Application: gvSIG desktop » gvSIG tools: ToolBox

package es.unex.sextante.statisticalMethods.covarianceMatrix;

import java.util.ArrayList;

import es.unex.sextante.additionalInfo.AdditionalInfoMultipleInput;

import es.unex.sextante.core.GeoAlgorithm;

import es.unex.sextante.core.Sextante;

import es.unex.sextante.dataObjects.IRasterLayer;

import es.unex.sextante.dataObjects.ITable;

import es.unex.sextante.exceptions.GeoAlgorithmExecutionException;

import es.unex.sextante.exceptions.RepeatedParameterNameException;

public class CovarianceMatrixAlgorithm

         extends

            GeoAlgorithm {

   public static final String COVARIANCES = "COVARIANCES";

   public static final String INPUT       = "INPUT";

   private static double      NODATA      = -9999999.;

   private ArrayList          m_RasterLayers;

   private IRasterLayer[]     m_Windows;

   private int                m_iNX, m_iNY;

   private double             m_dMean[];

   @Override

   public void defineCharacteristics() {

      setUserCanDefineAnalysisExtent(true);

      setGroup(Sextante.getText("Statistical_methods"));

      setName(Sextante.getText("Covariance_matrix"));

      try {

         m_Parameters.addMultipleInput(INPUT, Sextante.getText("Input_layers"), AdditionalInfoMultipleInput.DATA_TYPE_RASTER,

                  true);

         addOutputTable(COVARIANCES, Sextante.getText("Covariance_matrix"));

      }

      catch (final RepeatedParameterNameException e) {

         Sextante.addErrorToLog(e);

      }

   }

   @Override

   public boolean processAlgorithm() throws GeoAlgorithmExecutionException {

      int i, j;

      m_RasterLayers = m_Parameters.getParameterValueAsArrayList(INPUT);

      if (m_RasterLayers.size() == 0) {

         return false;

      }

      final Object[] values = new Object[m_RasterLayers.size()];

      final String sFields[] = new String[m_RasterLayers.size()];

      final Class iTypes[] = new Class[m_RasterLayers.size()];

      final String sTableName = "Matriz de covarianzas";

      final double dCovar[][] = new double[m_RasterLayers.size()][m_RasterLayers.size()];

      m_dMean = new double[m_RasterLayers.size()];

      this.adjustOutputExtent();

      m_Windows = new IRasterLayer[m_RasterLayers.size()];

      for (i = 0; i < m_RasterLayers.size(); i++) {

         m_Windows[i] = (IRasterLayer) m_RasterLayers.get(i);

         m_Windows[i].setWindowExtent(this.getAnalysisExtent());

         m_Windows[i].setInterpolationMethod(IRasterLayer.INTERPOLATION_BSpline);

         sFields[i] = m_Windows[i].getName();

         iTypes[i] = Double.class;

         m_dMean[i] = m_Windows[i].getMeanValue();

      }

      final ITable table = getNewTable(COVARIANCES, sTableName, iTypes, sFields);

      m_iNX = getAnalysisExtent().getNX();

      m_iNY = getAnalysisExtent().getNY();

      final int iTotal = (int) (m_RasterLayers.size() * m_RasterLayers.size() / 2.);

      int iCount = 0;

      for (i = 0; (i < m_RasterLayers.size() - 1) && setProgress(iCount, iTotal); i++) {

         dCovar[i][i] = 1.0;

         iCount++;

         for (j = i + 1; j < m_RasterLayers.size(); j++) {

            dCovar[i][j] = dCovar[j][i] = getCovar(i, j);

            iCount++;

         }

      }

      for (i = 0; i < m_RasterLayers.size(); i++) {

         for (j = 0; j < m_RasterLayers.size(); j++) {

            values[j] = new Double(dCovar[i][j]);

         }

         table.addRecord(values);

      }

      return !m_Task.isCanceled();

   }

   private double getCovar(final int i,

                           final int j) {

      int x, y;

      int iValues = 0;

      double dValuei, dValuej;

      double dSum = 0;

      for (y = 0; (y < m_iNY) && setProgress(y, m_iNY); y++) {

         for (x = 0; x < m_iNX; x++) {

            dValuei = m_Windows[i].getCellValueAsDouble(x, y);

            dValuej = m_Windows[j].getCellValueAsDouble(x, y);

            if (!m_Windows[i].isNoDataValue(dValuei) && !m_Windows[j].isNoDataValue(dValuej)) {

               dSum += (dValuei - m_dMean[i]) * (dValuej - m_dMean[j]);

               iValues++;

            }

         }

         if (iValues > 1) {

            return dSum / (iValues - 1);

         }

         else {

            return NODATA;

         }

      }

      return NODATA;

   }

}