Skip to content
Snippets Groups Projects
Commit 3707dd55 authored by Jonathan Guinet's avatar Jonathan Guinet
Browse files

ENH: KMeansSegmentationApplication enhancement using colorized output.

parent 35818efb
No related branches found
No related tags found
No related merge requests found
Hide whitespace changes
Inline Side-by-side
Applications/Classification/otbKMeansClassification.cxx
+ 215
− 16
View file @ 3707dd55
......@@ -33,6 +33,12 @@
#include "otbRAMDrivenTiledStreamingManager.h"
#include "otbRAMDrivenStrippedStreamingManager.h"
#include "otbChangeLabelImageFilter.h"
#include "itkLabelToRGBImageFilter.h"
#include "otbReliefColormapFunctor.h"
#include "itkScalarToRGBColormapImageFilter.h"
#include "otbWrapperTypes.h"
namespace otb
{
......@@ -104,6 +110,13 @@ private:
typedef FloatImageType::PixelType PixelType;
typedef UInt8ImageType LabeledImageType;
typedef UInt8VectorImageType VectorImageType;
typedef VectorImageType::PixelType VectorPixelType;
typedef UInt8RGBImageType RGBImageType;
typedef RGBImageType::PixelType RGBPixelType;
typedef LabeledImageType::PixelType LabelType;
......@@ -128,6 +141,20 @@ typedef itk::UnaryFunctorImageFilter<FloatVectorImageType,
LabeledImageType, KMeansFunctorType> KMeansFilterType;
// Manual label LUT
typedef otb::ChangeLabelImageFilter
<LabeledImageType, VectorImageType> ChangeLabelFilterType;
// Continuous LUT mapping
typedef itk::ScalarToRGBColormapImageFilter<LabeledImageType, RGBImageType> ColorMapFilterType;
typedef otb::Functor::ReliefColormapFunctor
<LabelType, RGBPixelType> ReliefColorMapFunctorType;
typedef otb::ImageMetadataInterfaceBase ImageMetadataInterfaceType;
class KMeansClassification: public Application
{
public:
......@@ -154,6 +181,25 @@ private:
SetDocAuthors("OTB-Team");
SetDocSeeAlso(" ");
// Build lut map
m_LutMap["Red"]=ColorMapFilterType::Red;
m_LutMap["Green"]=ColorMapFilterType::Green;
m_LutMap["Blue"]=ColorMapFilterType::Blue;
m_LutMap["Grey"]=ColorMapFilterType::Grey;
m_LutMap["Hot"]=ColorMapFilterType::Hot;
m_LutMap["Cool"]=ColorMapFilterType::Cool;
m_LutMap["Spring"]=ColorMapFilterType::Spring;
m_LutMap["Summer"]=ColorMapFilterType::Summer;
m_LutMap["Autumn"]=ColorMapFilterType::Autumn;
m_LutMap["Winter"]=ColorMapFilterType::Winter;
m_LutMap["Copper"]=ColorMapFilterType::Copper;
m_LutMap["Jet"]=ColorMapFilterType::Jet;
m_LutMap["HSV"]=ColorMapFilterType::HSV;
m_LutMap["OverUnder"]=ColorMapFilterType::OverUnder;
AddDocTag(Tags::Segmentation);
AddDocTag(Tags::Learning);
}
......@@ -166,27 +212,63 @@ private:
{
AddParameter(ParameterType_InputImage, "in", "Input Image");
SetParameterDescription("in","Input image filename.");
SetParameterDescription("in", "Input image filename.");
AddParameter(ParameterType_OutputImage, "out", "Output Image");
SetParameterDescription("out","Output image filename.");
SetParameterDescription("out", "Output image filename.");
AddParameter(ParameterType_RAM, "ram", "Available RAM");
SetDefaultParameterInt("ram", 256);
MandatoryOff("ram");
AddParameter(ParameterType_InputImage, "vm", "Validity Mask");
SetParameterDescription("vm","Validity mask. Only non-zero pixels will be used to estimate KMeans modes.");
