cornerEigenValsAndVecs()&cornerMinEigenVal()自定义角点检测

自定义角点检测

基于Harris与Shi-Tomasi特征点检测的自定义角点检测可以通过自己设定阈值来实现特征点的检测，使用起来更为灵活。

cornerEigenValsAndVecs()函数

函数定义如上图，该函数通过计算自相关矩阵M可以求出其特征值λ1和λ2，其中src为输入图像，dst为输出图像（宽度为输入图像6倍），blockSize为邻域大小，ksize为计算梯度算子的核尺寸大小。

cornerMinEigenVal()函数

函数定义如上图，该函数与cornerEigenValsAndVecs()函数作用相似，不同点在于，该函数输出图像dst与输入图像src同样大小，并且只能获得矩阵M的最小特征值。

自定义角点检测实验程序

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

using namespace cv;
using namespace std;
#define WINDOW_NAME1 "harris自定义检测"
#define WINDOW_NAME2 "tomasi自定义检测"

Mat srcImage, grayImage, harriRspImg, shiTomasiRsp;
double harris_min_rsp, harris_max_rsp;
int g_nThresh_harris = 60, g_nMaxThresh_harris = 100;
int g_nMax_count = 100;

// Shi-Tomasi
int g_nThreshTomasi = 40, g_nMaxThreshTomasi = 100;
double tomasi_min_rsp, tomasi_max_rsp;

void on_fCustomHarrsi(int, void*)
{
	if (g_nThresh_harris < 10)
		g_nThresh_harris = 10;

	Mat resultImage = srcImage.clone();

	float t = harris_min_rsp + (((double)g_nThresh_harris) / g_nMax_count) * (harris_max_rsp - harris_min_rsp);
	for (size_t row = 0; row < srcImage.rows; row++)
	{
		for (size_t col = 0; col < srcImage.cols; col++)
		{
			float v = harriRspImg.at<float>(row, col);
			if (v > t)
				circle(resultImage, Point(col, row), 2, Scalar(0, 0, 255),-1);
		}
	}
	imshow(WINDOW_NAME1, resultImage);
}

void on_fCustomTomasi(int, void*)
{
	if (g_nThreshTomasi < 20)
	{
		g_nThreshTomasi = 20;
	}
	Mat resultImage = srcImage.clone();
	float t = tomasi_min_rsp + ((double)g_nThreshTomasi / g_nMax_count) * (tomasi_max_rsp - tomasi_min_rsp);
	for (size_t row = 0; row < srcImage.rows; row++)
	{
		for (size_t col = 0; col < srcImage.cols; col++)
		{
			float v = shiTomasiRsp.at<float>(row, col);
			if (v > t)
				circle(resultImage, Point(col, row), 2, Scalar(0, 255, 255),-1);
		}
	}
	imshow(WINDOW_NAME2, resultImage);
}

int main()
{
	srcImage = imread("C:/Users/msi-/Desktop/picture/角点检测.png");
	cvtColor(srcImage, grayImage, COLOR_BGR2GRAY);

	// 计算特征值
	int blockSize = 3, ksize = 3;
	Mat harris_dst = Mat::zeros(srcImage.size(), CV_32FC(6)); // 特征结果
	harriRspImg = Mat::zeros(srcImage.size(), CV_32FC1); // 响应结果
	double k = 0.04;

	cornerEigenValsAndVecs(grayImage, harris_dst, blockSize, ksize, 4);

	// 计算响应
	for (size_t row = 0; row < harris_dst.rows; ++row)
	{
		for (size_t col = 0; col < harris_dst.cols; ++col)
		{
			double lambda1 = harris_dst.at<Vec6f>(row, col)[0]; // λ1
			double lambda2 = harris_dst.at<Vec6f>(row, col)[1]; // λ2
			// 根据公式计算响应值 R = λ1 * λ2 - k * pow( (λ1 + λ2), 2 )
			harriRspImg.at<float>(row, col) = lambda1 * lambda2 - k * pow((lambda1 + lambda2), 2);
		}
	}

	// 求出在这些响应点中求取最大值和最小值
	minMaxLoc(harriRspImg, &harris_min_rsp, &harris_max_rsp);

	namedWindow(WINDOW_NAME1, WINDOW_AUTOSIZE);
	createTrackbar("quality value:", WINDOW_NAME1, &g_nThresh_harris, g_nMaxThresh_harris, on_fCustomHarrsi);
	on_fCustomHarrsi(0, NULL);

	// 计算Shi-Tomasi最小特征值
	shiTomasiRsp = Mat::zeros(srcImage.size(), CV_32FC1);
	cornerMinEigenVal(grayImage, shiTomasiRsp, blockSize, ksize, 4);

	minMaxLoc(shiTomasiRsp, &tomasi_min_rsp, &tomasi_max_rsp);

	namedWindow(WINDOW_NAME2, WINDOW_AUTOSIZE);
	createTrackbar("tomasi quality value:", WINDOW_NAME2, &g_nThreshTomasi, g_nMaxThreshTomasi, on_fCustomTomasi);
	on_fCustomTomasi(0, NULL);

	waitKey(0);
	return 0;
}

结果图片：