Finding the gradient of a multivariable function via a spatial convolution.

Hi , I’m trying to write a function that can manually determine the gradient of a surface (represented by a multivariable function that I have created) at a particular point (x,y)

The code I've written is shown below.

function [pdx,pdy] = comp_gradient(x,y)

kernelx = [2.5,0,-2.5];

kernely = [2.5;0;-2.5];

z = cos(x/2).* cos(y) + (y/10) - (x/5);

pdx = conv2(kernelx,z,'same');

pdy = conv2(kernely,z,'same');

end

I'm not sure where to go from here, my output is clearly incorrect.

Any help would be greatly appreciated.