Measuring Material Inside a Building, Silo or Dome

An outdoor stockpile survey is a settled procedure. Fly a grid, hold position from satellites, tie it to ground control, build a surface, pick a base, get a volume. The method is described in drone stockpile measurement and it is reliable enough that most Alberta aggregate and municipal operations now buy it as routine.

Move the same pile indoors, into a salt shed, a grain terminal, a fertiliser building, a storage dome or a bunker, and most of that procedure stops working. There are no satellites. There is no ground control in the usual sense. The pile is bounded by a structure rather than by open ground, the floor under it cannot be seen, and the air may be full of the product. The question of how much is in there is still worth money, particularly at a fiscal year end, and the answer is still available by drone. It is just a different job.

What changes once you are under a roof

Four things, and each one has a consequence for the deliverable:

• No satellite positioning. An RTK or PPK workflow depends on a satellite signal and a correction stream, and neither reaches the inside of a steel or concrete building. The aircraft has to hold its position from what it can sense itself: visual odometry from downward and forward cameras, inertial sensors, and on some platforms a laser or ultrasonic height reference.
• No absolute frame. Without satellites there is no coordinate system in the output unless it is introduced deliberately. The capture is internally consistent but it is not in UTM, 3TM or any site grid, so the distinction between absolute and relative accuracy matters more here than almost anywhere else.
• Enclosed, cluttered airspace. Conveyors, loaders, roof structure, columns, chutes, dust collection ductwork and lighting all occupy the space the aircraft has to fly in, often with a few metres of clearance.
• A regulated space. Depending on the building and the product, the interior may be a confined space, a combustible dust environment, or both, which changes who may enter, what may be energised, and what has to happen before anyone or anything goes in.

What the method actually is

The aircraft is a collision-tolerant indoor platform rather than a survey drone. The defining features are a protective cage, so contact with a wall or a conveyor is an event the aircraft survives rather than a crash, and sensing designed for operation without satellites or external reference. UAV Imaging uses the Elios platform for this class of work, and the same capability is what makes it viable for confined space inspection.

A typical indoor volumetric capture runs as follows. The crew and the site agree an access point, a flight window and the isolation state of the building, and the operator plans a route that covers the full extent of the material plus the structure that bounds it. The pilot flies within visual line of sight where possible and on the live video feed where the geometry hides the aircraft, capturing overlapping imagery across the top and the visible flanks of the pile, including into the corners where the material meets the wall. The structure is captured as well, because the walls form part of the boundary of the volume. Scale is introduced by measuring known reference dimensions inside the building, typically structural spacings or placed targets surveyed by conventional means. Processing produces a surface of the material and a model of the enclosure, and the volume is the difference between the material surface and a base surface, exactly as outdoors.

The base surface is the part that gets interesting, and it is the usual source of disagreement. Outdoors you can often fly the ground before the material arrives. Indoors the floor is covered and frequently has never been surveyed. The honest options are the design floor elevation from the building drawings, a survey of the floor taken while the building was empty if one exists or can be scheduled, or an assumed flat floor at a stated elevation. Each produces a different number, and which one was used belongs in the report. The general problem is the same one described in base surface, and it is why two measurements of the same shed can differ without either being wrong, as explored in reconciling a drone volume against scale house tonnage.

What accuracy to expect, and how it is stated

This is where expectations need setting before the flight rather than after. An outdoor survey with ground control is normally quoted as a positional tolerance in centimetres. An indoor capture without satellite positioning cannot honestly be quoted that way, because there is no absolute frame to be accurate against. What can be quoted is a volumetric tolerance expressed as a percentage of the measured volume, derived from how well the surface and the scale reference were established on the day.

Several things on site drive where in that range a particular job lands, and all of them are worth discussing in advance:

• How much of the pile is actually visible. Material packed hard against a wall, tucked under a conveyor or sitting beneath roof structure is material the sensor cannot see, and the surface is interpolated across it.
• Whether the base surface is known or assumed. This is usually the largest single contributor to uncertainty, and an assumed floor can dominate everything else.
• Dust and lighting. A camera-based method needs texture it can match. Airborne dust and poor lighting both degrade that.
• Surface character. A loose, textured aggregate or salt surface matches well. A smooth, uniform, low-contrast surface such as fine powder or flour matches poorly.
• The scale reference. A surveyed reference dimension is better than a measured structural spacing, which is better than a drawing figure.

