How AccuracySuite Works

1. What is AccuracySuite?

AccuracySuite is a standalone computation tool that takes surveyed checkpoint coordinates (truth values) and model/measured coordinates at those same points, then computes positional accuracy metrics per ASPRS Edition 2. It is not a certification, not a licensed land survey, and not a substitute for a surveyor's stamp. It operates on coordinates and statistics only — no imagery, no point clouds, no photogrammetry processing.

2. The Standard

ASPRS Positional Accuracy Standards for Digital Geospatial Data, Edition 2 (formally published 2023, with Version 2 updates in 2024) defines:

• How to compute horizontal and vertical positional accuracy from independent checkpoints
• Standard accuracy classes (1, 2, 2.5, 5, 7.5, 10, 12.5, 15, 17.5, 20, 25, 30, 50, 100 cm)
• The 30-checkpoint threshold for "Tested to meet" vs "Produced to meet" language
• CE90 and LE90 confidence-level conversions
• Requirements for independent checkpoints of higher accuracy

3. Formulas

All formulas are from ASPRS Edition 2 conventions. No AI, no machine learning — pure deterministic mathematics.

Where n = number of qualifying independent checkpoints, and residuals are computed as (model − truth) in the selected coordinate system units, then converted to centimeters.

4. Accuracy Class Determination

The ASPRS accuracy class is the smallest standard class value that is greater than or equal to the computed RMSE. For example:

• RMSEr = 4.97 cm → Horizontal class = 5.0-cm (next class ≥ 4.97)
• RMSEz = 6.1 cm → Vertical class = 7.5-cm (next class ≥ 6.1)

Standard classes: 0.5, 1.0, 1.25, 2.0, 2.5, 5.0, 7.5, 10.0, 12.5, 15.0, 17.5, 20.0, 25.0, 30.0, 50.0, 100.0 cm.

5. "Tested to Meet" vs "Produced to Meet"

ASPRS Edition 2 distinguishes two confidence levels of accuracy reporting based on the number of independent checkpoints:

"Produced to meet" does NOT mean the data is less accurate — it means the confidence statement cannot assert the same statistical certainty with fewer independent observations.

6. Blunder Detection

Points with horizontal radial residuals (dR) exceeding 3× the dataset median are flagged as potential blunders. Blunders may indicate:

• Surveying errors at the checkpoint (wrong point measured)
• Processing errors (point misidentification in software)
• CRS or unit mismatches between truth and model data
• Transcription errors in coordinate values

Users can exclude flagged blunders from computation. Excluded points reduce the checkpoint count and may affect the Tested/Produced gate.

7. CRS, Units, and Datum

Correct CRS and unit selection is critical for accurate results.

Vertical datum (orthometric via geoid model, or ellipsoidal) must match between truth and model coordinates. Mixing ellipsoidal heights with orthometric checkpoints will produce incorrect vertical RMSE.

8. Versioning

Every report stamps two version identifiers for reproducibility:

Given the same inputs and the same formula bundle version, AccuracySuite will always produce identical outputs. This enables audit and reproducibility.

9. Limitations

• Results are limited by the accuracy of the supplied truth/checkpoint coordinates.
• CE90 and LE90 conversions assume residuals are approximately normally distributed.
• The tool does not assess checkpoint survey accuracy independently — the user must ensure checkpoints are of higher accuracy than the target class.
• No imagery, point cloud, or orthomosaic processing is performed.
• This is NOT a certification, NOT a licensed survey, and NOT a substitute for a surveyor's professional judgment.

10. References

• ASPRS. (2023). Positional Accuracy Standards for Digital Geospatial Data, Edition 2. American Society for Photogrammetry and Remote Sensing.
• ASPRS. (2024). Positional Accuracy Standards for Digital Geospatial Data, Edition 2, Version 2. American Society for Photogrammetry and Remote Sensing.