Strategies for processing images with 4D-Var data assimilation methods

Abstract : Data Assimilation is a well-known mathematical technic used, in environmental sciences, to improve, thanks to observation data, the forecasts obtained by meteorological, oceanographic or air quality simulation models. It aims to solve the evolution equations, describing the dynamics of the state variables, and an observation equation, linking at each space-time location the state vector and the observations. Data Assimilation allows to get a better knowledge of the actual system's state, named the reference. In this article, we first describe various strategies that can be applied in the framework of variational data assimilation to study various image processing issues. Second, we detail the mathematical setting and the analysis of pros and cons of each strategy for the issue of motion estimation. Last, results are provided on synthetic images and satellite acquisitions.
Document type :
Reports
Complete list of metadatas

Cited literature [13 references]  Display  Hide  Download

https://hal.inria.fr/inria-00546222
Contributor : Dominique Béréziat <>
Submitted on : Friday, December 17, 2010 - 7:09:31 PM
Last modification on : Thursday, March 21, 2019 - 1:00:34 PM
Long-term archiving on : Saturday, December 3, 2016 - 1:01:06 AM

File

RR-7495.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : inria-00546222, version 2

Citation

Isabelle Herlin, Dominique Béréziat, Nicolas Mercier. Strategies for processing images with 4D-Var data assimilation methods. [Research Report] RR-7495, INRIA. 2010, 23 p. ⟨inria-00546222v2⟩

Share

Metrics

Record views

541

Files downloads

153