二维高斯卷积核的Matlab和OpenCV实现

1、二维高斯函数
$$G(x,y)=\frac{1}{2\pi {\sigma }^2}\exp \left(-\frac{x^2+y^2}{2{\sigma }^2}\right)$$
2、坐标$(x,y)$（以5×5卷积核为例）

3、Matlab实现

sigma1=2;
k=5;    % kernel大小
pai=3.1415926;
kernel=zeros(k);
m=(k+1)/2;
sigma=2*sigma1*sigma1;
for i=-1*(k-1)/2:(k-1)/2
   for j=-1*(k-1)/2:(k-1)/2
      kernel(i+m,j+m)=(-1/(pai*sigma))*exp(-1*(i^2+j^2)/(sigma));
   end
end
kernel=kernel./sum(kernel,'all')  % 归一化

结果与Matlab函数fspecial()生成的高斯卷积核一致，卷积核如图所示：

4、OpenCV实现

int main()
{
	double sigma = 2, pai = 3.1415926;
	int k = 5, m = (k + 1) / 2, n = m - 1;	// k-高斯卷积核大小
	Mat kernel = Mat::zeros(Size(k, k), CV_64FC1);
	double *p, Ksum;
	sigma = 2 * sigma * sigma;
	for (int i = -1 * n; i <= n; i++)
	{
		p = kernel.ptr<double>(i + n);
		for (int j = -1 * n; j <= n; j++)
		{
			p[j + n] = (1 / (pai*sigma)) * exp(-1 * (i * i + j * j) / sigma);
		}
	}
	Ksum = sum(kernel).val[0];
	kernel = kernel.mul(1/Ksum);
	cout << kernel << endl;

	getchar();
	return 0;
}