Scan-to-scan deviation refers to the positional error between two overlapping scans after registration.
When we talk about accuracy of point cloud data, you typically get told the accuracy of the Laser Scanner itself ie from a single scan position which can be as accurate as +/- 1mm for a terrestrial laser scanner, with near perfect conditions.
Scan-to-scan deviation refers to the positional error between two overlapping scans after registration.
This is only relevant if in fact you are only having a single scan, but as soon as you have multiple scan positions there is a deviation factor to factor.
When a survey has been completed the collected point cloud data, gets registered together to build a full 3D picture of the site / asset. The registering process is accomplished by either:
Targets – Physical reference spheres and checkerboard targets are used to help triangulate the scanners movement from position to position.
Checkerboard Target
Cloud to Cloud – technology which aligns multiple 3D laser scans by matching overlapping data points, eliminating the need for physical targets.
After registering has been completed the operator will be given a report, depicting the Overall Cloud-to-Cloud Error, now this is an average across all the scan positions and most reports will also go into further details including the Overlap Percentage and the Confidence Percentage between each scan to scan.
When we talk about overlapping point cloud data, we are talking about the percentage of point cloud data which is in both clouds, this goes in to further analysis including the percentage of points which are found in each of the X, Y and Z axis.
The confidence score is an amalgamation of overlapping points, the registration error between the two scans, feature matching quality and sometimes will include other metrics like noise and point density.
High Overlap + Low Error → High Confidence
Low Overlap + High Error → Low Confidence
As most registration work is now done with cloud to cloud algorithms, we also need to check the checks, the cloud to cloud registration will try to find similarities in the point cloud data between scans, but if your scanning an area which doesn’t have a lot of overlapping features then confidence will be low and there will be a greater chance for mis-alignment.
But there have also been issues, with registration when there is a lot of overlapping points, but the area you have surveyed has a lot of similar areas, like a row of 8 tanks which are evenly spaced out, its easy to have a scan mis-place itself.
At Neonex Digital, we don’t just rely on automated registration—we manually verify all alignments to ensure precision and provide clients with detailed proof of accuracy.
Image showing two scans which have registered incorrectly to each other.
The true accuracy of your survey, isn’t purely depicted by the accuracy of the hardware, but is a combination:
(Scanner Accuracy x Calibration Quality x Environmental Factors) / (Scan-to-Scan Deviation + Registration Error)
When it comes to accuracy there is also a large factor around requirements, ie what accuracy is required for the job? Analysis and Design work requires a higher level of accuracy and confidence, where other jobs may only require a 3-5mm accuracy.
When it comes to accuracy, there is also a large factor around requirements—what level of accuracy is actually needed for the job? Analysis and design work demand higher accuracy and confidence, whereas other projects may only require 3-5mm accuracy.
Accuracy vs. Time vs. Cost
Accuracy takes time. If ultra-precise data isn’t necessary, we adjust our equipment or settings to prioritize speed, allowing us to complete the job faster and more cost-effectively. This ties directly into the business triangle of Quality, Time, and Cost:
At Neonex Digital, we tailor our approach to each project, ensuring the right balance of accuracy, efficiency, and budget. We take the registration process seriously—while technology has advanced, we believe manual verification is essential to maintaining quality, and we provide proof of accuracy to our clients.
