OpenCV_Find Basis F-Matrix and computeCorrespondEpilines(获取一对图像的基础矩阵及对应极线)

矫情吗；* 2022-06-17 22:38 82阅读 0赞

Codes:

// BasisMatrixCalculate.cpp : 定义控制台应用程序的入口点。
    //
    
    #include "stdafx.h"
    #include  <iostream>
    
    #include <vector>
    #include <opencv2/core/core.hpp>
    #include <opencv2/imgproc/imgproc.hpp>
    #include <opencv2/highgui/highgui.hpp>
    //引用cv::KeyPoint 特征检测器通用接口
    #include <opencv2/features2d/features2d.hpp> 
    # include <opencv2/calib3d/calib3d.hpp>
    #include <opencv2/nonfree/nonfree.hpp> //引用features2d.hpp中 SurfFeatureDetector
    #include <opencv2/legacy/legacy.hpp>  
    int main()
    {
    	//读取2张图像
    	cv::Mat image1 = cv::imread("../../aTestImage/church01.jpg", 0);
    	cv::Mat image2 = cv::imread("../../aTestImage/church03.jpg", 0);
    	if (!image1.data || !image2.data)
    		return 0;
    	//使用SURF特征 获取图像特征点
    	std::vector<cv::KeyPoint> keyPoints1;
    	std::vector<cv::KeyPoint> keyPoints2;
    	cv::SurfFeatureDetector surf(3000);
    	surf.detect(image1, keyPoints1);
    	surf.detect(image2, keyPoints2); //获取两幅图像的特征点
    
    	// //展示图像中的keyPoints
    	//cv::Mat imagekeyPt;
    	//cv::drawKeypoints(image1, keyPoints1, imagekeyPt, cv::Scalar(255, 255, 255), cv::DrawMatchesFlags::DRAW_RICH_KEYPOINTS);
    	//cv::namedWindow("image1SURFkeyPt");
    	//cv::imshow("image1SURFkeyPt", imagekeyPt);
    	//cv::drawKeypoints(image2, keyPoints2, imagekeyPt, cv::Scalar(255, 255, 255), cv::DrawMatchesFlags::DRAW_RICH_KEYPOINTS);
    	//cv::namedWindow("image2SURFkeyPt");
    	//cv::imshow("image2SURFkeyPt", imagekeyPt);
    
    	//通过特征点获取描述子
    	cv::SurfDescriptorExtractor  surfDesc;  // 构造描述子提取器
    	cv::Mat descriptors1, descriptors2;  //描述子记录局部强度差值/梯度变化/位置等信息
    	surfDesc.compute(image1, keyPoints1, descriptors1);
    	surfDesc.compute(image2, keyPoints2, descriptors2);
    
    	//匹配图像的描述子 descriptors
    	cv::BruteForceMatcher< cv::L2<float> > matcher; //构造匹配器
    	std::vector<cv::DMatch> matches; //匹配描述子
    	matcher.match(descriptors1, descriptors2, matches);
    	std::cout << "matches size= " << matches.size() << std::endl;
    
    	//选择部分描述子 使用 一对图像的基础矩阵 进行匹配观测
    	std::vector<cv::DMatch> partmatches; //部分匹配描述子 | 特征点对 的对应 索引结构体
    	partmatches.push_back(matches[14]);  //选择第n个匹配描述子
    	partmatches.push_back(matches[16]);
    	partmatches.push_back(matches[141]);
    	partmatches.push_back(matches[242]);
    	partmatches.push_back(matches[236]);
    	partmatches.push_back(matches[238]);
    	//partmatches.push_back(matches[100]);
    	partmatches.push_back(matches[200]);
    
    
    	//画出选择的匹配描述子 两幅图像连线
    	cv::Mat imagePartMatches;
    	cv::drawMatches(image1, keyPoints1, image2, keyPoints2,
    		partmatches, imagePartMatches, cv::Scalar(255, 255, 255));
    	cv::namedWindow("imagePartMatches");
    	cv::imshow("imagePartMatches", imagePartMatches);
    
    	//将一维向量的keyPoints点转换成二维的Point2f点
    	std::vector <int> pointIndexes1;//记录选择的匹配特征点的索引向量
    	std::vector <int> pointIndexes2;
    
    	for (std::vector<cv::DMatch>::const_iterator it = partmatches.begin();
    		it != partmatches.end(); ++it)
    	{
    		pointIndexes1.push_back(it->queryIdx); //查询图像1特征点索引
    		pointIndexes2.push_back(it->trainIdx);  //训练图像2特征点索引
    	}
    	std::vector <cv::Point2f>  selPoints1, selPoints2;
    	cv::KeyPoint::convert(keyPoints1, selPoints1, pointIndexes1);//将索引指定的 特征点 转换成2D点
    	cv::KeyPoint::convert(keyPoints2, selPoints2, pointIndexes2);
    	画出转换后的点二维点到原图像
    	std::vector<cv::Point2f>::const_iterator  it = selPoints1.begin();
    	//while (it != selPoints1.end())
    	//{
    	//	cv::circle(image1, *it, 3, cv::Scalar(255, 255, 255), 5);
    	//	++it;
    	//}
    	it = selPoints2.begin();
    	while (it != selPoints2.end())
    	{
    		cv::circle(image2, *it, 3, cv::Scalar(255, 255, 255), 2);
    		++it;
    	}
    	//cv::namedWindow("image1");
    	//cv::imshow("image1", image1);
    	//cv::namedWindow("image2");
    	//cv::imshow("image2", image2);
    
    	// 获取该对图像的基础矩阵 (使用7个匹配描述子matches) CV_FM_7POINT
    
    	cv::Mat fundemental = cv::findFundamentalMat(cv::Mat(selPoints1), cv::Mat(selPoints2),CV_FM_8POINT);//CV_FM_LMEDS
    	std::cout << "F-Matrix size= " << fundemental.rows << "," << fundemental.cols << std::endl;
    	
    	//使用基础矩阵 在对应图像上绘制外极线
    	std::vector<cv::Vec3f> lines1; //存储外极线
    	cv::computeCorrespondEpilines(cv::Mat(selPoints1), 1, fundemental, lines1);//获取图像1中的二维特征点 在图像2中对应的外极线
    	for (std::vector<cv::Vec3f>::const_iterator it = lines1.begin(); 
    		  it != lines1.end(); ++it)
    	{
    		cv::line(image2, 
    			cv::Point(0, -(*it)[2] / (*it)[1] ),
    			cv::Point(image2.cols ,  -( (*it)[2] + (*it)[0] * image2.cols )/(*it)[1] ),
    			cv::Scalar(255,255,255));
    	}
    	cv::namedWindow("Image2 Epilines");
    	cv::imshow("Image2 Epilines", image2);
    
    	cv::waitKey(0);
    	return 0;
    }

Result:

![Center][]

[Center]: /images/20220618/ae867ea610b64406a88d8f8050bc8d34.png