(opencv)ORB匹配算法

#include<opencv2/opencv.hpp>
#include<iostream>
using namespace std;
using namespace cv;
#include<vector>Mat img1 = imread("C:\\Users\\Administrator\\Desktop\\水面\\figa1.png");//读入图像
Mat img1_gray;
Mat img2 = imread("C:\\Users\\Administrator\\Desktop\\水面\\figa2.png");//读入图像
Mat img2_gray;
int thresh = 300;
RNG rng(12345);void PointsToKeyPoints(vector<Point2f>corner1, vector<KeyPoint>& keypoints_1)
{for (size_t i = 0; i < corner1.size(); i++) {keypoints_1.push_back(KeyPoint(corner1[i], 1.f));}
}//回调函数
void on_harris(int, void*)
{vector<Point2f>corner1;goodFeaturesToTrack(img1_gray, corner1, thresh, 0.01, 0.1, Mat(), 3, false, 0.04);vector<Point2f>corner2;goodFeaturesToTrack(img2_gray, corner2, thresh, 0.01, 0.1, Mat(), 3, false, 0.04);//cout << "检测到的角点数量为：" << corner1.size() << endl;//以随机颜色绘制出角点for (int i = 0; i < corner1.size(); i++){circle(img1, corner1[i], 4, Scalar(rng.uniform(0, 255), rng.uniform(0, 255), rng.uniform(0, 255)), 2, 8, 0);}imshow("【原始图像1】", img1);//以随机颜色绘制出角点for (int i = 0; i < corner2.size(); i++){circle(img2, corner2[i], 4, Scalar(rng.uniform(0, 255), rng.uniform(0, 255), rng.uniform(0, 255)), 2, 8, 0);}imshow("【原始图像2】", img2);//  调用cornerSubPix函数计算出亚像素角点的位置TermCriteria criteria = TermCriteria(TermCriteria::EPS + TermCriteria::MAX_ITER, 40, 0.001);cornerSubPix(img1_gray, corner1, Size(5, 5), Size(-1, -1), criteria);cornerSubPix(img2_gray, corner2, Size(5, 5), Size(-1, -1), criteria);//输出角点信息cout << corner1.size() << endl;cout << corner2.size() << endl;vector<KeyPoint> keypoints_1, keypoints_2;//关键点Mat descriptors_1, descriptors_2;//描述子Ptr<FeatureDetector> detector = ORB::create();Ptr<DescriptorExtractor> descriptor = ORB::create();Ptr<DescriptorMatcher> matcher = DescriptorMatcher::create("BruteForce-Hamming");//暴力匹配PointsToKeyPoints(corner1, keypoints_1);PointsToKeyPoints(corner2, keypoints_2);//--第二步根据角点位置计算BRIEF描述子descriptor->compute(img1, keypoints_1, descriptors_1);descriptor->compute(img2, keypoints_2, descriptors_2);//--第三步：对两幅图像中的BRIEF描述子进行匹配，使用Hamming距离vector<DMatch> matches;matcher->match(descriptors_1, descriptors_2, matches);//--第四步：匹配点对筛选auto min_max = minmax_element(matches.begin(), matches.end(),[](const DMatch& m1, const DMatch& m2) {return m1.distance < m2.distance; });double min_dist = min_max.first->distance;double max_dist = min_max.second->distance;printf("-- Mat dist : %f \n", max_dist);printf("-- Mat dist : %f \n", min_dist);//当描述子之间的距离大于两倍的最小距离时，即认为匹配有误。但是有时最小距离会非常小，所以要设置一个经验值30作为下限vector<DMatch> good_matches;for (int i = 0; i < descriptors_1.rows; i++) {if (matches[i].distance <= max(2 * min_dist, 30.0)) {good_matches.push_back(matches[i]);}}//--第五步：绘制匹配结果Mat img_match;Mat img_goodmatch;drawMatches(img1, keypoints_1, img2, keypoints_2, matches, img_match);drawMatches(img1, keypoints_1, img2, keypoints_2, good_matches, img_goodmatch);namedWindow("all matches", WINDOW_FREERATIO);namedWindow("good matches", WINDOW_FREERATIO);imshow("all matches", img_match);imshow("good matches", img_goodmatch);std::vector<Point2f> obj;std::vector<Point2f> scene;for (size_t i = 0; i < good_matches.size(); i++){//-- Get the keypoints from the good matchesobj.push_back(keypoints_1[good_matches[i].queryIdx].pt);scene.push_back(keypoints_2[good_matches[i].trainIdx].pt);}Mat H = findHomography(obj, scene, RANSAC);//-- Get the corners from the image_1 ( the object to be "detected" )std::vector<Point2f> obj_corners(4);obj_corners[0] = cvPoint(0, 0); obj_corners[1] = cvPoint(img1.cols, 0);obj_corners[2] = cvPoint(img1.cols, img1.rows); obj_corners[3] = cvPoint(0, img1.rows);std::vector<Point2f> scene_corners(4);perspectiveTransform(obj_corners, scene_corners, H);//-- Draw lines between the corners (the mapped object in the scene - image_2 )line(img_goodmatch, scene_corners[0] + Point2f(img1.cols, 0), scene_corners[1] + Point2f(img1.cols, 0), Scalar(0, 255, 0), 8);line(img_goodmatch, scene_corners[1] + Point2f(img1.cols, 0), scene_corners[2] + Point2f(img1.cols, 0), Scalar(0, 255, 0), 8);line(img_goodmatch, scene_corners[2] + Point2f(img1.cols, 0), scene_corners[3] + Point2f(img1.cols, 0), Scalar(0, 255, 0), 8);line(img_goodmatch, scene_corners[3] + Point2f(img1.cols, 0), scene_corners[0] + Point2f(img1.cols, 0), Scalar(0, 255, 0), 8);//-- Show detected matchesnamedWindow("Good Matches & Object detection",WINDOW_FREERATIO),imshow("Good Matches & Object detection", img_goodmatch);}int main()
{/*img1 = img1(Rect(1000, 0, 400, 400));img2 = img2(Rect(1000, 0, 400, 400));*/img1_gray = img1.clone();cvtColor(img1_gray, img1_gray, COLOR_RGB2GRAY);namedWindow("【原始图像1】", WINDOW_FREERATIO);img2_gray = img2.clone();cvtColor(img2_gray, img2_gray, COLOR_RGB2GRAY);namedWindow("【原始图像2】", WINDOW_FREERATIO);createTrackbar("【最大角点数】", "【原始图像1】", &thresh, 500, on_harris);on_harris(0, 0);waitKey(0);return 0;
}

