Stockpiles, pits and slopes, measured from the air

Stockpiles, pits and slopes, measured from the air

What open-pit mines, quarries and sand and gravel pits actually fly drones for, and how Omniris turns those flights into stockpile tonnage, volume change between surveys, slope and drainage maps and CAD surfaces, each with a record of how the number was made.

Summary

The United States produced about 1.65 billion US tons (an estimated 1.5 billion metric tons) of crushed stone in 2025 from an estimated 3,500 quarries, and about 960 million US tons (an estimated 870 million metric tons) of construction sand and gravel from an estimated 6,500 pits [1][2]. Every one of those operations has to know how much material is sitting in its stockpiles, how much it has dug, whether its walls and roads are safe, and whether the land it disturbs is being put back as its permit requires. Drones are now used for all of these: a review of published research on drones in mining found them used for topographic surveying, rock slope analysis, and monitoring of water and ecological restoration [3].

Omniris is a drone operations platform and a secure evidence vault. For a mine or quarry it does five jobs:

The sections below go mission by mission: what mines and quarries fly, what Omniris delivers for each, and how every number is made and recorded.

How mines and quarries use drones today

A systematic review of 65 published studies of drones in mining, from exploration to reclamation, found them used for geological and structural analysis, aerial geophysical survey, topographic surveying, rock slope analysis, working-environment analysis, underground surveying, and monitoring of soil, water, ecological restoration and ground subsidence [3]. Published comparisons of drone surveys at quarries and aggregate sites, such as [6] and [8], focus on stockpile volume. This section goes through each mission: what is flown, how often, what the buyer expects to receive, and which rules apply.

The missions at a glance

Stockpile inventory

Stockpile volume is the subject of the published drone comparisons at quarries and aggregate sites cited here [6][8]. Each survey gives a volume for every pile, and how often a site flies is set by its own inventory and accounting cycle. A drone survey is quicker than a traditional one and keeps people off the pile (both below) [6].

The buyer wants a volume and a tonnage for each pile, comparable from one survey to the next, and a record that will stand up when the inventory figure reaches the books. Volume turns into tonnage through a density. That density is either a handbook figure or one the site measured itself, and material swells when dug and compacts when placed, so the state of the material matters too.

Safety is a large part of the case. Without a drone, someone climbs the pile. The authors of a peer-reviewed comparison at a quarry call climbing the stockpile "the risky part" of the traditional survey, and note that injuries could be sustained where the pile holds large rock blocks [6]. MSHA's fatality alerts describe miners killed when stockpiles collapsed on them: in November 2022 a miner was engulfed in a coal stockpile, and MSHA's advice was to "make sure miners do not travel by foot on stockpiles or near the toe of stockpile slopes" [4]; in September 2022 a miner working near the toe of a pile to clear a feeder was engulfed when it collapsed [5]. Those accidents were not surveys, but the hazard is the same one a surveyor walking a pile faces.

Time. In a peer-reviewed comparison at a quarry, the drone survey of one pile took 35 minutes from setup to processed result, against 97 minutes for the traditional method [6].

Reconciliation

A survey volume is only half of reconciliation. The other half is what the scale house, belt scales and production records say moved. When the two disagree, the inventory is adjusted. Measuring more often means a disagreement is found at the next survey rather than at year-end. Researchers comparing survey methods on aggregate stockpiles note that stockpile owners want effective inventory volumes "to optimize profit and limit waste" [8].

Pits, benches and faces

The same flights that measure stockpiles map the pit itself: the floor, the benches and the faces. MSHA requires mining methods that keep walls, banks and slopes stable where people work or travel, and says that "when benching is necessary, the width and height shall be based on the type of equipment used" to clean the benches or scale the walls [9]. Loose material near the top of a pit or quarry wall must be sloped to the angle of repose or stripped back at least 10 feet (3 m) from the edge where people work or travel [10]. A current topographic survey of the pit, compared with the mine plan and with the last survey, is how an operator sees what was mined and whether the benches are where the plan says.

Highwalls and slopes

Under MSHA's rules for surface metal and non-metal mines, ground examination is part of every working day. Designated people experienced in examining for loose ground must examine work areas before work starts, after blasting and during the shift as conditions warrant, and "highwalls and banks adjoining travelways shall be examined weekly or more often if changing ground conditions warrant" [12]. Hazardous ground must be taken down or supported before other work or travel is permitted in the area, and until it is, the area must be posted against entry, with a barrier when it is left unattended [11]. Drones let an examiner see the upper face and the catch benches without standing under them. Research groups are building on that: the University of Arizona's Geotechnical Center of Excellence runs a NIOSH highwall safety project on rockfall and catch-bench design, and a project using drone photogrammetry to model surface water in open pits [23]. A drone flight adds to the examination but does not replace it, and periodic flights are not continuous slope monitoring.

