1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
|
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook XML V4.4//EN" "http://www.oasis-open.org/docbook/xml/4.4/docbookx.dtd">
<refentry id='pksvm'>
<refmeta>
<refentrytitle>pksvm</refentrytitle>
<manvolnum>1</manvolnum>
</refmeta>
<refnamediv>
<refname>pksvm</refname>
<refpurpose>classify raster image using Support Vector Machine</refpurpose>
</refnamediv>
<refsynopsisdiv id='synopsis'>
<cmdsynopsis>
<command>pksvm</command>
<arg choice='plain'><option>-t</option> <replaceable>training</replaceable></arg>
<arg choice='opt'><option>-i</option> <replaceable>input</replaceable></arg>
<arg choice='opt'><option>-o</option> <replaceable>output</replaceable></arg>
<arg choice='opt'><option>-cv</option> <replaceable>value</replaceable></arg>
<arg choice='opt'><replaceable>options</replaceable></arg>
<arg choice='opt'><replaceable>advanced options</replaceable></arg>
</cmdsynopsis>
</refsynopsisdiv>
<refsect1 id='description'>
<title>DESCRIPTION</title>
<para>
<command>pksvm</command> implements a support vector machine (SVM) to
solve a supervised classification problem.
The implementation is based on the open source C++ library libSVM
(http://www.csie.ntu.edu.tw/~cjlin/libsvm).
Both raster and vector files are supported as input.
The output will contain the classification result, either in raster or
vector format, corresponding to the format of the input.
A training sample must be provided as an OGR vector dataset that
contains the class labels and the features for each training point.
The point locations are not considered in the training step.
You can use the same training sample for classifying different images,
provided the number of bands of the images are identical.
Use the utility pkextract to create a suitable training sample, based
on a sample of points or polygons.
For raster output maps you can attach a color table using the option
<option>-ct</option>.
</para>
</refsect1>
<refsect1 id='options'>
<title>OPTIONS</title>
<variablelist>
<varlistentry>
<term><option>-t</option> <replaceable>filename</replaceable></term>
<term><option>--training</option> <replaceable>filename</replaceable></term>
<listitem>
<para>
Training vector file.
A single vector file contains all training features
(must be set as: b0, b1, b2,...) for all classes
(class numbers identified by label option).
Use multiple training files for bootstrap aggregation
(alternative to the <option>--bag</option> and
<option>--bagsize</option> options, where a random subset
is taken from a single training file)
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-i</option> <replaceable>filename</replaceable></term>
<term><option>--input</option> <replaceable>filename</replaceable></term>
<listitem>
<para>
input image
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-o</option> <replaceable>filename</replaceable></term>
<term><option>--output</option> <replaceable>filename</replaceable></term>
<listitem>
<para>
Output classification image
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-cv</option> <replaceable>value</replaceable></term>
<term><option>--cv</option> <replaceable>value</replaceable></term>
<listitem>
<para>
N-fold cross validation mode (default: 0)
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-tln</option> <replaceable>layer</replaceable></term>
<term><option>--tln</option> <replaceable>layer</replaceable></term>
<listitem>
<para>
Training layer name(s)
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-c</option> <replaceable>name</replaceable></term>
<term><option>--class</option> <replaceable>name</replaceable></term>
<listitem>
<para>
List of class names.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-r</option> <replaceable>value</replaceable></term>
<term><option>--reclass</option> <replaceable>value</replaceable></term>
<listitem>
<para>
List of class values (use same order as in
<option>--class</option> option).
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-of</option> <replaceable>GDALformat</replaceable></term>
<term><option>--oformat</option> <replaceable>GDALformat</replaceable></term>
<listitem>
<para>
Output image format (see also
<citerefentry>
<refentrytitle>gdal_translate</refentrytitle>
<manvolnum>1</manvolnum>
</citerefentry>).
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-f</option> <replaceable>format</replaceable></term>
<term><option>--f</option> <replaceable>format</replaceable></term>
<listitem>
<para>
Output ogr format for active training sample
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-co</option> <replaceable>NAME=VALUE</replaceable></term>
<term><option>--co</option> <replaceable>NAME=VALUE</replaceable></term>
<listitem>
<para>
Creation option for output file.
Multiple options can be specified.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-ct</option> <replaceable>filename</replaceable></term>
<term><option>--ct</option> <replaceable>filename</replaceable></term>
<listitem>
<para>
Color table in ASCII format having 5 columns:
id R G B ALFA (0: transparent, 255: solid)
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-label</option> <replaceable>attribute</replaceable></term>
<term><option>--label</option> <replaceable>attribute</replaceable></term>
<listitem>
<para>
Identifier for class label in training vector file.
