Main > Chinese Journal of Lasers >  Volume 47 >  Issue 12 >  Page 1204005 > Article
  • Abstract
  • Abstract
  • View Summary
  • Figures (10)
  • Tables (3)
  • Equations (0)
  • References (25)
  • Get PDF(in Chinese)
  • Paper Information
  • Received: Jun. 9, 2020

    Accepted: Jul. 27, 2020

    Posted: Dec. 1, 2020

    Published Online: Nov. 18, 2020

    The Author Email: Yang Ping (pingyang2516@163.com)

    DOI: 10.3788/CJL202047.1204005

  • Get Citation
  • Copy Citation Text

    Junlong Wu, Zhenghua Guo, Xianfeng Chen, Shuai Ma, Xu Yan, Licheng Zhu, Shuai Wang, Ping Yang. Three-Dimensional Measurement Method of Light Field Imaging Based on Deep Learning[J]. Chinese Journal of Lasers, 2020, 47(12): 1204005

    Download Citation

  • Category
  • measurement and metrology
  • Share
Chinese Journal of Lasers, Vol. 47, Issue 12, 1204005 (2020)

Three-Dimensional Measurement Method of Light Field Imaging Based on Deep Learning

Wu Junlong1,2,3, Guo Zhenghua1,2,3, Chen Xianfeng1,2,3, Ma Shuai1,2,3, Yan Xu1,2,3, Zhu Licheng1,2,3, Wang Shuai1,3, and Yang Ping1,3,*

Author Affiliations

  • 1Key Laboratory on Adaptive Optics, Chinese Academy of Sciences, Chengdu, Sichuan 610209, China
  • 2Institute of Optics and Electronics, Chinese Academy of Sciences, Chengdu, Sichuan 610209, China
  • 3University of Chinese Academy of Sciences, Beijing 100049, China

Abstract

To estimate the accurate disparity in weak texture region and fine structure region when the light field camera is used for three-dimensional measurement, a model of the light field depth estimation based on deep learning technology is proposed. Moreover, the relationship between the disparity and corresponding depth is also established. The proposed method is applied to a variety of complex scenes, and the experimental results show that the proposed method can accurately estimate the disparity information in the weak texture region and fine structure region, and leading to a good reconstruction of three-dimensional structure. The processing time of the proposed method is compressed to the order of 1s, which is 1 to 2 orders of magnitude lower than the traditional methods based on cost optimization.

keywords

Please Enter Your Email: