1i.maxlik(1) Grass User's Manual i.maxlik(1)
2
3
4
6 i.maxlik - An imagery function that classifies the cell spectral
7 reflectances in imagery data based on the spectral signature informa‐
8 tion generated by either i.cluster, i.class, or i.gensig.
9
11 imagery
12
14 i.maxlik
15 i.maxlik help
16 i.maxlik [-q] group=string subgroup=string sigfile=string class=string
17 [reject=string]
18
19 Flags:
20 -q Run quietly
21
22 Parameters:
23 group=string
24 Imagery group to be classified
25
26 subgroup=string
27 Subgroup containing image files to be classified
28
29 sigfile=string
30 Signatures to use for classification
31
32 class=string
33 Raster map to hold classification results
34
35 reject=string
36 Raster map to hold reject threshold results
37
39 i.maxlik is a maximum-likelihood discriminant analysis classifier. It
40 can be used to perform the second step in either an unsupervised or a
41 supervised image classification.
42
43 Either image classification methods are performed in two steps. The
44 first step in an unsupervised image classification is performed by
45 i.cluster; the first step in a supervised classification is executed by
46 the GRASS program i.class. In both cases, the second step in the image
47 classification procedure is performed by i.maxlik.
48
49 In an unsupervised classification, the maximum-likelihood classifier
50 uses the cluster means and covariance matrices from the i.cluster sig‐
51 nature file to determine to which category (spectral class) each cell
52 in the image has the highest probability of belonging. In a supervised
53 image classification, the maximum-likelihood classifier uses the region
54 means and covariance matrices from the spectral signature file gener‐
55 ated by i.class, based on regions (groups of image pixels) chosen by
56 the user, to determine to which category each cell in the image has the
57 highest probability of belonging.
58
59 In either case, the raster map layer output by i.maxlik is a classified
60 image in which each cell has been assigned to a spectral class (i.e., a
61 category). The spectral classes (categories) can be related to spe‐
62 cific land cover types on the ground.
63
64 The program will run non-interactively if the user specifies the names
65 of raster map layers, i.e., group and subgroup names, seed signature
66 file name, result classification file name, and any combination of non-
67 required options in the command line, using the form i.maxlik[-q]
68 group=name subgroup=name sigfile=name class=name [reject=name]
69
70 where each flag and options have the meanings stated below.
71
72 Alternatively, the user can simply type i.maxlik in the command line
73 without program arguments. In this case the user will be prompted for
74 the program parameter settings; the program will run foreground.
75
77 Flags:
78 -q Run quietly, without printing program messages to standard out‐
79 put.
80
81 Parameters:
82 group=name
83 The imagery group contains the subgroup to be classified.
84
85 subgroup=name
86 The subgroup contains image files, which were used to create the
87 signature file in the program i.cluster, i.class, or i.gensig to
88 be classified.
89
90 sigfile=name
91 The name of the signatures to be used for the classification.
92 The signature file contains the cluster and covariance matrices
93 that were calculated by the GRASS program i.cluster (or the
94 region means and covariance matrices generated by i.class, if
95 the user runs a supervised classification). These spectral sig‐
96 natures are what determine the categories (classes) to which
97 image pixels will be assigned during the classification process.
98
99 class=name
100 The name of a raster map holds the classification results. This
101 new raster map layer will contain categories that can be related
102 to land cover categories on the ground.
103
104 reject=name
105 The optional name of a raster map holds the reject threshold
106 results. This is the result of a chi square test on each dis‐
107 criminant result at various threshold levels of confidence to
108 determine at what confidence level each cell classified (catego‐
109 rized). It is the reject threshold map layer, and contains one
110 calculated confidence level for each classified cell in the
111 classified image. One of the possible uses for this map layer is
112 as a mask, to identify cells in the classified image that have
113 the lowest probability of being assigned to the correct class.
114
116 The maximum-likelihood classifier assumes that the spectral signatures
117 for each class (category) in each band file are normally distributed
118 (i.e., Gaussian in nature). Algorithms, such as i.cluster, i.class, or
119 i.gensig, however, can create signatures that are not valid distributed
120 (more likely with i.class). If this occurs, i.maxlik will reject them
121 and display a warning message.
122
123 This program runs interactively if the user types i.maxlik only. If the
124 user types i.maxlik along with all required options, it will overwrite
125 the classified raster map without prompting if this map existed.
126
128 GRASS Tutorial: Image Processing
129
130 i.class
131 i.cluster
132 i.gensig
133 i.group
134
136 Michael Shapiro, U.S.Army Construction Engineering Research Laboratory
137 Tao Wen, University of Illinois at Urbana-Champaign, Illinois
138
139 Last changed: $Date: 2006/04/13 18:50:10 $
140
141 Full index
142
143
144
145GRASS 6.2.2 i.maxlik(1)