(opencv)ORB匹配算法相关推荐

opencv ORB特征匹配
AKAZE 局部特征匹配级联分类器使用等比例缩放图片给图片加logo 鱼眼校正智能答卷识别 opencv滤镜效果灰度图像增强方式 opencv模板匹配基础知识点 ORB 算法使用 FAS ...
java opencv 模板匹配算法_OpenCV探索之路（九）：模板匹配
模板匹配的作用在图像识别领域作用可大了.那什么是模板匹配? 模板匹配,就是在一幅图像中寻找另一幅模板图像最匹配(也就是最相似)的部分的技术. 说的有点抽象,下面给个例子说明就很明白了. 在上面这幅全明 ...
opencv(ORB算法)实现相似度检测
直接上代码 # 自定义计算两个图片相似度函数 def img_similarity(img1_path, img2_path):""":param img1_path: ...
OpenCV模板匹配算法详解
本博客在https://www.cnblogs.com/zhaoweiwei/p/OpenVC_matchTemplate.html基础上进行更加详细的注解.当初有几个地方看的比较费劲,但是里面没有注 ...
OpenCV | ORB特征检测与描述
ORB (Oriented FAST and Rotated BRIEF) ORB基本上是FAST关键点检测器和Brief描述符的融合,并进行了许多修改以增强性能.首先,它使用FAST查找关键点,然后 ...
OpenCV ORB角点检测
ORB算法是FAST算法和BRIEF算法的结合,ORB可以用来对图像中的关键点快速创建特征向量,并用这些特征向量来识别图像中的对象. 实例化ORB orb = cv.ORB_create(nfeatu ...
学习OpenCV——ORB简化版Location加速版
根据前面surf简化版的结构,重新把ORB检测的代码给简化以下,发现虽然速度一样,确实能省好多行代码,关键是有 BruteForceMatcher<HammingLUT>matcher的帮 ...
opencv ORB角检测
2011年提出融合了Oriented FAST and Rotated BRIEF的优点参考文献: https://docs.opencv.org/3.4.3/d1/d89/tutorial_py_ ...
opencv orb
代码在git . ├── 1.png ├── 2.png ├── build ├── CMakeLists.txt └── orb_cv.cpp cd build cmake .. make ./or ...

(opencv)ORB匹配算法

(opencv)ORB匹配算法相关推荐

最新文章

热门文章