Carolyn J. Merry

1
Landsat-7 - Image Processing with MultiSpec
Carolyn J. Merry
2
(No Transcript)
3
Definition of Remote Sensing

• "Remote sensing is the practice of deriving
  information about the earth's land and water
  surfaces using images acquired from an overhead
  perspective, using electromagnetic radiation in
  one or more regions of the electromagnetic
  spectrum, reflected or emitted from the earths
  surface. (Campbell, 1996)

4
From Lillesand Kiefer, 2001
5
Electromagnetic Spectrum

• Remote sensing images are taken within specific
  spectral regions

6
Platforms Used toAcquire Remote Sensing Data

• Aircraft
• Low, medium high altitude
• Higher level of spatial detail
• Satellite
• Polar-orbiting, sun-synchronous
• 800-900 km altitude, 90-100 minutes/orbit
• Geo-synchronous
• 35,900 km altitude, 24 hrs/orbit
• stationary relative to Earth

7
Landsat-7 Satellite

• 705-km altitude
• 16-day repeat cycle
• 185 km swath width
• Descending node at 1000 - 15
• min.
• Whisk-broom scanner
• Radiometric resolution 28
• (256 levels)

8
From Jensen, 2001
9
Landsat-7 Satellite

• ETM sensor
• 30-m XS (for 6 bands) 60-m thermal
• 15-m pan band
• Image data (185 km by 185 km)
• 475 raw data 600 corrected data
• NASA developing a global archive of ETM

10
Atmospheric Absorption
11
(No Transcript)
12
(No Transcript)
13
Available Data for Buckeyes (OhioView Project)
OhioView is represented by ten Ohio universities
and partners, including NASA GRC, the USGS EROS
Data Center, OAI, and the Ohio Library and
Information Network (OhioLINK) The primary
mission for OhioView is to make remote sensing
imagery accessible to Ohioans and to fill the
knowledge gap in education about the use of these
valuable data sets.
14
OhioView Mirror Set _at_ OSUView
http//OSUView.ceegs.ohio-state.edu
SDE Server
IMS Server
15
Landsat Web Sites

• http//geo.arc.nasa.gov/sge/landsat/landsat.html
• http//landsat.gsfc.nasa.gov/
• http//landsat.usgs.gov/
• http//earthexplorer.usgs.gov
• http//glovis.usgs.gov
• http//www.ohioview.org/

16
Image Display
False Color Composite
17
Image display
Selected bands are remapped (stretched) to fit
the display device. The output image color space
is called a look-up table.
18
Image histogram
Entire image histogram
Original image
Pavement pixels only
19
Image Enhancement
Forest/Grassland
Water
Water
20
Standard Deviation Stretch
The standard deviation stretch is used to stretch
the output values using a normal distribution.
The result of this stretch is similar to what is
seen by the human eye.
21
Linear Contrast Stretch
The linear contrast stretch is used to stretch
the output values over a selected range using a
linear function. This method linearly
interpolates the pixel range to fit the output
space.
22
Linear contrast stretch
23
Histogram Equalization Stretch
The histogram equalization is a nonlinear stretch
that redistributes pixel values so that there is
approximately the same number of pixels across
the entire range. This results in a flat
histogram. The contrast is increased at the peaks
of the histogram and lessened at the tails. This
kind of stretch can separate pixels into several
distinct groups.
24
Histogram-equalized stretch
25
Level Slice Stretch
The level slice stretch is used to stretch the
output values based on a given number of
categories. It is similar to a linear contrast
stretch, but the levels are manually assigned.
26
Natural color composite 3,2,1
False color composite 4,3,2
27
Image classification

• Spectral pattern recognition
• Informational vs. spectral classes
• Supervised classification
• Training areas
• Unsupervised classification
• Clustering algorithm ISODATA

28
Spectral Reflectance Curve
High
Vegetation
Soil
Spectral Reflectance
Water
Low
Blue Green Red Near IR
Mid IR
Spectral Region
29
Reflectance from a leaf
From Avery Berlin, 1977
30
Unsupervised classification

• Analyst has minimal interaction
• Computer algorithm searches for natural, inherent
  groupings in remote sensing images
• Analyst determines categories for these spectral
  groups by comparing classified image to ground
  reference data

31
Unsupervised classification
Source Canadian Center for Remote Sensing
32
Multispec

• Developed at Purdue University free!
• Works on 512 by 512 images
• Simple image processing techniques
• Techniques today Delaware, OH area
• Image display
• Image classification
• Take home images of your school area

33
Delaware, Ohio 26 July 2000
TM band 1 Blue 0.45-0.52 mm
TM band 4 Near IR 0.75-0.90 mm
34
On-line tutorials in remote sensing

• Fundamentals of Remote Sensing - CCRS
• http//www.ccrs.nrcan.gc.ca/ccrs/learn/tutorials/f
  undam/fundam_e.html
• NASA Remote Sensing Tutorial
• http//rst.gsfc.nasa.gov/
• Remote Sensing Core Curriculum J. Jensen,
  Introductory Digital Image Processing
• http//www.cla.sc.edu/geog/rslab/Rscc/index.html