Skip to main navigation Skip to search Skip to main content

Monocular Dense 3D Reconstruction Algorithm Based on Inverse Depth Filter

  • Jingyun Duo
  • , Long Zhao*
  • *Corresponding author for this work
  • Beihang University

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In this paper, an inverse depth filter based real-time monocular 3D dense reconstruction algorithm is proposed. We construct the Gaussian-uniformly mixed probability distribution model by fusing the right and wrong observations in inverse depth filter, which effectively solves the “long tail” problem in conventional depth filter. Then a fast smoothing method is adopted to reduce the effects of noises and false observations, which greatly improves the accuracy and robustness of our method. Experiments are designed to evaluate the accuracy and effectiveness of the proposed algorithm, and the results show that our algorithm has good 3D dense reconstruction performance and high computational accuracy.

Original languageEnglish
Title of host publicationProceedings of 2020 Chinese Intelligent Systems Conference - Volume II
EditorsYingmin Jia, Weicun Zhang, Yongling Fu
PublisherSpringer Science and Business Media Deutschland GmbH
Pages203-211
Number of pages9
ISBN (Print)9789811584572
DOIs
StatePublished - 2021
EventChinese Intelligent Systems Conference, CISC 2020 - Shenzhen, China
Duration: 24 Oct 202025 Oct 2020

Publication series

NameLecture Notes in Electrical Engineering
Volume706 LNEE
ISSN (Print)1876-1100
ISSN (Electronic)1876-1119

Conference

ConferenceChinese Intelligent Systems Conference, CISC 2020
Country/TerritoryChina
CityShenzhen
Period24/10/2025/10/20

Keywords

  • 3D reconstruction
  • De-noising
  • Inverse depth
  • Monocular camera

Fingerprint

Dive into the research topics of 'Monocular Dense 3D Reconstruction Algorithm Based on Inverse Depth Filter'. Together they form a unique fingerprint.

Cite this