Satellite Image Characteristics

What to consider when selecting

satellite imagery
 Image Characteristics
 Spatial resolution
 Spectral characteristics
 Temporal characteristics
 Sensor sensitivity
 Program History
 Image surface area
 Multi-angle capability
 Tasking
 Price and licensing
 Browse options
 Processing Options
 4 sensor characteristics that affect
what can be seen in an image

 Spatial resolution
 Spectral characteristics
 Temporal characteristics
 Sensor sensitivity

Source: NASA
4 sensor characteristics that affect
what can be seen in an image

 Spatial resolution
 Spectral characteristics
 Temporal characteristics
 Sensitivity of the sensor
Often called the 4 types of resolution

 Spatial resolution
 Spectral resolution
 Temporal resolution
 Radiometric resolution
 Spatial resolution

 Often referred to simply as

resolution
 Size of an image pixel in
ground dimensions
 Usually represented by the
length of one side of a
square (i.e., 30m
resolution)
 Spatial Resolution Advice
 Moving from detection =>
identification => analysis requires
finer resolution
 Rule of thumb – select resolution
~1/10th the size of the feature you
want to examine
 Rule of thumb is not very useful –
strongly linked to feature
characteristics (contrast, location,
shape…)
 Get advice from others – experience
is invaluable
 High contrast between features
allows detection of sub-pixel sized
features
 Spectral characteristics
 Bandwidth
 Range of wavelengths (colors) detected by a particular band
 Band placement
 The portion of the electromagnetic spectrum detected by a
particular band
 Defined by the low and high wavelengths of the range or the by
center of the range
 Number of bands
 The number of bands imaged by the sensor
 Often grouped as panchromatic (single band), multispectral
(more than one band), or hyperspectral (usually over 100 bands)

 Spectral resolution definition is inconsistent and varies to

include some or all of the above characteristics
 Spectral characteristics for Landsat ETM+

Bandwidth – width of rectangles

Band placement – location of the rectangle along the x-axis
Number of bands – number of rectangles
Hyperspectral sensors provide more detailed spectral information
 Temporal characteristics

 The temporal frequency or minimum time a particular

feature can be recorded twice
 Repeat frequency can be shorter than the overpass frequency of
the platform if the sensor is pointable
 Just because the satellite is flying over a feature doesn’t mean
that feature will be imaged
 Short repeat interval increase chances of acquiring a “clean”
image
 The specific time and date a feature is imaged

 Note: Most satellite orbits do not pass over the poles so

these data can not always be acquired
Northern Madagascar seasonal change

July November
Northeastern Madagascar deforestation

1993 2000
 Sensitivity of the sensor

 Dynamic range of the sensor

 Upper and lower limits of the intensity of the signal
(reflected radiation) a sensor can measure
 Some sensors have a high and low gain setting
 Range of values that can represent the value of
a pixel (quantization)
 Often defined as the bit depth
 Common DN ranges are 0-255 (8-bit) and 0-1023
(10-bit)
8 bits (0-255) 6 bits (0–63)

3 bits (0-7) 1 bit (0-1)

 Program History

 When the program started

 How long the program ran
 Will the program continue
 Are historic images still available
 Image Surface Area

 The area on the ground covered by an image is

called the footprint
 Larger footprint increases chances that your
study area falls on fewer images
 Larger footprint usually reduces problems
associated with mosaicing multiple images (color
matching, geometric matching, different
seasons…)
 Multi-Angle Options

 Some sensors are pointable

 Improves the ability to target specific areas
instead of systematic collection
 Can be used to collect stereo pairs for creation
of DEMs
 Off-angle viewing distorts the image although
this is compensated for during processing
 Effectively shortens repeat interval
 Tasking

 Many satellite remote sensing systems can be

tasked to acquire imagery on a specific date or
time period
 Tasking can be expensive
 Read the small print of the agreement to
understand how “acceptable” (cloud cover, time
frame, number of tries…) is defined
 Price and Licensing

 Image prices range from free to >$100/km2

 Some imagery can be distributed free of charge
 Many commercial companies only license their
imagery so a user never “owns” the data
 Licensing agreements define how the imagery
can be used and distributed – read them
carefully before you buy
 Browse Options

 Most image providers have powerful browse

capabilities with tutorials
 Some systems have historic imagery scattered
around the world at ground receiving stations
 Regional centers sometimes have browse
capabilities for a variety of remotely sensed
imagery
 Researching available imagery can be time
consuming if you are interested in multiple
image types
 Processing Options

 Can be a bit confusing

 Most vendors provide a variety of radiometric (adjusting
the “colors”) and geometric (adjusting the image so it
matches a particular orientation or a specific map
projection) options
 The simplest processing uses on-board calibration and
orbital models
 Improved processing involves off-the-shelf ancillary data
such as a DEM or user supplied data to improve
geometry
 With appropriate software, skills, and data users can do
their own radiometric and geometric processing
 Even with pre-processed imagery a user may have to do
additional processing if the imagery is to be combined
with other spatial data
 Types of satellite imagery
 Passive sensors (use sunlight for energy source)
 Low resolution (>100 meter resolution)
 MODIS, AVHRR, SPOT Vegetation
 Moderate resolution (15 – 100 meter resolution)
 Landsat TM/ETM+, SPOT, ASTER, IRS
 High resolution (<15 meter resolution)
 IKONOS, Quickbird, OrbViewIRS, SPOT, Corona
 Active sensors (generate their own energy)
 Radar
 Radarsat, ERS, Envisat, Space Shuttle
 Lidar
 Mostly airborne platforms for now
 ICESat is only satellite lidar platform
MODIS (500m) – Composited using imagery
acquired from June – September 2001
Landsat ETM+ (30m) - 2 April 2002
ASTER (15m) - 8 November 2003
CORONA (5m) – 4 March 1967
IKONOS Pan merge (1m) – 29 April 2002
IKONOS zoomed
 Active systems

