Towards Accurate Photogrammetry Using Molded Markers
Author Information
Author(s): Auzmendi Iriarte Iñigo, Saez de Egilaz Oier, Gonzalez de Alaiza Martinez Pedro, Herrera Imanol
Primary Institution: IDEKO Research Center, Basque Research and Technology Alliance (BRTA)
Hypothesis
Can molded markers improve the accuracy and usability of photogrammetry systems?
Conclusion
The study demonstrates that molded markers can enhance the accuracy of photogrammetry systems, achieving a median accuracy of 0.170 mm/m.
Supporting Evidence
- The study achieved a median accuracy of 0.170 mm/m with molded markers.
- Polarimetric images improved detection rates significantly under poor lighting conditions.
- Data augmentation techniques enhanced the training of the detection model.
Takeaway
This study shows that using special molded markers instead of stickers can make measuring things in 3D easier and more accurate.
Methodology
A two-stage detection algorithm was developed, combining deep learning for initial detection and classical methods for accurate localization.
Limitations
The molded markers may not be suitable for highly curved surfaces or areas with extreme deformations.
Statistical Information
Confidence Interval
0.069 to 0.368 mm/m
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website