feat(street-level): place coverage on the bare earth on Google 3D

Draped ground polylines and clamped billboards land on the top of the
Google Photorealistic 3D mesh, so on a street under trees or an overpass
the coverage seemed to float at roof height.

At street zoom on Google 3D (camera below 1,400 m, back above 1,800 m)
the core switches to a `terrain` surface mode. groundCast.js samples the
existing terrain service on the ground-floor grid (0.001 deg), interpolates
bilinearly and adds 2 m; Mapillary lines become plain polylines at those
heights, and cones and the position marker stand on the street, so the
mesh hides what is behind buildings and trees. A tile draws draped first
and swaps to its cast lines once the heights arrive, keeping the old lines
until the new primitives are ready. Lookups are once per tile and aborted
when the tile is dropped; a geoid fallback counts as unknown and leaves
lines draped. Other map stacks and higher views are unchanged.

Providers get getSurface() and groundCaster in their context and an
optional setSurface(mode); getUIState() exposes `surface`. The QA gate
checks draped on Esri and terrain on Google 3D.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

CHANGELOG.md

@@ -8,6 +8,8 @@
 
 - Street Level: a provider-neutral street-level imagery layer modelled on the iD editor's photo overlay. One panel in the right rail carries a chip per imagery provider, shared 360°/flat and captured-since filters, coverage drawn on the globe from orbit down to a single street, image cones per sequence, and one embedded viewer whose adapter follows the provider of the open image; each active provider credits its imagery on the globe, and share links carry the provider switches and the filter. Mapillary is the first provider (free client token; CC BY-SA 4.0 imagery); Google Street View, KartaView and Panoramax are next. The panel is kept simple: the viewer opens at the top so the photo needs no scrolling, the header pill and the provider chips switch the layer, and SINCE is a slider. FOLLOW (the globe camera following the street-level view) is available on the Google 3D map only.
 
+- Street Level coverage no longer floats over trees and bridges on the Google 3D map. At street zoom, sequence lines, image cones and the position marker are placed on the bare earth from the terrain service (with a 2 m lift) instead of being draped on the top of the photoreal mesh, so buildings and tree canopies hide what is behind them; other maps and higher views keep draped lines. Opening a photo no longer flies the camera underground when the Google tiles under it have not loaded yet.
+
 - CCTV cameras whose bearing is a guess now say so. Packs mark bearings derived
   from a hash of the camera id as `headingConfidence: 'low'`, but nothing read the
   flag, so roughly 70% of a default catalog rendered like surveyed facings. The HUD