Deep-learning-based deflectometry for simultaneous multi-surface measurement of freeform refractive optics

Zhendong Wu, Daodang Wang, Jinchao Dou, Ming Kong, Lihua Lei, Rongguang Liang

Research output: Chapter in Book/Report/Conference proceedingConference contribution


Due to the highly general surface geometry of freeform optics, the measurement of freeform optical surfaces is still a challenging and rewarding issue. Here, we propose a simultaneous multi-surface measurement method based on deep learning for freeform refractive optics, in which the surfaces are reconstructed based on the transmitted wavefront measured with computer-Aided deflectometry. By adopting the deep learning approaches in geometrical error calibration and wavefront reconstruction, both the efficiency and robustness is significantly improved, and the surface measurement accuracy in the order of nanometers can be achieved. The proposed method provides an effective, robust and accurate way for testing freeform refractive optics with multiple surfaces and a large slope range

Original languageEnglish (US)
Title of host publicationOptical Design and Testing XI
EditorsYongtian Wang, Tina E. Kidger, Osamu Matoba, Rengmao Wu
ISBN (Electronic)9781510646391
StatePublished - 2021
Externally publishedYes
EventOptical Design and Testing XI 2021 - Nantong, China
Duration: Oct 10 2021Oct 12 2021

Publication series

NameProceedings of SPIE - The International Society for Optical Engineering
ISSN (Print)0277-786X
ISSN (Electronic)1996-756X


ConferenceOptical Design and Testing XI 2021


  • deep learning
  • deflectometry
  • system calibration
  • wavefront reconstruction.

ASJC Scopus subject areas

  • Electronic, Optical and Magnetic Materials
  • Condensed Matter Physics
  • Computer Science Applications
  • Applied Mathematics
  • Electrical and Electronic Engineering


Dive into the research topics of 'Deep-learning-based deflectometry for simultaneous multi-surface measurement of freeform refractive optics'. Together they form a unique fingerprint.

Cite this