Share Email Print
cover

Proceedings Paper

Adaptive merging of large datasets of a 3D measuring endoscope in an industrial environment
Author(s): Lennart Hinz; Markus Kästner; Eduard Reithmeier
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

A 3D measuring endoscope with a small measuring head and parallel arrangement of the fibers can be guided into forming plants and carry out precise measurements of geometries which are unreachable for most three-dimensional measuring systems. The data obtained can be used to quantify the wear of highly stressed structures and thus provide information for maintenance. Due to the compact sensor design and the required accuracy, optics with small working distance and a small measuring volume are used. In addition to in situ single measurements of highly stressed structures, over a hundred individual measurements are conceivable in order to convert large and complex geometries into point clouds. Besides the robust and accurate registration of all measurements, merging is one of the main causes of inaccurate measurement results. Conventional merging algorithms merge all points within a voxel into a single point. Due to the large overlap areas required for registration, points of diverse quality are averaged. In order to perform an improved adaptive merging, it is necessary to define metrics that robustly identify only the good points in the overlapping areas. On the one hand, the 2D camera sensor data can be used to estimate signal-based the quality of each point measured. Furthermore, the 3D features from the camera and projector calibration can evaluate the calibration of a triangulated point. Finally, the uniformity of the point cloud can also be used as a metric. Multiple measurements on features of a calibrated microcontour standard were used to determine which metrics provide the best possible merging.

Paper Details

Date Published: 1 April 2020
PDF: 14 pages
Proc. SPIE 11352, Optics and Photonics for Advanced Dimensional Metrology, 113520G (1 April 2020); doi: 10.1117/12.2553939
Show Author Affiliations
Lennart Hinz, Institute of Measurement and Automatic Control (Germany)
Markus Kästner, Institute of Measurement and Automatic Control (Germany)
Eduard Reithmeier, Institute of Measurement and Automatic Control (Germany)


Published in SPIE Proceedings Vol. 11352:
Optics and Photonics for Advanced Dimensional Metrology
Peter J. de Groot; Richard K. Leach; Pascal Picart, Editor(s)

© SPIE. Terms of Use
Back to Top
PREMIUM CONTENT
Sign in to read the full article
Create a free SPIE account to get access to
premium articles and original research
Forgot your username?
close_icon_gray