otbTrainOGRLayersClassifier.cxx 7.47 KB
Newer Older
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
/*=========================================================================
 Program:   ORFEO Toolbox
 Language:  C++
 Date:      $Date$
 Version:   $Revision$


 Copyright (c) Centre National d'Etudes Spatiales. All rights reserved.
 See OTBCopyright.txt for details.


 This software is distributed WITHOUT ANY WARRANTY; without even
 the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR
 PURPOSE.  See the above copyright notices for more information.

 =========================================================================*/
#include "otbWrapperApplication.h"
#include "otbWrapperApplicationFactory.h"

#include "otbOGRDataSourceWrapper.h"
#include "otbOGRFeatureWrapper.h"
#include "otbStatisticsXMLFileWriter.h"

#include "itkVariableLengthVector.h"
#include "otbStatisticsXMLFileReader.h"

#include "itkListSample.h"
#include "otbShiftScaleSampleListFilter.h"
#include "otbLibSVMMachineLearningModel.h"

#include <time.h>

namespace otb
{
namespace Wrapper
{
class TrainOGRLayersClassifier : public Application
{
public:
  typedef TrainOGRLayersClassifier Self;
  typedef Application Superclass;
  typedef itk::SmartPointer<Self> Pointer;
  typedef itk::SmartPointer<const Self> ConstPointer;
  itkNewMacro(Self)
;

  itkTypeMacro(TrainOGRLayersClassifier, otb::Application)
;

private:
  void DoInit()
  {
    SetName("TrainOGRLayersClassifier");
    SetDescription("Train a SVM classifier based on labeled geometries and a list of features to consider.");

    SetDocName("TrainOGRLayersClassifier");
    SetDocLongDescription("This application trains a SVM classifier based on labeled geometries and a list of features to consider for classification.");
    SetDocLimitations("Experimental. For now only shapefiles are supported. Tuning of SVM classifier is not available.");
    SetDocAuthors("David Youssefi during internship at CNES");
    SetDocSeeAlso("OGRLayerClassifier,ComputeOGRLayersFeaturesStatistics");
    AddDocTag(Tags::Segmentation);
  
    AddParameter(ParameterType_InputFilename, "inshp", "Name of the input shapefile");
    SetParameterDescription("inshp","Name of the input shapefile");

66 67
    AddParameter(ParameterType_InputFilename, "instats", "XML file containing mean and variance of each feature.");
    SetParameterDescription("instats", "XML file containing mean and variance of each feature.");
68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83

    AddParameter(ParameterType_OutputFilename, "outsvm", "Output model filename.");
    SetParameterDescription("outsvm", "Output model filename.");

    AddParameter(ParameterType_ListView,  "feat", "List of features to consider for classification.");
    SetParameterDescription("feat","List of features to consider for classification.");

    AddParameter(ParameterType_String,"cfield","Field containing the class id for supervision");
    SetParameterDescription("cfield","Field containing the class id for supervision. Only geometries with this field available will be taken into account.");
    SetParameterString("cfield","class");

  }

  void DoUpdateParameters()
  {
    if ( HasValue("inshp") )
OTB Bot's avatar
OTB Bot committed
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
      {
       const char * shapefile = GetParameterString("inshp").c_str();

       otb::ogr::DataSource::Pointer ogrDS;
       otb::ogr::Layer layer(NULL, false);

       OGRSpatialReference oSRS("");
       std::vector<std::string> options;
       
       ogrDS = otb::ogr::DataSource::New(shapefile, otb::ogr::DataSource::Modes::Read);
       std::string layername = itksys::SystemTools::GetFilenameName(shapefile);
       layername = layername.substr(0,layername.size()-4);
       layer = ogrDS->GetLayer(0);

       otb::ogr::Feature feature = layer.ogr().GetNextFeature();
       ClearChoices("feat");
       for(int iField=0; iField<feature.ogr().GetFieldCount(); iField++)
         {
           std::string key, item = feature.ogr().GetFieldDefnRef(iField)->GetNameRef();
           key = item;
           key.erase(std::remove(key.begin(), key.end(), ' '), key.end());
           std::transform(key.begin(), key.end(), key.begin(), tolower);
           key="feat."+key;
           AddChoice(key,item);
         }
109 110 111 112 113 114 115 116
      }
  }

  void DoExecute()
  {
    clock_t tic = clock();

