[Prev][Next][Index][Thread]

(meteorobs) Re: Automated video detection of meteors



Hi folks,

Pete wrote: 

> The other system developed by myself called MeteorScan currently scans
> videotapes or live video data for meteor events and records the image
> sequence to disk.  In addition if the field of view is calibrated
> beforehand, the software can associate the meteors to a known shower from
> the IMO working list.  The software only makes approximate estimates for
> speed, position, direction as it is trying to operate in real-time on a PC. 

MetRec has essentially the same features as Pete's MeteorScan. It inspects
video data in real time at 25/30 frames/s (PAL/NTSC) and saves the
appearance times of meteors. Optionally you can save images of the meteors
and compute parameters like equatorial coordinates, velocity, brightness,
shower membership etc. In MetRec these parameters are not estimates (maybe
the brightness, but certainly not the coordinates, velocity and shower
membership). Usually meteors are detected in several frames. The resulting
mean meteor path is accurate to a few minutes of arc. That is, you do not
reach photographic precision (which is impossible due to the limited
spatial resolution of the intensifier and video camera), but you are
certainly more accurate than visual observers. This holds especially for
meteor-storm like conditions, since the software has no problems to
detect - say - 5 meteors appearing in parallel at one instant.

MetRec has been using for an automatic video observatory since early this
year, and the number of active users is steadily growing. Like MeteorScan,
MetRec has been improving for a long time. The meteor detection rate is
close to what human observers detect when watching the video tape, and in
one test it even outperformed the manual tape inspection.

Even though MeteorScan and MetRec were developed more or less parallel and
independently, they share a lot of properties and functionality. As for
the differences: Both programs differ in the detection algorithm. Pete
and me have not yet tested, which algorithm is more efficient, but this is
still on our to-do-list. So far MeteorScan is only available on Mac for
different frame grabbers (?), but Pete is working on a PC-based version
(right?). MetRec runs on PCs under DOS and requires a Matrox PCI frame
grabber (preferably a Meteor II board).

Cheers, Sirko

--
************************************************************************** 
*  Dipl.-Inform. Sirko Molau                  *                          *
*  RWTH Aachen, Lehrstuhl fuer Informatik VI  *              __          *
*  Ahornstr. 55, D-52056 Aachen, Germany      *       " 2B v 2B "        *
*                                             *                          * 
*  phone: +49-241-8021615                     *             Shakespeare  *
*  fax  : +49-241-8888219                     *                          *
*  email: molau@informatik.rwth-aachendot de     *                          *
**************************************************************************
*  www  : http://www.informatik.rwth-aachendot de/I6/Colleagues/molau       *
**************************************************************************

To UNSUBSCRIBE from the 'meteorobs' email list, use the Web form at:
http://www.tiacdot net/users/lewkaren/meteorobs/subscribe.html

References: