[Talk-ca] My Android tablet

James Ewen ve6srv at gmail.com
Sat Apr 6 17:04:21 UTC 2013


On Fri, Apr 5, 2013 at 1:01 PM, John Kerr <johneddie.kerr at gmail.com> wrote:

> Today I tried some mapping with my android tablet. I was trying to trace the
> outline of a building by walking around it with OSMtracker. I have not
> played with my files yet because I did notice one thing while doing this and
> that is I was only accurate to 7 metres. That is not very accurate. So just
> a few minutes ago I downloaded Vespucci for Android and I hope to give it a
> try at the same task.

John, how will Vespucci increase the accuracy of the your Android device?

The accuracy is based on the quality of signals received from the GPS
satellites. Working close to the walls of a building will degrade the
signals available, and cause possibly even more inaccuracy. Changing
the software that is reading the position reported won't make the
reported position more accurate.

One thing you can do to decrease the inaccuracy is to get a clear
unobstructed view of the sky. Moving away from the buildings will do
that for you.

Have a look at these images I just created...

http://wiki.openstreetmap.org/wiki/File:Extension.png

A visual depiction of how using a sight line along a building face can
help to reduce the position ambiguity of a GPS unit. Standing close to
a building blocks the GPS signal reception, degrading position
accuracy even more than normal. Moving out away from the building gets
a clear sky view, and that combined with the extended length of the
wall can be used to reduce the error of the actual measured item. Draw
lines between the extended points, and use them as guides to draw the
actual building outline.

http://wiki.openstreetmap.org/wiki/File:Outline1.png

Image showing how using GPS positions captured by extending the
sightlines along buildings can then be used to draw the actual
building outline. More GPS locations are required to be captured, but
GPS position ambiguity is reduced due to being clear of the building
obstruction, and also due to the reduction in position error due to
mathematical angular error reduction. The further from the building
the GPS locations are recorded, less error is introduced into the
actual building corner position.

I added a bit to my Wiki page...

http://wiki.openstreetmap.org/wiki/User:Ve6srv#Extending_Sightlines_to_Reduce_GPS_Error


This technique will not remove all error, but can reduce the angular
errors when trying to lay out a building outline. You can still have
displacement errors (all measured points can be horizontally displace,
ie. the whole building might be 2 metres north of where it is supposed
to be), but your walls should be closer to true than what you can
achieve by walking the perimeter of the building.

Complex building shapes are harder to plot using this technique, but
you can extrapolate some of the wall locations using this process.

-- 
James
VE6SRV



More information about the Talk-ca mailing list