fix(infrastructure): test the ground, not the ellipsoid, for in-view sampling

An unsampled base sits at ellipsoid height 0. Over high terrain seen at an
angle that point projects below the screen while the ground above it is in
view, so the gate skipped ground the camera could see and never sampled it.
The gate now tests loaded globe terrain when there is some, and otherwise
the whole span of heights a sample can accept (0 to 9,000 m).

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

docs/CURRENT-STATE.md

@@ -1959,9 +1959,11 @@ during continuous motion. Disable/re-enable resets selection cleanly.
 Close-up stem sizing uses camera-to-feature distance without a 5 km minimum.
 Unchanged overlay entries do not request another frame; moved positions,
 changed membership and re-enabling still publish. Ground sampling waits for
-visible globe tiles to settle, samples only markers whose base is in the
-camera's view and above the horizon (an unseen base reads through the Earth
-and would retry every 2 s), and uses valid loaded terrain as a floor while
+visible globe tiles to settle, samples only markers whose ground is in the
+camera's view and above the horizon (unseen ground reads through the Earth
+and would retry every 2 s; until it is seen, a stem keeps its unsampled
+height, which can show for a stem whose base is just off screen), and uses
+valid loaded terrain as a floor while
 retaining rooftop and below-sea-level elevations. A hidden globe does not
 constrain photoreal geometry, and retry work remains bounded. Already sampled
 nearby markers follow higher settled terrain on existing bounded walks, without