[Talk-GB] MapRoulette challenge to connect isolated highways

[Rob Nickerson]

Sounds good to me - should have no false positives then unlike machine
learning / image recognition approaches. I look forward to seeing the
MapRoulette task set up. Any idea when this might be? Entirely selfish
question as I want to make sure I am free when it happens otherwise I might
not get a chance to participate given how quickly the last task was
completed!! :-)

Thanks,
*Rob*


On Tue, 17 Jan 2023 at 18:38, Marjan Van de Kauter <
Marjan.VanDeKauter at tomtom.com> wrote:

> Hi Rob,
>
> We identified the missing highway leads by spatially comparing TomTom’s
> in-house road centerlines with the OSM road network, and validating the
> results using the available imagery sources. We did not make use of machine
> learning or image recognition.
>
> Hope this helps!
>
> Marjan
>
>
>
> *From:* Rob Nickerson <rob.j.nickerson at gmail.com>
> *Sent:* Saturday 14 January 2023 0:24
> *To:* Talk-GB <talk-gb at openstreetmap.org>
> *Subject:* Re: [Talk-GB] MapRoulette challenge to connect isolated
> highways
>
>
>
> >Would you also be interested in some leads for highways that are missing
> from the map?
>
>
>
> Sure. No harm in having a look at these. Based on experience there are a
> lot of false negatives in this sort of exercise but we can take a look :-)
>
>
>
> It would be useful if you could describe where the data has come from. For
> example is it from machine learning / image recognition or from GPS trace
> data?
>
>
>
> Thanks,
>
> *Rob*
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-gb/attachments/20230117/d87c651f/attachment.htm>