.TH i.smap 1grass "" "GRASS 6.4.4" "Grass User's Manual"
.SH NAME
\fI\fBi.smap\fR\fR  - Performs contextual image classification using sequential maximum a posteriori (SMAP) estimation.
.SH KEYWORDS
imagery, classification, supervised, SMAP
.SH SYNOPSIS
\fBi.smap\fR
.br
\fBi.smap help\fR
.br
\fBi.smap\fR [\-\fBmq\fR] \fBgroup\fR=\fIname\fR \fBsubgroup\fR=\fIname\fR \fBsignaturefile\fR=\fIname\fR \fBoutput\fR=\fIname\fR  [\fBblocksize\fR=\fIinteger\fR]   [\-\-\fBoverwrite\fR]  [\-\-\fBverbose\fR]  [\-\-\fBquiet\fR] 
.SS Flags:
.IP "\fB\-m\fR" 4m
.br
Use maximum likelihood estimation (instead of smap)
.IP "\fB\-q\fR" 4m
.br
Run quietly
.IP "\fB\-\-overwrite\fR" 4m
.br
Allow output files to overwrite existing files
.IP "\fB\-\-verbose\fR" 4m
.br
Verbose module output
.IP "\fB\-\-quiet\fR" 4m
.br
Quiet module output
.PP
.SS Parameters:
.IP "\fBgroup\fR=\fIname\fR" 4m
.br
Name of input imagery group
.IP "\fBsubgroup\fR=\fIname\fR" 4m
.br
Name of input imagery subgroup
.IP "\fBsignaturefile\fR=\fIname\fR" 4m
.br
Name of file containing signatures
.br
Generated by i.gensigset
.IP "\fBoutput\fR=\fIname\fR" 4m
.br
Name for output raster map
.IP "\fBblocksize\fR=\fIinteger\fR" 4m
.br
Size of submatrix to process at one time
.br
Default: \fI128\fR
.PP
.SH DESCRIPTION
The \fIi.smap\fR program is used to segment
multispectral images using a spectral class model known as
a Gaussian mixture distribution.  Since Gaussian mixture
distributions include conventional multivariate Gaussian
distributions, this program may also be used to segment
multispectral images based on simple spectral mean and
covariance parameters.
.PP
\fIi.smap\fR has two modes of operation.  The first mode
is the sequential maximum a posteriori (SMAP) mode
[2].  The SMAP
segmentation algorithm attempts to improve segmentation
accuracy by segmenting the image into regions rather than
segmenting each pixel separately 
(see NOTES).
.PP
The second mode is the more conventional maximum likelihood (ML)
classification which classifies each pixel separately,
but requires somewhat less computation. This mode is selected with
the \fB-m\fR flag (see below).
.SH OPTIONS
.SS Flags:
.IP "\fB\-m\fR
.br
Use maximum likelihood estimation (instead of smap).
Normal operation is to use SMAP estimation (see
NOTES).
.IP "\fB\-q\fR 
.br
Run quietly, without printing messages about program
progress.  Without this flag, messages will be printed (to
stderr) as the program progresses.
.PP
.SS Parameters:
.IP "\fBgroup=\fR\fIname\fR 
.br
imagery group
.br
The imagery group that defines the image to be classified.
.IP "\fBsubgroup=\fR\fIname\fR 
.br
imagery subgroup
.br
The subgroup within the group specified that specifies the
subset of the band files that are to be used as image data
to be classified.
.IP "\fBsignaturefile=\fR\fIname\fR 
.br
imagery signaturefile
.br
The signature file that contains the spectral signatures (i.e., the
statistics) for the classes to be identified in the image.
This signature file is produced by the program
\fIi.gensigset\fR
(see NOTES).
.IP "\fBblocksize=\fR\fIvalue\fR 
.br
size of submatrix to process at one time
.br
default: 128
.br
This option specifies the size of the "window" to be used when
reading the image data. 
.PP
This program was written to be nice about memory usage
without influencing the resultant classification. This
option allows the user to control how much memory is used.
More memory may mean faster (or slower) operation depending
on how much real memory your machine has and how much
virtual memory the program uses.
.PP
The size of the submatrix used in segmenting the image has
a principle function of controlling memory usage; however,
it also can have a subtle effect on the quality of the
segmentation in the smap mode.  The smoothing parameters
for the smap segmentation are estimated separately for each
submatrix.  Therefore, if the image has regions with
qualitatively different behavior, (e.g., natural woodlands
and man-made agricultural fields) it may be useful to use a
submatrix small enough so that different smoothing
parameters may be used for each distinctive region of the
image.
