Ground Control, Checkpoints, and Accuracy

Center for Geospatial Analytics at North Carolina State University

Corey White

Objectives

• Relate flight GSD to the horizontal and vertical accuracy a UAS survey can reach
• Design ground control and checkpoint schemes
• Choose between GCP-based and RTK/PPK georeferencing
• Read an Agisoft or WebODM processing report
• Determine the level of detection (LoD)

• The accuracy lecture Topic 2 has been pointing at
• 2A: checkpoints are the real number. 2B: quote the checkpoint RMSE
• Today: where those numbers come from and what they permit
• Everything runs on their own 2B and 2C reports and the 9/16 Emlid checkpoints
• The lab validates the flight’s own products

The report number is not the accuracy

A worked example from the ASPRS standard (American Society for Photogrammetry and Remote Sensing 2024):

• The DSM fits the checkpoints with RMSE_V1 = 1 cm
• The checkpoints themselves were surveyed with an RTK rover: RMSE_V2 = 2 cm
• You may claim the quadrature sum: \(\text{RMSE}_V = \sqrt{1^2 + 2^2} = 2.24\) cm

• Open with the arithmetic: the software prints a fit
• Edition 1 required checkpoints 3x better, so their error was negligible
• Edition 2 relaxed that to 2x and makes you carry the term
• For the lab, RMSE V2 is the rover repeatability from the field day
• Habit for the hour: every claim says what it was measured against

What GSD buys you

• Section 1: the floor pixel size sets, and why you rarely stand on it

GSD sets a floor, not the accuracy

• Envelopes from Jiménez-Jiménez et al. (2021); best case from Sanz-Ablanedo et al. (2018)
• Vertical error runs about 2.5 x the horizontal error (Sanz-Ablanedo et al. 2018)
• With uncorrected navigation-grade georeferencing, errors reach meters and pixel size stops mattering

• Envelope: Jimenez-Jimenez 2021. Asymptotes: Sanz-Ablanedo 2018, 3465 GCP combinations over 1200 ha
• Teach the multipliers as a band
• The gap between the floor and what you get is georeferencing and systematic error
• Ask the class: Lake Wheeler was 2 to 3 cm GSD, what vertical accuracy do you predict?
• Write the predictions down, the lab checks them

Planning with GSD

flowchart LR
    A["What you must<br>measure"] --> B["The accuracy<br>that requires"]
    B --> C["GSD:<br>0.5 to 1.0 x that"]
    C --> D["Flying height<br>(Topic 3A)"]

• ASPRS recommends source imagery GSD of 0.5 to 1.0 x the target horizontal accuracy class: a 5 cm class wants 2.5 to 5 cm GSD (American Society for Photogrammetry and Remote Sensing 2024)
• Read backwards: expect RMSE_H of 1 to 2 x GSD from a well-run project
• The standard calls this a planning heuristic and says it may not be used to report accuracy

• Connects 3A’s GSD formula to a requirement instead of a preference
• The chain runs left to right, never the other way
• Quote the standard’s caveat: the table tells you what to fly

Ground control

• Section 2: how many points, where, and which ones you are allowed to judge with

How many GCPs, and where

• About 3 to 4 GCPs per 100 photos; below 1 per 100 accuracy degrades fast (Sanz-Ablanedo et al. 2018)
• Edge coverage first, then an even interior; points past that improve error structure rather than RMSE (Dai et al. 2023)

• Per-100-photos scales with overlap and altitude; per-hectare rules do not
• Dai 2023: after the first few GCPs, RMSE barely moves, spatial autocorrelation of the residual keeps improving
• That structure is what matters for differencing
• Our 9/16 design, 4 targets on a 78-photo flight, sits on the recommended density

Control at Lake Wheeler

Our own flight. Targets ring the site: a few become control, the rest stay checkpoints.

• Their own orthomosaic from 9/16, their own targets
• Point at the spread: edges first, then interior
• Ask which they would make control and why
• Every target they do not use as control is a checkpoint they get for free

Control is not validation

• Error reported on control points overstates accuracy by 3 x (dense control) to 4 to 8 x (sparse control) (Sanz-Ablanedo et al. 2018)
• ASPRS: checkpoints must be independent, excluded from the aerial triangulation, and away from any GCP (American Society for Photogrammetry and Remote Sensing 2024)

• The 3 to 8 times overestimation is the number to remember
• It is why the field protocol withheld most of the surveyed points
• If they keep one distinction from today, this is it
• The midterm asks it

RTK, PPK, and what they still need

• Section 3: direct georeferencing moved fast in five years
• Current state and the working recipe

Direct georeferencing today

Evidence	Result
ASPRS Addendum V (American Society for Photogrammetry and Remote Sensing 2024)	PPK reaches 1 to 2 cm horizontal, 2 to 4 cm vertical “when used properly”; RTK typically 5 to 10 cm
RTK aircraft, zero GCPs, purpose-built flight (Stott et al. 2020)	Vertical RMSE 6.6 cm over 2 km of river; adding 5 GCPs did not improve it
PPK versus GCP workflows (Zhang et al. 2019)	PPK matched GCP-workflow accuracy (MAE about 2 cm, RMSE about 3 cm); some missions carried a vertical bias that one GCP removed, improving vertical accuracy 20 to 30 percent
Repeat surveys (Nota et al. 2022)	Co-alignment of epochs gives sub-2 cm relative accuracy; ground control in at least one epoch still anchors absolute Z

• Stott’s zero-GCP result required a double grid flown in opposing directions at 20 degrees off nadir
• Against sparse checkpoints, vertical errors still clustered spatially to 0.14 m
• So distributed checkpoints stay mandatory
• Zhang: the vertical bias is near constant, so one known point kills it
• Nota matters for Topic 6: co-align epochs, put control in one of them

Where the standard disagrees

Working defaults: RTK or PPK for georeferencing, 1 or 2 GCPs to control the vertical datum, every other surveyed point withheld as a checkpoint

• ASPRS keeps GCPs “highly recommended” and gives them priority over airborne positions in the adjustment: camera centers have fewer epochs and weaker geoid handling than survey-grade ground occupations (American Society for Photogrammetry and Remote Sensing 2024)
• The journal results assume strong flight geometry; the standard is written for contracts and courts

• Do not resolve the tension, name it
• Recent results justify light control when the flight design is strong
• The standard assumes you must defend the number to a client
• A consultant should know both and say which regime they are in
• Our lab uses the recipe: 3 targets as control, the rest checkpoints

The base coordinate governs everything

• Base within 0.5 mm: an Emlid Reach RS2 base gave DEM vertical RMSE of 1.94 cm versus 1.89 cm for a Trimble Alloy with a choke-ring antenna (Famiglietti et al. 2021), at roughly a tenth of the price
• Antenna beats receiver: a low-cost dual-frequency board matched geodetic instruments in network RTK only with a geodetic-grade antenna; with a patch antenna it needed open sky (Janos and Kuras 2021)

• Connects the literature to the hardware they held
• Famiglietti is the fair Emlid-class citation; receivers have improved since
• Nobody has published RS4 versus state RTN: the class dataset is a credible short note
• Recruit interest for it here

Systematic error

• Section 4: the error that does not average away and that differencing cannot forgive

Doming

• Camera orientations off parallel by only 2 degrees produce a dome of about 0.2 m over 100 m (James and Robson 2014)
• Fixes in order of power: a known camera model with self-calibration off; oblique imagery (a convergent orbit reached under 2 mm in simulation); two flying heights; well-distributed control
• Best camera angles in the tested 0 to 35 degree range were 20 to 35 degrees, worth nearly 50 percent better accuracy in high relief (Nesbit and Hugenholtz 2019)
• A gently tilted camera (under 15 degrees) can make things worse on consumer wide-angle cameras with onboard lens correction (James et al. 2020)

• Recall the pitfalls table from 2A, here is the mechanism
• The 2020 caveat is the one practitioners miss
• “Just tilt the camera a bit” backfires below 15 degrees on exactly the cameras they fly
• More overlap does not fix doming: James 2020 showed relative insensitivity to overlap

Why systematic error matters more

• Random error averages away in aggregates; a 2 cm systematic bias does not
• Propagated uncertainty in change detection is dominated by spatially correlated and systematic error, not point precision (Anderson 2019)
• Blunder and bias screens from the standard (American Society for Photogrammetry and Remote Sensing 2024)

• The mean-error rule is the doming detector in a report
• A mean far from zero, or above a quarter of the RMSE, says biased
• Students run exactly this test on their DSM residuals in the lab

Reading the reports

• Section 5: the week’s stated purpose
• Put their own 2B and 2C reports on screen here
• Slides give the map, the reports give the territory

A processing report, page by page

Survey overview, DEM, survey data with the overlap figure, camera calibration with residuals.

• Walk the four panels before the comparison table
• Survey data page carries GSD, coverage, aligned images
• Camera calibration residual plot is where a bad lens model shows
• None of these four pages is an accuracy statement

Where you declare your ruler

Marker accuracy is RMSE_V2. The software weights the adjustment with it, then reports a fit that still does not include it.

