Ah, nice - and compared to the mapillary ones, has picked out school zone/hours/etc very well (at least in the example provided)
Will be great when https://github.com/openstreetmap/iD/issues/4851 gets into ID; possibly https://github.com/openstreetmap/iD/issues/5666 gets sorted out too. On Fri, Mar 1, 2019 at 4:14 AM Martijn van Exel <m...@rtijn.org> wrote: > Hi all, > > You will remember from my previous messages to this list that > OpenStreetCam held a competition to collect imagery in Australia (as well > as in New Zealand). The winners have their gift cards, but the community as > a whole gains as well. Our map team used the contributed imagery to > generate a large enough training set for our sign detection platform to > learn 45+ (and growing) Australia-specific sign types. You can use the > detected signs to map in JOSM (iD in the future) if you install the > OpenStreetCam plugin. > > New imagery captured will be analyzed for signs almost immediately. If you > go to your own trips you can click on the ‘detections’ tab to see results. > Example: https://openstreetcam.org/details/999263/314/detections > > Note that none of this data will be added to OSM automatically — it is > available for the community to use. > > Here’s a blog post with some more detail: > http://blog.improveosm.org/en/2019/02/openstreetcam-detect-signs-in-australia-newzealand/ > > > I hope that you will keep contributing to OSC to improve coverage and > generate more of this useful mapping data. > > Happy mapping, > > Martijn > _______________________________________________ > Talk-au mailing list > Talk-au@openstreetmap.org > https://lists.openstreetmap.org/listinfo/talk-au >
_______________________________________________ Talk-au mailing list Talk-au@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-au
