When a contractor submits a pay application for cut-and-fill work, the underlying volume calculation is only as defensible as the survey method that produced it.
UAV-based photogrammetry and LiDAR have both become standard tools for generating the digital terrain models that drive those calculations. Each can produce centimeter-level results under the right conditions. Each can also introduce errors that go undetected until quantities are disputed.
How Each Method Derives Volume
Both technologies generate point clouds used to build digital terrain models (DTMs). Volumes are calculated by comparing a pre-construction DTM to a post-construction DTM, or by comparing measured conditions against a design surface.
UAV photogrammetry uses overlapping aerial images and Structure-from-Motion (SfM) algorithms to reconstruct terrain geometry from matched photo features. Its accuracy is tied to image quality, ground control point (GCP) distribution, and site conditions at the time of flight.
LiDAR measures distances using pulsed laser signals, recording multiple returns per pulse where vegetation is present. Because range is derived from time-of-flight rather than image feature matching, accuracy depends primarily on the sensor’s GNSS receiver and inertial measurement unit (IMU) rather than surface texture or lighting.
Where Photogrammetry Performs — and Where It Doesn’t
On open, well-lit, recently graded sites, UAV SfM photogrammetry can match LiDAR at a significantly lower cost. Research published in Drones (MDPI) confirms centimeter-level absolute accuracy is achievable with sufficient ground control, and a comparative study published in Photogrammetric Engineering & Remote Sensing evaluating UAS photogrammetry against helicopter-based LiDAR found point clouds comparable to within centimeters on hard, uniform surfaces.
GCP density is the critical variable.
A peer-reviewed study in Remote Sensing demonstrated that vertical accuracy improves continuously with GCP count in ways horizontal accuracy does not; a meaningful asymmetry, since Z-axis precision drives cut-and-fill calculations.
Surveys relying solely on onboard RTK without independent GCPs carry vertical uncertainty that may not be apparent in the orthomosaic output.
Photogrammetry also fails to differentiate between vegetation and bare ground. The SfM algorithm reconstructs whatever the camera can see. On sites with residual ground cover, stockpiles with irregular shadows, or active earthmoving, the resulting terrain model can reflect surface clutter rather than actual grade.
Research published in PMC comparing LiDAR and SfM for bare-earth characterization found aerial LiDAR outperformed SfM in vegetated terrain, which is directly relevant to construction sites that haven’t been fully cleared.
LiDAR’s Ground Detection Advantage
LiDAR’s multiple-return capability allows processing software to filter vegetation and isolate ground points. On sites with residual brush, loose surface material, or shadow-obscured stockpile edges, LiDAR consistently produces a cleaner bare-earth DTM.
The FHWA Tech Brief on the Use of Small UAS for Construction Quantity Estimation identifies earthwork volume and stockpile calculation as primary UAS applications, and a comparative accuracy study in Infrastructures (MDPI) validated that high relative accuracy is achievable across LiDAR and photogrammetric methods alike when survey control is properly established.
What the Standards Require
The ASPRS Positional Accuracy Standards for Digital Geospatial Data define vertical accuracy using RMSEz and require a minimum of 20 independent checkpoints for validation. Volume calculations derived from data without independent checkpoint validation — or photogrammetric surveys lacking adequate GCPs — cannot be reliably assigned an accuracy class. Without a defensible accuracy class, the data is difficult to stand behind if quantities are challenged.
The Professional Liability Consideration
Method selection creates exposure that extends beyond the survey itself. If a pay application is disputed and the volume calculation rests on a photogrammetric survey conducted without adequate GCPs, over partially vegetated terrain, or in poor lighting, the surveyor or engineer of record may face questions about whether the method was appropriate, regardless of whether the final numbers are close.
Both methods, when properly executed and validated, can produce data that holds up. The distinguishing factor is documentation: GCP reports, accuracy checkpoints, sensor calibration records, and flight logs that allow independent review of data quality.
The survey that saves money on data collection but introduces volume uncertainty that exceeds the value of a disputed pay item is not a cost-efficient choice.