(default: label)
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-prior</option> <replaceable>value</replaceable></term>
<term><option>--prior</option> <replaceable>value</replaceable></term>
<listitem>
<para>
Prior probabilities for each class (e.g.,
<option>-prior</option> 0.3 <option>-prior</option> 0.3
<option>-prior</option> 0.2)
Used for input only (ignored for cross validation)
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-g</option> <replaceable>gamma</replaceable></term>
<term><option>--gamma</option> <replaceable>gamma</replaceable></term>
<listitem>
<para>
Gamma in kernel function
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-cc</option> <replaceable>cost</replaceable></term>
<term><option>--ccost</option> <replaceable>cost</replaceable></term>
<listitem>
<para>
The parameter C of C_SVC, epsilon_SVR, and nu_SVR
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-m</option> <replaceable>filename</replaceable></term>
<term><option>--mask</option> <replaceable>filename</replaceable></term>
<listitem>
<para>
Only classify within specified mask (vector or raster).
For raster mask, set nodata values with the option
<option>--msknodata</option>.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-msknodata</option> <replaceable>value</replaceable></term>
<term><option>--msknodata</option> <replaceable>value</replaceable></term>
<listitem>
<para>
Mask value(s) not to consider for classification.
Values will be taken over in classification image.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-nodata</option> <replaceable>value</replaceable></term>
<term><option>--nodata</option> <replaceable>value</replaceable></term>
<listitem>
<para>
Nodata value to put where image is masked as nodata
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-v</option> <replaceable>level</replaceable></term>
<term><option>--verbose</option> <replaceable>level</replaceable></term>
<listitem>
<para>
Verbose level
</para>
</listitem>
</varlistentry>
</variablelist>
<para>Advanced options</para>
<variablelist>
<varlistentry>
<term><option>-b</option> <replaceable>band</replaceable></term>
<term><option>--band</option> <replaceable>band</replaceable></term>
<listitem>
<para>
Band index (starting from 0, either use <option>--band</option>
option or use <option>--startband</option> to
<option>--endband</option>)
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-sband</option> <replaceable>band</replaceable></term>
<term><option>--startband</option> <replaceable>band</replaceable></term>
<listitem>
<para>
Start band sequence number
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-eband</option> <replaceable>band</replaceable></term>
<term><option>--endband</option> <replaceable>band</replaceable></term>
<listitem>
<para>
End band sequence number
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-bal</option> <replaceable>size</replaceable></term>
<term><option>--balance</option> <replaceable>size</replaceable></term>
<listitem>
<para>
Balance the input data to this number of samples for each class
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-min</option> <replaceable>number</replaceable></term>
<term><option>--min</option> <replaceable>number</replaceable></term>
<listitem>
<para>
If number of training pixels is less then min, do not take this
class into account (0: consider all classes)
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-bag</option> <replaceable>value</replaceable></term>
<term><option>--bag</option> <replaceable>value</replaceable></term>
<listitem>
<para>
Number of bootstrap aggregations (default is no bagging: 1)
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-bagsize</option> <replaceable>value</replaceable></term>
<term><option>--bagsize</option> <replaceable>value</replaceable></term>
<listitem>
<para>
Percentage of features used from available training features for
each bootstrap aggregation (one size for all classes, or a
different size for each class respectively
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-comb</option> <replaceable>rule</replaceable></term>
<term><option>--comb</option> <replaceable>rule</replaceable></term>
<listitem>
<para>
How to combine bootstrap aggregation classifiers
(0: sum rule, 1: product rule, 2: max rule).
