寻找Harris、Shi-Tomasi和亚像素角点

Harris、Shi-Tomasi和亚像素角点都是角点，隶属于特征点这个大类（特征点可以分为边缘、角点、斑点).

一、Harris角点检测是一种直接基于灰度图像的角点提取算法，稳定性较高，但是也可能出现有用信息丢失的情况。

函数：cornerHarris()

Harris corner detector.

The function runs the Harris corner detector on the image. Similarly to cornerMinEigenVal and cornerEigenValsAndVecs , for each pixel (x, y) it calculates a 2\times2 gradient covariance matrix M^{(x,y)} over a \texttt{blockSize} \times \texttt{blockSize}

（特征点计算方法）

Corners in the image can be found as the local maxima of this response map.

调用：

Mat srcGray = imread("e:/template/lena.jpg",IMREAD_GRAYSCALE);
     

         //进行角点检测
     

         Mat matHarris;
     

         cornerHarris(srcGray,matHarris,2,3,0.01);
     

         //显示
     

         Mat matDst;
     

         threshold(matHarris,matDst,0.00001,255,THRESH_BINARY);
     

         imshow("matDst",matDst);
     

         waitKey(0);

lena的结果：

二、Shi-Tomasi角点

一般认为是Harris的改进，因为当时提出的论文叫做《Good Features to Track》,所以这种角点再OpenCV中叫做goodFeatures

函数:goodFeaturesToTrack()

Determines strong corners on an image.

The function finds the most prominent corners in the image or in the specified image region, as described in ​​[154]​​

• Function calculates the corner quality measure at every source image pixel using the cornerMinEigenVal or cornerHarris .
• Function performs a non-maximum suppression (the local maximums in 3 x 3 neighborhood are retained).
• The corners with the minimal eigenvalue less than qualityLevel⋅maxx,yqualityMeasureMap(x,y)
• The remaining corners are sorted by the quality measure in the descending order.
• Function throws away each corner for which there is a stronger corner at a distance less than maxDistance.

The function can be used to initialize a point-based tracker of an object.

If the function is called with different values A and B of the parameter qualityLevel , and A > B, the vector of returned corners with qualityLevel=A will be the prefix of the output vector with qualityLevel=B .

调用：

Mat srcGray = imread("e:/template/lena.jpg",IMREAD_GRAYSCALE);    

//进行角点检测
    

        Mat matHarris;
    

        vector<Point2f> corners;//输出向量
    

        goodFeaturesToTrack(srcGray,corners,100,0.01,10,Mat(),3,false,0.04);    

        //显示
    

        Mat matDst = srcGray.clone();
    

        for (int i=0;i<corners.size();i++)    

        {
    

            circle(matDst,corners[i],2,Scalar(255));
    

        }
    

         
    

        imshow("matDst",matDst);
    

        waitKey(0);

结果：

可以看到，眼部、帽子上面的尖端这些的却是"GoodFeatures"的地方都被标注了出来

三、如果需要亚像素的角点，我们必须更进一步。

函数:cornerSubPix()

调用:需要注意现计算goodfeatures再算亚像素

Mat srcGray = imread("e:/template/lena.jpg",IMREAD_GRAYSCALE);
     

         //进行角点检测
     

         Mat matHarris;
     

         vector<Point2f> corners;//输出向量
     

         cv::goodFeaturesToTrack(
     

             srcGray,                              // Image to track
     

             corners,                          // Vector of detected corners (output)
     

             MAX_CORNERS,                       // Keep up to this many corners
     

             0.01,                              // Quality level (percent of maximum)
     

             5,                                 // Min distance between corners
     

             cv::noArray(),                     // Mask
     

             3,                                 // Block size
     

             false,                             // true: Harris, false: Shi-Tomasi     

             0.04                               // method specific parameter
     

             );
     

         cv::cornerSubPix(
     

             srcGray,                          // Input image
     

             corners,                          // Vector of corners (input and output)
     

             cv::Size(5, 5),      // Half side length of search window
     

             cv::Size(-1,-1),                   // Half side length of dead zone (-1=none)
     

             cv::TermCriteria(
     

                 cv::TermCriteria::MAX_ITER | cv::TermCriteria::EPS,
     

                 20,                              // Maximum number of iterations
     

                 0.03                             // Minimum change per iteration
     

                 )
     

             );
     

         //显示
     

         Mat matDst = srcGray.clone();
     

         for (int i=0;i<corners.size();i++)     

         {
     

             circle(matDst,corners[i],2,Scalar(255));
     

             cout<<"\t"<<"序号"<<i<<"亚像素坐标("<<corners[i].x<<","<<corners[i].y<<")"<<endl;
     

         }
     

          
     

         imshow("matDst",matDst);
     

         waitKey(0);

结果:

可以看到其计算处理小数点后面的值。

四、小结

角点虽然现在用的比较少了，但是作为基本的知识有必要了解；下一步的更为复杂的特征点模型都是基于角点的，它们之间有着一脉相承的关系。