Opencv基于SUFT算子选点，匹配，仿射变换纠正图片 RGB 3S博客

前言

之前写过一个基于仿射变换，进行图像旋转的例子。今天这个名字比较绕，其实就是SUFT算子选点，然后进行点的匹配，基于匹配计算变换矩阵，最后基于变换矩阵进行图片纠正，是Opencv应用于遥感的简单小尝试。

代码

/*
作者：山科_xxin
时间：2017-03-27 22:53:11
功能：<a href="https://www.ixxin.cn/tag/opencv/" title="查看更多关于Opencv的文章" target="_blank">Opencv</a>基于SUFI算子选点，匹配，<a href="https://www.ixxin.cn/tag/%e4%bb%bf%e5%b0%84%e5%8f%98%e6%8d%a2/" title="查看更多关于仿射变换的文章" target="_blank">仿射变换</a>纠正图片
类别：<a href="https://www.ixxin.cn/tag/opencv/" title="查看更多关于Opencv的文章" target="_blank">Opencv</a>函数练习
*/
#include <iostream>
#include <string>
#include <opencv2\opencv.hpp>  
#include <opencv2\highgui\highgui.hpp>
#include "opencv2/core/core.hpp" 
#include "opencv2/imgproc/imgproc.hpp"  
#include "opencv2/features2d/features2d.hpp"  
#include "opencv2/nonfree/nonfree.hpp"
#include "opencv2/legacy/legacy.hpp" 
using namespace std;
using namespace cv;


using namespace std;
using namespace cv;

int maopao(vector< DMatch > matches)
{
	int msize = matches.size();
	float tmp;
	for(int i = 0;i<msize-1;i++)
		for(int j = 0;j<msize-1-i;j++)
		{
			if(matches[j].distance>matches[j+1].distance)
			{
				tmp = matches[j].distance;
				matches[j].distance = matches[j+1].distance;
				matches[j+1].distance = tmp;
			}
		}
	for(int i = 0;i<msize;i++)
		cout<<matches[i].distance<<" ";
	return 0;
}
int main()
{
	cout<<"SUFT算法："<<endl;
	Mat srcImage1 = imread("G:\\opencv\\b1.JPG",1);  
    Mat srcImage2 = imread("G:\\opencv\\b11.jpg",1);  
    if( !srcImage1.data || !srcImage2.data )  
    { 
		cout<<"打开错误"<<endl; 
		return false; 
	}
	else
	{
		//if(srcImage2.data)
	}
	double t = getTickCount();
 
    int minHessian = 700
    SurfFeatureDetector detector( minHessian );
    std::vector<KeyPoint> keyPoint1, keyPoints2;
  

    detector.detect( srcImage1, keyPoint1 );  
    detector.detect( srcImage2, keyPoints2 ); 
	cout<<"图像1特征点个数:"<<keyPoint1.size()<<endl;  
    cout<<"图像2特征点个数:"<<keyPoints2.size()<<endl;
  

    SurfDescriptorExtractor extractor;  
    Mat descriptors1, descriptors2;  
    extractor.compute( srcImage1, keyPoint1, descriptors1 );  
    extractor.compute( srcImage2, keyPoints2, descriptors2 ); 
	t = ((double)getTickCount() - t)/getTickFrequency();
	cout<<"SIFT提取点算法用时："<<t<<"秒"<<endl;


  
 
    BruteForceMatcher< L2<float> > matcher;  
    std::vector< DMatch > matches;  
	 

    matcher.match( descriptors1, descriptors2, matches ); 

	maopao(matches);

	float min_dist = 100;
	float max_dist = -100;
	for(int i = 0;i<matches.size();i++)
	{  
        float dist = matches[i].distance;  
        if(dist < min_dist) min_dist = dist;  
        if(dist > max_dist) max_dist = dist;  
    }
	cout<<"Match个数："<<matches.size()<<endl;
	cout<<"最小距离"<<min_dist<<endl;
	cout<<"最大距离"<<max_dist<<endl;
	
	Point2f srcTri[3];
	Point2f dstTri[3];

	for(int i = 0;i<3;i++)
	{
		dstTri[i] = Point2f(keyPoint1.at(matches[i].queryIdx).pt.x,keyPoint1.at(matches[i].queryIdx).pt.y);
		srcTri[i] = Point2f(keyPoints2.at(matches[i].trainIdx).pt.x,keyPoints2.at(matches[i].trainIdx).pt.y);
	}



	vector<DMatch> goodMatches;  
    for(int i=0; i<3; i++)  
    {  
		goodMatches.push_back(matches[i]);
		cout <<"第一个图中的"<< matches[i].queryIdx<<"匹配了第二个图中的"<<matches[i].trainIdx<<endl;       
    }  
    //cout<<"goodMatch个数："<<goodMatches.size()<<endl;
	
	/*
	//限定点数
	int point;
	point = 30;
	nth_element(matches.begin(),    //初始位置  
        matches.begin() + (point-1),   //排序元素的位置  
        matches.end());     //终止位置  

    //移除之后所有的元素  
    matches.erase(matches.begin() + point, matches.end());  
	*/ 



    Mat imgMatches;  
    drawMatches( srcImage1, keyPoint1, srcImage2, keyPoints2,goodMatches, imgMatches,Scalar::all(-1),Scalar::all(-1),Mat(),2);//进行绘制  
	
	Mat warpMat( 2, 3, CV_32FC1 );
	Mat dstImage_warp = Mat::zeros( srcImage2.rows, srcImage2.cols,srcImage2.type() );


	warpMat = getAffineTransform( srcTri, dstTri);

	warpAffine(srcImage2, dstImage_warp, warpMat, dstImage_warp.size());

	namedWindow("MatchSUFT",1);
    imshow("MatchSUFT", imgMatches);
	imshow("还原",dstImage_warp);
	waitKey(0);

	return 0;

}