Paper
13 December 2021 Piston error detection technology for optical sparse aperture system via transfer learning
Jiawen Li, Rongzhu Zhang
Author Affiliations +
Proceedings Volume 12070, 10th International Symposium on Advanced Optical Manufacturing and Testing Technologies: Large Mirror and Telescopes; 1207003 (2021) https://doi.org/10.1117/12.2604013
Event: Tenth International Symposium on Advanced Optical Manufacturing and Testing Technologies (AOMATT 2021), 2021, Chengdu, China
Abstract
Optical sparse aperture systems often need to correct the piston error between sub-apertures to obtain high-resolution imaging quality. This paper constructs a deep convolutional neural network via transfer learning to perform piston error sensing on a single focal plane image. First, a batch of data sets that introduce piston error generates according to the principle of sparse aperture imaging. The image features are extracted through the pre-training network, and the mapping relationship between the features and the corresponding piston error values is established. When a given single image is sent to the network for processing, then get corresponding piston error value via search with the nearest neighbors. The simulation results verify the method's effectiveness and measure the piston error of the 1λ dynamic range. Because this detection method can perform piston error sensing tasks without introducing additional measuring equipment, it dramatically reduces the complexity and hardware cost of the optical system. It may be used as a new piston error detection technology in the future.
© (2021) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Jiawen Li and Rongzhu Zhang "Piston error detection technology for optical sparse aperture system via transfer learning", Proc. SPIE 12070, 10th International Symposium on Advanced Optical Manufacturing and Testing Technologies: Large Mirror and Telescopes, 1207003 (13 December 2021); https://doi.org/10.1117/12.2604013
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Imaging systems

Feature extraction

Point spread functions

Convolutional neural networks

Neural networks

Error analysis

Mirrors

Back to Top