Revision 14322 trunk/extensions/extRemoteSensing/src/org/gvsig/remotesensing/classification/ClassificationProcess.java
| ClassificationProcess.java | ||
|---|---|---|
| 45 | 45 |
import java.awt.geom.AffineTransform; |
| 46 | 46 |
import java.io.File; |
| 47 | 47 |
import java.io.IOException; |
| 48 |
import java.util.ArrayList; |
|
| 49 |
import java.util.Iterator; |
|
| 48 | 50 |
|
| 49 | 51 |
import javax.swing.JOptionPane; |
| 50 | 52 |
|
| ... | ... | |
| 53 | 55 |
import org.gvsig.gui.beans.incrementabletask.IncrementableEvent; |
| 54 | 56 |
import org.gvsig.gui.beans.incrementabletask.IncrementableListener; |
| 55 | 57 |
import org.gvsig.gui.beans.incrementabletask.IncrementableTask; |
| 56 |
import org.gvsig.gui.beans.table.exceptions.NotInitializeException; |
|
| 57 | 58 |
import org.gvsig.raster.buffer.RasterBuffer; |
| 58 | 59 |
import org.gvsig.raster.dataset.GeoRasterWriter; |
| 59 | 60 |
import org.gvsig.raster.dataset.IBuffer; |
| 60 | 61 |
import org.gvsig.raster.dataset.NotSupportedExtensionException; |
| 61 | 62 |
import org.gvsig.raster.dataset.RasterDriverException; |
| 63 |
import org.gvsig.raster.datastruct.ColorItem; |
|
| 64 |
import org.gvsig.raster.datastruct.ColorTable; |
|
| 62 | 65 |
import org.gvsig.raster.grid.Grid; |
| 66 |
import org.gvsig.raster.grid.GridPalette; |
|
| 67 |
import org.gvsig.raster.grid.filter.RasterFilterList; |
|
| 68 |
import org.gvsig.raster.grid.filter.RasterFilterListManager; |
|
| 69 |
import org.gvsig.raster.grid.filter.bands.ColorTableFilter; |
|
| 70 |
import org.gvsig.raster.grid.filter.bands.ColorTableListManager; |
|
| 71 |
import org.gvsig.raster.grid.filter.enhancement.LinearEnhancementFilter; |
|
| 72 |
import org.gvsig.raster.grid.filter.statistics.TailTrimFilter; |
|
| 73 |
import org.gvsig.raster.grid.roi.ROI; |
|
| 74 |
import org.gvsig.raster.hierarchy.IRasterRendering; |
|
| 63 | 75 |
import org.gvsig.raster.process.CancelEvent; |
| 64 | 76 |
import org.gvsig.raster.process.RasterTask; |
| 65 | 77 |
import org.gvsig.raster.process.RasterTaskQueue; |
| 66 | 78 |
import org.gvsig.rastertools.cutting.WriterBufferServer; |
| 67 |
import org.gvsig.rastertools.roi.ui.ROIManagerPanel; |
|
| 68 | 79 |
|
| 69 | 80 |
import Jama.Matrix; |
| 70 | 81 |
|
| ... | ... | |
| 87 | 98 |
private MapContext mapContext = null; |
| 88 | 99 |
private int percent = 0; |
| 89 | 100 |
private Thread blinker = null; |
| 90 |
private ROIManagerPanel roiPanel = null;
|
|
| 101 |
private ArrayList rois = null;
|
|
| 91 | 102 |
private IncrementableTask incrementableTask = null; |
| 92 | 103 |
private WriterBufferServer writerBufferServer = null; |
| 93 | 104 |
private String fileNameOutput = null; |
| ... | ... | |
| 106 | 117 |
* @param filenameOutput string con la ruta del fichero de salida |
| 107 | 118 |
* atransform del raster original |
| 108 | 119 |
* */ |
| 109 |
public ClassificationProcess(Grid imageGrid, ROIManagerPanel roisP, MapContext mapContext, String fileNameOutput,AffineTransform atransform){
|
|
| 120 |
public ClassificationProcess(Grid imageGrid, ArrayList rois, MapContext mapContext, String fileNameOutput,AffineTransform atransform){
|
|
| 110 | 121 |
inputGrid= imageGrid; |
| 111 |
roiPanel= roisP;
|
|
| 122 |
this.rois = rois;
|
|
| 112 | 123 |
this.mapContext=mapContext; |
| 113 | 124 |
this.fileNameOutput= fileNameOutput; |
| 114 | 125 |
this.atransform= atransform; |
| ... | ... | |
| 130 | 141 |
int iNX= inputGrid.getLayerNX(); |
| 131 | 142 |
|
| 132 | 143 |
try{
|
| 133 |
numClases = roiPanel.getTable().getRowCount();
|
|
| 144 |
numClases = rois.size();
|
|
| 134 | 145 |
// Se calculan las inversas de las matrices de Varianza-covarianza de todas las rois y se almacenan en inverseVarCovMAtrix |
| 135 | 146 |
Matrix Sinverse=null; |
| 136 | 147 |
inverseVarCovMatrix= new Matrix[numClases]; |
| 137 | 148 |
|
| 138 | 149 |
for (int i=0; i<numClases;i++){
|
| 139 |
S= new Matrix((roiPanel.getROI("ROI" + String.valueOf(i)).getRoi()).getVarCovMatrix());
|
|
| 150 |
S= new Matrix(((ROI)rois.get(i)).getVarCovMatrix());
|
|
| 140 | 151 |
Sinverse= S.inverse(); |
| 141 | 152 |
inverseVarCovMatrix[i]= Sinverse; |
| 142 | 153 |
} |
| ... | ... | |
| 224 | 235 |
if (endIndex < 0) |
| 225 | 236 |
endIndex = fileNameOutput.length(); |
| 226 | 237 |
lyr = FLyrRasterSE.createLayer(fileNameOutput.substring(fileNameOutput.lastIndexOf(File.separator) + 1, endIndex),new File(fileNameOutput), mapContext.getProjection()); |
| 238 |
|
|
| 239 |
setLeyend(lyr); |
|
| 240 |
|
|
| 227 | 241 |
mapContext.getLayers().addLayer(lyr); |
| 228 | 242 |
mapContext.endAtomicEvent(); |
| 229 | 243 |
mapContext.invalidate(); |
| ... | ... | |
| 236 | 250 |
e.printStackTrace(); |
| 237 | 251 |
} catch (IOException e) {
|
| 238 | 252 |
e.printStackTrace(); |
| 239 |
} catch (NotInitializeException e) {
|
|
| 240 |
e.printStackTrace(); |
|
| 241 |
} |
|
| 242 |
catch (InterruptedException e) {
|
|
| 253 |
} catch (InterruptedException e) {
|
|
| 243 | 254 |
Thread.currentThread().interrupt(); |
| 244 | 255 |
} catch (LoadLayerException e) {
|
| 245 | 256 |
JOptionPane.showMessageDialog((Component)PluginServices.getMainFrame(), |
| ... | ... | |
| 249 | 260 |
} |
| 250 | 261 |
|
| 251 | 262 |
} // Fin run() |
| 263 |
|
|
| 264 |
private void setLeyend(FLayer lyr) {
|
|
| 265 |
if(lyr instanceof IRasterRendering) {
|
|
| 266 |
IRasterRendering rendering = (IRasterRendering) lyr; |
|
| 267 |
ArrayList colorItems = new ArrayList(); |
|
| 268 |
ColorItem colorItem = null; |
|
| 269 |
int classValue = 0; |
|
| 270 |
for (Iterator iter = rois.iterator(); iter.hasNext();) {
|
|
| 271 |
ROI roi = (ROI) iter.next(); |
|
| 272 |
