Skip to main navigation Skip to search Skip to main content

Analysis of statistical properties of atmospheric turbulence-induced image dancing based on Hilbert transform and dense optical flow

  • Beihang University

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

Abstract

In this paper, the statistical properties of pixel displacements in turbulence degraded images are analyzed. Two main problems are addressed before that. One is the computation of pixel displacements. Dense optical flow is used since blur makes features like points and edges hard to track. The other one is selection of statistical samples. We use 2D-Hilbert transform to extract feature points, and only displacements at those points are considered. Statistical analysis includes distribution fitting, statistical parameters and normality test at different sample times and turbulence strengths. In the experiments, the method of computing distortions is first applied to simulated dataset to test its validity. Then this method of computing displacements and statistical analysis is applied to real-scene image sequences.

Original languageEnglish
Title of host publicationICSP 2016 - 2016 IEEE 13th International Conference on Signal Processing, Proceedings
EditorsYuan Baozong, Ruan Qiuqi, Zhao Yao, An Gaoyun
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages702-707
Number of pages6
ISBN (Electronic)9781509013449
DOIs
StatePublished - 2 Jul 2016
Event13th IEEE International Conference on Signal Processing, ICSP 2016 - Chengdu, China
Duration: 6 Nov 201610 Nov 2016

Publication series

NameInternational Conference on Signal Processing Proceedings, ICSP
Volume0

Conference

Conference13th IEEE International Conference on Signal Processing, ICSP 2016
Country/TerritoryChina
CityChengdu
Period6/11/1610/11/16

Keywords

  • Hilbert transform
  • atmospheric turbulence
  • image dancing
  • optical flow
  • statistical analysis

Fingerprint

Dive into the research topics of 'Analysis of statistical properties of atmospheric turbulence-induced image dancing based on Hilbert transform and dense optical flow'. Together they form a unique fingerprint.

Cite this