Notes of potential interest to instructors interested in image capture
or using the WWW for teaching
The production of these on-line course materials was partially funded
by The University of Wisconsin-Madison. The Division of
Instructional Technology asked us to record our experiences so that
others might benefit from our mistakes and discoveries.
(1) How to capture high-resolution color images from objects ranging
from a few microns to ~15 cm in diameter:
(i) Single chip color CCD cameras do not provide acceptable
resolution.
(ii) Although three chip color CCD cameras would probably
provide acceptable resolution, they cost approximately an order of
magnitude more than monochrome models.
- Consequently, we chose to use a single chip, monochrome CCD
camera. The model we selected was manufactured by SONY (XC-75:
cost including cables and power supply ~ US$1,000). The CCD is ~
7 cm in length).
- Full color images are collected using KODAK color
separation filters.
The appropriate filters are No's 29, 47, 61 (Cat. 149 5621, 149 5787,
and 149 5894). Cost for the three filters ~ US$60-.
- Image capture requires a video capture board. These boards range
in price and capabilities considerably. Because our intention is to
capture images of still objects, and because low light intensity is not
a problem, we selected a capture board manufactured by SCION
Corporation (LG-3). The board comes with a cable and a version of
NIH Image. Approximate cost US$1,000.
- To preserve the developers sanity, it is important that the computer used in
this process be as fast as possible. We are using a Macintosh PowerPC 8100 with 48
mB of RAM. Allocate as much memory as possible to Photoshop and NIH Image.
- The CCD must be connected to a lens. As we are using a lens with a
NIKON-type
mount, a connector piece is required. This is termed a
'C-mount-to-NIKON' connector. Cost ~ US$30.
- The CCD and lens are mounted on a copy stand. The copy stand camera
mount screws into the back of the CCD. The stand can be purchased
inexpensively from any number of companies. The stand we have chosen has a
small light table built into the base for imaging slides and other transparent
materials.
- The CCD can also be mounted on a microscope (binocular or
petrographic). A C-mount-to-NIKON converter may again be needed.
- Lens. Many choices are possible, depending upon the size range of
objects to be imaged. Our normal range is ~6 mm- 6 cm. We use a
macro SIGMA lens, 50 mm, F 2.8. This has proven to be an excellent
choice. A wide angle lens provides additional range for larger
samples.
A diagram of our set up will eventually be available.
The procedure we use is as follows:
- Using NIH Image (or similar software), start capturing. This provides
a real-time image on the computer screen, allowing focus and positioning of the
sample. Lighting of the sample is currently achieved using a pair of fiber optic
lights. This system probably could be improved significantly -
suggestions are welcome. All image capture is performed with the room lights off
to avoid spurious reflections.
- The color images are gathered in 3 passes. Hold the red filter (mounted in some
type of holder; we use a cardboard holder that resembles a slide frame) under the
camera lens. Select 'Stop Capturing'. This immediately results in a captured image.
For simplicity (see below) this image is saved as a TIFF? file named 'r' (without the
quotation marks). Repeat for green (save as 'g') and blue (save as 'b').
- Go to Photoshop (we used version 3.01, the commands should be the same for
version 2.5). Open all three filtered images.
Under 'Window' select 'Palettes', 'Show Channels', `Merge Channels'.
Then select 'RGB' and OK. (If you named your images 'r', 'g', 'b', as above,
the correct filtered images should be assigned to the correct
channels. The resulting merged image may need to be corrected for
brightness/contrast. It is unlikely, in our experience, that the color
will have to be modified.
The sequence of opening the r, g, b images followed by the series of
menu options becomes very tedious. It is possible to automate these
steps using a macro approach. We did this using a program called
Photomatic 2.0 (recently
released as freeware). Once this 'macro' is defined,
the reconstructed color image can be created by selecting the
appropriate routine from under the `Record' menu, option `Play'.
- Save as a JPEG document. We use the high quality option.
For line drawings that have been modified in Photoshop, save as a GIF file.
(2) Using the captured images:
Insert call outs for figures in text.
The text can be written in Microsoft word, saved as text only with
line breaks, copied to the appropriate directory with the images
(public_html on our UNIX server), and translated to html markup
language using the command:
txt2html <"txtfile.txt"> txtfile.html
The file will have to be tidied up, to optimize formatting.
In some cases, in-line images are required. These should be
GIF files. It is essential that the file size is small, as they load over
the web slowly. The in-line GIF files in the menu page of our
document are each about 20K. These were created from ~100 K jpeg files
in the following manner:
The image was opened using the `Graphic Converter' program. Under
Colors, select 256 colors. This will change the appearance of the
image slightly. Under `Resolution' change the dpi to a smaller value.
We use 20 dpi. This makes the image look very pixilated, but once it
is shrunk down (under `Size', we chose 24%), it looks fine. Then save
the file in a GIF format. Select the `Interlaced' option which allows the user to
access the scroll bars and move to and select items before the full resolution of each
of the graphics is obtained.
(3) Web site searches were done using standard web browsing
programs.
For the most part, we used `Lycos'.
(4) How we will use the www document in classroom teaching.
A format for the lecture notes has been selected to allow use of the
outline for each lecture in the classroom. To ensure that the more
detailed information and explanations are available (without cluttering the
outlines) we have moved wordy sections to
another level linked by active text.
(5) On-line test taking.
We do not anticipate using on-line examinations immediately. However, we have
provided trial examinations on-line. These incorporate a feedback
mechanism, so that the students learn where the flaws in their
understanding lie. The format of the two on-line tests is such that the
tests can be taken without immediate feedback - the student receives a score
indicating how they did ( x questions right out of y taken; information
regarding which questions were answered incorrectly is also available) - or with direct
feedback, the incorrect answer results in an immediate link to the part of
the notes that explains the correct answer.
The on-line tests are password protected. For more information, contact us
using the form on the introduction page.
We are currently investigating the possibility of offering the course by correspondence at some time in the future.
Feedback on how much demand for such an option there might be
would be appreciated.
Final comment. We gladly share our experiences and would
appreciate any comments, feedback, and suggestions.
Thanking you,
- Jill Banfield (jill@geology.wisc.edu)
- Phil Brown (pbrown@geology.wisc.edu)
- Heather Henkel