Multiframe super resolution reconstruction method based on light field angular images

  • Shubo Zhou
  • , Yan Yuan*
  • , Lijuan Su
  • , Xiaomin Ding
  • , Jichao Wang
  • *Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

The plenoptic camera can directly obtain 4-dimensional light field information from a 2-dimensional sensor. However, based on the sampling theorem, the spatial resolution is greatly limited by the microlenses. In this paper, we present a method of reconstructing high-resolution images from the angular images. First, the ray tracing method is used to model the telecentric-based light field imaging process. Then, we analyze the subpixel shifts between the angular images extracted from the defocused light field data and the blur in the angular images. According to the analysis above, we construct the observation model from the ideal high-resolution image to the angular images. Applying the regularized super resolution method, we can obtain the super resolution result with a magnification ratio of 8. The results demonstrate the effectiveness of the proposed observation model.

Original languageEnglish
Pages (from-to)189-195
Number of pages7
JournalOptics Communications
Volume404
DOIs
StatePublished - 1 Dec 2017

Keywords

  • Angular image
  • Observation model
  • Plenoptic imaging
  • Super resolution

Fingerprint

Dive into the research topics of 'Multiframe super resolution reconstruction method based on light field angular images'. Together they form a unique fingerprint.

Cite this