 Active sensors provide their own illumination

source whereas passive sensors, such as
Landsat, use the sun for illumination.
 Radar and lidar are the two most popular active
remote sensing systems.
 Active sensors can provide direct measurements
of vegetation structure.
 Currently a lot of research on developing active
systems and using active sensors for ecological
studies.
 LIDAR
 LIDAR = LIght Detection And Ranging
 Measures the distance between the sensor and a target
and strength of return
 Wavelengths from the blue through near-infrared are
used
 Lidar can be deployed as fixed or scanningsystems
 Best known for acquiring DEM information
 Can be used to measure vegetation height and structure
 Measures discrete portions of the returned signal and
more sophisticated systems can measure the entire
waveform of the returned signal
 Most systems are for airborne platforms although NASA’s
ICESat is a satellite sensor has been used to measure
tree height and other vegetation characteristics.
From: Lefsky, M.A., W.B. Cohen, G.G. Parker and D.J. Harding. 2002.
Lidar remote sensing for ecosystem studies. Bioscience 52(1) 19-30.
 RADAR
 RADAR = RAdio Detection And Ranging
 Radar systems measure the strength of the backscatter (portion of
energy received by the radar antenna) and the time delay between
when the signal was emitted and when it was received
 Radio wavelengths are much longer than visible or infrared
wavelengths and because of that they are penetrate vegetation
canopies and the longer wavelength radar systems can penetrate
the soil surface
 The way a radar signal interacts with depends on the object’s size,
shape, surface roughness, the angle of the incident microwave
energy, and dielectric constant
 Radar systems can emit and receive vertically and horizontally
polarized radio waves
 Radar systems are flown on aircraft and satellite
 Some microwave systems are passive instruments that measure
emitted radio waves. These systems are used to measure soil
moisture
 Radar is used to map land cover, measure vegetation structure, and
create DEMs
 Processing and interpreting radar imagery requires special training
 What if “traditional” imagery is not
available?
 Look for less common data sets such as:
 Aerial photography
 Corona imagery
 Airborne imagery
 AIRSAR
 AVIRIS
 Research satellites/sensors
 Hyperion
 Advanced Land Imager
 Astronaut photography
 RADAR
 LIDAR
 Modify analysis methods and/or study area
 Modify project objectives
Andros, Bahamas
Astronaut photo
Andros, Bahamas
Astronaut photo
 Landsat Scan Line Correction Error
 Broken on May 2003
 Error causes gaps in
the images that start
at the center and
become wider at the
edges
 Data is still being
obtained and sold for
$300 per image (half
the normal price)
 Landsat Scan Line Correction Error
 Data offered with “gaps filled” from older data

SLC-Off April 2004 Image April 2003 Image

SLC-Off “Gap-Filled” Image

Kruger National Park – Republic of South Africa
Landsat ETM+ (left) and ALI (right)
AIRSAR - Howland forest, Maine
AVIRIS - Howland forest, Maine
So - How do you decide what to use?

 Compare options using tables like the one at:

http://homepage.mac.com/alexandreleroux/arsist/
 Image availability and cost are the two primary
limiting factors
 Easy to get lost in the details – use common
sense
 Get advice from someone with experience
 Sources for data
 Global Land Cover Facility (GLCF)
 http://landcover.org
 USGS EarthExplorer
 http://edcsns17.cr.usgs.gov/EarthExplorer/
 USGS Global Visualization Viewer
 http://glovis.usgs.gov/
 EROS Data Gateway
 http://edcimswww.cr.usgs.gov/pub/imswelcome/plain.html
 USGS National Map Seamless Data Distribution System
 http://seamless.usgs.gov/
 Astronaut Photography
 http://eol.jsc.nasa.gov/default.htm
 MrSid GeoCover Landsat TM images
 https://zulu.ssc.nasa.gov/mrsid/
 Terraserver
 http://terraserver-usa.com
 Astronaut Photography
 http://eol.jsc.nasa.gov/default.htm
 NGA Raster Roam
 http://geoengine.nima.mil/geospatial/SW_TOOLS/NIMAMUSE/webinter/ras
t_roam.html
 GeoData.gov
 http://www.geodata.gov/gos
 More sources for data
 Tropical Rain Forest Information Center (TRFIC)
 TRIFC - http://bsrsi.msu.edu/trfic/home.html
 Landsat.org
 Landsat.org - http://www.landsat.org
 African Data Dissemination Service
 http://edcsnw4.cr.usgs.gov/adds/index.php
 Centre for Remote Imaging, Sensing and Processing (CRISP)
 http://www.crisp.nus.edu.sg/crisp.html
 ESA Earth Observation Earthnet Online
 http://earth.esa.int/
 SPOT Vegetation products
 http://free.vgt.vito.be/
 Free data for Canada
 http://geogratis.cgdi.gc.ca/clf/en
 Shuttle Radar Topography Mission (SRTM)
 ftp://edcsgs9.cr.usgs.gov/pub/data/srtm/ and GLCF
 ASTER Protected Area Archive
 http://asterweb.jpl.nasa.gov/APAA/default.htm
 Wim Bakker’s list of data directories and inventories
 http://www.itc.nl/~bakker/invdir.html