colorItem = new ColorItem(); |
|
| 273 |
colorItem.setColor(roi.getColor()); |
|
| 274 |
colorItem.setNameClass(roi.getName()); |
|
| 275 |
colorItem.setValue(classValue); |
|
| 276 |
colorItems.add(colorItem); |
|
| 277 |
classValue++; |
|
| 278 |
} |
|
| 279 |
ColorTable colorTable = new ColorTable(); |
|
| 280 |
colorTable.createPaletteFromColorItems(colorItems, false); |
|
| 281 |
|
|
| 282 |
RasterFilterList filterList = rendering.getRenderFilterList(); |
|
| 283 |
RasterFilterListManager manager = new RasterFilterListManager(filterList); |
|
| 284 |
ColorTableListManager cManager = (ColorTableListManager) manager.getManagerByClass(ColorTableListManager.class); |
|
| 252 | 285 |
|
| 253 |
|
|
| 254 |
|
|
| 286 |
filterList.remove(ColorTableFilter.class); |
|
| 287 |
|
|
| 288 |
((FLyrRasterSE)lyr).setLastLegend(null); |
|
| 289 |
|
|
| 290 |
filterList.remove(LinearEnhancementFilter.class); |
|
| 291 |
filterList.remove(TailTrimFilter.class); |
|
| 292 |
GridPalette gridPalette = new GridPalette(colorTable); |
|
| 293 |
cManager.addColorTableFilter(gridPalette); |
|
| 294 |
((FLyrRasterSE)lyr).setLastLegend(gridPalette); |
|
| 295 |
|
|
| 296 |
rendering.setRenderFilterList(filterList); |
|
| 297 |
} |
|
| 298 |
} |
|
| 299 |
|
|
| 300 |
|
|
| 255 | 301 |
/** |
| 256 | 302 |
* Metodo que implementa el clasificador de maxima probabilidad. |
| 257 | 303 |
* Para cada pixel, obtiene la calase que minimiza la expresion: -Ln(P(x))= Ln(|Si|)+Y'* inverse(Si)*Y |
| ... | ... | |
| 265 | 311 |
{
|
| 266 | 312 |
double[][] y = new double[inputGrid.getBandCount()][1]; |
| 267 | 313 |
for (int i=0;i<inputGrid.getBandCount();i++){
|
| 268 |
roiPanel.getROI("ROI" + String.valueOf(clase)).getRoi().setBandToOperate(i);
|
|
| 269 |
y[i][0]=pixelBandsValues[i]-(roiPanel.getROI("ROI" + String.valueOf(clase)).getRoi().getMeanValue());
|
|
| 314 |
((ROI)rois.get(clase)).setBandToOperate(i);
|
|
| 315 |
y[i][0]=pixelBandsValues[i]-((ROI)rois.get(clase)).getMeanValue();
|
|
| 270 | 316 |
} |
| 271 | 317 |
Y = new Matrix(y); |
| 272 | 318 |
result= Y.transpose().times((Matrix)inverseVarCovMatrix[clase]).times(Y); |
| ... | ... | |
| 300 | 346 |
{
|
| 301 | 347 |
double[][] y = new double[inputGrid.getBandCount()][1]; |
| 302 | 348 |
for (int i=0;i<inputGrid.getBandCount();i++){
|
| 303 |
roiPanel.getROI("ROI" + String.valueOf(clase)).getRoi().setBandToOperate(i);
|
|
| 304 |
y[i][0]=pixelBandsValues[i]-(roiPanel.getROI("ROI" + String.valueOf(clase)).getRoi().getMeanValue());
|
|
| 349 |
((ROI)rois.get(clase)).setBandToOperate(i);
|
|
| 350 |
y[i][0]=pixelBandsValues[i]-((ROI)rois.get(clase)).getMeanValue();
|
|
| 305 | 351 |
} |
| 306 | 352 |
Y = new Matrix(y); |
| 307 | 353 |
result= (Y.transpose().times(inverseVarCovMatrix[clase])).times(Y); |
| ... | ... | |
| 335 | 381 |
{
|
| 336 | 382 |
