RIT COLLEGE OF SCIENCE COURSE OUTLINE SIMG-463 COURSE TITLE

lemonadeviraginityAI and Robotics

Nov 6, 2013 (3 years and 9 months ago)

63 views


RIT COLLEGE OF SCIENCE

COURSE OUTLINE

SIMG
-
46
3




I.

COURSE TITLE

Digital Image Processing

III

-

Multispectral
Digital Image
Processing


II.

COURSE CATALOG DESCRIPTION

This course discusses the
digital image processing
concepts and algorithms used
for the analys
is of

hyperspectral, multispectral and multi
-
channel data in
remote

sensing and other application areas
. Concepts are covered at the theor
etical and
implementation level

using
current, popular commercial software packages and
high
-
level programming langua
ges

for examples,
homework
and programming
assignments. The requisite multivariate statistics will be presented as part of this
course as

an

extension

of the univariate statistics that the student
s

have previously
been exposed to
. Topics to be covered wi
ll include methods for supervised data
classification,
clustering algorithms and unsupervised classification,
multispectral
data transformations, data redundancy reduction techniques
,

image
-
to
-
image
rectification
, and data fusion for resolution enhancement
.

(Prerequisites: 1051
-
211

(or equivalent)
, 1016
-
351, 1061
-
352)


III.

OBJECTIVES OF THE COURSE

The objective of this course is to

provide the student with basic knowledge and
skills to
analyze

hyperspectral, multispectral and multi
-
channel

data using
commercia
lly available tools.
The student will use pre
-
existing algorithmic
implementations
in
commercially
-
available software packages as well as code
their own implementations
in a high
-
level
programming language such as IDL,
MATLAB, C++ or Java.


IV.

COURSE OUTLINE
:


Multivariate Statistics


Conditional probability


The normal probability distribution



Univariate case



Multivariate case


Statistical distance measures


Data

Types


Multi
-
channel data


Multispectral data


Hyperspectral data


Supervised Data Classifica
tion


Training


Minimum distance to the mean classifiers


Parallelepiped classifiers


Maximum likelihood classifiers



Bayesian assumptions



Linear discriminant functions


Mahalanobis distance


Spectral angle mapper (SAM)


Clustering and Unsupervised Clas
sification


Similarity metrics and clustering criteria


Iterative clustering algorithms (migrating means)



Seeding techniques



K
-
Means



ISODATA



Merging, splitting and deleting classes


Single pass techniques


Multispectral Data Transformations
/Data Re
dundancy Reduction


Eigenvector transformations


Principal components analysis


Kauth
-
Thomas (KT) tasseled cap transformation


Minimum Noise Fraction (MNF)


Image
-
to
-
Image Rectification


Multiple linear regression


Ground control point (GCP) selection


Map
ping polynomials


Analysis of variance and error characterization


Data Fusion


Multispectral resolution enhancement



Using color transformations



Radiometry preserving techniques


V.

INSTRUCTIONAL TECHNIQUES


Classroom lectures, assigned reading, homewo
rk, and programming exercises


VI.

METHODS OF EVALUATION


Homework and programming assignments, midterm exam, final exam


VII.

BIBLIOGRAPHY


Richards, J.A and X. Jia, Remote Sensing Digital Image Analysis, An
Introduction, 3rd Edition, Springer
-
Verlag, New

York, 1999.