Question about finding pits of a matrix?

o

[Image Analyst]

Johny - don't do that. Don't edit away your question. It really annoys people a lot.

[Tara Wagoner]

haha

Answer 1

[Matt J]

If I've understood what you're after, the whole task is doable in just a few lines,

 minval=min(map(:)); %minimum cell value

 pitmap=(map==minval); %all minima/pits

 pitmap([1,end],[1,end])=false;  %get rid of edge pits

 [pitrows,pitcols]=find(pitmap);   %pit coordinates

[Matt J]

So you're saying "pits" are local minima, not global ones? If so, you can use imregionalmin and some small modifications to my code above

 pitmap=imregionalmin(map);

 pitmap([1,end],[1,end])=false;  %get rid of edge pits

 [pitrows,pitcols]=find(pitmap);   %pit coordinates

 output=[pitrows,pitcols];

[Image Analyst]

I agree with Matt about using imregionalmin(). By the way, you might find this interesting: http://www.mathworks.com/matlabcentral/fileexchange/35452-finddepressions

[Johny]

This looks promising It seems this mregionalmin function uses 8-connected neighborhoods for 2D images, but I have to take into account the points on the input matrix which are themselves the minimum, aka offset of (0,0) which means that it should scan for the local minimum within the cell itself as well. does Imregionalmin do this?