For a tonnage figure there is one more term, and it is not a surveying term at all. Mass equals volume multiplied by density, and the density of stored bulk material varies with moisture, compaction and depth. A volume measured to within a few percent, converted using a density figure assumed rather than sampled, produces a tonnage with a larger uncertainty than the volume had. If the number is going into a financial record, sample the density or state the figure used and where it came from.

Dust, atmosphere and the quiet window

The operational constraint that most often decides whether an indoor capture works is not the aircraft. It is the air.

Airborne dust defeats a camera-based capture in two ways: it obscures the surface being measured, and it fills the frames with moving particles that corrupt the image matching. A capture flown while a loader is working the pile or while product is being conveyed in is usually not interpretable. What the flight needs is a quiet window rather than a full shutdown: material movement stopped, dust allowed to settle, and no traffic in the building for the duration of the capture. On most sites that is achievable within a shift change or a break, and it is far cheaper than an outage.

Lighting is the second condition. Indoor light levels are frequently well below what a camera wants, and the aircraft's own lighting covers only part of the problem. Opening a door, bringing in temporary lighting or scheduling around a time when the building lighting is fully on all help, and the operator should raise it during planning rather than discover it on arrival.

The atmosphere itself is a safety question before it is a data question. Many bulk products form combustible dust, and a building handling them has a hazard assessment and an electrical area classification that govern what equipment may be brought in. That is a site decision, not an operator decision. The questions to settle in advance are whether the interior is classified, whether the product presents a combustible dust hazard, whether a hot work or equipment entry permit is required, and what the building's own procedure says. Confirm the answers with the site's safety authority, and expect them to differ between a salt shed and a grain terminal.

What the drone replaces, and what it does not

The value of a collision-tolerant aircraft indoors is that it removes a person from a bad place. Nobody climbs the pile, nobody works from a man basket over stored material, nobody walks a bin floor that could cave, and nobody is suspended inside a dome to get a tape on something. The exposure reduction is the main reason this work gets bought, and the same logic applies to inspection work that would otherwise need scaffolding.

What it does not do is make the space stop being a confined space. If entry into that building or vessel required isolation, lockout, atmospheric testing, a permit and an attendant before, those requirements are set by the space and the product rather than by who or what is going in. In practice the aircraft often reduces a full entry to a doorway launch, which is a real and substantial improvement, but the determination belongs to the site's own procedure. The parallel point about outage work is made in scheduling drone inspection inside a shutdown window: a drone replaces the entry, never the isolation.

When a drone is the wrong answer

Four cases where the honest answer is no, or not yet:

• A sealed silo or bin with no safe launch and recovery point. If there is no way in that does not itself require an entry, the drone has not solved anything. A fixed level sensor or a through-wall scanning method is the better answer.
• An active combustible dust hazard that the site cannot stand down. If the atmosphere cannot be made safe for equipment, the flight does not happen.
• A space that has not been cleared. No atmospheric test, no permit, no flight.
• A tolerance tighter than the method supports. Where a contract or a regulator demands a tighter figure than an indoor relative capture can deliver, the measurement needs a different method or a surveyed base and reference, and that should be agreed before the flight rather than discovered in the report.

One more worth naming: a building that is routinely emptied is often better measured empty. If the floor can be surveyed once while the building is bare, every subsequent capture inherits a known base surface, and the uncertainty in the annual number drops substantially for the cost of a single extra visit.

What to settle before booking

Six questions. What is the product and does it present a combustible dust hazard. Is the interior a confined space under your procedure, and what does entry require. Where can an operator launch and recover without entering. Can material movement be stopped for the duration of the capture, and for how long. Is there a surveyed floor elevation or a set of as-built drawings for the base surface. And is the output wanted as a volume, a tonnage, or both, because a tonnage needs a density figure and somebody has to own where that number came from.

Answer those six and an indoor volumetric capture is a routine half day. Leave them to the morning of the flight and it becomes a site visit that produces a conversation instead of a number.