Acta Optica Sinica, Volume. 43, Issue 21, 2111002(2023)

Reconstruction of Non-Line-of-Sight Depth Data Using Genetic Algorithm-Lucy-Richardson Based on Time of Flight Camera

Yujie Fang1,2,3, Xia Wang1,2,3、*, and Binghua Su2,3
Author Affiliations
  • 1Key Laboratory of Optoelectronic Imaging Technology and System, Ministry of Education, School of Optics and Photonics, Beijing Institute of Technology, Beijing 100081, China
  • 2Beijing Institute of Technology, Zhuhai 519085, Guangdong , China
  • 3Key Laboratory of Intelligent Detection in Complex Environment of Aerospace, Land, and Sea , Zhuhai 519085, Guangdong , China
  • show less
    Figures & Tables(12)
    Influence of surface relay characteristics of materials on non-line-of-sight (NLOS) imaging
    Flow chart of target reconstruction in NLOS imaging
    Experimental scenes and their locations
    Measured targets and their corresponding depth images collected by flat mirror
    Scattering characteristics of PP plate and PMMA plate and raw depth data collected from different material relay surfaces. (a) PP plate and part of its BRDF curve; (b) PMMA plate and part of its BRDF curve; (c) mirror; (d) ceramic tile; (e) PP plate; (d) PMMA plate
    Fitness curves of approximate optical model solutions for three materials
    Approximate solution for relay surface obtained by genetic algorithm. (a) Approximate solution for optical properties of ceramic tile; (b) approximate solution for optical properties of PP plate; (c) approximate solution for optical properties of PMMA plate
    Row depth data, average data, error data, and reconstruction results under three sizes of fbs characteristics with different targets using PP plate as relay surface. (a) Gypsum cone; (b) smooth plastics; (c) natural complex scenes
    Row depth data, average data, error data, and reconstruction results under three sizes of fbs characteristics with different targets using PMMA plate as relay surface. (a) Gypsum cone; (b) smooth plastics; (c) natural complex scenes
    • Table 1. Basic parameter setting for experimental scenario

      View table

      Table 1. Basic parameter setting for experimental scenario

      ParameterResolution /pixelModulation /MHz

      Illumination power /W

      Depth accuracy /mmTarget size /cm

      Focal length /mm

      Value320×240

      46

      55

      Avg 8520×2022
    • Table 2. Mean square error after NLOS image reconstruction based on PP plate relay surface

      View table

      Table 2. Mean square error after NLOS image reconstruction based on PP plate relay surface

      Raw dataDenoised7×7 model10×10 model15×15 model
      Gypsum cone0.03530.01940.01720.01870.0105
      Smooth plastics0.04630.03520.00940.01050.0103
      Natural complex scenes0.05930.04910.01930.01950.0253
    • Table 3. Mean square error after NLOS image reconstruction based on PMMA plate relay surface

      View table

      Table 3. Mean square error after NLOS image reconstruction based on PMMA plate relay surface

      Raw dataDenoised7×7 model10×10 model15×15 model
      Gypsum cone0.02460.01690.00750.00720.0074
      Smooth plastics0.03420.02070.00790.01050.0078
      Natural complex scenes0.05620.02410.00740.00750.0074
    Tools

    Get Citation

    Copy Citation Text

    Yujie Fang, Xia Wang, Binghua Su. Reconstruction of Non-Line-of-Sight Depth Data Using Genetic Algorithm-Lucy-Richardson Based on Time of Flight Camera[J]. Acta Optica Sinica, 2023, 43(21): 2111002

    Download Citation

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

    Category: Imaging Systems

    Received: Apr. 24, 2023

    Accepted: Jun. 14, 2023

    Published Online: Nov. 16, 2023

    The Author Email: Wang Xia (angelniuniu@bit.edu.cn)

    DOI:10.3788/AOS230870

    Topics