Floor Modes & Empty Scans
Every volume is the space between two surfaces: the top of the material, which the sensor reads directly, and the floor underneath it, which disappears the moment material is piled on top. The top surface is measured. The floor is modeled, and how well it is modeled is the biggest single factor in the accuracy of the number.
Why the floor matters
Section titled “Why the floor matters”A floor error does not average out and does not look wrong. Get the floor 10 cm too low and every cell in the zone reports 10 cm too much material — about 118 m³ of product that does not exist across a 1,180 m² footprint. It is present in every reading, stable scan to scan, and nothing in the data draws attention to it.
That is why the floor gets its own page while the top surface gets a paragraph. The setting lives on the zone form and in the 3D scene’s zone properties panel, under the heading Ground Detection.
The two floor options
Section titled “The two floor options”Dynamic
Section titled “Dynamic”The platform picks out the floor returns it can see — bare ground inside the zone outline, and the ring just outside it — and fits one continuous plane through them, extending it under the material. It re-decides at every measurement, so a sensor that gets nudged or re-aimed is absorbed without anyone touching a setting.
This is what a zone starts on, and it is a stable way to run a zone indefinitely. It needs some bare floor in view: a bay that fills wall to wall with no ground showing anywhere gives it nothing to work from.
Sloped floors are measured faithfully at any angle — a graded bay or a bunker that falls away toward a drain is measured as the surface it is.
Empty scan
Section titled “Empty scan”You clear the material out and scan the bay empty. Whatever is standing is stored cell by cell as the zone’s baseline: the real floor contour, and any permanent structures — bay and bunker walls, retaining edges, curbs, pillars, dividers. Every later measurement is read against that baseline.
Structures matter here. They return points exactly as product does, so without a baseline they are either counted as material or have to be masked out by hand. With one, they appear in both the baseline and the live scan and cancel out.
Because nothing has to be visible at measurement time, a bay that fills edge to edge measures exactly as well as an empty one.
Choosing between them
Section titled “Choosing between them”Dynamic is a solid, well-supported choice, not a fallback, and there is no accuracy penalty for leaving a zone on it where bare ground is visible around the pile.
An empty scan is more accurate where you can take one, and it is the clear choice in two cases: the bay has permanent structures inside the outline, or its floor is uneven enough — ruts, worn hollows, trenches, drainage falls — that one plane would misrepresent it. A plane has to average those. A baseline records the 6 cm hollow as 6 cm deep, in the cell where it actually is.
| Your situation | Use |
|---|---|
| Open pile with visible ground around it | Dynamic |
| Bay or bunker with walls, retaining edges, curbs, pillars, or dividers inside the footprint | Empty scan |
| Uneven, rutted, or trenched floor | Empty scan |
| Bay that fills wall to wall and is never emptied | Empty scan, taken during a turnaround |
Most sites end up with a mix, and that is the right outcome.
-
Start on dynamic. It needs no setup and it is a legitimate long-term answer for any zone with visible ground around the pile.
-
Draw the footprint slightly wider than the material reaches. That ring of bare ground is direct floor evidence, and it is the cheapest accuracy you will ever buy.
-
Rank your zones for a baseline. Anything with structures inside the outline goes first. Genuinely uneven floors go next.
-
Capture the empty state on the zones you can empty. Clear the material, leave every permanent structure standing, and run the capture from the zone.
-
Put re-capture on a schedule, tied to the events that invalidate a baseline — regrading, sensor work, a fixture moved, redrawing the outline.
How dynamic detection thinks
Section titled “How dynamic detection thinks”At every measurement the platform sorts the returns it has and picks out the ones that are floor — bare ground inside the zone outline, and the ring of ground just outside it. Those chosen points are then connected into one continuous surface, which carries on underneath the material where nothing could be seen at all. That extension is the method: the floor under the pile is not measured, it is inferred from the floor around it.
It also keeps the floor it settled on last time, and only moves off it when the new evidence is clearly stronger. That is what keeps a zone’s readings comparable day to day instead of drifting a few cubic meters in each direction as the returns vary.
Two checks sit on top of the election:
- A floor cannot float. A candidate that would sit above real measured points is rejected — it would imply negative material.
- The surrounding ground gets a vote. If the ground around the zone sits materially lower than the chosen floor, the two disagree, and the platform throws the scan out rather than fitting a plausible floor to a covered bay and reporting a confident, low, wrong number.
The trade-off
Section titled “The trade-off”| Dynamic | Empty scan | |
|---|---|---|
| Decides the floor | Every measurement | Once, at capture |
| Tolerates a moved or re-aimed sensor | Yes — it re-fits | No — re-capture needed |
| Tolerates a completely buried floor | No — it needs visible ground | Yes — nothing needs to be visible |
| Goes wrong when | It cannot see enough floor | The bay changes and the scan is not retaken |
The difference that matters is in the last row. If dynamic cannot see enough floor, the zone stops producing new numbers and you notice. A stale baseline keeps producing numbers that look completely normal, so put re-capture on a schedule rather than waiting for something to look wrong.
See it work
Section titled “See it work”The visualization below has two tracks, and the fastest way to see the difference is to move to the same step in each. Dynamic shows the returns the platform picks out as floor, and those points becoming one continuous surface that carries on under the material. Empty scan shows the same bay scanned empty, with the wall still standing, stored as the surface every later measurement is read against.