• This dialog is where the survey accuracy of the control goes in
• Marker accuracy 0.005 m says the targets are good to 5 mm
• Camera accuracy 10 m is the UX5 log: navigation grade, correctly downweighted
• Wrong numbers here mis-weight the whole adjustment
• Metashape still does not carry marker accuracy into the reported RMSE; you do that

The same facts in two reports

Quantity	Agisoft Metashape report	WebODM quality report
GSD	Survey Data: ground resolution	Overview: average GSD
Coverage and photos	Survey Data: coverage area, aligned images	Overview: area, reconstructed images
Overlap	Camera locations and image overlap figure	Survey Data overlap heatmap
Camera model	Calibrated values with uncertainties and residual plot	Camera parameters, per-band
Camera position error	Camera Locations table: X, Y, Z error	GPS/geolocation details, 3D errors
Control fit	Ground Control Points table: control RMSE	GCP errors section (when GCPs used)
Checkpoint error	Check Points rows in the same table	Absent unless checkpoints were declared

• Walk each row on the projector with a real pair of reports
• Trap 1: both tools print camera-position errors that students read as map accuracy
• Trap 2: WebODM reports only what it was given, so no declared checkpoints means no accuracy at all
• That silence does not mean the map is accurate
• The lab’s comparison table is this slide filled in with their numbers

What neither report tells you

• Whether the checkpoints were independent; the software trusts your labeling
• The survey accuracy of the control and checkpoints, so neither can do the quadrature for you
• Whether the error is spatially structured: a mean near zero can hide a dome that cancels
• Anything about the parts of the site with no checkpoints

Report vocabulary (American Society for Photogrammetry and Remote Sensing 2024): report RMSE_H and RMSE_V (RMSE_3D if asked); Edition 2 drops the old 95 percent confidence reporting; 30 checkpoints is the minimum for a statistical claim, and Section 7.16.1 gives the wording for when you have fewer

• Section 7.16.1 applies to nearly every project they will run
• Small sites cannot afford 30 checkpoints
• The answer is truth in reporting: give the RMSE, state the count, claim no confidence the sample cannot support
• The lab report requires that sentence

What can you measure

• Section 6: the same flight either answers the question or it does not
• The arithmetic decides

The level of detection

\[\text{LoD} = 1.96 \sqrt{\sigma_{z1}^2 + \sigma_{z2}^2} \approx 2.8\,\sigma_z\]

• Thresholding is for gross change only; it biases net change, where random errors already cancel (Anderson 2019)
• Spatially variable LoD can exclude from a quarter to more than half of observed volumetric change (Wheaton et al. 2010)

• Every accuracy number from the last fifty minutes gets plugged in here
• Wheaton’s Feshie numbers: 24 to 31 percent of change excluded in dry years, 57 to 59 in wet
• Anderson’s correction: gross change yes, net change no; error propagation is not thresholding
• Zhang 2019 reached sub-5 cm thresholds with PPK flights below 45 m

Trees or erosion

Question	Binding constraint	Verdict at 3 cm GSD, 5 cm sigma-z
Count the trees	Detection in the orthomosaic	Easy; accuracy barely matters
Measure canopy height	Canopy and ground surfaces, not GSD: flying height 80 to 120 m showed no significant effect on tree height RMSE (1.8 to 3.2 m) (Grybas and Congalton 2022)	Expect meters of error regardless of pixel size
Detect a 30 cm rill or gully	LoD about 14 cm	Yes, cleanly
Quantify 2 cm sheet erosion	LoD, dominated by systematic error	No; needs sub-2 cm sigma-z, near-ground flights, and bias control

• For trees the limiting error lives in the canopy and the terrain underneath, so more GSD buys nothing
• For micro-topography the limiting error is the whole chain built today
• Base coordinate, control, calibration, bias all show up in the LoD
• Students argue these verdicts with their own numbers in part three of the lab

Key terms

Checkpoint: surveyed, withheld, judges the map; also called check point, validation point, independent test point

Direct georeferencing: camera positions from RTK or PPK doing the work of control; also called GNSS-aided or GCP-free

Level of detection (LoD): the smallest elevation change distinguishable from error at a stated confidence

Doming: broad systematic DEM deformation from the self-calibration ambiguity; also called bowling when inverted

RMSE_H, RMSE_V, RMSE_3D: Edition 2’s reporting quantities, replacing RMSEx/RMSEy/RMSEz and the 95 percent statistics

Quadrature: accuracies add as the square root of summed squares; the product carries the checkpoint survey error

• The synonym lists matter: they will hit every variant in the literature
• Midterm vocabulary

Wrap-up

Assignment 3B: Validate the Lake Wheeler flight: establish the base coordinate three ways, evaluate the flight’s orthomosaic and DSM against the checkpoints you collected, and rule on what the flight can measure

Midterm 11/4: the control versus checkpoint distinction, the quadrature rule, and the LoD arithmetic are all fair game

• Recap the chain: GSD sets the floor
• Georeferencing and systematic error set the distance above it
• Checkpoints measure that distance
• Quadrature keeps the claim defensible
• LoD turns the chain into a yes or no about the question you care about
• The lab data is their own field work from 9/16

References

American Society for Photogrammetry and Remote Sensing. 2024. ASPRS Positional Accuracy Standards for Digital Geospatial Data, Edition 2, Version 2. American Society for Photogrammetry; Remote Sensing. https://publicdocuments.asprs.org/PositionalAccuracyStd-Ed2-V2.

