[Osmf-talk] google Open Buildings usage request

Sat Jul 31 13:19:31 UTC 2021

Hi All,

Great thread here.

I did a quick check for Lagos Nigeria on GEE as well. It's a great one.
However,I realized the accuracy is dependent on the density of the building
footprints in the area of interest.

For dense areas,the predictions are not accurate but for sparse area,the
predictions are mostly accurate,yet with some overlapping and projection
errors.

I think correcting this kind of data would take a longer time that manually
drawing the boundaries on OSM.

Overall,it's a great start,I believe if Google could pull this,what to come
would be superb.

But please, don't merge with existing OSM data.

Africa is rapidly developing and most built-up area lack pattern which can
be very difficult for an AI model trained with perhaps non African building
data to make good predictions.

Thanks

On Sat, Jul 31, 2021, 8:37 AM Ralf Bernhardt <raabee at gmx.de> wrote:

> Hi,
>
> Google can replace AI-generated data when ever new images or more
> advanced algorithms become available. But that won't work for OSM. In
> the long run, you can't add more data than you can maintain. Large parts
> of Tanzania today are a bad example of the impact of too much AI. And
> there we have local mappers importing masses of outdated data. It seems
> that people have become too accustomed to this workflow and trust AI
> more than their own eyes.
>
> Ralf
>
>
> Am 31.07.21 um 08:25 schrieb dfjkman at gmail.com:
> > Hi all,
> >
> > I would just like to add that even if the AI was accurate there are a
> number
> > of problems as regards Africa that are missed by those sitting outside.
> > Parts of Africa are developing rapidly so the major cities and towns are
> > changing constantly expanding  as well as new buildings going up within
> the
> > cities as well. What you map today will be out of date within 6 months
> let
> > alone one year. The satellite imagery available is way out of date as it
> is
> > anyway. Lusaka has many new roads that do not appear in any imagery so
> using
> > Google Maps for navigation can lead to some interesting journeys
> especially
> > in the outskirts.
> >
> > Another point is touched upon by Craig, 'If the same thing had happened
> in
> > Germany there would have been a riot on this channel'. Africa is treated
> as
> > the training ground for new mappers, this is all well and good, new
> mappers
> > are welcome and needed, but the large areas they map and errors
> introduced
> > are many and varied and can take considerable time and effort to correct
> and
> > as a result nobody bothers. Nobody actually asks the Africans what they
> want
> > or whether they mind this mess being created in their backyard, judging
> by
> > the response to this thread they don't want it. No wonder many in Africa
> > feel they are just the guineapigs for the rest of the world.
> >
> > Dave
> >
> > -----Original Message-----
> > From:
> >
> b.kilhu+ytrwslnopdfujmrtjzkvsxweizfjncifmzdwwypiihjzdikulpnvql+- at gmail.com
> > <
> b.kilhu+ytrwslnopdfujmrtjzkvsxweizfjncifmzdwwypiihjzdikulpnvql+- at gmail.com
> >
> >
> > Sent: 31 July 2021 00:52
> > Cc: osmf-talk at openstreetmap.org
> > Subject: Re: [Osmf-talk] google Open Buildings usage request
> >
> > See The Machine Learning Enabler and MapWith AI (RapiD) road prediction
> from
> > Facebook:
> > https://www.hotosm.org/updates/the-machine-learning-enabler/
> > https://mapwith.ai/
> >
> >
> > On Sat, Jul 31, 2021 at 12:05 AM Craig Allan <allan at iafrica.com> wrote:
> >> Dear List,
> >> As an African mapper, working mainly in Chad, DRC and Kenya, I am
> >> perpetually scared that some well-meaning AI professional in a far-off
> >> wealthy country is going to proudly overwrite Africa with dirty data.
> >>
> >> In that regard I note that this latest 'Open Buildings' building
> >> recognition project seems to be good work, at the current level of AI
> >> image recognition. That is very cool and the field of work is
> >> fascinating.  But the quality is not yet good enough for OSM purposes.
> >> The authors of the paper are open about this and are working hard to
> >> improve their accuracy.
> >>
> >> One day, when AI image recognition consistently produces work that
> >> cannot be distinguished from the work of an experienced human then we
> >> should certainly look at using AI to classify features.  Until then I
> >> vote *NO!!!!* to any overwrite.
> >>
> >> An African example of the worst case already happening is Cameroon,
> >> where an AI enthusiast blanketed the country with bad vegetation
> >> coverage data that nobody will ever bother to fix, because it looks
> >> 'done', whereas it is not. It is dirty data, which far worse than
> >> blank.   If the same thing had happened in Germany there would have been
> >> a riot on this channel.
> >>
> >> I also find that if you politely write to mass editors to request
> >> correction, there is no recourse. The mass editor can't fix the errors
> >> because that involves hand editing millions of points and ways and
> >> they don't have the resources to do that. The alternatives are to dump
> >> the whole import, which is impossible, or fix it yourself.
> >>
> >> I would be quite happy with AI produced work being loaded up as a
> >> digitising aid on another layer, like OpenTopo, and I expect that that
> >> such limited use as reference-only material would probably be easier
> >> to licence.
> >>
> >> Craig Allan
> >> osm: cRaIgalLAn
> >>
> >>
> >> On 2021/07/30 09:31, Enock Seth Nyamador wrote:
> >>> ...
> >>> In Ghana alone, there is enough of very bad mappings from remote
> >>> mapping in recent years that we wouldn't want to see more added.
> >>
> >> _______________________________________________
> >> osmf-talk mailing list
> >> osmf-talk at openstreetmap.org
> >> https://lists.openstreetmap.org/listinfo/osmf-talk
> > _______________________________________________
> > osmf-talk mailing list
> > osmf-talk at openstreetmap.org
> > https://lists.openstreetmap.org/listinfo/osmf-talk
> >
> >
> > _______________________________________________
> > osmf-talk mailing list
> > osmf-talk at openstreetmap.org
> > https://lists.openstreetmap.org/listinfo/osmf-talk
>
> _______________________________________________
> osmf-talk mailing list
> osmf-talk at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/osmf-talk
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/osmf-talk/attachments/20210731/ec2696e1/attachment.htm>