Key takeaways
- A siding capture has two owners of permission. The airspace is one conversation and the ground inside a rail corridor is a separate one with the railway, on their rules and their lead time.
- A private spur on land you control is a different problem from track the railway owns. Settle which side of the interchange point the work sits on before anything else.
- The switching schedule sets the date. A siding with cars spotted on it is a siding you cannot measure, and the clear window is frequently a few hours rather than a day.
- The loadout structure is the technical obstacle: it blocks the view of the pile beneath it, it shades and occludes the pad, and its steel degrades GNSS right where the crew wants to stand.
- What the capture settles well is stored volume by commodity, pad and pile geometry, the working footprint, drainage, and track centreline and siding length as built.
- What it does not settle is anything inside a bin or a car, rail condition below the ballast surface, track geometry to an engineering standard, or the legal boundary.
- Live track is treated as bystanders that cannot be cleared, the same as traffic on a highway, which shapes the flight lines rather than the schedule.
Rail-served sites are measured less often than they should be. A grain loadout, an aggregate siding, a frac sand transload, a fertilizer shed with a spur: all of them hold inventory worth knowing, all of them have a pad whose geometry decides throughput, and most of them have a track layout whose only record is a drawing from whenever the spur went in.
The reason they get skipped is that a siding looks complicated to fly, and it is, but not for the reason people assume. The airspace is usually the easy part. The hard parts are the ground, the schedule and the structure.
Two permissions, and the airspace is the simpler one
Any capture that touches a rail corridor needs the airspace question answered and the ground question answered, and they are answered by different parties. That split is covered in detail for sites beside a highway, road allowance or rail corridor and the short version is worth repeating here: an airspace authorisation does not give anybody the right to stand on railway property, and permission to be on railway property does not resolve the airspace.
The first question on a siding job is therefore a property question. Where is the interchange point, and which side of it does the work sit on.
- Entirely on your own land, on your own spur. The simplest case. You control access, launch and recovery, and the railway's interest is in what happens if a flight strays over their corridor or their equipment.
- Partly on railway property. Their rules apply to the ground and usually to overflight of their track and equipment, they will have their own process and their own protection requirements, and that process has a lead time measured in weeks rather than days.
- Adjacent, flying nothing over their property. Still worth notifying, because an aircraft working next to a live corridor is their safety concern whether or not it crosses the boundary.
We do not publish any railway's specific process, because they differ between operators and they change. The planning assumption that holds is that railway permission is the item with the longest lead time on the job, so it goes first and the flight date gets set after it comes back.
The switching schedule, not the forecast
On most captures the weather sets the date. On a siding, the cars do.
A string of cars spotted on the siding occludes the track, the pad beside it, part of the pile and the loadout approach. Nothing in processing recovers ground that was under a hopper car. So the capture goes in a window when the siding is clear, and the length of that window depends on the operation: between a pull and the next spot, in a lull between unit trains, or on a day the facility is not loading at all.
Practical consequences:
- The window is short. Plan the flight to fit the clear period with margin, which usually means a tighter block and a decision in advance about what gets dropped if the window closes early.
- Somebody at the facility owns the call. The person who knows when the siding is clear is a loadout operator or a logistics coordinator, not the person who signed the purchase order. Get them into the scheduling conversation directly.
- A partial capture is a legitimate plan. Pile and pad in one window, track and approach in another, provided both are tied to the same ground control so they combine. That is a case for the same permanent control and fixed processing frame that any repeat capture benefits from.
- Cars on an adjacent track are not a problem. They are a fixed object to fly around. Cars on the track being measured are.
The loadout structure is the real technical problem
A loadout is a tall, narrow, steel structure with a conveyor or a spout, usually sitting right where the measurement is wanted. It causes four separate problems:
- Occlusion of the pile. Material directly under a stacker, a conveyor or a spout is not visible from any camera angle. That portion of the volume has to be modelled rather than measured, and the modelling assumption has to be stated. On a pile actively being built this is the same cutoff problem as measuring a pile while loaders are on it.
- Shadow. A tall structure throws a hard shadow across the pad for most of the day, and a shadowed region with no texture is a region photogrammetry struggles to reconstruct. Time of day is a planning input, not an afterthought.
- GNSS degradation. Steel and a large vertical structure produce multipath and sky blockage in exactly the area the crew needs good position, both for the aircraft and for ground control. On a site where correction quality is marginal to begin with the piece on working without a correction stream or with blocked sky applies directly.
- Obstacle clearance. The structure, any overhead conveyor gallery, lighting masts and the service lines feeding the site all sit inside the flight volume. They are mapped before the flight, not discovered during it, and an energised line gets the treatment described for energised site hazards.
What a siding capture is actually good for
Four things, in roughly the order clients ask for them:
- Stored volume by commodity. The core reason most of these flights happen. Multiple piles of different material, each needing its own base surface and its own density factor before it becomes a tonnage, which is where a reconciliation against scale house records either closes or does not. Piles against a wall, under a conveyor or in a bunker carry more uncertainty than a free-standing cone, and that belongs in the report rather than in a footnote.
- Pad geometry and working footprint. How much pad is actually available, where the pile toes have crept, where the loader turning radius is constrained, how the stacking plan compares to the pad that was built. This is the analysis that turns into a capital decision about whether to extend the pad or change the handling sequence.
- Track layout as built. Siding length, centreline position in a stated coordinate system, the switch locations, car spot capacity against the length the facility believes it has, and clearances from the track to structures, stockpile toes and parked equipment. On older spurs the as-built differs from the drawing often enough that it is worth checking before anybody plans a change.
- Drainage and surface condition. Where water goes on the pad and around the ballast, surface rutting and contamination, ballast fouling visible at the surface, and vegetation encroaching on the corridor. A dated record of this is also what an insurer or an auditor asks for after the fact.
Where product is inside a bin, dome or shed rather than out on the pad, the measurement moves indoors and becomes a different job, covered in indoor bulk storage volume capture.
The limits
Worth saying out loud before anybody scopes a flight around them:
- Nothing inside a car or a bin. A loaded hopper car reads as a closed box.
- Not a track geometry inspection. Gauge, cross level, alignment and surface to an engineering standard are measured by track geometry equipment, not by a camera overhead. A capture gives position, layout and surface condition.
- Nothing below the ballast surface. Subgrade condition, tie condition under the ballast and drainage structures are not visible.
- Not a boundary. Where the railway's property line or the easement actually runs is determined by an Alberta Land Surveyor. A capture shows occupation and use, which is a different fact from title, the same caution that applies to any due diligence capture.
- Not a substitute for the railway's own clearance process. If a change is planned near their track, their engineering review is the thing that governs and our data is an input.
Scoping one of these properly
A siding job that goes smoothly has the following settled before anybody drives out:
- Which track is which, who owns it, and where the interchange point sits.
- Railway notification or permission started early, with their conditions written into the plan.
- A named contact at the facility who knows the switching schedule, and an agreed clear window with a fallback window.
- Commodity list with the base surface and density basis for each pile agreed in advance.
- Obstacle inventory: loadout, galleries, masts, overhead lines, with heights.
- Coordinate system, vertical datum and geoid model stated, and control that will still be there next year if this becomes an annual measurement.
- A statement in the report of what was occluded and how it was treated.
None of that is exotic. It is the ordinary difference between a capture that produces a number the facility can take to its auditor and one that produces a pretty picture of a train.
