Forums

spatial filtering

Started by Marinus Lamabelawa October 29, 2010
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
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