MEDIA CYBERNETICS TECHNICAL SUPPORT
Frequently Asked Questions Technical Service Bulletins Mailing List Archive Contact Us!
Technical Support Home Product Updates Device Drivers Media Cybernetics Home

RE: ImagePro> how do dilate/erode work?

From: Brian Kuyatt (brian@mediacy.com)
Date: 05/25/99 

"Brian Kuyatt" <brian@mediacy.com> writes:

Hi Jill,
	There is not an easy answer to your question.  Whole books are written on
image processing.  I may suggest purchasing Dr. John Russ' book called "The
Image Processing Handbook" now in the 3rd Version (CRC press--or can be
ordered from our Website to amazon.com).  John is a consultant for us and
teaches our Image Analysis Tutorial.  you might want to look on our Website
for these classes.
	I teach the Intro to Image-Pro Plus classes and we cover some of the filter
methods in that class.  John's is a more advanced class.
	Basically, eroding a white (binary) object will eliminate pixels and
dilating a white (binary) object will add pixels to the edge.  To what
extent this happens depends on the kernel structure (square, cross, row
circle, column, etc.), the kernel size (2x2, 3x3, 5x5, 7x7, etc.) and on how
many passed the kernel makes over these objects (e.g., more passes for an
erosion filter causes more pixels to be lost).
	To make this more complex, grayscale image processing deals with pixels
differently than with binary images.
	I suggest taking one of the courses and/or purchasing an image processing
book to understand how kernels and image processing work.
	Hope this helps.
	==========================================
Brian L. Kuyatt
Quality Assurance Analyst/Technical Support Engineer
Media Cybernetics, L.P. "The Imaging Experts"
(301) 495-3305        FAX  (301) 495-5964
brian@mediacy.com  or  techsupport@mediacy.com
http://www.mediacy.com
==========================================



> -----Original Message-----
> From: owner-imagepro-users@atlantech.net
> [mailto:owner-imagepro-users@atlantech.net]On Behalf Of Jill Schmidt
> Sent: Tuesday, May 25, 1999 2:14 PM
> To: imagepro-users@mediacy.com
> Subject: ImagePro> how do dilate/erode work?
>
>
> Jill Schmidt <jills@ocean.washington.edu> writes:
>
>
> Hi,
> Can someone please tell me how these filters actually work?  For
> simplicity, assume a binary image, white grains on black background.
>
> For dilate, does the center value of the kernel change to white if any
> part of the kernel is on a white pixel?  Or does the whole kernel turn
> white if any part of the kernel is on a white pixel?  Is erode the
> reverse, i.e. if any part of the kernel is on a black pixel it turns the
> center pixel black?
>
> Thanks!!  Jill
>
>
> On Tue, 25 May 1999, Jill Schmidt wrote:
>
> > Jill Schmidt <jills@ocean.washington.edu> writes:
> >
> >
> > Hi Charles,
> >
> > No, they aren't holes, they are "canyons", i.e., the phenomenon only
> > occurs when the really bright area touches the edge of the particle.
> > Jill
> >
> >
> ------------------------------------------------------------------
> ------------
> > 	Jill L. Schmidt			Phone:	(206) 685-1926
> > 	University of Washington	FAX:	(206) 543-0275
> > 	Oceanography Box 357940		e-mail:	jills@ocean.washington.edu
> > 	Seattle, WA 98195-7940
> > 	 http://www.ocean.washington.edu/people/grads/jills/Schmidt.html
> >
> ------------------------------------------------------------------
> ------------
> >
> > On Tue, 25 May 1999, Batishko, Charles R wrote:
> >
> > > "Batishko, Charles R" <optiker.crb@pnl.gov> writes:
> > >
> > > Jill...
> > >
> > > It sounds like it may be due to the "fill holes" option
> (OPTIONS/FILL HOLES in
> > > the COUNT/SIZE window).  Are you familiar with that?  For
> example, picture a
> > > donut whose gray level is 150, and whose center is 255, on a
> background of 100.
> > > Let's assume you set the high and low threshold manually to
> something like 125
> > > to 175.  Let's also assume that you haven't checked the "fill
> holes" option.
> > > The outer perimeter of the donut will be outlined as
> expected, but the bright
> > > donut hole will also be outlined because it is not in the
> range you selected.
> > > If you select the "fill holes" option, then the donut hole
> will be treated a
> > > part of the donut, even though it is not in the range.
> > >
> > > Media Cy guys...please correct me if that's not right.
> > >
> > > Chuck
> > > Chuck Batishko
> > > PNNL - Sensors & Measurement Systems Group
> > > optiker.crb@pnl.gov
> > > http://www.pnl.gov/dcbpweb/batishko/batishko_cr.html
> > >
> > > -----Original Message-----
> > > From:	Jill Schmidt [SMTP:jills@ocean.washington.edu]
> > > Sent:	Monday, May 24, 1999 11:53 AM
> > > To:	imagepro-users@mediacy.com
> > > Subject:	ImagePro> count/measure
> > >
> > > Jill Schmidt <jills@ocean.washington.edu> writes:
> > >
> > >
> > > Hi,
> > >
> > > I am having a problem with counting bright particles in
> grey-scale images
> > > (using Count). In particular, if a particle has a section in
> it that is
> > > "too white" (around 255 on an 8 bit image), the outline of
> the particle
> > > traces around that white section, as if the white section
> were equal to
> > > the background (leaving big indentations in the outline).
> Anyone know what
> > > the problem is?  Everything works great for the particles that do not
> > > contain areas that are bright white.
> > >
> > > Thanks,
> > > Jill
> > >
> > >
> ------------------------------------------------------------------
> ------------
> > > 	Jill L. Schmidt			Phone:	(206) 685-1926
> > > 	University of Washington	FAX:	(206) 543-0275
> > > 	Oceanography Box 357940		e-mail:	jills@ocean.washington.edu
> > > 	Seattle, WA 98195-7940
> > > 	 http://www.ocean.washington.edu/people/grads/jills/Schmidt.html
> > >
> ------------------------------------------------------------------
> ------------
> > >
> > > ---------------------------------
> > > To unsubscribe from this list, send an email to
> majordomo@mediacy.com with
> > > 'unsubscribe imagepro-users' as the body of the message.
> > > ---------------------------------
> > > To unsubscribe from this list, send an email to
> majordomo@mediacy.com with
> > > 'unsubscribe imagepro-users' as the body of the message.
> > >
> >
> > ---------------------------------
> > To unsubscribe from this list, send an email to
> majordomo@mediacy.com with
> > 'unsubscribe imagepro-users' as the body of the message.
> >
>
> ---------------------------------
> To unsubscribe from this list, send an email to majordomo@mediacy.com with
> 'unsubscribe imagepro-users' as the body of the message.
>

---------------------------------
To unsubscribe from this list, send an email to majordomo@mediacy.com with
'unsubscribe imagepro-users' as the body of the message.

This archive was generated by hypermail 2b29 : 08/04/00 EDT