Date: 07 Feb 91 15:41:45-PST
From: Vision-List moderator Phil Kahn <Vision-List-Request@ADS.COM>
Errors-to: Vision-List-Request@ADS.COM
Reply-to: Vision-List@ADS.COM
Subject: VISION-LIST digest 10.7
To: Vision-List@ADS.COM

VISION-LIST Digest    Thu Feb 07 15:41:46 PDT 91     Volume 10 : Issue 7

 - Send submissions to Vision-List@ADS.COM
 - Send requests for list membership to Vision-List-Request@ADS.COM
 - Access Vision List Archives via anonymous ftp to ADS.COM

Today's Topics:

 Helicopter Flight Image Database
 Image processing systems
 Bubble size and movement measurement
 Sensory Data Integration and Fusion
 Bayes Classification Method
 Re: video WORM
 Re: video WORM

----------------------------------------------------------------------

Date: Tue, 5 Feb 91 10:40:48 -0800
From: Banavar Sridhar <bsridhar@ames.arc.nasa.gov>
Subject: Helicopter Flight Image Database

HELICOPTER FLIGHT IMAGE DATABASE

NASA Ames Research Center has developed a Helicopter Flight Image
Database. The database includes 512*512 8-bit images, helicopter
motion information, and true range measurements for selected objects.
This database is available to the participants of the IEEE Workshop on
Visual Motion, IEEE Special Workshop on Passive Ranging, and other
vision researchers. The images can be obtained via anonymous FTP from
vision@windchime.arc.nasa.gov. If you have any questions, please contact:
Banavar Sridhar (phone:415-604-5450, email:bsridhar@ames.arc.nasa.gov) or 
Phillip Smith (phone: 415-604-5469, email: pnsmith@ames.arc.nasa.gov)

------------------------------

Date: 1 Feb 91 01:57:34 GMT
From: rwang@caip.rutgers.edu (Ruye Wang)
Subject: image processing systems
Organization: Rutgers Univ., New Brunswick, N.J.

I plan to buy a small but sophisticated image processing system for
research use. I want to have the basic functions such as input, output
display, and processing (with some special hardware to speed up the
processing), also there should be a software package available for
standard image processing tasks (such as fft, convolution, edge
detection, etc.)  Moreover, I would like to have a friendly
programming environment so that I can develop my own application
programs (fortran, c). And I hope to keep the price below, say, $80k.

Now, any one can give me some basic idea about what commercial
available system to buy? Some one suggested to consider Stardent 500
or 3000. Is this a good choice? If so, what image I/O devices should I
get? I would like to have hard copies of the image displayed on the
monitor, I also like to have an image input camera which can not only
scan photos, but also take picture of the 3D scene as the input. Are
Eikonix and Dunn good choices?

If some one can provide the name, model, manufacturer (with phone
number), and estimated price of the devices he suggests me to
consider, I would really appreciate a LOT!

I heard there is some publication called "Advanced Imaging" and the
December issue last year is a "Buyers Guide 1991". But I failed to
find their number to contact them. Any one knows where I can find this
issue?

If anybody can give me any help, please send me email to:
	rwang@hmcvax.claremont.edu

MANY THANKS in advance !!!!

------------------------------

Date: Mon, 4 Feb 91 11:05:14 GMT
From: Netherwood P J <cs_s424@ux.kingston.ac.uk>
Subject: Bubble size and movement measurement
Organisation: Kingston Polytechnic

I am trying to find vision software to measure the size of air bubbles
in water and trace their movement between frames. If you know of any
software to perform this or any similar software that can be applied to
this then email direct to me.

Paul Netherwood
Kingston Polytechnic

janet: p.j.netherwood@uk.ac.kingston  
internet: p.j.netherwood@kingston.ac.uk

------------------------------

Date: Mon, 4 Feb 91 00:28:37 GMT
From: raj@utkux1.utk.edu (Rajendra)
Subject: Sensory Data Integration and Fusion
Organization: University of Tennessee Computing Center, Knoxville

Hi, 
	I would appreciate if someone could mail me a list of references
or pointers to any on-going or previous work related to following 
topics:
1). Data integration and fusion using neural networks
2). Data integration and fusion using fuzzy logic

Thanks in advance,

Rajendra Patil

------------------------------

Date: 6 Feb 91 17:57:00 EDT
From: chettri@minnie.me.udel.edu
Subject: Bayes Classification Method

I am currently trying to learn about Bayes Classification for remotely
sensed data. I have 4 channels of data and ground truth for each pixel
in the image.

