Number of the records: 1
Subaperture stitching computation time optimization using a system of linear equations
- 1.0569037 - ÚFP 2023 RIV US eng J - Journal Article
Stašík, Marek - Psota, Pavel - Lédl, Vít - Kredba, Jan
Subaperture stitching computation time optimization using a system of linear equations.
Applied Optics. Roč. 60, č. 27 (2021), s. 8556-8568. ISSN 1559-128X. E-ISSN 2155-3165
R&D Projects: GA MŠMT(CZ) EF16_026/0008390
Institutional support: RVO:61389021
Keywords : stitching * linear equations * optimization
OECD category: Optics (including laser optics and quantum optics)
Impact factor: 1.905, year: 2021
Method of publishing: Limited access
https://opg.optica.org/ao/viewmedia.cfm?uri=ao-60-27-8556&seq=0
Measurement of large or aspheric optical surface shapes as a single aperture using interferometry is problematic for many reasons. A typical problem is the numerical aperture limitation of the interferometer transmission element and the surface slope deviation of aspheres. This deviation typically causes vignetting and spatial aliasing on the camera. A solution is subaperture measurement and subsequent subaperture stitching. A stitching algorithm, in principle, uses overlaps between subapertures to eliminate aberrations of each subaperture to obtain a full aperture for further analysis. This process is computation time demanding and requires optimization in order to obtain a result in a reasonable time to reduce, in turn, the overall manufacturing time. In this paper, a novel, to the best of our knowledge, and fast stitching method based on a system of linear equations is proposed and mathematically described. The developed method was compared with other algorithms, and theoretical computation complexity was calculated and compared. The method was tested practically, with real data measured on spherical surfaces using QED ASI (QED Technologies aspheric stitching interferometer) and an experimental interferometer, and the results are presented. Stitching quality was quantified for results and compared to other algorithms.
Permanent Link: https://hdl.handle.net/11104/0340348
Number of the records: 1