Anderson, Scott W. 2019. “Uncertainty in Quantitative Analyses of Topographic Change: Error Propagation and the Role of Thresholding.” Earth Surface Processes and Landforms 44 (5): 1015–33. https://doi.org/10.1002/esp.4551.

Dai, Wen, Ruibo Qiu, Bo Wang, et al. 2023. “Enhancing UAV-SfM Photogrammetry for Terrain Modeling from the Perspective of Spatial Structure of Errors.” Remote Sensing 15 (17): 4305. https://doi.org/10.3390/rs15174305.

Famiglietti, Nicola Angelo, Gianpaolo Cecere, Carmine Grasso, Antonino Memmolo, and Annamaria Vicari. 2021. “A Test on the Potential of a Low Cost Unmanned Aerial Vehicle RTK/PPK Solution for Precision Positioning.” Sensors 21 (11): 3882. https://doi.org/10.3390/s21113882.

Grybas, Heather, and Russell G. Congalton. 2022. “Evaluating the Impacts of Flying Height and Forward Overlap on Tree Height Estimates Using Unmanned Aerial Systems.” Forests 13 (9): 1462. https://doi.org/10.3390/f13091462.

James, Mike R., Gilles Antoniazza, Stuart Robson, and Stuart N. Lane. 2020. “Mitigating Systematic Error in Topographic Models for Geomorphic Change Detection: Accuracy, Precision and Considerations Beyond Off-Nadir Imagery.” Earth Surface Processes and Landforms 45 (10): 2251–71. https://doi.org/10.1002/esp.4878.

James, Mike R., and Stuart Robson. 2014. “Mitigating Systematic Error in Topographic Models Derived from UAV and Ground-Based Image Networks.” Earth Surface Processes and Landforms 39 (10): 1413–20. https://doi.org/10.1002/esp.3609.

Janos, Daniel, and Przemysław Kuras. 2021. “Evaluation of Low-Cost GNSS Receiver Under Demanding Conditions in RTK Network Mode.” Sensors 21 (16): 5552. https://doi.org/10.3390/s21165552.

Jiménez-Jiménez, Sergio Iván, Waldo Ojeda-Bustamante, Mariana de Jesús Marcial-Pablo, and Juan Enciso. 2021. “Digital Terrain Models Generated with Low-Cost UAV Photogrammetry: Methodology and Accuracy.” ISPRS International Journal of Geo-Information 10 (5): 285. https://doi.org/10.3390/ijgi10050285.

Nesbit, Paul Ryan, and Christopher H. Hugenholtz. 2019. “Enhancing UAV-SfM 3D Model Accuracy in High-Relief Landscapes by Incorporating Oblique Images.” Remote Sensing 11 (3): 239. https://doi.org/10.3390/rs11030239.

Nota, E. W., W. Nijland, and T. de Haas. 2022. “Improving UAV-SfM Time-Series Accuracy by Co-Alignment and Contributions of Ground Control or RTK Positioning.” International Journal of Applied Earth Observation and Geoinformation 109: 102772. https://doi.org/10.1016/j.jag.2022.102772.

Sanz-Ablanedo, Enoc, Jim H. Chandler, José Ramón Rodríguez-Pérez, and Celestino Ordóñez. 2018. “Accuracy of Unmanned Aerial Vehicle (UAV) and SfM Photogrammetry Survey as a Function of the Number and Location of Ground Control Points Used.” Remote Sensing 10 (10): 1606. https://doi.org/10.3390/rs10101606.

Stott, Eilidh, Richard D. Williams, and Trevor B. Hoey. 2020. “Ground Control Point Distribution for Accurate Kilometre-Scale Topographic Mapping Using an RTK-GNSS Unmanned Aerial Vehicle and SfM Photogrammetry.” Drones 4 (3): 55. https://doi.org/10.3390/drones4030055.

Wheaton, Joseph M., James Brasington, Stephen E. Darby, and David A. Sear. 2010. “Accounting for Uncertainty in DEMs from Repeat Topographic Surveys: Improved Sediment Budgets.” Earth Surface Processes and Landforms 35 (2): 136–56. https://doi.org/10.1002/esp.1886.

Zhang, He, Emilien Aldana-Jague, François Clapuyt, Florian Wilken, Veerle Vanacker, and Kristof Van Oost. 2019. “Evaluating the Potential of Post-Processing Kinematic (PPK) Georeferencing for UAV-Based Structure-from-Motion (SfM) Photogrammetry and Surface Change Detection.” Earth Surface Dynamics 7 (3): 807–27. https://doi.org/10.5194/esurf-7-807-2019.