Hi Marinus,
It sounds like you're trying to get the hang of 2-D convolution in general.
The general premise is the same as with 1-D convolution: slide one signal
across the other, and sum at each step. The difference is that you now have two
dimensions to slide across. So, you can just think of computing it as sliding
your convolution mask (for example, your Roberts Cross matrix) across your image
data across each row, summing as you go.
A good way to experiment with this would be to try writing some simple Matlab
code that works the way you *think* it should work, then see if your results
match the output of conv2(), which is Matlab's standard 2-D convolution
function.
Hope this helps - feel free to come back with more questions.
--Tom
thx form moderator
>
>i'm a new member in this group & beginner in matlab
>may i question about problem in spatial filtering in image processing
>how we count convolution method using robert cross gradient operator
>example : robert operator
>
> -1
>
> 0
>
> �0
>
> 1
>
>image :
>
> A
>
> B
>
> C
>
> D
>
> E
>
> F
>
> G
>
> H
>
> I
>
> J
>
> K
>
> L
>
> M
>
> N
>
> O
>
> P
>
>new image in : A11 = A(-1)+B(0)+E(0)+F(1)
>
>or� A11 = A(-1)+B(1)+E(0)+F(0)
>
>or it's not exactly ? how with sobel operator
>mybe help me
>
>Best Rgrds
>
>Marinus