    const char * shapefile = GetParameterString("inshp").c_str();
117
    const char * XMLfile = GetParameterString("instats").c_str();
118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135
    const char * modelfile = GetParameterString("outsvm").c_str();

    typedef double ValueType;
    typedef itk::VariableLengthVector<ValueType> MeasurementType;
    typedef itk::Statistics::ListSample <MeasurementType> ListSampleType;
    typedef otb::StatisticsXMLFileReader<MeasurementType> StatisticsReader;
  
    typedef unsigned int LabelPixelType;
    typedef itk::FixedArray<LabelPixelType,1> LabelSampleType;
    typedef itk::Statistics::ListSample <LabelSampleType> LabelListSampleType;
  
    typedef otb::Statistics::ShiftScaleSampleListFilter<ListSampleType, ListSampleType> ShiftScaleFilterType;

    StatisticsReader::Pointer statisticsReader = StatisticsReader::New();
    statisticsReader->SetFileName(XMLfile);

    MeasurementType meanMeasurementVector = statisticsReader->GetStatisticVectorByName("mean");
    MeasurementType stddevMeasurementVector = statisticsReader->GetStatisticVectorByName("stddev");
136
   
OTB Bot's avatar
OTB Bot committed
137 138
    otb::ogr::DataSource::Pointer source = otb::ogr::DataSource::New(shapefile, otb::ogr::DataSource::Modes::Read);
    otb::ogr::Layer layer = source->GetLayer(0);
139 140 141 142 143 144
    bool goesOn = true;
    otb::ogr::Feature feature = layer.ogr().GetNextFeature();

    ListSampleType::Pointer input = ListSampleType::New();
    LabelListSampleType::Pointer target = LabelListSampleType::New();
    const int nbFeatures = GetSelectedItems("feat").size();
145 146 147

    input->SetMeasurementVectorSize(nbFeatures);
   
148 149
    if(feature.addr())
      while(goesOn)
OTB Bot's avatar
OTB Bot committed
150
       {
151
        if(feature.ogr().IsFieldSet(feature.ogr().GetFieldIndex(GetParameterString("cfield").c_str())))
OTB Bot's avatar
OTB Bot committed
152 153 154 155 156 157 158 159 160 161 162 163
           {
             MeasurementType mv; mv.SetSize(nbFeatures);
             
             for(int idx=0; idx < nbFeatures; ++idx)
              mv[idx] = feature.ogr().GetFieldAsDouble(GetSelectedItems("feat")[idx]);
             
             input->PushBack(mv);
             target->PushBack(feature.ogr().GetFieldAsInteger(GetParameterString("cfield").c_str()));
           }
         feature = layer.ogr().GetNextFeature();
         goesOn = feature.addr() != 0;
       }
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

    ShiftScaleFilterType::Pointer trainingShiftScaleFilter = ShiftScaleFilterType::New();
    trainingShiftScaleFilter->SetInput(input);
    trainingShiftScaleFilter->SetShifts(meanMeasurementVector);
    trainingShiftScaleFilter->SetScales(stddevMeasurementVector);
    trainingShiftScaleFilter->Update();
  
    ListSampleType::Pointer listSample;
    LabelListSampleType::Pointer labelListSample;

    listSample = trainingShiftScaleFilter->GetOutput();
    labelListSample = target;

    ListSampleType::Pointer trainingListSample = listSample;
    LabelListSampleType::Pointer trainingLabeledListSample = labelListSample;

    typedef otb::LibSVMMachineLearningModel<ValueType,LabelPixelType> LibSVMType;
    LibSVMType::Pointer libSVMClassifier = LibSVMType::New();
    libSVMClassifier->SetInputListSample(trainingListSample);
    libSVMClassifier->SetTargetListSample(trainingLabeledListSample);
    libSVMClassifier->SetParameterOptimization(true);
    libSVMClassifier->SetC(1.0);
    libSVMClassifier->SetKernelType(LINEAR);
    libSVMClassifier->Train();
    libSVMClassifier->Save(modelfile);

    clock_t toc = clock();

    otbAppLogINFO( "Elapsed: "<< ((double)(toc - tic) / CLOCKS_PER_SEC)<<" seconds.");
    }

};
}
}

OTB_APPLICATION_EXPORT(otb::Wrapper::TrainOGRLayersClassifier)