1i.smap(1) Grass User's Manual i.smap(1)
2
3
4
6 i.smap - Performs contextual image classification using sequential
7 maximum a posteriori (SMAP) estimation.
8
10 imagery
11
13 i.smap
14 i.smap help
15 i.smap [-mq] group=string subgroup=string signaturefile=string [block‐
16 size=integer] output=string [--overwrite] [--verbose] [--quiet]
17
18 Flags:
19 -m
20 Use maximum likelihood estimation (instead of smap)
21
22 -q
23 Run quietly
24
25 --overwrite
26 Allow output files to overwrite existing files
27
28 --verbose
29 Verbose module output
30
31 --quiet
32 Quiet module output
33
34 Parameters:
35 group=string
36 imagery group
37
38 subgroup=string
39 imagery subgroup
40
41 signaturefile=string
42 imagery signaturefile
43
44 blocksize=integer
45 size of submatrix to process at one time
46 Default: 128
47
48 output=string
49 output raster map
50
52 The i.smap program is used to segment multispectral images using a
53 spectral class model known as a Gaussian mixture distribution. Since
54 Gaussian mixture distributions include conventional multivariate Gauss‐
55 ian distributions, this program may also be used to segment multispec‐
56 tral images based on simple spectral mean and covariance parameters.
57
58 i.smap has two modes of operation. The first mode is the sequential
59 maximum a posteriori (SMAP) mode [2]. The SMAP segmentation algorithm
60 attempts to improve segmentation accuracy by segmenting the image into
61 regions rather than segmenting each pixel separately (see NOTES).
62
63 The second mode is the more conventional maximum likelihood (ML) clas‐
64 sification which classifies each pixel separately, but requires some‐
65 what less computation. This mode is selected with the -m flag (see
66 below).
67
69 Flags:
70 -m
71 Use maximum likelihood estimation (instead of smap). Normal
72 operation is to use SMAP estimation (see NOTES).
73
74 -q
75 Run quietly, without printing messages about program progress.
76 Without this flag, messages will be printed (to stderr) as the
77 program progresses.
78
79 Parameters:
80 group=name
81 imagery group
82 The imagery group that defines the image to be classified.
83
84 subgroup=name
85 imagery subgroup
86 The subgroup within the group specified that specifies the sub‐
87 set of the band files that are to be used as image data to be
88 classified.
89
90 signaturefile=name
91 imagery signaturefile
92 The signature file that contains the spectral signatures (i.e.,
93 the statistics) for the classes to be identified in the image.
94 This signature file is produced by the program i.gensigset (see
95 NOTES).
96
97 blocksize=value
98 size of submatrix to process at one time
99 default: 128
100 This option specifies the size of the "window" to be used when
101 reading the image data.
102
103 This program was written to be nice about memory usage without influ‐
104 encing the resultant classification. This option allows the user to
105 control how much memory is used. More memory may mean faster (or
106 slower) operation depending on how much real memory your machine has
107 and how much virtual memory the program uses.
108
109 The size of the submatrix used in segmenting the image has a principle
110 function of controlling memory usage; however, it also can have a sub‐
111 tle effect on the quality of the segmentation in the smap mode. The
112 smoothing parameters for the smap segmentation are estimated separately
113 for each submatrix. Therefore, if the image has regions with qualita‐
114 tively different behavior, (e.g., natural woodlands and man-made agri‐
115 cultural fields) it may be useful to use a submatrix small enough so
116 that different smoothing parameters may be used for each distinctive
117 region of the image.
118
119 The submatrix size has no effect on the performance of the ML segmenta‐
120 tion method.
121
122 output=name
123 output raster map.
124 The name of a raster map that will contain the classification
125 results. This new raster map layer will contain categories that
126 can be related to landcover categories on the ground.
127
129 If none of the arguments are specified on the command line, i.smap will
130 interactively prompt for the names of the maps and files.
131
133 The SMAP algorithm exploits the fact that nearby pixels in an image are
134 likely to have the same class. It works by segmenting the image at
135 various scales or resolutions and using the coarse scale segmentations
136 to guide the finer scale segmentations. In addition to reducing the
137 number of misclassifications, the SMAP algorithm generally produces
138 segmentations with larger connected regions of a fixed class which may
139 be useful in some applications.
140
141 The amount of smoothing that is performed in the segmentation is depen‐
142 dent of the behavior of the data in the image. If the data suggests
143 that the nearby pixels often change class, then the algorithm will
144 adaptively reduce the amount of smoothing. This ensures that exces‐
145 sively large regions are not formed.
146
147 The module i.smap does not support MASKed or NULL cells. Therefore it
148 might be necessary to create a copy of the classification results using
149 e.g. r.mapcalc.
150
151 r.mapcalc MASKed_map=classification results
152
154 C. Bouman and M. Shapiro, "Multispectral Image Segmentation using a
155 Multiscale Image Model", Proc. of IEEE Int'l Conf. on Acoust., Speech
156 and Sig. Proc., pp. III-565 - III-568, San Francisco, California, March
157 23-26, 1992. C. Bouman and M. Shapiro 1994, "A Multiscale Random Field
158 Model for Bayesian Image Segmentation", IEEE Trans. on Image Process‐
159 ing., 3(2), 162-177" McCauley, J.D. and B.A. Engel 1995, "Comparison of
160 Scene Segmentations: SMAP, ECHO and Maximum Likelyhood", IEEE Trans. on
161 Geoscience and Remote Sensing, 33(6): 1313-1316.
162
164 i.group for creating groups and subgroups
165
166 r.mapcalc to copy classification result in order to cut out MASKed sub‐
167 areas
168
169 i.gensigset to generate the signature file required by this program
170
172 Charles Bouman, School of Electrical Engineering, Purdue University
173 Michael Shapiro, U.S.Army Construction Engineering Research Laboratory
174
175 Last changed: $Date: 2008-01-08 14:16:37 +0100 (Tue, 08 Jan 2008) $
176
177 Full index
178
179 © 2003-2008 GRASS Development Team
180
181
182
183GRASS 6.3.0 i.smap(1)