Haul roads and dump points

Berms or guardrails are required on the banks of roadways "where a drop-off exists of sufficient grade or depth to cause a vehicle to overturn or endanger persons in equipment", and they must be at least mid-axle height of the largest self-propelled mobile equipment that usually travels the road [13]. Openings are allowed for drainage [13]. Dumping locations need berms, bumper blocks or similar restraints where there is a hazard of overtravel or overturning [14]. The Bureau of Mines' 1977 manual on haul-road design covers vertical and horizontal alignment, cross slope, drainage, lane width and runaway-vehicle provisions [15]. A drone survey gives the road's elevation along its length and across it, which is what grade, cross-slope and berm-height checks are read from.

Water, impoundments and tailings

Water decides where a pit floods, where roads wash out and where sediment leaves the site. At coal mines, water, sediment and slurry impoundments that meet MSHA's size or hazard criteria must be examined "for appearances of structural weakness and other hazardous conditions" at intervals not exceeding 7 days, with instruments monitored on the same schedule, by a qualified person, unless the District Manager approves a longer interval [17]. At metal and non-metal mines, a water or silt retaining dam whose failure would create a hazard must be "of substantial construction and inspected at regular intervals" [16]. For tailings, the Global Industry Standard on Tailings Management, published on 5 August 2020, sets 15 principles and 77 auditable requirements, and says "a comprehensive monitoring system must support the full implementation of the 'observational method'" [18]. Peer-reviewed work shows how drones fit: a three-summer drone photogrammetry campaign over a tailings impoundment at a sub-Arctic mine built a surface model each year to track settlement and displacement, and its authors judged drone monitoring "sufficiently accurate for supporting impoundment management operations" [19]. Within such a system a drone survey supplies topography, beach and embankment change and a visual record; instruments supply pore pressure and movement.

Reclamation and permits

Mined land has to be put back. For coal mines, the federal performance standards require disturbed areas to be graded to the approximate original contour, highwalls, spoil piles and depressions to be eliminated, and slopes kept no steeper than the angle of repose or the slope that achieves a long-term static safety factor of 1.3 [20]. Revegetation is judged against written standards for ground cover, production or stocking, and passes at not less than 90 percent of the standard, over a responsibility period of at least five years where annual average precipitation exceeds 26.0 inches (660 mm) and ten years where it does not [21]. Quarries and pits are reclaimed under state programmes, which differ from state to state and are not covered here. A drone survey gives the regraded surface, its slopes and the extent of disturbed and revegetated ground, flight after flight, for as long as the permit runs.

Blasting

Before a shot, blasters need the true shape of the face to know the burden in front of each hole. A paper published by the International Society of Explosives Engineers reports that at a granite quarry with elevated air blast readings, blasts profiled in 3D by drone showed improved air blast levels compared with blasts profiled in 2D, helping the operator manage its readings and keep its licence to operate in the community [22]. After a shot, the muckpile and the new face can be flown again.

Who flies

A site can fly with its own staff pilots or hire a drone service provider. Omniris supports either route (see Getting the flights flown).

Accuracy buyers expect

For pile volumes, published studies put a well-flown drone survey within a few percent of the reference. At a quarry, a drone survey of one pile with five ground control points was within 2.6% of the mill operator's figure for the pile, and the traditional tape survey within 1.3%. Both were inside the ±3% maximum allowable error the authors cite [6]. A comparison against laser scanning found every photogrammetric model within 4%, with ground control improving accuracy [7]. A study of aggregate stockpiles found total-station surveys generally under-reported volumes and were impractical on finer aggregates, terrestrial laser scanning outperformed them, drone surveys showed promise particularly when flown lower with greater overlap, and accuracy varied by aggregate type [8]. For positions, the reference is the ASPRS Positional Accuracy Standards (Edition 2, first issued in August 2023), which report accuracy as RMSE at checkpoints, need at least 30 checkpoints for a product accuracy assessment, and want the checkpoints at least twice as accurate as the product they test [24].

Use case by use case: what Omniris does

Each mission from the previous section, matched to what Omniris delivers from the flight.

Stockpile inventory, step by step

Every map Omniris stitches from an aerial flight comes with two elevation models: a surface model (DSM), the top of everything including the piles, and a terrain model (DTM), the modelled bare ground. A 3D model built from the flight provides the same two surfaces. Volumes are measured on them.