SetParameterDescription("vm", "Validity mask. Only non-zero pixels will be used to estimate KMeans modes.");
AddParameter(ParameterType_Int, "ts", "Training set size");
SetParameterDescription("ts", "Size of the training set.");
SetDefaultParameterInt("ts", 100);
MandatoryOff("ts");
AddParameter(ParameterType_Int, "nc", "Number of classes");
SetParameterDescription("nc","number of modes, which will be used to generate class membership.");
SetDefaultParameterInt("nc", 3);
SetParameterDescription("nc", "number of modes, which will be used to generate class membership.");
SetDefaultParameterInt("nc", 5);
AddParameter(ParameterType_Int, "maxit", "Maximum number of iterations");
SetParameterDescription("maxit", "Maximum number of iterations.");
SetDefaultParameterFloat("maxit", 1000);
SetDefaultParameterInt("maxit", 1000);
MandatoryOff("maxit");
AddParameter(ParameterType_Float, "ct", "Convergence threshold");
SetParameterDescription("ct", "Convergence threshold for class centroid (L2 distance, by default 0.01).");
SetDefaultParameterFloat("ct", 0.01);
SetDefaultParameterFloat("ct", 0.0001);
MandatoryOff("ct");
AddParameter(ParameterType_Filename, "outmeans", "Centroid filename");
SetParameterDescription("outmeans", "save label centroid in txt file.");
MandatoryOff("outmeans");
SetParameterRole("outmeans", Role_Output);
AddParameter(ParameterType_Group, "mapping", "Color mapping");
SetParameterDescription("mapping", "Save colorized image and LUT in txt file.");
AddParameter(ParameterType_Filename, "mapping.labelout","Output Colorized Image");
SetParameterDescription("mapping.labelout", "Output colorized image filename.");
SetParameterRole("mapping.labelout", Role_Output);
AddParameter(ParameterType_Filename, "mapping.lutout", "Output LUT filename.");
AddParameter(ParameterType_Choice, "mapping.lut", "Look-up tables");
SetParameterDescription("mapping.lut", "Available look-up tables. RGBIntensity remap label with class centroids, which are reweighted into [0 255]");
AddChoice("mapping.lut.rgbintensity", "RGBIntensity");
AddChoice("mapping.lut.red", "Red");
AddChoice("mapping.lut.green", "Green");
AddChoice("mapping.lut.blue", "Blue");
AddChoice("mapping.lut.grey", "Grey");
AddChoice("mapping.lut.hot", "Hot");
AddChoice("mapping.lut.cool", "Cool");
AddChoice("mapping.lut.spring", "Spring");
AddChoice("mapping.lut.summer", "Summer");
AddChoice("mapping.lut.autumn", "Autumn");
AddChoice("mapping.lut.winter", "Winter");
AddChoice("mapping.lut.copper", "Copper");
AddChoice("mapping.lut.jet", "Jet");
AddChoice("mapping.lut.hsv", "HSV");
AddChoice("mapping.lut.overunder", "OverUnder");
AddChoice("mapping.lut.relief", "Relief");
SetParameterInt("mapping.lut", 0);
MandatoryOff("mapping");
// Doc example parameter settings
SetDocExampleParameterValue("in", "qb_RoadExtract.img");
......@@ -194,7 +276,7 @@ private:
SetDocExampleParameterValue("ts", "1000");
SetDocExampleParameterValue("nc", "5");
SetDocExampleParameterValue("maxit", "1000");
SetDocExampleParameterValue("ct", "0.01");
SetDocExampleParameterValue("ct", "0.0001");
SetDocExampleParameterValue("out", "ClassificationFilterOuptut.tif");
}
......@@ -374,10 +456,10 @@ private:
message.clear();
