Lecture 0: Course Introduction
Course outline
• Objectives:
o Important, currently available remote sensing techniques and sensors for the earth sciences
o Physics of (imaging) spectroscopy and other Earth observation methods and the use of
spectral libraries to aid image interpretation;
o Interpretation of spatial patterns and time series data in remote sensing;
o Use of current desktop- and cloud-based image processing tools
o How to get access to earth observation imagery
o To critically evaluate remote sensing products
ENVI software
• Image processing system
• Preprocessing: geocoding, radiometric correction, noise reduction
• Classification, spectral unmixing, ortho-rectification, GIS-capabilities
,Lecture 1: Basics of Remote Sensing
Aerial Photo Interpretation
• Human eye evaluates 7 criteria
o Association
o Pattern
o Shadow
o Shape
o Size
o Texture
o Tone
o Spectral/colour
Color production
• Three screen colours
o Red
o Green
o Blue
• Combine three different bands in one image
o We can assign bands that we cannot see to a colour we can
• Often transform (stretch) information to use color space
effectively
o No stretch, linear stretch, histogram stretch, special stretch
o Stretching may result in an unnatural look, but it is just for viewing it
Adding different wavelengths
• By adding different wavelengths to the
colours on screen, you can visualize
different features of objects.
Electromagnetic spectrum
,Remote Sensing System
A. Energy source or Illumination
B. Atmosphere
C. Interaction with the target
D. Recording of energy by the sensor
E. Transmission, reception and processing
F. Interpretation and analysis
G. Application
Sensors
• Airborne and spaceborne
• Spectral bands
o Number and spectral position
o Band width
• Spatial resolution (or pixel size)
• Wide or narrow Field of View (The area the sensor
can scan in 1 movement)
• Orbit: geostationary & polar
• Active and passive systems
• Multispectral = multiple bands that show some wavelengths
• Hyperspectral = almost every wavelength of the spectrum is shown
Data products
• Almost entire archive available with standard processing
o Geometric correction
o Sensor calibration (toa radiance)
o Calibrate for incoming light (toa reflectance)
o Atmospheric and terrain correction (surface reflectance)
o Quality flags (water, snow, haze labels)
Radiometric Calibration
• Sensor sensitivity (DN into radiance)
o 𝐿(𝜆) = 𝐺 ∗ 𝐷𝑁 + 𝐵
▪ DN: digital count in the specific spectral sensor band
▪ G: calibration slope for the specific sensor band (channel gain)
▪ B: calibration offset for zero radiance for that sensor band (bias, channel offset)
, • Radiance into reflectance
𝜋𝐿(𝜆)
o 𝑅(𝜆) = 1
𝐸0 (𝜆)( 2 )cos(𝜃0 )
𝑟
▪ 𝐸0 : the solar constant in the band pass of the sensor
▪ 𝑟: the normalized Earth-Sun distance (in astronomical units ~1.0)
▪ 𝜃0 : the solar zenith angle at the image centre (i.e. seasonal position of the sun)
▪ 𝜋: pi = 3.14159265
Radiometric Corrections
• Not all satellites measure within the same wavelengths, slight differences might need corrections
o Not needed in the non-transparent gaps of the atmosphere
Viewing an Image, interpretation and analysis
• There are four ways to view an image
o Histogram
o Feature space
o Image space
o Spectral signature
• The properties of a multispectral image can be used to
classify an image or make a masking (e.g. NDVI)
o Can be used for a range of usages:
▪ Spectral signature
▪ Wild fires
▪ Storms
▪ Humanitarian aid
▪ Biomass production
▪ Water quality
▪ Deforestation
▪ Change detection (glacier)
Course outline
• Objectives:
o Important, currently available remote sensing techniques and sensors for the earth sciences
o Physics of (imaging) spectroscopy and other Earth observation methods and the use of
spectral libraries to aid image interpretation;
o Interpretation of spatial patterns and time series data in remote sensing;
o Use of current desktop- and cloud-based image processing tools
o How to get access to earth observation imagery
o To critically evaluate remote sensing products
ENVI software
• Image processing system
• Preprocessing: geocoding, radiometric correction, noise reduction
• Classification, spectral unmixing, ortho-rectification, GIS-capabilities
,Lecture 1: Basics of Remote Sensing
Aerial Photo Interpretation
• Human eye evaluates 7 criteria
o Association
o Pattern
o Shadow
o Shape
o Size
o Texture
o Tone
o Spectral/colour
Color production
• Three screen colours
o Red
o Green
o Blue
• Combine three different bands in one image
o We can assign bands that we cannot see to a colour we can
• Often transform (stretch) information to use color space
effectively
o No stretch, linear stretch, histogram stretch, special stretch
o Stretching may result in an unnatural look, but it is just for viewing it
Adding different wavelengths
• By adding different wavelengths to the
colours on screen, you can visualize
different features of objects.
Electromagnetic spectrum
,Remote Sensing System
A. Energy source or Illumination
B. Atmosphere
C. Interaction with the target
D. Recording of energy by the sensor
E. Transmission, reception and processing
F. Interpretation and analysis
G. Application
Sensors
• Airborne and spaceborne
• Spectral bands
o Number and spectral position
o Band width
• Spatial resolution (or pixel size)
• Wide or narrow Field of View (The area the sensor
can scan in 1 movement)
• Orbit: geostationary & polar
• Active and passive systems
• Multispectral = multiple bands that show some wavelengths
• Hyperspectral = almost every wavelength of the spectrum is shown
Data products
• Almost entire archive available with standard processing
o Geometric correction
o Sensor calibration (toa radiance)
o Calibrate for incoming light (toa reflectance)
o Atmospheric and terrain correction (surface reflectance)
o Quality flags (water, snow, haze labels)
Radiometric Calibration
• Sensor sensitivity (DN into radiance)
o 𝐿(𝜆) = 𝐺 ∗ 𝐷𝑁 + 𝐵
▪ DN: digital count in the specific spectral sensor band
▪ G: calibration slope for the specific sensor band (channel gain)
▪ B: calibration offset for zero radiance for that sensor band (bias, channel offset)
, • Radiance into reflectance
𝜋𝐿(𝜆)
o 𝑅(𝜆) = 1
𝐸0 (𝜆)( 2 )cos(𝜃0 )
𝑟
▪ 𝐸0 : the solar constant in the band pass of the sensor
▪ 𝑟: the normalized Earth-Sun distance (in astronomical units ~1.0)
▪ 𝜃0 : the solar zenith angle at the image centre (i.e. seasonal position of the sun)
▪ 𝜋: pi = 3.14159265
Radiometric Corrections
• Not all satellites measure within the same wavelengths, slight differences might need corrections
o Not needed in the non-transparent gaps of the atmosphere
Viewing an Image, interpretation and analysis
• There are four ways to view an image
o Histogram
o Feature space
o Image space
o Spectral signature
• The properties of a multispectral image can be used to
classify an image or make a masking (e.g. NDVI)
o Can be used for a range of usages:
▪ Spectral signature
▪ Wild fires
▪ Storms
▪ Humanitarian aid
▪ Biomass production
▪ Water quality
▪ Deforestation
▪ Change detection (glacier)
