Dot-coded structured light for accurate and robust 3D reconstruction

Authors:
  • FGFeifei Gu
    Feifei Gu
  • HCHuazhao Cao
    Huazhao Cao
  • ZSZhan Song
    Zhan Song
  • PXPengju Xie
    Pengju Xie
  • JZJuan Zhao
    Juan Zhao
  • JLJing Liu
    Jing Liu

Abstract
Speckle dots have the advantage of easy projection, which makes them good candidate features of structured light (SL) cameras, such as Kinect v1. However, they generally yield poor accuracy due to block matching. To improve their accuracy, this paper proposes a dot-coded SL, the coding information of which is added into dot distribution. Some of the dots are arranged regularly to provide easy-to-locate corner features, while others are specially designed to form different shapes of unique identification. A Gaussian-cross module and a simplified ResNet have been proposed to conduct robust decoding. Various experiments are performed to verify the accuracy and robustness of our framework.
Funding Information
  • Key-Area Research and Development Program of Guangdong Province, China (2019B010149002)
  • Natural Science Foundation of Guangdong Province (2020A1515010486, 2018A030310668)
  • Natural Science Foundation of Shenzhen Municipality (JCYJ20190806171403585, JSGG20191129114035610)
  • Dongguan City Core Technology Research Frontier Project (2019622101001)

This publication has 32 references indexed in Scilit: