[
Top
] [
Up
] [Prev] [
Next
]
[
Up
] [
Tutorial
]
mmtutor V1.0 30abr2001
mm010why - Why Mathematical Morphology.
Synopsis
mm010why
Description
Linear Systems
Any
linear and translation invariant
image operator must satisfy the following:
They can be implemented by the convolution
where h is the
kernel
(also called
Point Spread Function - PSF
) of the convolution, which characterizes
any
linear translation invariant operator
Morphological Dilation and Erosion
Dilation
is any operator that commutes with union (max):
Erosion
is any operator that commutes with intersection (min):
Structural dilation
is any translation invariant dilation:
Structural erosion
is any translation invariant erosion:
The structural dilation and erosion are characterized by the
structuring element
b
Convolution x Dilation/Erosion
Note that the summation in convolution is replaced by the Max and Min in dilation and erosion, and the multiplication is replaced by addition and subtraction.
Compositions of Operators
Composing linearly two convolutions can be replaced by a single convolution
Composing dilation, erosion, union, intersection and negation:
can generate
any
digital image operator
More complicated morphological operators can be designed by the composition of the 5 primitive operators above.
[
Top
] [
Up
] [Prev] [
Next
]
[
Up
] [
Tutorial
]
Copyright (c) 1998-20001 by SDC Information SystemsCopyright (c) 1998-20001 by SDC Information Systems