[OSM-talk] city accuracy

[Tom Higgy]

I was thinking on the bus today that it'd be useful if JOSM had some
options for dealing with DOP in GPS traces. Maybe showing points in a
different colour depending on their DOP, so the more accurate points,
which can be more accurately relied upon, are most obvious, fading as
the DOP increases.

At least eventually enough points of reliable accuracy could be recorded
to see clearly where stuff actually is (assuming the DOP is an accurate
figure when there's lot of reflected signals around).

Could this work?

Obviously it won't work for points returned by the API as this
information isn't stored AFAIK. So it'd have to be done from traces
directly to get anything useful.

I also thought I'll have to try and learn Java sometime...

Cheers,

Franc Carter wrote:
> 
> I have had a look at a few places in Sydney on Yahoo Imagery with josm 
> and parts of it
> are out by a fair bit - roads go at inconsistent angles which is the 
> giveaway. Obviously
> getting good traces in amongst all those tall buildings is pretty unlikely.
> 
> So, one idea that springs to mind is to start with a set of good traces 
> just outside the
> city, align the imagery to that and then work my way in to the city 
> aligning the new
> images to the ones already downloaded and so on.
> 
> However, this will only work if the distortion on the Yahoo Imagery 
> changes slowly,
> and I don't know if it does . . .
> 
> So, does this seem reasonable and does anybody have a better way of getting
> some known good points to work from ?
> 
> cheers
> 
> -- 
> Franc
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> talk mailing list
> talk at openstreetmap.org
> http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk