<div dir="ltr"><div dir="ltr"><br></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Wed, 19 Aug 2020 at 14:52, Jass Kurn <<a href="mailto:jasskurn@gmail.com">jasskurn@gmail.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr"><div dir="ltr"><br></div><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Wed, 19 Aug 2020 at 13:58, Colin Smale <<a href="mailto:colin.smale@xs4all.nl" target="_blank">colin.smale@xs4all.nl</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div style="font-size:10pt;font-family:Verdana,Geneva,sans-serif">
<p><span style="font-family:monospace;font-size:10pt">Possibly even better that StreetView imagery is data that has been imported directly from OS, such as OS Boundary-Line for the admin boundaries. This is probably the closest we can get to cm-level accuracy - even though they don't give us the full resolution, the base points such as tripoints where boundaries meet are likely to be pretty damn accurate. I would recommend using these as a kind of calibration point to sanity-check imagery alignment and other data based on less accurate GPS positioning (e.g. from any consumer-grade GPS kit).</span></p></div></blockquote><div style="margin:0px;padding:0px 0px 20px;width:1204px;font-family:Roboto,RobotoDraft,Helvetica,Arial,sans-serif;font-size:medium"><div><div id="gmail-m_-3281766524178245122gmail-:1bv" style="font-size:0.875rem;direction:ltr;margin:8px 0px 0px;padding:0px"><div id="gmail-m_-3281766524178245122gmail-:1bu" style="overflow:hidden;font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:small;line-height:1.5;font-family:Arial,Helvetica,sans-serif"><div dir="ltr"><div class="gmail_quote"><div><div dir="ltr">I've been coincidently wrestling with this issue of offsets for the last two days. New Bing imagery is resulting in very detailed and useful mapping (e.g. solar panels) but imagery is nearly always out by a problematic amount. I also feel the best source for offsets is not Streetview, but the vector data from OS OpenData. I've done some experimenting over the last two days, and my favourite source for alignment is "OS OpenData Local - Vector". Within those downloaded files is a data set called "FunctionalSite" which is primarily the boundaries of educational sites. They're excellent for alignment because the file is not too big, school sites are common, and the boundaries are commonly thin fences which are easy to align with Bing imagery, bringing errors to a trivial amount. </div><div dir="ltr"><br></div><div>I think the secondary useful task is to make offsets available using the "The Imagery Offset Database". It's been around for a long time, but is now way more useful due to the issues being discussed. <br><br></div></div></div></div></div></div></div></div></div></div></blockquote><div>
I think in an ideal world it would be good for the offset to be based on a record of features detected within the imagery itself. If we were able to generate these repeatably, then it may be possible to use the same control points across different imagery and have a chance of self-correcting after updates. In the real world this is probably too computationally expensive for OSM editing software, even for the relatively "simple" features used in panorama stitching software. I also have no idea if the license on the imagery we have access to would allow this. </div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr"><div class="gmail_quote"><div style="margin:0px;padding:0px 0px 20px;width:1204px;font-family:Roboto,RobotoDraft,Helvetica,Arial,sans-serif;font-size:medium"><div><div id="gmail-m_-3281766524178245122gmail-:1bv" style="font-size:0.875rem;direction:ltr;margin:8px 0px 0px;padding:0px"><div id="gmail-m_-3281766524178245122gmail-:1bu" style="overflow:hidden;font-variant-numeric:normal;font-variant-east-asian:normal;font-stretch:normal;font-size:small;line-height:1.5;font-family:Arial,Helvetica,sans-serif"><div dir="ltr"><div class="gmail_quote"><div><div>With imagery likely to become more detailed, with more high quality tracing of the high quality but misaligned imagery, I think we'll need a more formal approach to "Align imagery before tracing guidelines"</div></div><font color="#888888"><div><br></div><div>Jass</div></font></div></div><div></div><div></div></div><div></div></div><div style="border-bottom-left-radius:1px;border-bottom-right-radius:1px;padding:0px;width:auto;background:rgb(242,242,242);margin:0px"></div></div></div><br><div> </div></div></div>
_______________________________________________<br>
Talk-GB mailing list<br>
<a href="mailto:Talk-GB@openstreetmap.org" target="_blank">Talk-GB@openstreetmap.org</a><br>
<a href="https://lists.openstreetmap.org/listinfo/talk-gb" rel="noreferrer" target="_blank">https://lists.openstreetmap.org/listinfo/talk-gb</a><br>
</blockquote></div></div>