Local maxima in a matrix

I need to locate the peaks and valleys of numbers given in a large matrix. I need to compare a value in the matrix with the 8 surrounding values to determine if it is either a peak (maximum) or a valley (minimum). The matrix that I am using is 100X100 and the coordinates of the value will then need to be displayed in a table. For example: [3 6 2 5; 4 9 1 5; 7 3 5 4] has a peak of 9 at (2,2) and a valley of 1 at (3,2). Any ideas would be greatly appreciated.

Answer 1

[Image Analyst]

If you have the IMage Processing Toolbox, you are really in luck because there are two functions meant for exactly this situation: imregionalmax() and imregionalmin(). Here is how to use them:

grayImage = [3 6 2 5; 4 9 1 5; 7 3 5 4] 
regionalMaxima = imregionalmax(grayImage)
regionalMinima = imregionalmin(grayImage)

In the command window, observe the results:

grayImage =
     3     6     2     5
     4     9     1     5
     7     3     5     4
regionalMaxima =
     0     0     0     0
     0     1     0     0
     0     0     0     0
regionalMinima =
     1     0     0     0
     0     0     1     0
     0     0     0     0

If you want rows and columns of the max or mins, just use find():

[rows, columns] = find(regionalMaxima)

rows =
     2
columns =
     2

Answer 2

[Azzi Abdelmalek]

a=[3 6 2 5; 4 9 1 5; 7 3 5 4]
n=size(a);
[mina,idx1]=min(a(:))
[ii1,jj1]=ind2sub(n,idx1)
[maxa,idx2]=max(a(:))
[ii2,jj2]=ind2sub(n,idx2)

[Mohammad Abouali]

This is not the answer. This gives the global min/max. He is asking for local one.