Is there a good source for learning about these techniques for
classification of images. I'm interested in practical implementation
and theoretical aspects with emphasis on practical implementation.

If there is sufficient interest mail directly to list.  Otherwise send
me mail personally.

Samir Chettri

------------------------------

Date: Fri, 1 Feb 91 9:30:33 MET DST
From: bellutta@irst.it (Paolo Bellutta)
Organization: I.R.S.T. 38050 POVO (TRENTO) ITALY
Subject: Re: video WORM

I made quite an extensive research in the past about that, and probably for
the same reason.

What I've found is:

Abekas A60: digital 4:2:2 component (like D1), records on RAM, with 
            Ethernet and TCP/IP can be seen as a server, 60sec. max for
            K$ 200

Sony LVR6000/LVS 6000: analog component, records on a WORM, rs232 for remote
            control, 24min per side, K$24

Panasonic LQ 4000: analog magneto-optical, I think analog 
            composite, about the same as a LD, should cost something like
            K$10.

I do have more info about the Abekas and the Sony and I've seen both of
them. I only read about the Panasonic and don't have any other info about
that. Prices are quoted for Italy, prices in the US may be as 3 times less.

Paolo Bellutta
I.R.S.T.		vox:	+39 461 814417
loc. Pante' di Povo	fax:	+39 461 810851
38050 POVO (TN)		e-mail:	bellutta@irst.uucp
ITALY				bellutta%irst@uunet.uu.net

------------------------------

From: crossley@SLCS.SLB.COM
Date: Fri, 1 Feb 91 09:04:49 CST
Subject: Re: video WORM

Response to request for info on WORM analog video disc equipment:

We are in process of evaluating these for our image capture system, and
so far have looked only at the Sony offering.  I believe Panasonic offer
competing boxes, but we haven't had a chance to look at those yet.

Info on the Sony system:

A basic acquisition/playback system consists of two boxes: one is
the controller and i/o electronics, the other is the mechanical drive
for the disc.  Each box costs about USD10K, but may be discounted,
depending on the vendor-customer relationship.  Each controller can
supervise up to 10 disc drive units.  Two types of disc drives are 
available: one record/playback, the other playback only, but the price 
is about the same.  Each box has physical dimensions about like other Sony
broadcast equipment offerings for Betacam or Umatic: rack mount width,
10-1/2 inches high, about 20 inches deep, one-man lift but hardly
portable in the accepted sense.

The medium is a 12-inch disc that looks for all the world like the big
brother of a Mac diskette.  Cost is about USD300 per disc, recording
time is a little over 20 minutes per side.

Video i/o standard interfaces are provided for composite, RGB with 
sync-on-green or separate sync, R-Y/Y/B-Y, hardware switch selectable.  
A PAL interface is available as an option.  Internally the format is
Y-R/Y/B-Y, as you'd probably expect.

Two separate control interfaces are provided, one RS232 for computer control,
one for an edit controller.  Thus you don't have to go patching around with
cables to have the system working with both a computer graphics system and 
a video editing system.

Audio is all digital internally (since we don't have a digital sound system,
I didn't look to see if they provided an external digital audio patch).

Sony provided a demo disc which of course showed pretty images, including
some very effective still-frames.  Any frame can be called up by frame number 
using front-panel pushbuttons, or under external control.  Disc edge-to-edge 
seek time is 1/2-second maximum.  The still-frames are rock-steady and clean,
except for interlace flicker when a frame contains elements that have 
large motion between fields.  This could be avoided in principle by selecting 
a single field and interpolating the other field, but the system doesn't 
provide this capability.

We measured the performance of the system using a wide variety of test
signals, and I won't list all the details, but the major points were:

(1)  The composite i/o system is about as good as can be done;

(2)  RGB i/o is good, but Y-R/Y/B-Y is best, as would be expected;

(3)  Signal-to-noise ratio on the particular unit we had was a little under
     spec (49db measured, spec is 51 I think);

(4)  Frequency response is drooping a bit at 2MHz, is between 3 and 4db down
     at 4MHz.  This makes it about comparable with Betacam, I think.

Hope this helps -- Tony.

------------------------------

End of VISION-LIST digest 10.7
************************
