Font Size: a A A

Statistic Model And Simulation Of Infrared Cloudy Sky Images

Posted on:2006-10-09Degree:MasterType:Thesis
Country:ChinaCandidate:C LiuFull Text:PDF
GTID:2178360182469195Subject:Pattern Recognition and Intelligent Systems
Abstract/Summary:PDF Full Text Request
This thesis gives a thorough research of infrared cloudy sky images often seen as backgrounds of remote sensing images and sets up the corresponding models. Furthermore, based on these models, we develop various algorithms to obtain the simulated infrared cloudy sky images. This research provides tasks such as remote sensing image processing and target recognition with pre-knowledge of background clutters. First in this paper, classical methods are first used to analyze statistical features such as Fourier power spectrum, grey distribution and etc. Their models are given separately. We also design corresponding algorithms to do simulation experiments both in space domain and frequency domain. Second, we use markov random field to do the work of modeling and simulation. Last but not least we introduce multi-resolution method to analysis the target images. The marginal statistics of sub band images are researched by decomposing the target image using Steerable Pyramid wavelets before we use Bessel K function to model those frequency bands separately and carry out experiments based on the models. In order to get more statistic features involved, the wavelet parameters of infrared cloudy sky images are directly put on the table after that. A set of self-contained features is extracted and simulation experiments are done by feature matching. At last we make the conclusion that the models proposed are effective and the simulation algorithms are useful.
Keywords/Search Tags:model, simulation, power spectrum, grey distribution, Steerable Pyramid decomposition, marginal statistics, Bessel K function, feature matching
PDF Full Text Request
Related items