快速最近邻匹配

这是一个完整的利用Suft快速近邻匹配的程序

#include <opencv2/nonfree/nonfree.hpp>
#include <opencv2/nonfree/features2d.hpp>
#include <opencv.hpp>
#include <opencv2/imgproc/imgproc.hpp>
#include <iostream>
#include <opencv2/highgui/highgui.hpp>
#include <opencv2/features2d/features2d.hpp>
#include <opencv2/core/core.hpp>
#include <vector>
#include <math.h>
#include <opencv2/legacy/legacy.hpp>
using namespace std;
using namespace cv;
double Calcoefficient(const Mat &src1 , const Point2f key1 , const Mat &src2 , const Point2f key2) ;//用于计算匹配点之间的相关系数
    Mat src1 ; Mat src2 ;
    src1 = imread("zhugeleft.jpg") ;
    src2 = imread("zhugeright.jpg") ;
    if (!src1.data || !src2.data)
    {
        cout<<"读取图片失败！\n" ;
        return ;
    }
        SurfFeatureDetector dector(400) ;
    vector<KeyPoint> keypoints_1 , keypoints_2 ; //保存特征点向量
    dector.detect(src1 , keypoints_1) ; //检测SURF特征
    dector.detect(src2 , keypoints_2) ;

    //计算描述子
    SurfDescriptorExtractor extractor ;
    Mat descriptors_1 , descriptors_2 ;
    extractor.compute(src1 , keypoints_1 , descriptors_1) ;
    extractor.compute(src2 , keypoints_2 , descriptors_2) ;

      //Flan特征匹配算法
    FlannBasedMatcher matcher ;
    vector<DMatch> matches ;
    matcher.match(descriptors_1 , descriptors_2 , matches) ;
    cout<<"快速最近邻搜索匹配获得的匹配点数为："<<matches.size()<<endl ;
    double max_dist = 0 ; double min_dist = 100 ;  

    //计算关键点之间最大最小欧式距离
    for (int i = 0 ; i < descriptors_1.rows ; i++)
    {
        double dist = matches[i].distance ;
        if (dist < min_dist)
        {
            min_dist = dist ;
        }
        if (dist > max_dist)
        {
            max_dist = dist ;
        }
    }
    cout<<"关键点之间欧式距离最小最大值分别为：max_dist = "<<max_dist<<"\tmin_dist = "<<min_dist<<endl ;

    //选择欧式距离小于2倍最小距离的匹配作为优良匹配
    vector<DMatch> good_matches ;
    for (int i = 0 ; i < descriptors_1.rows ; i++)
    {
        if (matches[i].distance < 2 * min_dist)
        {
            good_matches.push_back(matches[i]) ;
        }
    }

    Mat m_matches ;
    m_matches.create(src1.rows , 2 * src1.cols , src1.type()) ;
    Mat imageROI ;
    imageROI = m_matches(Rect(0 , 0 , src1.cols , src1.rows)) ;
    Mat imageROI2 ;
    imageROI2 = m_matches(Rect(src1.cols , 0 , src2.cols , src2.rows)) ;

    addWeighted(imageROI , 0.0 , src1 , 1.0 , 0 , imageROI) ;
    addWeighted(imageROI2 , 0.0 , src2 , 1.0 , 0 , imageROI2) ;

    cout<<"输出匹配点坐标：\n" ;
    for (int i = 0 ; i < good_matches.size() ; ++i)
    {
        int index1 = good_matches[i].queryIdx ;
        int index2 = good_matches[i].trainIdx ;
        circle(m_matches , keypoints_1[index1].pt , 10 , Scalar(0 , 0 , 255) , 5) ;
        circle(m_matches , Point(src1.cols + keypoints_2[index2].pt.x , keypoints_2[index2].pt.y) , 10 , Scalar(0 , 0 , 255) , 5) ;
        line(m_matches , keypoints_1[index1].pt , Point(keypoints_2[index2].pt.x + src1.cols , keypoints_2[index2].pt.y) , Scalar::all(-1) , 10 , 8) ;
        double coefficent = Calcoefficient(src1 , keypoints_1[index1].pt , src2 , keypoints_2[index2].pt) ;
        cout<<keypoints_1[index1].pt<<"\t"<<keypoints_2[index2].pt<<"\t匹配系数为："<<coefficent ;
        cout<<endl ;
    }


    int good_count = good_matches.size() ;
    cout<<"优良匹配数量为："<<good_count<<endl ;
    namedWindow("m_matches" , 0) ;
    imshow("m_matches" , m_matches) ;
    waitKey(0) ;

程序运行结果;

结果是不是很好啊！