Hi, I'm not sure why you would want to use imagery that doesn't show a large percentage of buildings. The offset between the two images seems pretty small, and while undesirable, not that big of a problem.
Dan On Tue, Nov 8, 2016 at 2:12 PM, Chad Blevins <cblev...@usaid.gov> wrote: > I modified the instructions in the task 2102 > <http://tasks.hotosm.org/project/2102>to hopefully help clarify the > imagery problem. This it was originally mapped using Bing, I suggest we > validate using that source and use the service for cloudy or poor image > quality areas. I'm open to feedback of why this may or may not be a good > approach. > > The idea of using one image source for new mappers is great, but so many > parts of the world are cloudy. It would be nice if the TM allowed cells to > be marked as "cloudy", making it much easier to take Joost's approach. > > - Chad > > On Mon, Nov 7, 2016 at 12:15 PM, joost schouppe <joost.schou...@gmail.com> > wrote: > >> Ooh, sorry to jump to conclusions. It looked like a straightforward >> problem at first glance. >> >> We've had similar challenges, where cloudy spots are mapped with >> different imagery. Not only was there considerable change on the ground, >> but also it was impossible to align the pictures because of the hilly >> terrain. I'm afraid there's no easy solution. >> >> It does some likely that offering several imagery sources to newer >> mappers is challenging. It might be better to just map what's possible with >> one image, then create a new task to map what's left. Though that has it's >> own problems too. >> >> 2016-11-07 15:17 GMT+01:00 Chad Blevins <cblev...@usaid.gov>: >> >>> Hi everyone, >>> >>> Thanks for the note, I created task 2102 and also provided the >>> additional imagery service. Unfortunately certain areas of the custom >>> image service are offset from Bing and existing OSM data. For this >>> reason I purposely didn't load the custom service into the task, but >>> offered it in the instructions to supplement clouds, etc. These tasks >>> are designed to get new mappers involved with OSM. Offset imagery >>> significantly raises the bar of entry and can introduce data quality >>> issues. >>> >>> I've since noticed areas with significant change (buildings, trees, >>> etc.) between the two images. The service I provided is several years >>> newer than Bing. This highlights one major challenge with introducing >>> new imagery in OSM. It would be great if traced features were >>> automatically tagged with the date/year of the service used. >>> >>> Before creating a custom service I check Bing for date/year, cloud >>> cover, and quality. If Mapbox isn't a good supplement then I search >>> for other scenes from the DG archive, and do my best to find a recent >>> cloud free scene that aligns with Bing. I'm open to suggestions on >>> the best way to proceed. >>> >>> Thanks, >>> Chad >>> >>> >>> >>> > On Nov 7, 2016, at 7:21 AM, Blake Girardot HOT/OSM < >>> blake.girar...@hotosm.org> wrote: >>> > >>> > Hi Joost, Mike, Daniel, >>> > >>> > Joost is of course correct. In looking at the project in the TM, it is >>> > not clear which imagery should be preferred for mapping, it might be >>> > on purpose the other imagery is not in the imagery field. >>> > >>> > I sent an email via OSM to the project creator to see if we can >>> > clarify the instructions a bit and put the other imagery in the >>> > imagery field if it should be used. >>> > >>> > Thank you for pointing it out Daniel. >>> > >>> > Regards, >>> > Blake >>> > ---------------------------------------------------- >>> > Blake Girardot >>> > Humanitarian OpenStreetMap Team >>> > skype: jblakegirardot >>> > HOT Core Team Contact: i...@hotosm.org >>> > >>> > >>> > On Mon, Nov 7, 2016 at 11:32 AM, joost schouppe >>> > <joost.schou...@gmail.com> wrote: >>> >> That shouldn't be necessary though. >>> >> >>> >> When you set up a task in the tasking manager, there is a tab to set >>> the >>> >> background imagery. Doing this ensures the custom imagery is loaded as >>> >> default in ID. I don't know who managed the task, but they should be >>> >> informed that this is a really useful thing to do. >>> >> >>> >> >>> >> _______________________________________________ >>> >> HOT mailing list >>> >> HOT@openstreetmap.org >>> >> https://lists.openstreetmap.org/listinfo/hot >>> > >>> > _______________________________________________ >>> > HOT mailing list >>> > HOT@openstreetmap.org >>> > https://lists.openstreetmap.org/listinfo/hot >>> >> >> >> >> -- >> Joost @ >> Openstreetmap <http://www.openstreetmap.org/user/joost%20schouppe/> | >> Twitter <https://twitter.com/joostjakob> | LinkedIn >> <https://www.linkedin.com/pub/joost-schouppe/48/939/603> | Meetup >> <http://www.meetup.com/OpenStreetMap-Belgium/members/97979802/> >> > > > > -- > *Chad Blevins* > GeoCenter > U.S. Global Development Lab > USAID > 202-712-0464 > Internal GeoCenter Website (USAID Only) > <https://sites.google.com/a/usaid.gov/usaidgeocenter/home> > -- Dan
_______________________________________________ HOT mailing list HOT@openstreetmap.org https://lists.openstreetmap.org/listinfo/hot
