{
  "id": "t01_static_pyramid",
  "order": 1,
  "tag": "artifact",
  "eyebrow": "01 \u00b7 Static scene",
  "title": "A stationary pyramid produces depth-like patterns",
  "question": "If nothing moves and nothing is buried, does the method still draw structures at depth?",
  "finding": "The reconstructed pipeline produces layered patterns although the simulated scene has no motion or subsurface. Mean best window-fit scores are 0.306 inside the ground footprint and 0.308 outside it.",
  "limitations": "One stylized scene and one random seed. The footprint is a ground-plane mask, not a scatterer attribution map: layover mixes locations. A high fit score is not a probability of a buried structure.",
  "method": "Scene: 348,713 point scatterers (course edges with aspect-dependent 'flash', diffuse face scatter, rocks) plus speckle clutter, synthesised into a 1280 \u00d7 1024 SLC. Pipeline: K = 50 reference/offset sub-aperture pairs (half-band, 88 Hz offset snapped to 4 FFT bins), 32 \u00d7 32 complex cross-correlation to 1/1000 px, global common-mode removal, \u03bb\u209b = 0.48 m. Grid: 54,000 pixels at 1 m, run time 4.4 min.",
  "figures": [
    {
      "file": "results/t01_static_pyramid/figs/t01_slc.png",
      "caption": "<b>The input.</b> A simulated X-band single-look complex image at 0.25 m pixels of a motionless stepped pyramid on flat desert clutter. Layover folds the near face and apex toward the radar (up). Nothing exists below the ground plane. The orange and green lines are the transects analysed below; the dashed box is the 2-D grid."
    },
    {
      "file": "results/t01_static_pyramid/figs/t01_transect_az.png",
      "caption": "<b>Azimuth transect.</b> Top: image amplitude along the line (the two bright stretches are the side faces). Middle and bottom: the two published depth-focusing variants applied to the same 50-step trajectories. Both fill the whole depth axis with blobs, streaks and layered bands. The desert between the faces is as 'structured' as the pyramid."
    },
    {
      "file": "results/t01_static_pyramid/figs/t01_transect_rg.png",
      "caption": "<b>Through the near face and desert.</b> Both focusing variants produce depth-like patterns along this range transect. The image coordinate mixes height and ground range through layover."
    },
    {
      "file": "results/t01_static_pyramid/figs/t01_maps.png",
      "caption": "<b>Where the best fits occur.</b> The input amplitude, maximum window-fit score, and winning model depth on a 1 m target grid. High scores also occur away from the pyramid. The winning depth is a processing output, not a surveyed location."
    },
    {
      "file": "results/t01_static_pyramid/figs/t01_trajectories.png",
      "caption": "<b>Three static trajectories.</b> Reference/offset patch shifts at a near-face, side-face, and desert location. None of the simulated scatterers moves. The curves show that changing the sub-aperture can change the measured patch displacement."
    }
  ],
  "metrics": {
    "grid_pixels": 54000,
    "runtime_s": 266.1179869174957,
    "mean_best_adjR2_pyramid": 0.30645185708999634,
    "mean_best_adjR2_desert": 0.3080204129219055,
    "frac_adjR2_gt_0p5_pyramid": 0.013728078446162549,
    "frac_adjR2_gt_0p5_desert": 0.01346188830271057,
    "trajectory_rms_px": 0.007339921294636644,
    "trajectory_rms_mm": 1.834980323659161,
    "scatterers": 348713
  },
  "date": "2026-09-19",
  "provenance": {
    "python": "3.11.3",
    "numpy": "2.4.6",
    "source_sha256": "d2e13a5df79b23e4d60e99c1833dd1c5be37fc2459bfd6239482cdedc0623fc1"
  }
}