1. Draw the boundary. Open Map Tools, choose Volume, and click out the pile's outline along its toe, where the pile meets the ground. The panel's guidance says why this matters: a snug boundary, drawn the same way every time, keeps repeat measurements consistent. Show elevation underlay draws the survey's elevation model on the map while you draw, which makes the toe much easier to see on uniform material than the photo map alone.
2. Choose the base. The volume is the material between the surface and a base, your model of the ground under the pile. Triangulated (the default) follows sloping ground from the boundary's own elevations. Best-fit plane suits a tilted, flat pad. Lowest point and Highest point put a flat base at the lowest or highest boundary point (the second is for pits and ponds, measuring the void below). Custom elevation takes a known floor or bench elevation. Design surface measures against an engineering surface (below). After a compute, each boundary point's base elevation can be overridden, and Align vertices to lowest sets them all to the lowest in one click. That covers piles against a wall, in a bin, on a bench, or overlapping another pile.
3. Choose the material. Pick a density from the preset list (sand, gravel, crushed stone, topsoil, wet clay, riprap, asphalt millings, demolished concrete, bituminous coal and wood chips, each a nominal handbook figure) or type your own. A preset is labelled preset (nominal); a density you enter is labelled site-tested, and the label travels with every tonnage figure. Optionally record the material's state (bank, loose or compacted) and enter a swell/shrink factor: above 1 for material that bulks when dug (1.12 means 12% fluff), below 1 for compacted fill. The state is a label kept with the measurement. The factor multiplies the tonnage (volume × density × factor), which turns a loose or compacted density into an in-place one; for a pile measured as it stands, use the pile's own loose density and leave the factor at 1.
4. Compute. With a flat base (lowest point, highest point or a custom elevation), a quick preview marked approximate updates as you adjust the boundary. The recorded number comes only from Compute volume, which works through the elevation model cell by cell at full resolution. The result card gives cut (for a stockpile, the pile), fill, net and the tonnage, each with its ± uncertainty, and the area measured, in cubic yards and US tons by default (metric in the account's units setting).
5. Save it on the site. Save measurement turns the boundary and its settings into a named measurement. Attached to a site, it measures itself on every survey of that site in your account that covers the whole boundary: the flights already there, and every new map or 3D model of the site from then on, usually within minutes of it finishing. A flight that clips only a corner of the pile is left out, because its number would look like the pile shrank.

The results appear where a site manager already looks. The Site Navigator's Volumes section draws every saved boundary on the map and lists each measurement with one line per flight date and a volume over time chart. The site's page in the Site Manager shows the latest value of each measurement, with a printable report for each and a spreadsheet (CSV) of every result. That is your stockpile inventory as of the latest flight.

What moved: change between surveys

Two ways to compare dates, for two questions.

• A saved measurement's Compare epochs. Pick an earlier and a later survey and Omniris differences the two surfaces inside the boundary directly: how much material was added and how much was removed, each with its ± uncertainty. Because each surface was measured independently, the ± of a comparison carries both flights' error and is larger than a single flight's.
• Compare in the Site Navigator. No saved measurement needed. Choose two dates and an area: the whole survey, a zone drawn on the site, a measurement's boundary, or an area you draw on the spot. You get material added, removed and net, a downloadable comparison report and spreadsheet, and a change heatmap, red where the surface rose and blue where it dropped. Visual mode puts two dates side by side with a fade, a blink or a swipe, for seeing what changed before measuring it.

Every comparison is graded for how well the two flights line up vertically. A pair counts as anchored when the later flight was aligned to the earlier one, or when both were stitched against the same surveyed ground-control site. An anchored pair is graded from its recorded alignment error (RMSE): green at 1.6 ft (0.5 m) or less, amber up to 6.6 ft (2.0 m), and red above that; an anchored pair with no alignment error recorded is amber. Any other pair, including two RTK flights with no shared ground control, still computes, labelled estimate-grade. This matters at a mine more than almost anywhere: a single flight's pile volume is internally consistent without ground control, but the difference between two flights inherits any vertical drift between them, and drift reads as phantom material gained or lost. Fly with ground control, or align the re-flight to the earlier one, when the answer is a number someone will book. Comparisons between dates run between stitched maps, not 3D models, so two different reconstructions of the same ground are never mistaken for change.

Change alerts

A comparison answers once. A change alert keeps asking. From any elevation comparison on a site, press Alert me and set a threshold on net volume change, material added or material removed, in cubic yards (the default is 654 yd³ (500 m³)). Whenever a new flight lands on the site, Omniris compares it with the flight before it over the same area and emails you, and up to ten colleagues who can read the site, when the threshold is reached. Checks run every six hours.

