Function buildTopDownFrame

buildTopDownFrame(
    points: readonly { lat: number; lng: number }[],
    gridPx?: number,
    marginM?: number,
): TopDownFrame | null
Build the synthetic top-down orthographic frame over a set of stations. Higher surfaces render NEARER (smaller depth), so the depth grid's front surface is the roofline. Ground resolution follows the bbox (a 12 km line across a 1024px grid ≈ 12 m/px — coarser than the DEM but right for a "what stands here" line).
Parameters
- points: readonly { lat: number; lng: number }[]
  The stations (lng/lat degrees; ≥1).
- gridPx: number = 1024
  Grid size in pixels for the LONG bbox axis (default 1024, min 64).
- marginM: number = 30
  Padding around the stations' bbox, metres (default 30). The grid's ground resolution is the padded span over gridPx, so a tight probe wants a small margin: two stations 6 m apart at the default become a 66 m box (~1 m/px at gridPx: 64 — coarser than a doorway), at marginM: 2 a 10 m one.
Returns TopDownFrame | null
The TopDownFrame, or null for an empty/degenerate input.