i.pysptools.unmix - GRASS GIS manual

NAME

i.pysptools.unmix - Extract endmembers from imagery group and perform spectral unmixing using pysptools

KEYWORDS

SYNOPSIS

i.pysptools.unmix

i.pysptools.unmix --help

i.pysptools.unmix [-n] input=name [output=name] [prefix=string] [endmembers=name] endmember_n=integer [extraction_method=string] [unmixing_method=string] [maxit=integer] [--overwrite] [--help] [--verbose] [--quiet] [--ui]

Flags:

-n: Do not use Automatic Target Generation Process (ATGP)
--overwrite: Allow output files to overwrite existing files
--help: Print usage summary
--verbose: Verbose module output
--quiet: Quiet module output
--ui: Force launching GUI dialog

Parameters:

input=name [required]: Input imagery group
output=name: Text file storing endmember information for i.spec.unmix
prefix=string: Prefix for resulting raster maps
endmembers=name: Vector map representing identified endmembers
endmember_n=integer [required]: Number of endmembers to identify
extraction_method=string: Method for endmember extraction; Options: FIPPI, PPI, NFINDR; Default: NFINDR
unmixing_method=string: Algorithm for spectral unmixing; Options: FCLS, UCLS, NNLS; Default: FCLS
maxit=integer: Maximal number of iterations for endmember extraction (default=3*number of bands)

DESCRIPTION
NOTES
EXAMPLE
REQUIREMENTS
REFERENCES
SEE ALSO
AUTHORS

DESCRIPTION

i.pysptools.unmix extracts endmembers from imagery group and performs spectral unmixing using pysptools. The module creates an endmember text file and endmember raster maps.

The module is a wrapper around the pysptools Python library, that integrates its functionality for Endmember Extraction and Spectral Unmixing into GRASS GIS.

It requires that the Python libraries pysptools and scikit-learn are installed (see below).

Supported algorithms for Endmember Extraction are:

NFINDR: N-FINDR endmembers induction algorithm after Winter (1999), that also makes use of an Automatic Target Generation Process (ATGP) (Plaza & Chang 2006). (Default)
FIPPI: Fast Iterative Pixel Purity Index after Chang (2006)
PPI: Pixel Purity Index

Supported algorithms for Spectral Unmixing are:

FCLS: Fully Constrained Least Squares (FCLS): Estimates endmember abundance per pixel with the constraint that values are non-negative and sum up to one per pixel (Default)
UCLS: Unconstrained Least Squares (UCLS): Estimates endmember abundance per pixel in an unconstrained way
NNLS: Non-negative Constrained Least Squares (NNLS): Estimates endmember abundance per pixel with the constraint that values are non-negative

NOTES

Number of endmembers to extract (endmember_n) is supposed to be lower than the number of bands in the imagery group. Only the PPI method can extract more endmembers than there are bands in the imagery group.

EXAMPLE

Example for the North Carolina sample dataset:

# Create list of bands excluding thermal bands
bands=`g.list type=raster pattern="lsat7_2002*" exclude="lsat7_2002_6?" separator=','`
echo "$bands"

# Create imagery group
i.group group=lsat_2002 input="$bands"

# set computation region
g.region raster=lsat7_2002_10 -p

# Extract Endmembers and perform spectral unmixing using pysptools
# resulting in an endmember text file and raster maps (here: 5 endmember)
i.pysptools.unmix input=lsat_2002 endmembers=endmembers endmember_n=5 \
  output=spectrum.txt prefix=lsat_spectra --v

# Compare to result from i.spec.unmix addon
i.spec.unmix group=lsat7_2002 matrix=sample/spectrum.dat result=lsat7_2002_unmix \
  error=lsat7_2002_unmix_err iter=lsat7_2002_unmix_iterations

REQUIREMENTS

python-cvxopt (install through system software management)
python-matplotlib (install through system software management)
python-scikit-learn (install through system software management)
python-scipy (install through system software management)
pysptools library
scikit-learn library

REFERENCES

Chang, C.-I. 2006: A fast iterative algorithm for implementation of pixel purity index. Geoscience and Remote Sensing Letters, IEEE, 3(1): 63-67.
Plaza, A. & Chang, C.-I. 2006: Impact of Initialization on Design of Endmember Extraction Algorithms. Geoscience and Remote Sensing, IEEE Transactions. 44(11): 3397-3407.
Winter, M. E. 1999: N-FINDR: an algorithm for fast autonomous spectral end-member determination in hyperspectral data. Presented at the Imaging Spectrometry V, Denver, CO, USA, (3753): 266-275.