The alert page outlines and numbers the areas that changed, largest first, with each one's size, average height change and volume, and whether the ground rose or fell. An area counts only when its change is larger than the surveys can resolve: more than three times the survey's resolution, more than twice the alignment error between the flights, and more than any minimum height you set, over at least 100 ft² (9.3 m²) unless you set a different minimum. Areas are traced only for anchored pairs; an estimate-grade pair fires on its number but gets no outlined areas. Each area can be turned into a finding, or the alert can do that every time it fires. For a highwall, a waste dump or an embankment, this is a standing question: tell me where this ground moved by more than I can explain.

Measuring against the plan

Load the design into the vault and pick Design surface as a volume's base: a DXF of 3D faces or 3D polylines, a LandXML TIN surface, a design GeoTIFF, or an imported design model whose chosen objects' top faces form the surface. The result is cut and fill against the design, for a pit shell, a final bench, a pad or a reclamation contour. Compare's Design mode does the same against an imported design model for a whole survey and draws the cut and fill on the map. A design file is measured the way a design model is: the survey is converted into the file's own coordinate system and heights cell by cell, so a design in your State Plane zone in feet with NAVD88 heights is placed and compared correctly against a survey in UTM metres with GPS heights. What the file declares is used as it is. When the file leaves something out, the Volume tool asks for it: the coordinate system with the same picker as Export to CAD (a State Plane zone in its legal unit, UTM, or the survey's own), the vertical datum with the same choices as Import as Design Model, and, for a NAD83(2011) design in California, its epoch. The answers are saved with the measurement, and its report says which values the file declared and which were stated, and by whom. Heights that can't be converted reliably are refused with the reason, as is a design that lands nowhere near the survey. The Design, progress and integrations paper covers design models in full.

Slopes, contours, ponding and drainage

From a map's terrain or surface model, Terrain analysis makes, at full resolution:

• Contours at the interval you choose (1 ft by default, with a heavier line every fifth), each line carrying its elevation in feet and metres and the vertical datum.
• Slope in percent grade (the default) or degrees, on a fixed colour scale so dates can be compared by eye. Click the map to read a value such as 10.0% grade (5.7°).
• Aspect, the compass direction each slope faces.
• Ponding: where water would stand, with each pond's maximum depth in inches, area in ft² and acres, and volume in cubic yards and US gallons.
• Drainage flow lines above a minimum contributing area in acres.

On a haul road the slope map shows grade along and across the road; on a pit floor or a regraded dump, ponding shows where water will collect before the rain does. Use the terrain model for ground contours: contours on the surface model follow equipment and piles.

Profiles, roads and berms

The Distance tool gives a line's horizontal length, its slope distance and its length over the ground at 16 ft (5 m) resolution. The Profile tool draws the elevation along a line as a chart, with a readout under the cursor, gaps where there is no data, and a CSV of the samples. Along a haul road that is the grade; across a road it is the cross slope; across a berm it is the berm's height above the running surface, ready to set against the mid-axle height of the largest equipment using the road.

Faces in 3D

A crosshatch or perimeter flight builds a textured 3D model of a face or a pile. In the 3D viewer, Measure reports path length, straight-line distance and vertical difference separately, so a face's height is never confused with a diagonal. Face area reports a face's area, perimeter and tilt from horizontal (0° flat, 90° vertical), which is the face angle. These tools are offered only on models the build placed on the Earth (georeferenced). A 3D model's surface and terrain models also work in the map's volume tool, labelled with their source. Gaussian splats, the other 3D product, are for viewing; measurements are taken on the 3D model or the map. See the Digital twins paper.

LiDAR

Where vegetation hides the ground, at a reclamation site or the edge of a pit, a LiDAR sensor sees through it. Omniris processes LAS or LAZ files from any LiDAR sensor into a classified point cloud (ground, vegetation, buildings, wires, noise), a terrain model, a surface model, a canopy height model and contours, viewable in a 3D viewer that streams the cloud. Classes can be corrected by hand and the terrain products rebuilt. The volume tool, Compare, terrain analysis and CAD export work from photo surveys (stitched maps and 3D models), not from LiDAR terrain models. See the LiDAR and point clouds paper.

Reclamation and revegetation

Regrading is measured like any other earthwork: the regraded surface against the approved contour as a design surface, its slopes on the slope map, and the area of disturbed ground with the Area tool. Revegetation needs a multispectral camera (Multispectral paper). A multispectral flight produces an NDVI vegetation map, and an NDRE map when the camera has a red-edge band; Band Analytics adds other indices, and Compare measures how an index changed between two flights over a zone, including how much of the area improved, declined or stayed the same. Omniris shows the vegetation and how it is changing; the success determination under the permit, made from its own sampling standard, stays with the regulator.

