Investigation of 3D Scanning Data Quality
Islam, Shaiful (2025)
Islam, Shaiful
2025
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2025122938992
https://urn.fi/URN:NBN:fi:amk-2025122938992
Tiivistelmä
This study investigates how selected 3D scanning parameters influence the dimensional accuracy of a Fused Deposition Modelling printed test part. The study focuses on three factors: object colour, ambient light level and mesh setting in the scanning software. A simple trapezoidal test part was modelled in SolidWorks and printed in four colours (copper, blue, red and white) using a Creality K2 Plus FDM printer with identical process parameters. The parts were scanned with an EinScan HX hand-held 3D scanner in laser mode under three light intensity levels (0 lux, 4–9 lux and 45–77 lux). For each scan, two STL meshes were generated in EXScan HX using different meshing options: one without filtering or smoothing (NONEMESH) and one with medium filtering, smoothing and spike removal (MEDMESH). The meshes were compared to the original CAD model in GOM Inspect using both full surface colour maps and section-wise deviation analysis along selected edges. Deviation values were exported to Excel, where average deviation and average absolute deviation were calculated for each of the 36 unique scan setups.
The results show that object colour has the strongest influence on accuracy, followed by light level, while the mesh setting has a smaller but systematic effect. Copper and blue parts give clearly smaller mean absolute deviations than red and especially white parts. Across all setups, the mean absolute deviations mostly fall between 90 and 230 µm (0.09–0.23 mm). The best combination where the copper part was scanned at the high light level (45–77 lux) with MEDMESH and it reaches a mean absolute deviation of 86.71 µm, whereas the worst combination is the white part scanned at the same light level with NONEMESH, which reaches 233.42 µm.
These values are comparable to previously reported dimensional errors for workflows that combine FDM printing and optical 3D scanning. The study provides practical guidelines for improving 3D-scanning workflows in the Arcada lab, but the conclusions are limited to the four tested colours and the two meshing strategies used.
The results show that object colour has the strongest influence on accuracy, followed by light level, while the mesh setting has a smaller but systematic effect. Copper and blue parts give clearly smaller mean absolute deviations than red and especially white parts. Across all setups, the mean absolute deviations mostly fall between 90 and 230 µm (0.09–0.23 mm). The best combination where the copper part was scanned at the high light level (45–77 lux) with MEDMESH and it reaches a mean absolute deviation of 86.71 µm, whereas the worst combination is the white part scanned at the same light level with NONEMESH, which reaches 233.42 µm.
These values are comparable to previously reported dimensional errors for workflows that combine FDM printing and optical 3D scanning. The study provides practical guidelines for improving 3D-scanning workflows in the Arcada lab, but the conclusions are limited to the four tested colours and the two meshing strategies used.
