利用特征匹配寻找已知物体位置

版权声明：本文为博主原创文章，未经博主允许不得转载。 https://blog.csdn.net/xxzxxzdlut/article/details/72962900

使用匹配寻找已知物体位置，并用方框标识出来。
使用劳氏算法进行特征点排除的时候，先进行的是匹配目标，然后是匹配标版。顺序弄错会drawmatches出错。
主要函数：

• 寻找透视变换：findHomography(InputArray src,//源平面上对应点
  InputArray dst,//目标平面上对应点
  int method=0,//计算方法，默认0，CV_RANSAC，CV_LMEDS三种
  double ransacReprojThreshold=3,//允许重投影误差值，src，dst若以像素为单位，取1到10
  OutputArray mask=noArray()//设置输出掩码)

• 进行透视矩阵变换：perspectiveTransform(InputArray src,//原图像，需为双通道或三通道浮点型图像，每个元素是二维或三维可转换矢量
  OutputArray dest,//保存运算结果
  InputArray m //变换矩阵，3*3或4*4浮点型矩阵)

#include<opencv2/core/core.hpp>
#include<opencv2/features2d/features2d.hpp>
#include<opencv2/highgui/highgui.hpp>
#include<opencv2/calib3d/calib3d.hpp>
#include<opencv2/xfeatures2d/nonfree.hpp>
#include<opencv2/opencv.hpp>
#include<iostream>
using namespace std;
using namespace cv;
using namespace cv::xfeatures2d;

int main()
{
    Mat srcImage1 = imread("1.jpg",1);
    Mat srcImage2 = imread("2.jpg",1);
    if (!srcImage1.data || !srcImage2.data)
    {
        cout << "读取出错" << endl;
        return false;
    }
    Mat srcImage1_gray, srcImage2_gray;
    cvtColor(srcImage1,srcImage1_gray,CV_BGR2GRAY);
    cvtColor(srcImage2,srcImage2_gray,CV_BGR2GRAY);

    Ptr<SurfFeatureDetector> Detector = SurfFeatureDetector::create(100);
    vector<KeyPoint> keypoints_object, keypoints_scene;
    Mat descriptor_object, descriptor_scene;
    Detector->detectAndCompute(srcImage2_gray,Mat(),keypoints_scene,descriptor_scene);
    vector<Mat>sc_collection(1,descriptor_scene);
    FlannBasedMatcher matcher;
    matcher.add(sc_collection);
    matcher.train();

    Detector->detectAndCompute(srcImage1_gray,Mat(),keypoints_object,descriptor_object);
    vector<vector<DMatch>>matches;
    matcher.knnMatch(descriptor_object,matches,2);

    vector<DMatch>googmatches;
    for (unsigned int i = 0; i < matches.size(); i++)
    {
        if (matches[i][0].distance < 0.6*matches[i][1].distance)
            googmatches.push_back(matches[i][0]);
    }

    Mat img_matches;
    drawMatches(srcImage1_gray,keypoints_object, srcImage2_gray, keypoints_scene, googmatches,img_matches);

    vector<Point2f>obj;
    vector<Point2f>scene;

    for (unsigned int i = 0; i < googmatches.size(); i++)
    {
                                     obj.push_back(keypoints_object[googmatches[i].queryIdx].pt);
        scene.push_back(keypoints_scene[googmatches[i].trainIdx].pt);
    }

    Mat H = findHomography(obj,scene,CV_RANSAC);
    vector<Point2f>obj_corners(4);
    obj_corners[0] = cvPoint(0,0);
    obj_corners[1] = cvPoint(srcImage1_gray.cols,0);
    obj_corners[2] = cvPoint(srcImage1_gray.cols, srcImage1_gray.rows);
    obj_corners[3] = cvPoint(0,srcImage1_gray.rows);

    vector<Point2f>scene_corners(4);
    perspectiveTransform(obj_corners,scene_corners,H);

    line(img_matches,
        scene_corners[0]+Point2f(static_cast<float>(srcImage1_gray.cols),0),
        scene_corners[1]+Point2f(static_cast<float>(srcImage1_gray.cols),0),
        Scalar(255,0,123),4);
    line(img_matches,
        scene_corners[1] + Point2f(static_cast<float>(srcImage1_gray.cols), 0),
        scene_corners[2] + Point2f(static_cast<float>(srcImage1_gray.cols), 0),
        Scalar(255, 0, 123), 4);
    line(img_matches,
        scene_corners[2] + Point2f(static_cast<float>(srcImage1_gray.cols), 0),
        scene_corners[3] + Point2f(static_cast<float>(srcImage1_gray.cols), 0),
        Scalar(255, 0, 123), 4);
    line(img_matches,
        scene_corners[3] + Point2f(static_cast<float>(srcImage1_gray.cols), 0),
        scene_corners[0] + Point2f(static_cast<float>(srcImage1_gray.cols), 0),
        Scalar(255, 0, 123), 4);

    imshow("寻找物体位置",img_matches);
    waitKey(0);
    return 0;
}