在opencv的features2d中实现了SIFT和SURF算法,可以用于图像特征点的自动检测。具体实现是采用SurfFeatureDetector/SiftFeatureDetector类的detect函数检测SURF/SIFT特征的关键点,并保存在vector容器中,最后使用drawKeypoints函数绘制出特征点。
实验所用环境是opencv2.4.0+vs2008+win7,测试图片是:
实验代码如下。这里需要注意SurfFeatureDetector是包含在opencv2/nonfree/features2d.hpp中,所以应该include这个头文件,并且在“项目属性->链接器->输入->附加依赖项”中加入库文件:opencv_nonfree240d.lib。
/** * @SURF特征点检测并绘制特征点 * @author holybin */ #include <stdio.h> #include <iostream> #include "opencv2/core/core.hpp" #include "opencv2/highgui/highgui.hpp" //#include "opencv2/features2d/features2d.hpp" #include "opencv2/nonfree/features2d.hpp" //SurfFeatureDetector实际在该头文件中 using namespace cv; using namespace std; int main( int argc, char** argv ) { Mat src = imread( "D:\\opencv_pic\\cat3d120.jpg", 0 ); //Mat src = imread( "D:\\opencv_pic\\cat0.jpg", 0 ); if( !src.data ) { cout<< " --(!) Error reading images "<<endl; return -1; } //1--初始化SURF检测算子 int minHessian = 400; SurfFeatureDetector detector( minHessian ); //2--使用SURF算子检测特征点 vector<KeyPoint> keypoints; detector.detect( src, keypoints ); //3--绘制特征点 Mat keypointImg; drawKeypoints( src, keypoints, keypointImg, Scalar::all(-1), DrawMatchesFlags::DEFAULT ); imshow("SURF keypoints", keypointImg ); cout<<"keypoint number: "<<keypoints.size()<<endl; waitKey(0); return 0; }
检测结果:
同样的,使用SIFT特征描述子进行特征点检测的过程类似,只不过换成了SiftFeatureDetector类,实验代码如下:
/** * @SIFT特征点检测并绘制特征点 * @author holybin */ #include <stdio.h> #include <iostream> #include "opencv2/core/core.hpp" #include "opencv2/highgui/highgui.hpp" //#include "opencv2/features2d/features2d.hpp" #include "opencv2/nonfree/features2d.hpp" //SiftFeatureDetector实际在该头文件中 using namespace cv; using namespace std; int main( int argc, char** argv ) { Mat src = imread( "D:\\opencv_pic\\cat3d120.jpg", 0 ); //Mat src = imread( "D:\\opencv_pic\\cat0.jpg", 0 ); if( !src.data ) { cout<< " --(!) Error reading images "<<endl; return -1; } //1--初始化SIFT检测算子 //int minHessian = 400; SiftFeatureDetector detector;//( minHessian ); //2--使用SIFT算子检测特征点 vector<KeyPoint> keypoints; detector.detect( src, keypoints ); //3--绘制特征点 Mat keypointImg; drawKeypoints( src, keypoints, keypointImg, Scalar::all(-1), DrawMatchesFlags::DEFAULT ); imshow("SIFT keypoints", keypointImg ); cout<<"keypoint number: "<<keypoints.size()<<endl; waitKey(0); return 0; }
从检测结果可以看出,SURF算子检测到的特征点远远多于SIFT算子,至于检测的精确度如何,后面试试利用SIFT和SURF算子进行特征点匹配来区分。