This course will focus in Semi-Automatic Classification Algorithm: The differences between Minimum Distance, Maximum Likelihood, and Spectral Angle Mapper based on remotely-sensed data
Semi-Automatic Classification Algorithm: The differences between Minimum Distance, Maximum Likelihood, and Spectral Angle Mapper
1. Fatwa Ramdani
Geoenvironment, Earth Science, Grad. School of Science
Remote sensing e-course
Semi-Automatic Classification Algorithm: The
differences between Minimum Distance, Maximum
Likelihood, and Spectral Angle Mapper
2. Outline
• This course will focus in Semi-Automatic Classification
Algorithm: The differences between Minimum Distance,
Maximum Likelihood, and Spectral Angle Mapper based on
remotely-sensed data; Landsat 8 OLI. The methods how to
analyze and exploit the Landsat 8 OLI information for Land Use
mapping will be illustrated in QGIS open source software.
• In final section will be follow with the exercise and questions to
allow student expand their understanding.
3. Course Goal and Objectives
• Understand the concept of Semi-Automatic Classification
Algorithm
• Understand algorithm in QGIS open source software
• Understand the differences between Minimum Distance,
Maximum Likelihood, and Spectral Angel Mapper algorithm
4. Intended Audience
• University student with basic level of
knowledge in Remote Sensing studies
• Course Requirements:
– Internet access
– QuantumGIS software (http://www.qgis.org/en/site/forusers/download.html)
– Downloaded data
5. Semi-Automatic Classification Algorithm
General algorithm of imagery classification
Raw DN
Conversion into TOA
using DOS method
TOA
Band set
Knowledge of
the study area
Producing ROI
(sampling of training data)
Running Semi-Automatic
Classification
Land cover
classification
Accuracy
assessment
Statistic
calculation
Manual input
6. Minimum Distance
The minimum distance technique uses
the mean vectors of each endmember
and calculates the Euclidean distance
from each unknown pixel to the mean
vector for each class. All pixels are
classified to the nearest class unless a
standard deviation or distance threshold
is specified, in which case some pixels
may be unclassified if they do not meet
the selected criteria.
Reference
Richards, J.A., 1999, Remote Sensing Digital Image
Analysis, Springer-Verlag, Berlin, p. 240.
7. Maximum Likelihood
Maximum likelihood classification assumes
that the statistics for each class in each band
are normally distributed and calculates the
probability that a given pixel belongs to a
specific class. Unless you select a probability
threshold, all pixels are classified. Each pixel is
assigned to the class that has the highest
probability (that is, the maximum likelihood).
If the highest probability is smaller than a
threshold you specify, the pixel remains
unclassified.
Reference
Richards, J.A., 1999, Remote Sensing Digital Image Analysis,
Springer-Verlag, Berlin, p. 240.
Instead based on training class multispectral distance
measurements, the maximum likelihood decision rule
is based on probability.
The maximum likelihood procedure assumes that
each training class in each band are normally
distributed (Gaussian).
The probability of a pixel belonging to each of a
predefined set of X classes is calculated, and the pixel
is then assigned to the class for which the probability
is the highest.
8. Spectral Angle Mapper
Spectral Angle Mapper (SAM) is a physically-based spectral classification that uses an n-D angle to match pixels
to reference spectra. The algorithm determines the spectral similarity between two spectra by calculating the
angle between the spectra and treating them as vectors in a space with dimensionality equal to the number of
bands. This technique, when used on calibrated reflectance data, is relatively insensitive to illumination and
albedo effects. Endmember spectra used by SAM can come from ASCII files or spectral libraries, or you can
extract them directly from an image (as ROI average spectra). SAM compares the angle between the
endmember spectrum vector and each pixel vector in n-D space.
Small angles between the two spectrums indicate high similarity and high angles indicate low similarity. This
method is not affected by solar illumination factors, because the angle between the two vectors is independent
of vectors length.
SAM classification assumes reflectance data. However, if you use radiance data, the error is generally not
significant because the origin is still near zero.
Reference
Kruse, F. A., A. B. Lefkoff, J. B. Boardman, K. B. Heidebrecht, A. T. Shapiro, P. J. Barloon, and A. F. H. Goetz, 1993, “The Spectral
Image Processing System (SIPS) - Interactive Visualization and Analysis of Imaging spectrometer Data.” Remote Sensing of the
Environment, v. 44, p. 145 - 163.
9. Activities!
• Check your computer spec, if 64bit then install the WinPhython first
• Download QGIS and install the Semi-Automatic Classification Plugin, run
your QGIS and click Plugins – Manage and Install Plugins..
• Learn step-by-step the algorithm of Semi-Automatic Classification
• Compare the result between three different method!
10. Algorithm
Raw DN
Conversion into TOA
using DOS method
TOA
Band set
Knowledge of
the study area
Producing ROI
(sampling of training data)
Running Semi-Automatic
Classification
Land cover
classification
Accuracy
assessment
Statistic
calculation
Manual input
11. Exercise!
• Explore the DN and TOA values of different
land cover!
• Produce the scatter plot and signature plot of
land cover and analyse it!
15. Quiz?
• Which method is the best one? Why?
• What are the advantages and the
disadvantages of the each method?
• What is the difference between Land Use and
Land Cover?
16. Resources
• Kruse, F. A., A. B. Lefkoff, J. B. Boardman, K. B. Heidebrecht, A. T. Shapiro, P. J. Barloon, and A.
F. H. Goetz, 1993, “The Spectral Image Processing System (SIPS) - Interactive Visualization and
Analysis of Imaging spectrometer Data.” Remote Sensing of the Environment, v. 44, p. 145 -
163.
• Richards, J.A., 1999, Remote Sensing Digital Image Analysis, Springer-Verlag, Berlin, p. 240.
• http://fatwaramdani.wordpress.com/2014/06/26/land-use-classification-using-qgis/
Read more from Luca Congedo, the author of the Semi-Automatic Classification Plugin for QGIS,
here
• http://fromgistors.blogspot.pt/2014/06/land-cover-classification-using-SCP-3.html