21 Sep 05 Lecture 4

1
EOS 840 - Hyperspectral Imaging Systems
Class Instructors Dr. Richard B.
Gomez rgomez_at_gmu.edu Dr. Ronald G.
Resmini ronald.g.resmini_at_boeing.com
Lecture 4 (21 September 05)
George Mason University School of Computational
Sciences Fall Semester 2005 31 August 7
December
2
Contact Information
Ron Resmini v 703-735-3899 ronald.g.resmini_at_boein
g.com (Please put EOS840 in Subject Line) Office
hours by appointment
3
Outline

• Review/context/the thread...
• The atmosphere
• Atmospheric compensation
• Empirical Line Method (ELM)
• The ENVI s/w system
• An application Mt St. Helens
• Thinking about spectra
• Using ENVI
• Your semester project status
• Reading assignment

4
Review/Context/A Thread...

• HSI RS is based on the measurement of a physical
  quantityas a function of wavelength its
  spectroscopy
• HSI is based on discerning/measuring the
  interaction oflight (photons, waves) with matter
• The sun is the source
• Earth RS scenarios involve the atmosphere
• There are complex interactions in the atmosphere
• There are complex interactions between light and
  targetsof interest in a scene
• There are complex interactions between light,
  targets ofinterest, and the atmosphere
• Theres a lot (lots!) of information in the
  spectra

5
Review/Context/A Thread...Why?

• A stronger focus on HSI applications emphasis on
  earth RS
• You understand the problem have done the
  research and planning
• Data are collected by sensor systems (and you,
  too!)
• The data are calibrated (spectrally and
  radiometrically)
• Other corrections/fixes may need to be applied
• Apply atmospheric correction (maybe not all the
  time, though)
• Apply information extraction algorithms
• Based on your conception of spectral data
• Utilize spectral libraries
• Apply information and/or data fusion
• Apply geometric processing e.g.,
  orthorectification
• Product generation report writing
• Disseminate, publish, archive
• Plan and conduct future data collection and
  exploitation
• Repeathaving learned more from the process!

6
The Atmosphere
7

• Atmospheric compensation VNIR/SWIR
• What you must do
• Multiplicative and additive terms
• Strategies
• Scene-based (w/ and w/o ground truth)
  RT-modeling combination of scene-based and
  RT-modeling
• Slides 38 to 48 from M. Goforth presentation
• The empirical line method (ELM)
• Using ENVI
• Impact of dropping terms from the big equation
• Iteration yes you can...and should!
• Emissive later in the semester...unless...

8
An Application Mt. St. Helens
Resmini, R.G., Sunshine, J.M., Tompkins, S., and
Farrand, W.H., 1997.Mapping of alteration
mineralogy and fumarole indicators at Mt. St.
Helens.Proceedings of the Twelfth International
Conference on Applied GeologicRemote Sensing,
17-19 November, Denver, Colo., ERIM
International, Inc.,publ., pp. II-457 to
II-464. Resmini, R.G., Sunshine, J.M., Tompkins,
S., and Farrand, W.H., 1997.Airborne imaging
spectrometer data of Mt. St. Helens volcano.EOS,
Trans., A.G.U., v. 78, no. 17, p. S329.
But first, the RS processing flow...
9
(No Transcript)
10
Thinking About Spectra (Using ENVI)
11

• Spectral parameterization
• Albedo/brightness
• Band depth
• Band width
• Band shape/superimposed features
• Spectral slope
• Spectral indices
• Derivative spectroscopy
• Wavelet transform
• Combinations
• Pre-processing transforms e.g., SSA
• All must have a physical basis!
• Tie all observations to physical reality!!

12

• Collections of spectra in hyperspace
• Resmini (2003)
• Density
• Volume
• Orientation
• Clustering/ of clusters/cluster spacing
• Mixing trends (binary, ternary, linear,
  non-linear)
• Eigenvectors/eigenvalues
• etc...
• Still vastly uncharted territory with lots of
  potential for understanding HSI...imho

13
Your Semester Project Status
A Few Project Challenges For You

• N-P Theory sensitivity to spatial/spectral
  subsets
• When is spectral mixing linear v. non-linear?
  I.e., is this evidentfrom the spectra?
• Measure the volume of hyperspace actually
  occupied by real HSI data
• Hughes Phenomenon. It gets mis-applied
  because...
• Spectral angle between spectra and filter
  vectors is the separabilitygreater than angles
  derived from a confusion matrix analysis of
  aspectral library? Use, also, a measure of SCR

14
Reading Assignment

• Read chapter 1 in van der Meer and de Jong (2003)
• You may be called on to lead a discussion
• Im not looking for the right answer Im looking
  forhow youd think about/approach/attack
  problemsand issues in RS
• Theres no grade and no right/wrong were all
  adultsand I just want to have an adult-level,
  grad-schoolseminar-like discussion

15
Backup Slides
16
Some Additional Items...

• Conversion between cm-1 and mm
• A bit more on quantum mechanics
• Spatial resolution
• Spectral resolution and spectral sampling

17
Linewidth Full Width at Half Maximum