Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Master → Chine → 2002 → A Study of Hyperspectral Remote Sensing Data Preprocessing and Spectral Re-building in Desertification Areas

Northeast Forestry University (2002)

A Study of Hyperspectral Remote Sensing Data Preprocessing and Spectral Re-building in Desertification Areas

DuHuaQiang

Titre : A Study of Hyperspectral Remote Sensing Data Preprocessing and Spectral Re-building in Desertification Areas

Auteur : DuHuaQiang

Grade : Master Thesis 2002

University : Northeast Forestry University

Résumé
In this study, the National Natural Science Foundation Project desertified areas Spectral reconstruction and quantitative inversion \Compared with the conventional remote sensing , hyperspectral remote sensing data processing and identification of target objects requires the use of some new technology and means . According to the the desertified areas the actual situation and the characteristics of the hyperspectral remote sensing data , data preprocessing scheme , derived from the theory and application of point of view the spectroradiometer reconstruction model , discussed and analyzed the pros and cons of data preprocessing and the accuracy and applicability of the spectral reconstruction model , achieved some results in the theory and practical application . The main work and innovation of this study include the following aspects : 1 . From the spectral characteristics of the statistical characteristics of the image and signal-to-noise ratio three aspects of hyperspectral remote sensing data band selection method . 2 . Tangent correction of image , we propose a based edge histogram matching radiation correction algorithm to correct differences in brightness of the images on the edge of the column and the center column . 3 . Try the application of wavelet analysis in the processing of hyperspectral remote sensing image noise . 4 . Feature spectral library of typical objects in the study area , respectively coverings at the center wavelength of the selected band spectral resampling . 5. Analysis of the spectral characteristics of the typical surface features in a spectral library , in particular, the vegetation of the spectral library spectra do derivative naturalization - and naturalization - derivative calculated after three treatments , to eliminate the same feature background , illumination differences spectral differences caused by such factors . 6. According to the theory of the composition of the image , on the basis of strict simulation of atmospheric radiative processes , the derivation and use of a linear model to reconstruct Spectral obtained satisfactory results for the target objects directly identify its quantitative retrieval of information laid foundation.

Mots clés : Desertification Hyperspectral Remote Sensing Data preprocessing Wavelet Analysis Spectroscopy Spectroradiometer reconstruction

Présentation (Dissertation topics)

Page publiée le 8 septembre 2014, mise à jour le 13 septembre 2017