svn-gvsig-desktop / trunk / extensions / extRemoteSensing / src / org / gvsig / remotesensing / classification / ClassificationParallelepipedProcess.java @ 18829
History | View | Annotate | Download (9.15 KB)
1 |
/* gvSIG. Sistema de Informaci?n Geogr?fica de la Generalitat Valenciana
|
---|---|
2 |
*
|
3 |
* Copyright (C) 2006 Instituto de Desarrollo Regional and Generalitat Valenciana.
|
4 |
*
|
5 |
* This program is free software; you can redistribute it and/or
|
6 |
* modify it under the terms of the GNU General Public License
|
7 |
* as published by the Free Software Foundation; either version 2
|
8 |
* of the License, or (at your option) any later version.
|
9 |
*
|
10 |
* This program is distributed in the hope that it will be useful,
|
11 |
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
12 |
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
13 |
* GNU General Public License for more details.
|
14 |
*
|
15 |
* You should have received a copy of the GNU General Public License
|
16 |
* along with this program; if not, write to the Free Software
|
17 |
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307,USA.
|
18 |
*
|
19 |
* For more information, contact:
|
20 |
*
|
21 |
* Generalitat Valenciana
|
22 |
* Conselleria d'Infraestructures i Transport
|
23 |
* Av. Blasco Iba?ez, 50
|
24 |
* 46010 VALENCIA
|
25 |
* SPAIN
|
26 |
*
|
27 |
* +34 963862235
|
28 |
* gvsig@gva.es
|
29 |
* www.gvsig.gva.es
|
30 |
*
|
31 |
* or
|
32 |
*
|
33 |
* Instituto de Desarrollo Regional (Universidad de Castilla La-Mancha)
|
34 |
* Campus Universitario s/n
|
35 |
* 02071 Alabacete
|
36 |
* Spain
|
37 |
*
|
38 |
* +34 967 599 200
|
39 |
*/
|
40 |
|
41 |
package org.gvsig.remotesensing.classification; |
42 |
|
43 |
import java.util.ArrayList; |
44 |
|
45 |
import org.gvsig.fmap.raster.layers.FLyrRasterSE; |
46 |
import org.gvsig.raster.buffer.RasterBuffer; |
47 |
import org.gvsig.raster.dataset.IBuffer; |
48 |
import org.gvsig.raster.grid.GridException; |
49 |
import org.gvsig.raster.grid.roi.ROI; |
50 |
|
51 |
import com.iver.cit.gvsig.project.documents.view.gui.View; |
52 |
|
53 |
/** ClassificationParallelepipedProcess implementa el m?todo de clasificaci?n de
|
54 |
* paralelepipedos o hipercubos.
|
55 |
*
|
56 |
* @see ClassificationGeneralProcess
|
57 |
* @author Alejandro Mu?oz Sanchez (alejandro.munoz@uclm.es)
|
58 |
* @author Diego Guerrero Sevilla (diego.guerrero@uclm.es)
|
59 |
* @version 19/10/2007
|
60 |
*/
|
61 |
|
62 |
public class ClassificationParallelepipedProcess extends ClassificationGeneralProcess{ |
63 |
|
64 |
private int bandCount = 0; |
65 |
double varianza[][] = null; |
66 |
double medias[][] = null; |
67 |
int defaultClass =0; |
68 |
double stevcoef =0; |
69 |
|
70 |
|
71 |
/**
|
72 |
* M?todo que implementa el clasificador por paralelepipedos.
|
73 |
*
|
74 |
* @param array de tipo byte con los valores del pixel en cada una de las bandas
|
75 |
* @return clase a la que pertenece el pixel
|
76 |
*/
|
77 |
|
78 |
public int getPixelClassForTypeByte(byte[] pixelBandsValues) { |
79 |
for (int clase=0; clase<numClases; clase++) |
80 |
{ |
81 |
boolean inClass= true; |
82 |
for(int i=0; i< bandCount; i++) |
83 |
{ |
84 |
if(!((medias[clase][i]-stevcoef*Math.sqrt(varianza[clase][i])<pixelBandsValues[i]) && |
85 |
(pixelBandsValues[i] <medias[clase][i]+stevcoef*Math.sqrt(varianza[clase][i]))))
|
86 |
{ |
87 |
inClass= false;
|
88 |
} |
89 |
} |
90 |
if(inClass==true) |
91 |
return clase;
|
92 |
} |
93 |
|
94 |
return defaultClass;
|
95 |
} // Fin del metodo
|
96 |
|
97 |
|
98 |
|
99 |
/**
|
100 |
* M?todo que implementa el clasificador por paralelepipedos.
|
101 |
*
|
102 |
* @param array de tipo short con los valores del pixel en cada una de las bandas
|
103 |
* @return primera clase a la que pertenece el pixel.
|
104 |
*/
|
105 |
|
106 |
public int getPixelClassForTypeShort(short[] pixelBandsValues) { |
107 |
for (int clase=0; clase<numClases; clase++) |
108 |
{ |
109 |
boolean inClass= true; |
110 |
for(int i=0; i< bandCount; i++) |
111 |
{ |
112 |
if(!((medias[clase][i]-stevcoef*Math.sqrt(varianza[clase][i])<pixelBandsValues[i]) && |
113 |
(pixelBandsValues[i] <medias[clase][i]+stevcoef*Math.sqrt(varianza[clase][i]))))
|
114 |
{ |
115 |
inClass= false;
|
116 |
} |
117 |
} |
118 |
if(inClass==true) |
119 |
return clase;
|
120 |
} |
121 |
|
122 |
return defaultClass;
|
123 |
} |
124 |
|
125 |
|
126 |
/**
|
127 |
* M?todo que implementa el clasificador por paralelepipedos.
|
128 |
*
|
129 |
* @param array de tipo int con los valores del pixel en cada una de las bandas
|
130 |
* @return primera clase a la que pertenece el pixel.
|
131 |
*/
|
132 |
public int getPixelClassForTypeInt(int[] pixelBandsValues) { |
133 |
for (int clase=0; clase<numClases; clase++) |
134 |
{ |
135 |
boolean inClass= true; |
136 |
for(int i=0; i< bandCount; i++) |
137 |
{ |
138 |
if(!((medias[clase][i]-stevcoef*Math.sqrt(varianza[clase][i])<pixelBandsValues[i]) && |
139 |
(pixelBandsValues[i] <medias[clase][i]+stevcoef*Math.sqrt(varianza[clase][i]))))
|
140 |
{ |
141 |
inClass= false;
|
142 |
} |
143 |
} |
144 |
if(inClass==true) |
145 |
return clase;
|
146 |
} |
147 |
|
148 |
return defaultClass;
|
149 |
} |
150 |
|
151 |
|
152 |
|
153 |
/**
|
154 |
* M?todo que implementa el clasificador por paralelepipedos.
|
155 |
*
|
156 |
* @param array de tipo float con los valores del pixel en cada una de las bandas
|
157 |
* @return primera clase a la que pertenece el pixel.
|
158 |
*/
|
159 |
public int getPixelClassForTypeFloat(float[] pixelBandsValues) { |
160 |
for (int clase=0; clase<numClases; clase++) |
161 |
{ |
162 |
boolean inClass= true; |
163 |
for(int i=0; i< bandCount; i++) |
164 |
{ |
165 |
if(!((medias[clase][i]-stevcoef*Math.sqrt(varianza[clase][i])<pixelBandsValues[i]) && |
166 |
(pixelBandsValues[i] <medias[clase][i]+stevcoef*Math.sqrt(varianza[clase][i]))))
|
167 |
{ |
168 |
inClass= false;
|
169 |
} |
170 |
} |
171 |
if(inClass==true) |
172 |
return clase;
|
173 |
} |
174 |
|
175 |
return defaultClass;
|
176 |
} |
177 |
|
178 |
|
179 |
/**
|
180 |
* M?todo que implementa el clasificador por paralelepipedos.
|
181 |
*
|
182 |
* @param array de tipo double con los valores del pixel en cada una de las bandas
|
183 |
* @return primera clase a la que pertenece el pixel.
|
184 |
*/
|
185 |
public int getPixelClassForTypeDouble(double[] pixelBandsValues) { |
186 |
for (int clase=0; clase<numClases; clase++) |
187 |
{ |
188 |
boolean inClass= true; |
189 |
for(int i=0; i< bandCount; i++) |
190 |
{ |
191 |
if(!((medias[clase][i]-stevcoef*Math.sqrt(varianza[clase][i])<pixelBandsValues[i]) && |
192 |
(pixelBandsValues[i] <medias[clase][i]+stevcoef*Math.sqrt(varianza[clase][i]))))
|
193 |
{ |
194 |
inClass= false;
|
195 |
} |
196 |
} |
197 |
if(inClass==true) |
198 |
return clase;
|
199 |
} |
200 |
|
201 |
return defaultClass;
|
202 |
} |
203 |
|
204 |
|
205 |
/** Metodo que recoge los parametros del proceso de clasificacion de
|
206 |
* por paralelepipedos
|
207 |
* <LI>rasterSE: Capa de entrada para la clasificaci?n</LI>
|
208 |
* <LI> rois: lista de rois</LI>
|
209 |
* <LI> bandList:bandas habilitadas </LI>
|
210 |
* <LI>view: vista sobre la que se carga la capa al acabar el proceso</LI>
|
211 |
* <LI>filename: path con el fichero de salida</LI>
|
212 |
* <LI>stevcoef: coeficiente que establece el de desviacion maxima</LI>
|
213 |
*/
|
214 |
public void init() { |
215 |
|
216 |
rasterSE= (FLyrRasterSE)getParam("layer");
|
217 |
