自定义阈值的两种角点检测

阈值的设定主要是通过R值矩阵中的R值大小来确定的：

通过阈值来确定需要的角点R值的范围

R值矩阵的计算参看:https://www.cnblogs.com/Jack-Elvis/p/11640931.html

 

 harris和shiTomasi两种自定义阈值的角点检测代码如下：

#include <opencv2/opencv.hpp>
#include <iostream>

#include <math.h>
using namespace cv;
using namespace std;
const char* harris_win = "Custom Harris Corners Detector";
const char* shitomasi_win = "Custom Shi-Tomasi Corners Detector";

Mat src, gray_src;
// harris corner response
Mat harris_dst, harrisRspImg;
double harris_min_rsp;
double harris_max_rsp;
// shi-tomasi corner response

Mat shiTomasiRsp;
double shitomasi_max_rsp;
double shitomasi_min_rsp;
int sm_qualitylevel = 30;  //shiTomasiRsp矩阵的拖动变量
// quality level
int qualityLevel = 30;    //harrisRspImg R矩阵的拖动变量
int max_count = 100; 

void CustomHarris_Demo(int, void*);
void CustomShiTomasi_Demo(int, void*);


int main(int argc, char** argv) {
    src = imread("L:/6.jpg");
    if (src.empty()) {
        printf("could not load image...
");
        return -1;
    }
    namedWindow("input image", CV_WINDOW_AUTOSIZE);
    imshow("input image", src);
    cvtColor(src, gray_src, COLOR_BGR2GRAY);


    //1.harris方法 计算特征值：
    int blockSize = 3;
    int ksize = 3;
    double k = 0.04;
    harris_dst = Mat::zeros(src.size(), CV_32FC(6));
    harrisRspImg = Mat::zeros(src.size(), CV_32FC1);
    cornerEigenValsAndVecs(gray_src, harris_dst, blockSize, ksize, 4);
    // 计算gray_src响应,            输出herris_dst特征矩阵

    for (int row = 0; row < harris_dst.rows; row++) {
        for (int col = 0; col < harris_dst.cols; col++) {
        double lambda1 = harris_dst.at<Vec6f>(row, col)[0]; //harris_dst:M特征矩阵中的特征值：λ1
        double lambda2 = harris_dst.at<Vec6f>(row, col)[1]; //harris_dst:M特征矩阵中的特征值：λ2
        harrisRspImg.at<float>(row, col) = lambda1*lambda2 - k*pow((lambda1 + lambda2), 2);
        //harrisRspImg：R矩阵   R  =  det(M)-k(traceM)2  //平方
        }
    }
    minMaxLoc(harrisRspImg, &harris_min_rsp, &harris_max_rsp, 0, 0, Mat());
    //寻找harrisRspImg矩阵中R的最大最小值: harris_min_rsp  harris_max_rsp
    namedWindow(harris_win, CV_WINDOW_AUTOSIZE);
    createTrackbar("Quality Value:", harris_win, &qualityLevel, max_count, CustomHarris_Demo);
    CustomHarris_Demo(0, 0);



    //2. shiTomasi方法 计算最小特征值：
    shiTomasiRsp = Mat::zeros(src.size(), CV_32FC1);  //定义R矩阵类型
    cornerMinEigenVal(gray_src, shiTomasiRsp, blockSize, ksize, 4);
    // 输出shiTomasiRsp特征矩阵；即R矩阵
    minMaxLoc(shiTomasiRsp, &shitomasi_min_rsp, &shitomasi_max_rsp, 0, 0, Mat());
    //找出shiTomasiRsp矩阵最大最小值
    namedWindow(shitomasi_win, CV_WINDOW_AUTOSIZE);
    createTrackbar("Quality:", shitomasi_win, &sm_qualitylevel, max_count, CustomShiTomasi_Demo);
    CustomShiTomasi_Demo(0, 0);

    waitKey(0);
    return 0;
}

void CustomHarris_Demo(int, void*) {
    if (qualityLevel < 10) {
        qualityLevel = 10;
    }
    Mat resultImg = src.clone();
    float t = harris_min_rsp + (((double)qualityLevel) / max_count)*(harris_max_rsp - harris_min_rsp);
    //阈值t =  R矩阵最小值  +   百分之qualityLevel（滑动条）*  R矩阵的值得范围（max-min）
    for (int row = 0; row < src.rows; row++) {
        for (int col = 0; col < src.cols; col++) {
            float v = harrisRspImg.at<float>(row, col); //提取R矩阵的值给v
            if (v > t) {                               //画出v>t的点
                circle(resultImg, Point(col, row), 2, Scalar(0, 0, 255), 2, 8, 0);
            }
        }
    }

    imshow(harris_win, resultImg);
}

//下面函数与上面CustomHarris_Demo函数一样
void CustomShiTomasi_Demo(int, void*) {
    if (sm_qualitylevel < 20) {
        sm_qualitylevel = 20;
    }

    Mat resultImg = src.clone();
    float t = shitomasi_min_rsp + (((double)sm_qualitylevel) / max_count)*(shitomasi_max_rsp - shitomasi_min_rsp);
    for (int row = 0; row < src.rows; row++) {
        for (int col = 0; col < src.cols; col++) {
            float v = shiTomasiRsp.at<float>(row, col);
            if (v > t) {
                circle(resultImg, Point(col, row), 2, Scalar(0, 0, 255), 2, 8, 0);
            }
        }
    }
    imshow(shitomasi_win, resultImg);
}

结果：

1.harris                                                                                          2.shiTomasi