From: John Halleck (John.Halleck_at_utah.edu)
Date: Fri Feb 09 2001 - 21:54:00 CET
Received: (from mdom_at_localhost) by karto.ethz.ch (8.9.3/8.9.3/SuSE Linux 8.9.3-0.1) id VAA11072 for cavexml-outgoing; Fri, 9 Feb 2001 21:53:21 +0100 Received: from cor.oz.cc.utah.edu (nahaj_at_cor.oz.cc.utah.edu [155.99.2.2]) by karto.ethz.ch (8.9.3/8.9.3/SuSE Linux 8.9.3-0.1) with ESMTP id VAA11068 for <cavexml_at_cartography.ch>; Fri, 9 Feb 2001 21:53:19 +0100 Received: from localhost (nahaj_at_localhost) by cor.oz.cc.utah.edu (8.9.2/8.9.2) with ESMTP id NAA29520 for <cavexml_at_cartography.ch>; Fri, 9 Feb 2001 13:54:01 -0700 (MST) Date: Fri, 9 Feb 2001 13:54:00 -0700 (MST) From: John Halleck <John.Halleck_at_utah.edu> To: cavexml_at_cartography.ch Subject: Re: Other areas that haven't been discussed. In-Reply-To: <3A844064.1010606_at_aic.nrl.navy.mil> Message-ID: <Pine.GSO.4.05.10102091347540.11013-100000@cor.oz.cc.utah.edu> Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-cavexml_at_karto.baug.ethz.ch Precedence: bulk Reply-To: cavexml_at_cartography.ch
On Fri, 9 Feb 2001, Ralph Hartley wrote:
> Date: Fri, 09 Feb 2001 14:09:24 -0500
> From: Ralph Hartley <hartley_at_aic.nrl.navy.mil>
> Reply-To: cavexml_at_cartography.ch
> To: cavexml_at_cartography.ch
> Subject: Re: Other areas that haven't been discussed.
>
> One more thing.
>
> It would be good to have an optional "reliability" attribute for any
> data. The relation between reliability and resolution would be similar
> to that between accuracy and precision.
>
> Possible values for reliability could be "godgiven", "good", "normal",
> "bad", and "error", with only the first and last having real meaning.
> If a measurement is marked reliability="godgiven" any program should
> treat it as fixed, and apply no corrections, adjustments, etc. to it.
> Loop closures should consider its variance to be zero. It is the users
If you have more than one point with variance zero, then the least
squares adjustment has a singular matrix, and the adjustment is
therefore invalid.
I don't think you understand the ramifications of zero variance.
> responsibility to make sure that there are never errors in data marked
> this way, or at least that he wants it to be treated as if it is
> absolutely correct. You would use this, for example, to set the
> coordinates of the datum to (0,0,0), which by definition is exact. It
> should be used very sparingly.
If you have a single control point, then the weighting of it is
going to cancel out, and a numericaly stable adjustement is not
going to move it anyway.
May I suggest that the details of LS adjustments is probably best
discussed off line, or by reference to any good book on the topic.
I recommend either Mikhail and Gracie's "Analysis and adjustment
of survey measurements" or Wolf and Ghilani's "Adjustment Computations".
> If a measurement is marked reliability="error" then all programs MUST
> ignore the measurement, giving the same results regardless of the
> measurement's value. A program may include information about the
> measurement in summaries, historical information etc. but the actual
> value of the measurement must not affect anything else. This is intended
> to be applied to measurements that are known to be incorrect.
> Setting reliability="error" is much better, from a data safety point of
> view, than deleting bad measurements from the file, because it is not so
> irreversible.
Survey books cover the correct ways of dealing with the problem.
Let's not put a hack in to do things some other way.
> Programs may, but need not, give more weight to data marked good than to
> normal, or to data marked normal than to bad. How this is done, if at
> all, is totally at the program's discretion.
Then I miss the point of having it at all.
> Ralph Hartley
>
This archive was generated by hypermail 2b30 : Thu Mar 01 2001 - 18:00:00 CET