Also used to aggregate classes with rc option.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-cb</option> <replaceable>filename</replaceable></term>
<term><option>--classbag</option> <replaceable>filename</replaceable></term>
<listitem>
<para>
Output for each individual bootstrap aggregation
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-prob</option> <replaceable>filename</replaceable></term>
<term><option>--prob</option> <replaceable>filename</replaceable></term>
<listitem>
<para>
Probability image.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-offset</option> <replaceable>value</replaceable></term>
<term><option>--offset</option> <replaceable>value</replaceable></term>
<listitem>
<para>
Offset value for each spectral band input features:
<literal>refl[band]=(DN[band]-offset[band])/scale[band]</literal>
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-scale</option> <replaceable>value</replaceable></term>
<term><option>--scale</option> <replaceable>value</replaceable></term>
<listitem>
<para>
Scale value for each spectral band input features:
<literal>refl=(DN[band]-offset[band])/scale[band]</literal>
(use <literal>0</literal> if scale min and max in each band to
<literal>-1.0</literal> and <literal>1.0</literal>)
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-svmt</option> <replaceable>type</replaceable></term>
<term><option>--svmtype</option> <replaceable>type</replaceable></term>
<listitem>
<para>
Type of SVM (C_SVC, nu_SVC,one_class, epsilon_SVR, nu_SVR)
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-kt</option> <replaceable>type</replaceable></term>
<term><option>--kerneltype</option> <replaceable>type</replaceable></term>
<listitem>
<para>
Type of kernel function (linear,polynomial,radial,sigmoid)
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-kd</option> <replaceable>value</replaceable></term>
<term><option>--kd</option> <replaceable>value</replaceable></term>
<listitem>
<para>
Degree in kernel function
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-c0</option> <replaceable>value</replaceable></term>
<term><option>--coef0</option> <replaceable>value</replaceable></term>
<listitem>
<para>
Coef0 in kernel function
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-nu</option> <replaceable>value</replaceable></term>
<term><option>--nu</option> <replaceable>value</replaceable></term>
<listitem>
<para>
The parameter nu of nu-SVC, one-class SVM, and nu-SVR
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-eloss</option> <replaceable>value</replaceable></term>
<term><option>--eloss</option> <replaceable>value</replaceable></term>
<listitem>
<para>
The epsilon in loss function of epsilon-SVR
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-cache</option> <replaceable>number</replaceable></term>
<term><option>--cache</option> <replaceable>number</replaceable></term>
<listitem>
<para>
<ulink url="http://pktools.nongnu.org/html/classCache.html">Cache</ulink>
memory size in MB
(default: 100)
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-etol</option> <replaceable>value</replaceable></term>
<term><option>--etol</option> <replaceable>value</replaceable></term>
<listitem>
<para>
the tolerance of termination criterion
(default: 0.001)
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-shrink</option></term>
<term><option>--shrink</option></term>
<listitem>
<para>
Whether to use the shrinking heuristics
</para>
</listitem>
</varlistentry>
<varlistentry>
<term><option>-na</option> <replaceable>number</replaceable></term>
<term><option>--nactive</option> <replaceable>number</replaceable></term>
<listitem>
<para>
Number of active training points
</para>
</listitem>
</varlistentry>
</variablelist>
</refsect1>
<refsect1 id='example'>
<title>EXAMPLE</title>
<example>
<para>
Classify input image input.tif with a support vector machine.
A training sample that is provided as an OGR vector dataset.
It contains all features (same dimensionality as input.tif) in its
fields (please check
<citerefentry>
<refentrytitle>pkextract</refentrytitle>
<manvolnum>1</manvolnum>
</citerefentry>
on how to obtain such a file from a "clean" vector file containing
locations only).
A two-fold cross validation (cv) is performed (output on screen).
The parameters cost and gamma of the support vector machine are set
to 1000 and 0.1 respectively.
A colourtable (a five column text file: image value, RED, GREEN,
BLUE, ALPHA) has also been provided.
</para>
<screen>
<command>pksvm</command> <option>-i</option> <replaceable>input.tif</replaceable> <option>-t</option> <replaceable>training.sqlite</replaceable> <option>-o</option> <replaceable>output.tif</replaceable> <option>-cv</option> <replaceable>2</replaceable> <option>-ct</option> <replaceable>colourtable.txt</replaceable> <option>-cc</option> <replaceable>1000</replaceable> <option>-g</option> <replaceable>0.1</replaceable>
</screen>
</example>
<example>
<para>
Classification using bootstrap aggregation.
The training sample is randomly split in three subsamples
(33% of the original sample each).
</para>
<screen>
<command>pksvm</command> <option>-i</option> <replaceable>input.tif</replaceable> <option>-t</option> <replaceable>training.sqlite</replaceable> <option>-o</option> <replaceable>output.tif</replaceable> <option>-bs</option> <replaceable>33</replaceable> <option>-bag</option> <replaceable>3</replaceable>
</screen>
</example>
<example>
<para>
Classification using prior probabilities for each class.
The priors are automatically normalized.
The order in which the options <option>-p</option> are provide should
respect the alphanumeric order of the class names (class 10 comes
before 2...)
</para>
<screen>
<command>pksvm</command> <option>-i</option> <replaceable>input.tif</replaceable> <option>-t</option> <replaceable>training.sqlite</replaceable> <option>-o</option> <replaceable>output.tif</replaceable> <option>-p</option> <replaceable>1</replaceable> <option>-p</option> <replaceable>1</replaceable> <option>-p</option> <replaceable>1</replaceable> <option>-p</option> <replaceable>1</replaceable> <option>-p</option> <replaceable>1</replaceable> <option>-p</option> <replaceable>1</replaceable> <option>-p</option> <replaceable>1</replaceable> <option>-p</option> <replaceable>1</replaceable> <option>-p</option> <replaceable>1</replaceable> <option>-p</option> <replaceable>1</replaceable> <option>-p</option> <replaceable>1</replaceable> <option>-p</option> <replaceable>0.2</replaceable> <option>-p</option> <replaceable>1</replaceable> <option>-p</option> <replaceable>1</replaceable> <option>-p</option> <replaceable>1</replaceable>
</screen>
</example>
</refsect1>
</refentry>
|