Findings, reports and Orion

A cracked berm, a washed-out ditch or a slumping face becomes a finding: a pin or outline on the map with a category, severity, status, assignee, photos and a custody history, created by hand, from AI detection, or from a change alert's areas of change. Findings export as GeoJSON, CSV or KML and reach other systems through the findings API and webhooks (Findings paper). AI detection is open-vocabulary: you describe what to look for, review and correct what it finds, and turn results into findings (AI detection paper). A situation report is one PDF telling a site's story between two dates, with the change, the findings and the source of every image.

Orion, the Omniris assistant, can answer questions about a site's surveys in plain language. It can measure a volume inside a saved boundary or a site zone, compare two flights and report material added and removed with the alignment grade, read an elevation profile, and report how tall things stand above the ground on a finished map. It reports in cubic yards and feet first, and it sees exactly what you are allowed to see (Orion paper).

Accuracy and standards

Two kinds of accuracy matter at a mine, and they answer different questions.

• Relative accuracy is how true the map is to itself. A single flight's pile volume, a bench height or a road grade relies on it.
• Absolute accuracy is how true the map is to the world. Comparing two flights, measuring against a design, or handing coordinates to a surveyor relies on it, and ordinary drone GPS is not good enough for it. The measuring tools say so: comparisons between dates and against designs need ground control, RTK or alignment on both surveys.

Ground control, RTK and alignment

Omniris covers the ground-control workflow from end to end (Stitching paper):

• Import surveyed points from the files survey equipment and software produce (Emlid Reach, Trimble Access, Leica Captivate, Pix4D, Metashape, OpenDroneMap, plain CSV and others) in the survey's own coordinate system, including State Plane in US survey feet. NAVD88 heights are converted with the GEOID18 model before the stitch, and a height whose datum is unknown is flagged rather than guessed.
• Mark the targets in the photos. Printed coded (ArUco) targets are found automatically, and a target marked once is predicted in the other photos. A control site is reused on every later flight of the same ground, so permanent targets around a quarry are surveyed once; each new flight's photos are marked again, automatically for coded targets and with predicted positions for the rest.
• RTK flights: you state the coordinate frame of the corrections, and whether the base sat on a known point, once for the site, and it applies to every flight on that site unless a flight overrides it.
• Align a re-flight to an earlier one when there is no fresh control, so the two dates compare against each other.

Checkpoints: measured, not assumed

A checkpoint is a surveyed target left out of the fit and used only to measure the result. Import a file as checkpoints (or switch any surveyed point between GCP and checkpoint, with the switch recorded), mark it like a GCP, and the finished map reports its measured accuracy: horizontal, vertical and 3D RMSE at the checkpoints, feet first, with each point's error east, north and up, the difference between the map's surface and the surveyed height, and a tag on checkpoints outside the area the GCPs enclose. Evaluate checkpoints re-measures a finished map without stitching again.

The ASPRS standards describe how such an assessment is done: RMSE at checkpoints, at least 30 of them for a product accuracy assessment, each at least twice as accurate as the product [24]. Omniris reports the RMSE at the checkpoints you give it. It does not assign an ASPRS accuracy class, and whether the checkpoints are enough is for the surveyor to decide.

Volume accuracy

Published studies put well-flown drone stockpile volumes within a few percent of a reference survey [6][7]. Omniris attaches a ± to every volume, from an error per elevation cell that grows with the survey's cell size, added up over the cells measured. It reports how much of the boundary had no elevation, and warns when a stockpile boundary looks wrong. What it cannot measure is the boundary itself: where the toe is drawn, and whether it is drawn the same way each month, usually matters more than anything in the calculation.

Heights and datums

Every contour file, terrain analysis and CAD export names the vertical datum its heights are in: orthometric heights above EGM96, ellipsoidal heights with their frame, or heights relative — vertical datum unknown when the flight cannot establish one. For an RTK flight on a base set up over an unsurveyed point, the output says that absolute position may be off by metres.

Evidence and records

An inventory figure ends up in the books, an as-mined volume in a royalty statement, and a slope picture in an incident file. Each one needs to show where it came from. Omniris keeps that record for every number in this paper (Evidence integrity paper).

