Analyse de texture et lissage optimal des images thermographiques par satellite

Abstract : Satellites are observing the Earth and acquire images of surface temperature. Study of these images request numerical processing. This work analyzes various statistical images, descriptive as well as analytical, that may help in interpreting and possibly improving images. The data are presented. It is shown that descriptive methods such as histograms, may help in visualizing images and in obtaining a first classification of their content. Texture of images is then studied using the concept of random function with support in R2. By applying data analysis, a method for segmenting images into zones of homogeneous texture is built. This method is used to develop an algorithm for filtering the noise affecting the images. It is based on an optimal mobile average and is adaptative to the local texture. All algorithms have been implemented in a computer and several application examples are shown.
Document type :
Theses
Complete list of metadatas

https://pastel.archives-ouvertes.fr/pastel-00949046
Contributor : Lucien Wald <>
Submitted on : Wednesday, February 19, 2014 - 9:05:06 AM
Last modification on : Thursday, February 7, 2019 - 1:32:19 AM
Long-term archiving on : Monday, May 19, 2014 - 11:01:26 AM

Identifiers

  • HAL Id : pastel-00949046, version 1

Citation

Michel Albuisson. Analyse de texture et lissage optimal des images thermographiques par satellite. Sciences de l'ingénieur [physics]. Université Pierre et Marie Curie - Paris VI, 1976. Français. ⟨pastel-00949046⟩

Share

Metrics

Record views

366

Files downloads

168