double[][] y = new double[inputGrid.getBandCount()][1]; |
| 337 | 383 |
for (int i=0;i<inputGrid.getBandCount();i++){
|
| 338 |
roiPanel.getROI("ROI" + String.valueOf(clase)).getRoi().setBandToOperate(i);
|
|
| 339 |
y[i][0]=pixelBandsValues[i]-(roiPanel.getROI("ROI" + String.valueOf(clase)).getRoi().getMeanValue());
|
|
| 384 |
((ROI)rois.get(clase)).setBandToOperate(i);
|
|
| 385 |
y[i][0]=pixelBandsValues[i]-((ROI)rois.get(clase)).getMeanValue();
|
|
| 340 | 386 |
} |
| 341 | 387 |
Y = new Matrix(y); |
| 342 | 388 |
result= (Y.transpose().times(inverseVarCovMatrix[clase])).times(Y); |
| ... | ... | |
| 369 | 415 |
{
|
| 370 | 416 |
double[][] y = new double[inputGrid.getBandCount()][1]; |
| 371 | 417 |
for (int i=0;i<inputGrid.getBandCount();i++){
|
| 372 |
roiPanel.getROI("ROI" + String.valueOf(clase)).getRoi().setBandToOperate(i);
|
|
| 373 |
y[i][0]=pixelBandsValues[i]-(roiPanel.getROI("ROI" + String.valueOf(clase)).getRoi().getMeanValue());
|
|
| 418 |
((ROI)rois.get(clase)).setBandToOperate(i);
|
|
| 419 |
y[i][0]=pixelBandsValues[i]-((ROI)rois.get(clase)).getMeanValue();
|
|
| 374 | 420 |
} |
| 375 | 421 |
Y = new Matrix(y); |
| 376 | 422 |
result= (Y.transpose().times(inverseVarCovMatrix[clase])).times(Y); |
| ... | ... | |
| 389 | 435 |
return clasefinal; |
| 390 | 436 |
} |
| 391 | 437 |
|
| 392 |
|
|
| 438 |
|
|
| 393 | 439 |
/** |
| 394 | 440 |
* Metodo que implementa el clasificador de maxima probabilidad. |
| 395 | 441 |
* Para cada pixel, obtiene la calase que minimiza la expresion: -Ln(P(x))= Ln(|Si|)+Y'* inverse(Si)*Y |
| ... | ... | |
| 398 | 444 |
* @return clase a la que pertenece el pixel (por el metodo de maxima probabilidad) |
| 399 | 445 |
* */ |
| 400 | 446 |
public int getPixelClassForTypeDouble(double pixelBandsValues[]){
|
| 401 |
double probabilidades[]=new double[roiPanel.getTable().getHeight()];
|
|
| 447 |
double probabilidades[]=new double[numClases];
|
|
| 402 | 448 |
for (int clase=0; clase<numClases;clase++) |
| 403 | 449 |
{
|
| 404 | 450 |
double[][] y = new double[inputGrid.getBandCount()][1]; |
| 405 | 451 |
for (int i=0;i<inputGrid.getBandCount();i++){
|
| 406 |
roiPanel.getROI("ROI" + String.valueOf(clase)).getRoi().setBandToOperate(i);
|
|
| 407 |
y[i][0]=pixelBandsValues[i]-(roiPanel.getROI("ROI" + String.valueOf(clase)).getRoi().getMeanValue());
|
|
| 452 |
((ROI)rois.get(clase)).setBandToOperate(i);
|
|
| 453 |
y[i][0]=pixelBandsValues[i]-((ROI)rois.get(clase)).getMeanValue();
|
|
| 408 | 454 |
} |
| 409 | 455 |
Y = new Matrix(y); |
| 410 | 456 |
result= (Y.transpose().times(inverseVarCovMatrix[clase])).times(Y); |
| ... | ... | |
| 423 | 469 |
return clasefinal; |
| 424 | 470 |
} |
| 425 | 471 |
|
| 426 |
|
|
| 427 | 472 |
public RasterBuffer getGridResult(){
|
| 428 | 473 |
return rasterResult; |
| 429 | 474 |
} |
Also available in: Unified diff