• Every measurement is on the record. Creating, editing, evaluating and deleting a saved volume measurement is sealed into the tamper-evident custody record, with who did it and when. Editing a boundary or a base recomputes the past results in the background, so the history always matches the current definition. The record keeps the old and new value of each setting the edit changed, and each recomputed flight's volumes and tonnage before and after.
• The report shows what was recorded. A measurement's report renders the stored result: the surface used, the flight date, the resolution, the base and its settings, the boundary, the share with no data, who computed it and when, the volume ± uncertainty, the tonnage with its preset or site-tested label, and an accuracy statement. It shows cubic yards and US tons with cubic metres and tonnes beside them, whatever the reader's preference. A comparison report adds the alignment grade.
• The inputs are in the vault. The photos, the map and its elevation models, the design file a volume was measured against, and the change maps are files with their own custody histories and SHA-256 fingerprints.
• Certificates anyone can check. A signed, timestamped custody certificate can be checked by an auditor online, or offline with free standard tools, without relying on Omniris.
• Legal holds and retention. A hold on a file, folder, job, site, account or organization freezes the evidence; a retention policy disposes of it at the end of its period with a signed certificate of what was deleted.

For an operator facing an MSHA citation, a contractor dispute over hauled quantities, or an auditor testing year-end inventory, that record answers the first question asked: can you show how this number was made?

Getting the flights flown

Planning the flight

The Omniris mission planner turns an outline on the map into a flight the pilot imports into the aircraft's own app (Mission planning paper). Five patterns cover mine work: a grid for maps and elevation models, a crosshatch with the camera tilted for 3D models of faces, perimeter rings facing inward for stockpiles and structures, a corridor along a haul road, and an orbit around a single object. The planner works out overlap, ground detail and motion blur from the real camera, and estimates photos, time and batteries.

A pit is the hard case for altitude. FAA rules cap a small drone at 400 feet (122 m) above ground level, except within 400 feet of a structure [25], and a drone told to fly at a set height flies at that height above its take-off point, not above the pit floor. Terrain follow gives every waypoint its own height to hold a steady height above the ground, from a worldwide terrain model or, far better in a working pit, from your own previous survey's surface model. A minimum clearance above the modelled surface, 33 ft (10 m) by default, is kept everywhere. When the ground under the area varies by 49 ft (15 m) or more, the planner recommends turning it on.

Every saved mission is checked against FAA airspace along its whole route, including the floors of Class B, C, D and E airspace, and Omniris works out what LAANC authorization to request when one is needed, which matters for quarries near airports. Weather is scored hour by hour against the aircraft's limits, and a Part 107 pre-flight checklist is saved with the job. Omniris plans and checks the flight; the remote pilot in command flies it and remains responsible for it.

Your pilots, or hired ones

Omniris supports both ways a site gets its flights flown.

• In-house. Your own pilots are members of your organization. They upload, stitch and measure in the same vault the site managers read, and the field app works on a phone at the site. Saved volume measurements and their automatic re-measurement read only the maps and 3D models stitched in the measuring person's own account, so the person who keeps the site's measurements should run the stitches. Compare between dates on a site also takes in other people's stitched maps of that site when you can read the site.
• Hired through the marketplace. Post a job with the site, the deliverables and the dates, and choose a pilot from those who apply or offer it to one directly. A job can recur weekly, every two weeks, monthly, quarterly or every N days, weeks or months, for a fixed number of flights or until stopped. Each flight is its own job with its own flight window, delivered and billed when it is completed. That fits whatever inventory cycle the site runs on.

A hired pilot's results arrive in your vault as a read-only copy of the delivered folder, each file with a custody history that starts with where it came from. You review everything first, and downloads unlock when you accept. To measure a hired pilot's flight, stitch the delivered photos in your own account; a saved volume measurement on the site then measures that map on its own. The marketplace, payments and payouts are part of the Omniris cloud and the data center edition. On-premises installations do not include them (Sharing and delivery paper).

Integrations and delivery

CAD export in State Plane. The Export to CAD dialog looks up the site's State Plane zone and starts in its legal unit (U.S. survey feet in most states, international feet in a few, metres where the law never adopted a foot). It shows the horizontal frame the map is in, the datum shift it will apply and its published accuracy, and the vertical datum. A map made without RTK or ground control is in WGS84/ITRF, which differs from NAD83(2011) by a metre or more and changes over time, so the export shifts it as of the flight date, letting the coordinates line up with surveyed NAD83 points. A map made with NAD83 ground control is never shifted twice. UTM and the survey's own coordinate system are offered too.

Sharing. Members of your organization work in its shared vault, and a folder can be shared with another Omniris user by a sharing code. For a consultant or a landowner who will never sign in, a folder, a 3D model or a site's map can be published on a public link; a folder link can carry a password and an expiry, and be view only or view and download. The Omniris desktop app opens the same sites, maps and measurements; the mobile field app carries jobs, files and findings on site.

Mine planning, fleet management, scale-house and inventory accounting systems take Omniris results through the files above.

