Acta Optica Sinica, Volume. 45, Issue 8, 0811001(2025)

Augmented Reality Freeform Prism Optical System Based on Back Propagation Neural Network

Yi Zhang1,2,3 and Da Mu1,2,3、*
Author Affiliations
  • 1School of Optoelectronic Engineering, Changchun University of Science and Technology, Changchun 130022, Jilin , China
  • 2Key Laboratory of Optoelectronic Measurement and Optical Information Transmission Technology, Ministry of Education, Changchun University of Science and Technology, Changchun 130022, Jilin , China
  • 3Zhongshan Institute, Changchun University of Science and Technology, Zhongshan 528403, Guangdong , China
  • show less
    Figures & Tables(12)
    Schematic diagram of three-dimensional structure of AR optical system. (a) Free-form prism; (b) overall optical system with compensation prism added
    Structural constraints of free-form prism system
    Dataset generation process
    Deep learning network architecture
    Deep learning network training result
    Average RMS spot size results of 100 random samples
    Comparison of before and after optimization. (a) Model prediction results; (b) results after further optimization
    Optical path RMS point array results for deep learning network model training samples
    Selection of optimized projection optical path RMS point array results for deep learning network model training samples
    MTF results of optimized projection optical system
    Three-dimensional image optimized with compensating prism added
    Optimization of mesh distortion map of perspective optical system after selecting training examples for deep learning network models
    Tools

    Get Citation

    Copy Citation Text

    Yi Zhang, Da Mu. Augmented Reality Freeform Prism Optical System Based on Back Propagation Neural Network[J]. Acta Optica Sinica, 2025, 45(8): 0811001

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: Imaging Systems

    Received: Jan. 13, 2025

    Accepted: Feb. 10, 2025

    Published Online: Apr. 16, 2025

    The Author Email: Da Mu (muda@cust.edu.cn)

    DOI:10.3788/AOS250483

    CSTR:32393.14.AOS250483

    Topics