f> curve;
        Ptr<GenericDescriptorMatcher> gdm =
        new VectorDescriptorMatcher( descriptorExtractor, descriptorMatcher );
        evaluateGenericDescriptorMatcher( img1, img2, H1to2,
                                         keypoints1, keypoints2, 0, 0, curve, gdm );

        Point2f firstPoint = *curve.begin();
        Point2f lastPoint = *curve.rbegin();
        int prevPointIndex = -1;
        cout << "1-precision = " << firstPoint.x << "; recall = " << firstPoint.y << endl;
        for( float l_p = 0; l_p <= 1 + FLT_EPSILON; l_p+=0.05f ){
            int nearest = getNearestPoint( curve, l_p );
            if( nearest >= 0 ){
                Point2f curPoint = curve[nearest];
                if( curPoint.x > firstPoint.x && curPoint.x < lastPoint.x &&
                   nearest != prevPointIndex ){
                    cout << "1-precision = " << curPoint.x << "; recall = " << curPoint.y << endl;
                    prevPointIndex = nearest;
                }
            }
        }
        cout << "1-precision = " << lastPoint.x
        << "; recall = " << lastPoint.y << endl << ">" << endl;
    }

    vector<int> queryIdxs( filteredMatches.size() ), trainIdxs( filteredMatches.size() );
    for( size_t i = 0; i < filteredMatches.size(); i++ ){
        queryIdxs[i] = filteredMatches[i].queryIdx;
        trainIdxs[i] = filteredMatches[i].trainIdx;
    }

    if( !isWarpPerspective && ransacReprojThreshold >= 0 ){
        cout << "< Computing homography (RANSAC)

";
        vector<Point2f> points1; KeyPoint::convert(keypoints1, points1, queryIdxs);
        vector<Point2f> points2; KeyPoint::convert(keypoints2, points2, trainIdxs);
        H1to2 = findHomography( Mat(points1), Mat(points2), CV_RANSAC, ransacReprojThreshold );
        cout << ">" << endl;
    }

    Mat drawImg;
    if( !H1to2.empty() ){   // filter outliers
        vector<char> matchesMask( filteredMatches.size(), 0 );
        vector<Point2f> points1; KeyPoint::convert(keypoints1, points1, queryIdxs);
        vector<Point2f> points2; KeyPoint::convert(keypoints2, points2, trainIdxs);
        Mat points1t; perspectiveTransform(Mat(points1), points1t, H1to2);

        double maxInlierDist = ransacReprojThreshold < 0 ? 3 : ransacReprojThreshold;
        for( size_t i1 = 0; i1 < points1.size(); i1++ ){
            if( norm(points2[i1] - points1t.at<Point2f>((int)i1,0)) <= maxInlierDist ) // inlier
                matchesMask[i1] = 1;
        }
        // draw inliers
        drawMatches( img1, keypoints1, img2, keypoints2,
                    filteredMatches, drawImg, CV_RGB(0, 255, 0), CV_RGB(0, 0, 255), matchesMask
#if DRAW_RICH_KEYPOINTS_MODE
                    , DrawMatchesFlags::DRAW_RICH_KEYPOINTS
#endif
                    );

#if DRAW_OUTLIERS_MODE
        // draw outliers
        for( size_t i1 = 0; i1 < matchesMask.size(); i1++ )
            matchesMask[i1] = !matchesMask[i1];
        drawMatches( img1, keypoints1, img2, keypoints2, filteredMatches,  drawImg,
                    CV_RGB(0, 0, 255), CV_RGB(255, 0, 0), matchesMask,
                    DrawMatchesFlags::DRAW_OVER_OUTIMG | DrawMatchesFlags::NOT_DRAW_SINGLE_POINTS );
#endif

        cout << "Number of inliers: " << countNonZero(matchesMask) << endl;
    }else
        drawMatches( img1, keypoints1, img2, keypoints2, filteredMatches, drawImg );

    return  [UIImage UIImageFromCVMat: drawImg];
}

UIImage * getMatch( Mat & img1, Mat & img2){
    cv::initModule_nonfree();  ////if use SIFT or SURF

    bool isWarpPerspective = 0;
    double ransacReprojThreshold = 3;

    cout << "< Creating detector, ";
    Ptr<FeatureDetector> detector = FeatureDetector::create( "SURF" );  //"HARRIS"
    cout << " descriptor extractor ";
    Ptr<DescriptorExtractor> descriptorExtractor = DescriptorExtractor::create( "SURF");
    cout << " and descriptor matcher

";
    Ptr<DescriptorMatcher> descriptorMatcher = DescriptorMatcher::create( "BruteForce");
    int mactherFilterType = getMatcherFilterType( "CrossCheckFilter");

    bool eval = false;
    cout << ">" << endl;

    if( detector.empty() || descriptorExtractor.empty() || descriptorMatcher.empty()  ){
        cout<<"Can not create " <<
        (detector.empty() ? "detector" : descriptorExtractor ? "descriptor exstractor":"descriptor")
        << " matcher of given types" <<endl;
        exit(-1);
    }

    if( img1.empty() || (!isWarpPerspective && img2.empty()) ){
        cout << "Can not read images" << endl;
        exit(-1);
    }

    cout << endl << "< Extracting keypoints from first image

";
    vector<KeyPoint> keypoints1;
    detector->detect( img1, keypoints1 );
    cout << keypoints1.size() << " points"  << ">" << endl;

    cout << "< Computing descriptors for keypoints from first image

";
    Mat descriptors1;
    descriptorExtractor->compute( img1, keypoints1, descriptors1 );
    cout << ">" << endl;

    RNG rng = theRNG();

    UIImage *retImg = doIteration( img1, img2, isWarpPerspective, keypoints1, descriptors1,
                       detector, descriptorExtractor, descriptorMatcher, mactherFilterType, eval,
                       ransacReprojThreshold, rng);
    detector.release();
    descriptorExtractor.release();
    descriptorMatcher.release();
    img1.release();
    img2.release();

    return retImg;
}

发表于 2014-06-12 18:06 何清龙阅读(3541) 评论(0) 编辑收藏引用所属分类: 开源项目

常用链接

留言簿(16)

随笔分类(49)

随笔档案(61)

文章分类(18)

文章档案(34)

嵌入式

软件工程

算法

搜索

最新评论

评论排行榜