Fields t, range (2 cols), meters, raw slant range (no transform, no filtering). range > 0 across the whole dataset: 35 files, 578,645 rows, 0 zeros/negatives/NaN, global min 0.01 m, 1,821 rows < 0.05 m, 30,393 < 0.10 m. 1 cm quantization, ~30 Hz.
Two unflagged sentinels (no NaN/zero marker):
the low ~0.01 m floor marks a no-return read (e.g. transition_1 reads 27.13 / 0.01 / 27.07 at ~27 m altitude);
the high spike (mars_3 84.85 m, the global max and README “Max Measured Range”) is also spurious — mars_3’s realistic max is ~18.2 m, and the Lidar-Lite v3’s ~40 m spec cannot return 84 m.
Reject both high and low outliers.
The 4 h static set is 68% of all LRF rows; any naive global statistic is dominated by it.
On mars_4/mars_8, lrf_range.csv is a flight-window crop of the longer lrf_range_data.csv.
On the transition runs the fiducial boards lie on the floor plane (a floor-plane fit holds to ~34 mm RMS / 66 mm p95 over the board-supported hull), and the down-facing LRF is co-mounted and aligned with the nadir nav camera, so the LRF nadir range is essentially the nav-camera-to-board-plane depth. It pins the metric scale of a single-tag fiducial fix — whose along-optical-axis range is otherwise weak — and supplies an independent vertical witness on the planar floor segments. See Camera-LRF Calibration Tests.