i need to find peaks and valleys in a large array and store them in another array. help?

my code so far is this

load('elevation_data');
e=elevation_data;
maxarray=[];
minarray=[];

[n,m]=size(elevation_data);
for j=2:n-1
    for k=2:m-1
        if elevation_data(j,k)>max([e(j-1,k),e(j-1,k-1),e(j-1,k+1),e(j,k+1),e(j,k-1),e(j+1,k-1),e(j+1,k),e(j+1,k-1)]);
                elevation_data(j,k)=maxarray;
        elseif elevation_data(j,k)<min([e(j-1,k),e(j-1,k-1),e(j-1,k+1),e(j,k+1),e(j,k-1),e(j+1,k-1),e(j+1,k),e(j+1,k-1)]);
            elevation_data(j,k)=minarray;
        end
    end
end

Answer 1

[Matt J]

Could you just use imregionalmax() and imregionalmin()?

[Daniel]

i guess so, i am a beginner to matlab though and we are on for loops right now. is there a way to do the same thing with arrays and for loops?

[Matt J]

If you have to use loops, your code thus far looks pretty close. The only problem I see is

elevation_data(j,k)=maxarray;

elevation_data(j,k)=minarray;

Since maxarray and minarray are always empty, I don't know what those lines are supposed to accomplish. Why would you even be trying to modify elevation_data(j,k)? You're trying to produce a new array based on it.

[Daniel]

oh okay so if i wanted to store the value of the max and mins based on the surrounding values, would how would i do that? i need them to be in arrays but with the mins and maxs in different ones.

so really the question is how do i store the values into the arrays from the if statements?

[Matt J]

One thing you could do is to first pre-allocate minarray and maxarray as arrays the same size of elevation_data

minarray=nan(size(elevation_data));
maxarray=nan(size(elevation_data));

You haven't said what should go in the places which are neither mins nor maxs, so I assume them to be NaN.

Then in the loop, you assign the min/max value to these arrays instead of to elevation_data.

Answer 2

[Andrei Bobrov]

as imregionalmax and imregionalmin :

A = elevation_data;
a1 = inf(size(A) + 2);
a2 = -a1;
a1(2:end-1,2:end-1) = A;
a2(2:end-1,2:end-1) = A;
s = size(A);
i1 = bsxfun(@plus,(1:s(1))',(0:s(2)-1)*(s(1)+2));
i2 =  bsxfun(@plus,(0:2).',(0:2)*(s(1)+2));
idx = bsxfun(@plus,i1(:),i2(:).');
idx1 = idx(:,5);
idx2 = idx(:,[1:4,6:end]);
mins = reshape(all(bsxfun(@le,a1(idx1),a1(idx2)),2),size(A));
maxs = reshape(all(bsxfun(@ge,a2(idx1),a2(idx2)),2),size(A));