[Prev][Next][Index][Thread]
(meteorobs) Re: All-sky and Intensifier Systems
Malcolm,
> The surplus Gen II MCPs I've tried have noticeable distortion, making
> the real-time data reduction much harder, say if you want to use Hough
> transforms.
I am not quite sure what what your stating here. As far as "detecting"
meteor tracks using a Hough transform, there is a small performance loss
because a long track is curved on a fish eye lens, but because of the wide
field of view, most tracks are short and well represented by a straight
line. The long tracks are typically due to fireballs which are bright and
thus the Hough transform can still detect these despite the curvature since
it only needs a short straight bright segment to declare a detection. The
lens distortions are not that severe that a meteor track can be well
approximated by a series of short straight segments.
For "data reduction" however, the story is a little different. Real-time
radiant association would be a problem as the curved track would be
difficult to trace back to a known radiant for testing. However, the
distortions in the lens can be pre-calibrated and the image reconstructed to
take out the distortions resulting in a straight line track on a gnomic
projected field of view. This does require some computing horsepower but is
it necessary to data reduce the track in real-time when a post-collection
analysis in non-real-time would be acceptable. For the 1999 Air Force
Leonid flux measurements we had such a real-time requirement for detecting
and reporting radiant associations of Leonids from video. I am not aware of
such stringent requirements (real-time track reconstruction/association) for
more typical video meteor work that is being done.
Anyway with the processing capabilities of todays PCs, even track
reconstruction is becoming a possibility in real-time.
Pete
peter.s.gural@saic.com
To UNSUBSCRIBE from the 'meteorobs' email list, use the Web form at:
http://www.tiacdot net/users/lewkaren/meteorobs/subscribe.html