cascafico wrote on 20.05.15 09:27:
I don't know if there's a general wiki for fresh mappers; if so, it should point out the use of shadows casted by artifacts (stright dark lines) and comparison with reference elements (ie: trees). As far as I've seen, in northern Nepal most of buildings are out of tree canopies and we should be able to detect their shadows. Besides, a collapsed building often shows the rectangular shape of roof, but no shadow.
In pre desaster Bing imagery of Nepal I have seen several of those "shape of roof, but no shadow" but usually interpret them in a different way: let's assume the house has been built against a hillside, then the midday sun would be shining at the front of the house and the back of the house is "leaning" against the hill. No shadow if the roof is rather flat. The same type of building built in a rather flat surrounding would logically produce a large shadow. Just remember the typical time of day when satellite imagery is shot (before noon) and if the imagery wasn't processed in a way so it produces "dull" light you can get a good idea of the topography.
Besides, I often wonder how many buildings in darker areas (such as close to trees or shadowed by topography) are overlooked by mappers who traced the same area earlier. When I look at the history of correctly identified objects nearby I mostly see edits done in iD. I suspect that iD's default setting for the brightness of background imagery is just too dark for most monitors. OTOH I do a lot of panoramic photography and when stitching images I sometimes need to manually spot control points in overlapping image pairs shot with differing yaw/pitch/roll values. So I'm sort of trained to spot tiny details.
Cheers, Carl _______________________________________________ HOT mailing list [email protected] https://lists.openstreetmap.org/listinfo/hot
