View (tag)
From SnapMap
view is a powerful tag that sets up several things about a photograph's location and geometry in a single declaration. They should always be used when you want to locate a photo better than with just latitude, longitude, heading and altitude.
Contents |
Slingshot views
Main article: slingshot view
This type of view consists of three lat-longs, called position, left and right. As well as camera location, which is expressed explicitly by the position lat-long, we can easily derive the camera heading and the field of view. It is ideal for photos taken near to the ground. (Aerial photos are better tagged with PhotoOverlay or point-match views.)
Catapult views are highly recommended for photos taken near ground level.
KML PhotoOverlays
Main article: PhotoOverlay view
Using Google Earth, you can load a photograph, then set up a virtual camera to be capturing the same view virtually as the original camera that took the photograph. It's not trivial, but nor is it very hard to get close.
In Google Earth, the feature is called a PhotoOverlay.
It seems the only aspect of photographic geometry that is not modelled is lens distortion. With modern cameras, lenses and software, one can correct for those things of course.
PhotoOverlay views are recommended for oblique aerial photos.
Views from Google Street View
Main article: Views from Google Street View
If you can match a photo’s view well using Google Street View, then the URL that Google gives you for the view is an excellent basis for setting up a view for the photo in SnapMap.
To add the view, open Google Maps Street View in a new window and set it up as accurately as you can, including choosing a zoom setting. Click 'Link' (top right) within the Street View UI. Copy this to the clipboard. Create a new tag, with ‘view’ as the key and this long Street View URL as the ‘value’. Upon adding the tag, you should see the latitude, longitude and heading immediately update. Elevation is also set.
Point-match
Main article: point-match view
A point-match view is a view tag that matches up corresponding points between the world and the image. At least 5 corresponding points are needed for a photograph where all points lie on a horizontal plane.
Point-match views would be highly recommended for oblique aerial photos, if only we could get the code to work.
