On 3 June 2012 20:13, Fae <faenwp@gmail.com> wrote:

On 3 June 2012 12:45, Kolossos <tim.alder@s2002.tu-chemnitz.de> wrote:
> The GPS-attitude from cameras has such an high noise level and is in some
> cases completely wrong so that it's IMO not useful. This is also the reason
> why the openstreetmap project collect these data.
>
> For me more important is the heading parameter, we have an usage of that on
> para's map and some cameras/smartphones has an compass, for all other
> cameras it would be nice if an user can add this parameter on the map with
> upload-wizzard.
>
> Greetings Kolossos

I agree, noise and also lag are a big issue (my camera often adds old
GPS data to new photos before the GPS has had time to update itself,
which can take several minutes from switching on). However we should
consider design and metadata coping for future cameras with better
accuracy. Perhaps on batch upload a user can have some options to
ignore parts of the EXIF data that may be inaccurate or might be felt
to be invading their privacy.

As an example of the heading data, my photo here
<http://commons.wikimedia.org/wiki/File:Shop_window_with_mini_couture_mannequin.jpg>
has compass data unused in the EXIF (from a Sony HXV5). I always find
the altitude data fails to be correctly interpreted (I have to delete
what I think is the direction in degrees from the altitude string). It
may be that the system might benefit from partially parsing with a bit
more intelligence based on knowing the camera model rather than
assuming standardization that may not be pragmatic. This may be
especially useful in adding some estimate of camera accuracy for EXIF
GPS information.

Fae

_______________________________________________
Commons-l mailing list
Commons-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/commons-l