-------- Original Message --------
Subject: RE: idea about image acquisition
Date: Fri, 16 Sep 2011 22:33:52 -0400
From: Sarah Degroot <[email protected]>
To: [email protected] <[email protected]>
I use a flat bed scanner to capture images of leaves. In general it
works well; however I have had some trouble with artifacts from shadows,
particularly for small, lobed leaves when scanned on a white background.
If I scan against a black background the shadows don't show, but
(depending on the species) sometimes hairs on the leaves show and add a
different artifact. So far my simplest solution is to scan each leaf
twice, once with a white background (so the hairs don't show) and once
with a black background (so the shadows don't show). However, when
small, hairy leaves produce shadows, neither background works very well.
I'd love to hear ideas about how to get around this.
Thanks,
Sarah De Groot
sarah.degroot[AT]cgu.edu
________________________________________
From: morphmet [[email protected]]
Sent: Friday, September 16, 2011 11:22 AM
To: morphmet
Subject: Re: idea about image acquisition
-------- Original Message --------
Subject: Re: idea about image acquisition
Date: Fri, 16 Sep 2011 14:08:41 -0400
From: Matt Burton-Kelly <[email protected]>
To: [email protected]
I use this method for imaging freshwater mussels for outline analysis.
It works great if you have specimens that don't have large
pseudocardinal teeth, which lift the one edge of the valve and therefore
introduce distortion into the outline shape. That being said, I haven't
tested how much distortion there is and whether it has any effect on the
analyses.
Matt
On Sep 16, 2011, at 1:01 PM, morphmet wrote:
-------- Original Message --------
Subject: idea about image acquisition
Date: Thu, 15 Sep 2011 18:20:00 -0400
From: David Thulman <[email protected]>
To: [email protected]
I have used a decidedly low-tech method for image acquisition by using a
flat-bed scanner. I've used this for bifacial archaeological artifacts
(but they have a lenticular cross-section that looks like a fish). The
scanner I've used is an Epson 4180 Prefection (cost was less than $100),
but most scanners these days with a CCD should work fine. The scanner
has a depth of field of at least 1 cm, and probably more than 2 cm. The
error was less than .01 mm (tested with a digital caliper). The images
were scanned at 600 dpi, which eliminated the need for a scale for the
analysis I did. I used the images for a traditional morphometric
analysis (length, width, ratios, etc.), but recently uploaded them into
tpsDIG2 and have started reanalyzing with GM.
Laying a fish on the scanner platen may get messy, but its an
alternative that should produce precise images at low cost and fuss for
the right kind of specimens.
David Thulman
Department of Anthropology
George Washington University
