DS4203 Multispectral Signal Analysis Syllabus:
DS4203 Multispectral Signal Analysis Syllabus – Anna University PG Syllabus Regulation 2021
COURSE OBJECTIVES:
To know the basics of hyperspectral sensors and applications.
To know the concept of mutual information.
To provide knowledge of independent component analysis
To familiarize the student with SVM,MRF
UNIT I HYPERSPECTRAL SENSORS AND APPLICATIONS
Introduction, Multi-spectral Scanning Systems (MSS),Hyperspectral Systems, Airborne sensors, Spaceborne sensors, Ground Spectroscopy, Software for Hyperspectral Processing, Applications, Atmosphere and Hydrosphere, Vegetation, Soils and Geology, Environmental Hazards and Anthropogenic Activity.
UNIT II MUTUAL INFORMATION
A Similarity Measure for Intensity Based Image Registration: Introduction, Mutual Information Similarity Measure, Joint Histogram Estimation Methods, Two-Step Joint Histogram Estimation, One-Step Joint Histogram Estimation, Interpolation Induced Artifacts, Generalized Partial Volume Estimation of Joint Histograms, Optimization Issues in the Maximization of MI.
UNIT III INDEPENDENT COMPONENT ANALYSIS
Introduction, Concept of ICA, ICA Algorithms, Preprocessing using PCA, Information Minimization Solution for ICA, ICA Solution through Non-Gaussianity Maximization, Application of ICA to Hyperspectral Imagery, Feature Extraction Based Model, Linear Mixture Model Based Model, An ICA algorithm for Hyperspectral Image Processing, Applications using ICA.
UNIT IV SUPPORT VECTOR MACHINES
Introduction, Statistical Learning Theory, Empirical Risk Minimization, Structural Risk Minimization, Design of Support Vector Machines, Linearly Separable Case, Linearly Non-Separable Case, NonLinear Support Vector Machines, SVMs for Multiclass Classification, Classification based on Decision Directed Acyclic Graph and Decision Tree Structure, optimization Methods, Applications using SVM.
UNIT V MARKOV RANDOM FIELD MODELS
Introduction, MRF and Gibbs Distribution, Random Field and Neighborhood ,Cliques, Potential and Gibbs Distributions, MRF Modeling in Remote Sensing Applications, Optimization Algorithms, Simulated Annealing, Metropolis Algorithm, Iterated Conditional Modes Algorithm
COURSE OUTCOMES:
CO1: Select appropriate hyperspectral data for a particular application
CO2: Understand basic concepts of data acquisition tasks required for multi and hyperspectral data analysis.
CO3:Understand basic concepts of image processing tasks required for multi and hyperspectral data analysis
CO4: Learn techniques for classification of multi and hyperspectral data.
CO5:Learn techniques for analysis of multi and hyperspectral data.
TOTAL:45 PERIODS
REFERENCES
1. Pramod K. Varshney, Manoj K. Arora, “Advanced Image Processing Techniques for Remotely Sensed Hyperspectral Data”, Springer, 2013.
2. S. Svanberg, “Multi-spectral Imaging– from Astronomy to Microscopy – from Radio waves to Gamma rays”, Springer Verlag, 2009
3. AAPO HYVÄRINEN, UHA KARHUNEN and ERKKI OJA,” Independent Component Analysis” John Wiley & Sons, 2001.
4. Ingo Steinwart,Andreas Christmann,”Support Vector Machines”, Springer-Verlag New York,2008.