.PP
The submatrix size has no effect on the performance of the
ML segmentation method.
.IP "\fBoutput=\fR\fIname\fR
.br
output raster map.
.br
The name of a raster map that will contain the
classification results.  This new raster map layer will
contain categories that can be related to landcover
categories on the ground.
.PP
.SH INTERACTIVE MODE
If none of the arguments are specified on the command line,
\fIi.smap\fR will interactively prompt for the names of
the maps and files.
.SH NOTES
The SMAP algorithm exploits the fact that nearby pixels in
an image are likely to have the same class.  It works by
segmenting the image at various scales or resolutions and
using the coarse scale segmentations to guide the finer
scale segmentations.  In addition to reducing the number of
misclassifications, the SMAP algorithm generally produces
segmentations with larger connected regions of a fixed
class which may be useful in some applications.
.PP
The amount of smoothing that is performed in the
segmentation is dependent of the behavior of the data in
the image.  If the data suggests that the nearby pixels
often change class, then the algorithm will adaptively
reduce the amount of smoothing.  This ensures that
excessively large regions are not formed.
.PP
The module \fIi.smap\fR does not support MASKed or NULL cells. Therefore 
it might be necessary to create a copy of the classification results 
using e.g. r.mapcalc:
\fC
.DS
.br
r.mapcalc "MASKed_map = classification_results"
.br
.DE
\fR
.SH EXAMPLE
Supervised classification of LANDSAT
\fC
.DS
.br
g.region rast=lsat7_2002_10 \-p
.br
.br
# store VIZ, NIR, MIR into group/subgroup
.br
i.group group=my_lsat7_2002 subgroup=my_lsat7_2002 \(rs
.br
  input=lsat7_2002_10,lsat7_2002_20,lsat7_2002_30,lsat7_2002_40,lsat7_2002_50,lsat7_2002_70
.br
.br
# Now digitize training areas "training" with the digitizer
.br
# and convert to raster model with v.to.rast
.br
v.to.rast training out=training use=cat labelcolumn=label
.br
.br
# calculate statistics
.br
i.gensigset trainingmap=training group=my_lsat7_2002 subgroup=my_lsat7_2002 \(rs
.br
            signaturefile=my_smap_lsat7_2002 maxsig=5
.br
.br
i.smap group=my_lsat7_2002 subgroup=my_lsat7_2002 signaturefile=my_smap_lsat7_2002 \(rs
.br
       output=lsat7_2002_smap_classes
.br
.br
# Visually check result
.br
d.mon x0
.br
d.rast.leg lsat7_2002_smap_classes
.br
.br
# Statistically check result
.br
r.kappa \-w classification=lsat7_2002_smap_classes reference=training
.br
.DE
\fR
.SH REFERENCES
.RS
.IP
C. Bouman and M. Shapiro,
\(dqMultispectral Image Segmentation using a Multiscale Image Model", 
\fIProc. of IEEE Int'l Conf. on Acoust., Speech and Sig. Proc.,\fR
pp. III-565 - III-568, San Francisco, California, March 23-26, 1992.
.IP
C. Bouman and M. Shapiro 1994,
\(dqA Multiscale Random Field Model for Bayesian Image Segmentation",
\fIIEEE Trans. on Image Processing., 3(2), 162-177" 
(PDF)\fR
.IP
McCauley, J.D. and B.A. Engel 1995,
\(dqComparison of Scene Segmentations: SMAP, ECHO and Maximum Likelyhood",
\fIIEEE Trans. on Geoscience and Remote Sensing, 33(6): 1313-1316.\fR
.RE
.SH SEE ALSO
\fIi.group\fR
for creating groups and subgroups
.PP
\fIr.mapcalc\fR
to copy classification result in order to cut out MASKed subareas
.PP
\fIi.gensigset\fR
to generate the signature file required by this program
.SH AUTHORS
Charles Bouman, 
School of Electrical Engineering, Purdue University
.PP
Michael Shapiro,
U.S.Army Construction Engineering 
Research Laboratory
.PP
\fILast changed: $Date: 2013-10-27 23:30:43 +0100 (Sun, 27 Oct 2013) $\fR
.PP
Full index
.PP
© 2003-2014 GRASS Development Team