شماره راهنما :
1156 دكتري
پديد آورنده :
عبدالملكي، مهدي
عنوان :
شناسايي الگوي رفتار طيفي كانسارهاي مس پورفيري با استفاده از داده‌هاي سنجش از دور فرا طيفي و روش‌هاي هوشمند
گرايش تحصيلي :
اكتشاف معدن
محل تحصيل :
اصفهان: دانشگاه صنعتي اصفهان، دانشكده معدن
صفحه شمار :
چهارده، ۱۴۰ص.: مصور، جدول، نمودار
استاد راهنما :
مرتضي طبايي، نادر فتحيان‌پور
توصيفگر ها :
انتخاب موجك مادر بهينه , تبديل موجك سه بعدي , تحليل مؤلفه‌هاي اصلي , تصاوير فرا طيفي , شبكه‌هاي عصبي , كانسارهاي مس پورفيري , ماشين بردار پشتيبان , نقشه‌بردار زاويه طيفي
استاد داور :
همايون صفايي، رضا جعفري، حسن طباطبايي
تاريخ ورود اطلاعات :
1396/12/19
كد ايرانداك :
ID1156 دكتري
چكيده انگليسي :
141Degree PhDTitle Spectral pattern recognition of porphyry copper deposits using hyperspectral remote sensing data and intelligence techniquesAuthor Mehdi AbdolmalekiSupervisor Dr Morteza Tabaei Dr Nader FathinpourDepartment Mining EngineeringDate February 5 2018Language PersianAbstract The increasing demands for base metals such as copper and also the reduction of thesuperficial resources of these elements have led to more advanced and sophisticatedmethods in exploratory studies of these mineral deposits Remote sensing is considered asan accurate rapid and low cost observatory system for mapping surface mineralalterations Replacing the classical methods with intelligent processing technique isnecessary to achieve the approperiate accuracy The objective of the current study is to provide a spectral pattern using intelligenttechniques and hyperspectral images for exploring the porphyry copper mineralization The study area covered by Hyperion data contains two well known porphyry copperdeposits Darrehzar and Sarcheshmeh After geometric and radiomeric corrections wavelettransform was used to denoising Hyperion image with implementing hard and softthreshold filtering To select optimum base wavelet energy criterion and matching shape criterion wereimplemented on three wavelet series covering Daubechie db symlet sym and coiflet coif High ranking base wavelets in mentioned criteria coif1 db3 and db7 wererecommended to be utilized in hyperspectral image classification Neural Networkclassifier was used for classification Based on classification error matrix and fieldmeasured data db7 introduced as an appropriate base wavelet for detecting porphyrycopper deposits The performance of two feature extraction methods including the Discrete WaveletTransform DWT and Principal Component Analysis PCA were compared The firstmethod is a powerful decomposition tool and a popular time frequency method onhyperspectral data and a second one is a widely used technique in feature extraction ofmineral mapping The effect of the number of the selected features extracted by DWT andPCA techniques on the various classification methods were evaluated Neural Network NN Support Vector Machines SVM and Spectral Angle Mapper SAM classifiers are
استاد راهنما :
مرتضي طبايي، نادر فتحيان‌پور
استاد داور :
همايون صفايي، رضا جعفري، حسن طباطبايي
