Edge-preserving rain removal for light fiele images based on RPCA

Cheen Hau Tan, Jie Chen, Lap Pui Chau

Research output: Chapter in book / Conference proceedingConference article published in proceeding or bookAcademic researchpeer-review

5 Citations (Scopus)

Abstract

Rain deteriorates outdoor vision and causes challenge for most vision based intelligent systems. In this paper we propose a method to efficiently remove the rain present in light field data. Firstly, the sub-view image sequence is globally aligned to the central view. Robust Principle Component Analysis (RPCA) are then applied to decompose the sequence into two parts, i.e., the low-rank data, and the sparse data. The decomposed sparse data contains both rain streaks and scene disparity edges. We propose to compute a dark view image to estimate the non-rain disparity edges, and the remaining part of the decomposed sparse data will be considered as rain. The disparity edges will then be added back to the low-rank data. The proposed method produces satisfactory rain removal visual results, and can efficiently preserve the light field perspective disparity at the same time.

Original languageEnglish
Title of host publication2017 22nd International Conference on Digital Signal Processing, DSP 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781538618950
DOIs
Publication statusPublished - 3 Nov 2017
Externally publishedYes
Event2017 22nd International Conference on Digital Signal Processing, DSP 2017 - London, United Kingdom
Duration: 23 Aug 201725 Aug 2017

Publication series

NameInternational Conference on Digital Signal Processing, DSP
Volume2017-August

Conference

Conference2017 22nd International Conference on Digital Signal Processing, DSP 2017
Country/TerritoryUnited Kingdom
CityLondon
Period23/08/1725/08/17

Keywords

  • Dark View Image
  • Light Field
  • Rain Removal
  • Robust PCA

ASJC Scopus subject areas

  • Signal Processing

Fingerprint

Dive into the research topics of 'Edge-preserving rain removal for light fiele images based on RPCA'. Together they form a unique fingerprint.

Cite this