Cloud, data center and on-premises

Omniris runs in four places, and every tool in this paper is part of each:

For a mining company that wants survey data, designs and evidence kept on its own network, on-premises keeps the processing and the vault in house. An on-premises installation connected to the Omniris cloud can send a single stitching job there when it chooses. Security, encryption and access control are described in the Security paper.

What the product claims, and what it doesn't

A mine's numbers end up in the books, in permits and in safety files, so Omniris states what its results are and are not:

• A measurement, not a licensed survey. Omniris computes and records. Where a state, a permit or a contract requires a licensed surveyor's seal, the surveyor supplies it. Omniris reports RMSE at the checkpoints you give it and does not assign an ASPRS accuracy class.
• The boundary is yours. You draw each stockpile boundary, and Omniris re-measures it on every survey from then on; a pile that moves or changes shape needs its boundary redrawn. The ± on a volume is the surface's uncertainty; where the toe is drawn, and whether it is drawn the same way each time, is not part of it.
• Change between dates needs anchored flights. A comparison between two flights without shared ground control or alignment still computes, labelled estimate-grade, because vertical drift between the flights reads as material gained or lost.
• An aid to ground examination, not a replacement. Slope maps, change between flights, alerts and 3D models help an examiner. They are periodic surveys, not continuous wall-movement monitoring or a slope-stability analysis, and they do not take the place of the examinations MSHA requires.
• Surveys alongside the instruments. For impoundments and tailings, Omniris supplies topography, change and a visual record; piezometer, inclinometer and prism readings, freeboard and reporting under the tailings standard remain in the site's monitoring system.
• Regulators decide. Regraded surfaces, slopes and vegetation indices are evidence for a reclamation file. Whether a permit's grading or revegetation success standard is met is the regulator's determination, from the sampling the permit specifies.
• Pilots fly. Omniris writes the mission and checks it; the remote pilot in command flies the aircraft and remains responsible for it. Airspace data is advisory, and NOTAMs and temporary flight restrictions must be checked before each flight.
• Commerce is cloud and data center only. The marketplace, payments and payouts are part of the Omniris cloud and the data center edition, not of on-premises installations.

Sources

1. U.S. Geological Survey, Mineral Commodity Summaries 2026, "Stone (Crushed)". https://pubs.usgs.gov/periodicals/mcs2026/mcs2026-stone-crushed.pdf
2. U.S. Geological Survey, Mineral Commodity Summaries 2026, "Sand and Gravel (Construction)". https://pubs.usgs.gov/periodicals/mcs2026/mcs2026-sand-gravel.pdf
3. S. Park, Y. Choi, "Applications of Unmanned Aerial Vehicles in Mining from Exploration to Reclamation: A Review", Minerals 10(8), 663 (2020). https://doi.org/10.3390/min10080663
4. Mine Safety and Health Administration, Fatality Alert, 5 November 2022 (engulfment in a coal stockpile). https://www.msha.gov/data-reports/fatality-reports/2022/november-5-2022-fatality/fatality-alert
5. Mine Safety and Health Administration, Fatality Alert, 28 September 2022 (collapsed stockpile). https://www.msha.gov/data-reports/fatality-reports/2022/september-28-2022-fatality/fatality-alert
6. J. Matsimbe, W. Mdolo, C. Kapachika, I. Musonda, M. Dinka, "Comparative utilization of drone technology vs. traditional methods in open pit stockpile volumetric computation: A case of Njuli quarry, Malawi", Frontiers in Built Environment 8 (2022). https://www.frontiersin.org/journals/built-environment/articles/10.3389/fbuil.2022.1037487/full
7. K. Kokamägi, K. Türk, N. Liba, "UAV photogrammetry for volume calculations", Agronomy Research 18(3) (2020). https://dspace.emu.ee/items/2f487a65-7795-448e-b819-d15be41fa076
8. D. J. A. Davis, N. S. Guy, "An Assessment of Point Cloud Data Acquisition Techniques for Aggregate Stockpiles and Volumetric Surveys", Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci. XLVIII-M-3-2023, 65–69 (2023). https://isprs-archives.copernicus.org/articles/XLVIII-M-3-2023/65/2023/
9. 30 CFR § 56.3130, Wall, bank, and slope stability. https://www.law.cornell.edu/cfr/text/30/56.3130
10. 30 CFR § 56.3131, Pit or quarry wall perimeter. https://www.law.cornell.edu/cfr/text/30/56.3131
11. 30 CFR § 56.3200, Correction of hazardous conditions. https://www.law.cornell.edu/cfr/text/30/56.3200
12. 30 CFR § 56.3401, Examination of ground conditions. https://www.law.cornell.edu/cfr/text/30/56.3401
13. 30 CFR § 56.9300, Berms or guardrails. https://www.law.cornell.edu/cfr/text/30/56.9300
14. 30 CFR § 56.9301, Dump site restraints. https://www.law.cornell.edu/cfr/text/30/56.9301
15. W. W. Kaufman, J. C. Ault, Design of Surface Mine Haulage Roads — A Manual, U.S. Bureau of Mines Information Circular 8758 (1977). https://trid.trb.org/View/69865
16. 30 CFR § 56.20010, Retaining dams. https://www.law.cornell.edu/cfr/text/30/56.20010
17. 30 CFR § 77.216-3, Water, sediment, or slurry impoundments and impounding structures; inspection requirements. https://www.law.cornell.edu/cfr/text/30/77.216-3
18. International Council on Mining and Metals, Global Industry Standard on Tailings Management (2020). https://www.icmm.com/en-gb/our-principles/tailings/global-industry-standard-on-tailings-management
19. A. Rauhala, A. Tuomela, C. Davids, P. M. Rossi, "UAV Remote Sensing Surveillance of a Mine Tailings Impoundment in Sub-Arctic Conditions", Remote Sensing 9(12), 1318 (2017). https://doi.org/10.3390/rs9121318
20. 30 CFR § 816.102, Backfilling and grading: General requirements. https://www.law.cornell.edu/cfr/text/30/816.102
21. 30 CFR § 816.116, Revegetation: Standards for success. https://www.law.cornell.edu/cfr/text/30/816.116
22. C. Rosen, T. Rockley, "Controlling Air Blast with Drone 3D Profiling", International Society of Explosives Engineers. https://onemine.org/documents/controlling-air-blast-with-drone-3d-profiling
23. University of Arizona, Geotechnical Center of Excellence, research efforts (NIOSH Highwall Safety Project; drone photogrammetry for surface hydrology in open pit mines). https://mining.arizona.edu/gce/research
24. D. B. Zilkoski, "ASPRS approves edition 2 of the ASPRS Positional Accuracy Standards for Digital Geospatial Data", GPS World, 6 December 2023. https://www.gpsworld.com/asprs-approves-edition-2-of-the-asprs-positional-accuracy-standards-for-digital-geospatial-data/
25. 14 CFR § 107.51, Operating limitations for small unmanned aircraft. https://www.law.cornell.edu/cfr/text/14/107.51

Regulation text is cited from the Legal Information Institute's reproduction of the Code of Federal Regulations; check the official eCFR before relying on it for a legal purpose.

Glossary

Angle of repose

The steepest angle at which loose material stays put without sliding.

Bank, loose and compacted

The three states of earth material: in place in the ground, broken and dug (bulkier), and placed and compacted as fill (denser). The same tons occupy a different volume in each.

Base surface

The surface a volume is measured down to: the modelled ground under a pile, a flat elevation, or a design surface.

Bench

A horizontal step cut into a pit wall, which catches falling material and gives equipment access.

Berm

A ridge of material along the edge of a road or dump point that keeps vehicles from going over a drop-off. MSHA requires roadway berms at least mid-axle height of the largest equipment that usually uses the road.

Checkpoint

A surveyed target left out of the map's fit and used only to measure its accuracy.

Cut and fill

In a single-flight volume, material above the base (cut) and the space below it (fill). Between two flights, Omniris labels material added as cut and material removed as fill, and change alerts name them material added and material removed.

DSM (surface model)

A height map of the top of everything: ground, piles, equipment, vegetation.

DTM (terrain model)

A height map of the modelled bare ground.

DXF and LandXML

Standard exchange files for CAD and civil design software. LandXML carries a surface as a TIN.

GCP (ground control point)

A surveyed target on the ground that the map is pinned to.

GSD (ground sample distance)

The size of one pixel or elevation cell on the ground, for example 0.16 ft (5 cm).

Highwall

The steep face of a pit or quarry left by excavation.

Impoundment

A structure that holds back water, sediment, slurry or tailings.

RMSE

Root mean square error: the standard way to state accuracy, the square root of the average squared error at the checkpoints.

RTK

Real-time kinematic positioning, which corrects the drone's GPS in flight from a base station or network.

State Plane

The set of map coordinate systems, one or more per US state, that surveyors and engineers use for local work.

Swell (or shrink) factor

The ratio by which material's volume grows when dug (above 1) or shrinks when compacted (below 1).

TIN

Triangulated irregular network: a surface made of triangles joining surveyed points.

Toe

The bottom edge of a pile or slope, where it meets the ground.

Omniris vertical paper · Mining and aggregates · October 2026. Related feature papers: Stitching · Site management · Digital twins · LiDAR and point clouds · Evidence integrity · Mission planning.