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

Accueil du site → Master → Chine → 2008 → Study on Automatic Extraction of Remote Sense Information of Karst Rock-desertification Area

Guizhou Normal (2008)

Study on Automatic Extraction of Remote Sense Information of Karst Rock-desertification Area

Zuo Li Hui

Titre : Study on Automatic Extraction of Remote Sense Information of Karst Rock-desertification Area

Auteur : Zuo Li Hui

Grade : Master’s Theses 2008

Université : Guizhou Normal

Résumé
Rocky Desertification in Guizhou distribution and its causes and rocky desertification area of ecological environment comprehensive management mode has been basically mature , but based on how to use the remote sensing images in the determination of the spectral reflectance characteristics of different backgrounds , different rocky desertification rocky desertification information automatically extracted and quantitative analysis is currently no related research . The thesis aims to explore a use of remote sensing images automation , high efficiency , high precision feature information extracted from of rocky desertification complex context , reveal different levels of the rocky desertification spectral reflectance differences to better use remote sensing technology to obtain rocky desertification information , for rocky desertification governance provide a theoretical basis . The main research contents are as follows : ( 1) the spectral characteristics of the different rocky desertification and its background feature research ; ( 2 ) automatic extraction of the design of the model based on artificial neural network classification Rocky Desertification surface information ; (3) quantitative accuracy evaluation of rocky desertification information extraction . The paper presents the the Rocky Desertification automatic classification based on neural network classification method . Select Bijie demonstration area of the Duck Pond Stonebridge watershed study area , using a typical post - classification to artificial neural network classification , and to carry out the field of remote sensing survey , in order to improve and validate the classification accuracy . The results show that : The method classification to avoid interference and influence of the other extra information , which can improve the classification accuracy .

Mots clés : Karst Rocky Desertification Neural Networks Mixing Spectral Model Spectral characteristics

Présentation (Dissertationtopic )

Page publiée le 12 septembre 2014, mise à jour le 5 décembre 2017