rois = (ArrayList)getParam("rois"); |
218 |
view=(View)getParam("view"); |
219 |
filename= getStringParam("filename");
|
220 |
bandList = (int[])getParam("bandList"); |
221 |
stevcoef= ((Double)getParam("dev")).doubleValue(); |
222 |
|
223 |
bandCount = bandList.length; |
224 |
numClases = rois.size(); |
225 |
defaultClass= numClases; |
226 |
medias= new double[numClases][bandCount]; |
227 |
varianza= new double[numClases][bandCount]; |
228 |
|
229 |
// Calculo de estadisticas de clases
|
230 |
for (int clase=0; clase<numClases; clase++) |
231 |
for (int i=0;i<bandCount;i++){ |
232 |
((ROI)rois.get(clase)).setBandToOperate(bandList[i]); |
233 |
try {
|
234 |
medias[clase][i]=((ROI)rois.get(clase)).getMeanValue(); |
235 |
varianza[clase][i]=((ROI)rois.get(clase)).getVariance(); |
236 |
} catch (GridException e) {
|
237 |
e.printStackTrace(); |
238 |
} |
239 |
} |
240 |
} |
241 |
|
242 |
|
243 |
|
244 |
/** Proceso de clasificacion por paralelepipedos*/
|
245 |
public void process() throws InterruptedException { |
246 |
|
247 |
setGrid(); |
248 |
withdefaultClass=true;
|
249 |
rasterResult= RasterBuffer.getBuffer(IBuffer.TYPE_BYTE, inputGrid.getRasterBuf().getWidth(), |
250 |
inputGrid.getRasterBuf().getHeight(), 1, true); |
251 |
int c=0; |
252 |
int iNY= inputGrid.getLayerNY();
|
253 |
int iNX= inputGrid.getLayerNX();
|
254 |
|
255 |
bandCount = inputGrid.getBandCount(); |
256 |
int inputGridNX = inputGrid.getNX();
|
257 |
int datType = inputGrid.getRasterBuf().getDataType();
|
258 |
|
259 |
// Caso buffer tipo byte
|
260 |
if (datType == RasterBuffer.TYPE_BYTE){
|
261 |
byte data[]= new byte[bandCount]; |
262 |
for(int i=0; i<iNY;i++){ |
263 |
for(int j=0; j<iNX;j++){ |
264 |
inputGrid.getRasterBuf().getElemByte(i, j, data); |
265 |
c= getPixelClassForTypeByte(data); |
266 |
rasterResult.setElem(i, j, 0,(byte) c); |
267 |
|
268 |
} |
269 |
percent = i*100/inputGridNX;
|
270 |
} |
271 |
} |
272 |
// Caso buffer tipo short
|
273 |
if (datType == RasterBuffer.TYPE_SHORT){
|
274 |
short data[]= new short[bandCount]; |
275 |
for(int i=0; i<iNY;i++){ |
276 |
for(int j=0; j<iNX;j++){ |
277 |
inputGrid.getRasterBuf().getElemShort(i, j, data); |
278 |
c= getPixelClassForTypeShort(data); |
279 |
rasterResult.setElem(i, j, 0,(byte) c); |
280 |
|
281 |
} |
282 |
percent = i*100/inputGridNX;
|
283 |
} |
284 |
} |
285 |
|
286 |
// Caso buffer tipo int
|
287 |
if (datType == RasterBuffer.TYPE_INT){
|
288 |
int data[]= new int[bandCount]; |
289 |
for(int i=0; i<iNY;i++){ |
290 |
for(int j=0; j<iNX;j++){ |
291 |
inputGrid.getRasterBuf().getElemInt(i, j, data); |
292 |
c= getPixelClassForTypeInt(data); |
293 |
rasterResult.setElem(i, j, 0,(byte) c); |
294 |
|
295 |
} |
296 |
percent = i*100/inputGridNX;
|
297 |
} |
298 |
} |
299 |
|
300 |
// Caso buffer tipo float
|
301 |
if (datType == RasterBuffer.TYPE_FLOAT){
|
302 |
float data[]= new float[bandCount]; |
303 |
for(int i=0; i<iNY;i++){ |
304 |
for(int j=0; j<iNX;j++){ |
305 |
inputGrid.getRasterBuf().getElemFloat(i, j, data); |
306 |
c= getPixelClassForTypeFloat(data); |
307 |
rasterResult.setElem(i, j, 0,(byte) c); |
308 |
|
309 |
} |
310 |
percent = i*100/inputGridNX;
|
311 |
} |
312 |
} |
313 |
|
314 |
// Caso buffer tipo double
|
315 |
if (datType == RasterBuffer.TYPE_DOUBLE){
|
316 |
double data[]= new double[bandCount]; |
317 |
for(int i=0; i<iNY;i++){ |
318 |
for(int j=0; j<iNX;j++){ |
319 |
inputGrid.getRasterBuf().getElemDouble(i, j, data); |
320 |
c= getPixelClassForTypeDouble(data); |
321 |
rasterResult.setElem(i, j, 0,(byte) c); |
322 |
|
323 |
} |
324 |
percent = i*100/inputGridNX;
|
325 |
} |
326 |
} |
327 |
|
328 |
writeToFile(); |
329 |
} |
330 |
|
331 |
|
332 |
} |