I need a graph to be smooth

Here is the code I have:

x=1:10;

y=1:21;

z=[0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 -9 0 0 0 0; 0 0 0 0 -5 -7 0 0 0 0; 0 0 -10 -25 -15 -5 0 0 0 0; 0 -20 -32 -46 -27 -7 0 0 0 0; 0 -17 -61 -82 -21 0 0 0 0 0; 0 -12 -28 -52 -30 -12 0 0 0 0; 0 -22 -43 -79 -53 -30 -12 0 0 0; 0 -15 -38 -63 -92 -68 -33 0 0 0; 0 0 -15 -37 -61 -49 -22 0 0 0; 0 0 0 -17 -56 -30 -17 0 0 0; 0 0 0 -15 -56 -30 -17 0 0 0; 0 0 -15 -32 -54 -94 -79 -31 0 0; 0 0 -21 -15 -25 -52 -37 -20 0 0; 0 0 -15 -40 -70 -49 -20 -15 0 0; 0 0 -11 -23 -14 -20 -46 -29 -14 0; 0 0 0 -10 -27 -39 -25 -15 0 0; 0 0 0 -15 -32 -56 -26 -10 0 0; 0 0 0 -10 -17 -25 -15 0 0 0; 0 0 0 0 -20 -10 0 0 0 0; 0 0 0 0 -5 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0];

surf(z);

xlabel('width')

ylabel('length')

title('Depth of Paugus Bay')

I tried adding this to smooth it, but it wasn't as smooth as I needed:

n = 3;

n2 = (n-1)/2;

[x,y] = meshgrid((1:n) - n2);

K = 1./(1 + sqrt((x - n2).^2 + (y - n2).^2));

K = K/sum(K,'all');

zhat = conv2(K,z)./conv2(ones(size(z)),K);

surf(zhat)

Answer 1

[John D'Errico]

You asked this question before. I answered it. In fact, you even show the code I wrote. And you accepted my answer then as I recall.

It you want it to be smoother, then the simple answer is to use a larger smoothing kernel!

x=1:10;
y=1:21;
z=[0 0 0 0 0 0 0 0 0 0; 0 0 0 0 0 -9 0 0 0 0; 0 0 0 0 -5 -7 0 0 0 0; 0 0 -10 -25 -15 -5 0 0 0 0; 0 -20 -32 -46 -27 -7 0 0 0 0; 0 -17 -61 -82 -21 0 0 0 0 0; 0 -12 -28 -52 -30 -12 0 0 0 0; 0 -22 -43 -79 -53 -30 -12 0 0 0; 0 -15 -38 -63 -92 -68 -33 0 0 0; 0 0 -15 -37 -61 -49 -22 0 0 0; 0 0 0 -17 -56 -30 -17 0 0 0; 0 0 0 -15 -56 -30 -17 0 0 0; 0 0 -15 -32 -54 -94 -79 -31 0 0;  0 0 -21 -15 -25 -52 -37 -20 0 0; 0 0 -15 -40 -70 -49 -20 -15 0 0; 0 0 -11 -23 -14 -20 -46 -29 -14 0; 0 0 0 -10 -27 -39 -25 -15 0 0; 0 0 0 -15 -32 -56 -26 -10 0 0; 0 0 0 -10 -17 -25 -15 0 0 0; 0 0 0 0 -20 -10 0 0 0 0; 0 0 0 0 -5 0 0 0 0 0; 0 0 0 0 0 0 0 0 0 0];
surf(z);
xlabel('width')
ylabel('length')

n = 5;    % NOTE THE DIFFERENCE!
n2 = (n-1)/2;
[x,y] = meshgrid((1:n) - n2);
K = 1./(1 + sqrt((x - n2).^2 + (y - n2).^2));
K = K/sum(K,'all');
zhat = conv2(K,z)./conv2(ones(size(z)),K);
surf(zhat)

Or you might build a gaussian blur instead. Or a variation of Savitsky-Golay filter, in 2-d. The above is easy, because it can be done almost trivially.

[Shayleah Gunning-Lavoie]

Yes I did, I just wasn't sure on how to make it more smooth until now. Thank you for the explination!