30.01.2013 Views

Shih_Image_Processing_and_Mathematical_Morpholo.pdf

Shih_Image_Processing_and_Mathematical_Morpholo.pdf

Shih_Image_Processing_and_Mathematical_Morpholo.pdf

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

26 <strong>Image</strong> <strong>Processing</strong> <strong>and</strong> <strong>Mathematical</strong> <strong>Morpholo</strong>gy<br />

3.1 Grayscale Dilation <strong>and</strong> Erosion<br />

Grayscale images can be represented as binary images in a three- dimensional<br />

space, with the third dimension representing brightness. Grayscale images<br />

can then be viewed as three-dimensional surfaces, with the height at each<br />

point equal to the brightness value. We begin with defi nitions of the top<br />

surface of a set <strong>and</strong> the umbra of a surface. Given a set A in E N , the top surface<br />

of A is a function defi ned on the projection of A onto its fi rst (N – 1) coordinates,<br />

<strong>and</strong> the highest value of the N-tuple is the function value (or gray<br />

value). We assume that for every x � F, {y |(x, y) � A} is topologically closed,<br />

<strong>and</strong> also that the sets E <strong>and</strong> K are fi nite.<br />

Defi nition 3.1: Let A Õ E N <strong>and</strong> F = {x � E N-1 |for some y � E, (x, y) � A}. The<br />

top or top surface of A, denoted by T[A]: F Æ E, is defi ned as<br />

T[A](x) = max{y | (x, y) � A}. (3.1)<br />

Defi nition 3.2: Let F Õ E N-1 <strong>and</strong> f : F Æ E. The umbra of f, denoted by U[ f ],<br />

U[ f ] Õ F × E, is defi ned as<br />

U[ f ] = {(x, y) � F × E | y £ f(x)}. (3.2)<br />

Example 3.1: Let f = {2, 1, 0, 2}. Show the umbra U[ f ].<br />

0 1 2 3<br />

0 2 1 0 2<br />

f<br />

3 0 0 0 0<br />

2 1 0 0 1<br />

1 1 1 0 1<br />

0 1 1 1 1<br />

. 1 1 1 1<br />

. 1 1 1 1<br />

. 1 1 1 1<br />

–∞ 1<br />

0 1 2 3<br />

1 1 1<br />

U[ f ]<br />

The top surface of a set <strong>and</strong> the umbra of a function are essentially inverses<br />

of each other. In the above example, the top surface T[A] is equal to the input<br />

function f where A = U[ f ]. That is, the one-dimensional grayscale input<br />

function f maps the one-dimensional coordinates to its function values. The<br />

umbra extends the value “1” from the function value (as coordinates) of

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!