message << std::endl;
for (unsigned int i = 0; i < nbClasses; ++i)
for (unsigned int i = 0; i < nbClasses; i++)
{
message << "Class " << i << ": ";
for (unsigned int j = 0; j < sampleSize; ++j)
for (unsigned int j = 0; j < sampleSize; j++)
{
message << std::setw(8) << initialMeans[i * sampleSize + j] << " ";
}
......@@ -412,10 +494,10 @@ private:
message.clear();
message << "Estimated centroids are: " << std::endl;
message << std::endl;
for (unsigned int i = 0; i < nbClasses; ++i)
for (unsigned int i = 0; i < nbClasses; i++)
{
message << "Class " << i << ": ";
for (unsigned int j = 0; j < sampleSize; ++j)
for (unsigned int j = 0; j < sampleSize; j++)
{
message << std::setw(8) << estimatedMeans[i * sampleSize + j] << " ";
}
......@@ -429,8 +511,7 @@ private:
/*******************************************/
/* Classification */
/*******************************************/
otbAppLogINFO("-- CLASSIFICATION --" << std::endl);
/*******************************************/otbAppLogINFO("-- CLASSIFICATION --" << std::endl);
// Finally, update the KMeans filter
KMeansFunctorType functor;
......@@ -450,17 +531,135 @@ private:
m_KMeansFilter->SetFunctor(functor);
m_KMeansFilter->SetInput(m_InImage);
// optional saving option -> lut
if (IsParameterEnabled("outmeans"))
{
std::ofstream file;
file.open(GetParameterString("outmeans").c_str());
for (unsigned int i = 0; i < nbClasses; i++)
{
for (unsigned int j = 0; j < sampleSize; j++)
{
file << std::setw(8) << estimatedMeans[i * sampleSize + j] << " ";
}
file << std::endl;
}
file.close();
}
// colormapping
if (IsParameterEnabled("mapping"))
{
m_ParamOut = OutputImageParameter::New();
m_ParamOut->SetPixelType(ImagePixelType_uint8);
m_ParamOut->SetFileName(GetParameterString("mapping.labelout"));
std::string lut = GetParameterString("mapping.lut");
if (lut == "RGBIntensity")
{
// find RGB
ImageMetadataInterfaceType::Pointer
metadataInterface = ImageMetadataInterfaceFactory::CreateIMI(m_InImage->GetMetaDataDictionary());
std::vector<unsigned int> RGBIndex = metadataInterface->GetDefaultDisplay();
otbAppLogINFO(" RGB index are "<<RGBIndex[0]<<" "<<RGBIndex[1]<<" "<<RGBIndex[2]<<std::endl);
// map intensity
m_CustomMapper = ChangeLabelFilterType::New();
m_CustomMapper->SetInput(m_KMeansFilter->GetOutput());
const unsigned int nbCompRGB = 3;
m_CustomMapper->SetNumberOfComponentsPerPixel(nbCompRGB);
EstimatorType::ParametersType weithedMeans = estimatedMeans;
std::vector<double> minVal(nbCompRGB);
std::vector<double> maxVal(nbCompRGB);
for (unsigned int i = 0; i < nbCompRGB; i++)
{
minVal[i] = estimatedMeans[i];
maxVal[i] = estimatedMeans[i];
for (unsigned int j = 0; j < nbClasses; j++)
{
double value = estimatedMeans[j * sampleSize + i];
std::cout << value << " " << std::endl;
if (value > maxVal[i]) maxVal[i] = value;
if (value < minVal[i]) minVal[i] = value;
}
}
for (unsigned int j = 0; j < nbClasses; j++)
{
VectorPixelType color(nbCompRGB);
color.Fill(0);
for (unsigned int i = 0; i < nbCompRGB; i++)
{
color[RGBIndex[i]] = static_cast<PixelType> (((estimatedMeans[j * sampleSize + i] - minVal[i]) / (maxVal[i]
- minVal[i])) * 255.0);
}
otbAppLogINFO("Adding color mapping " << j << " -> [" << (int) color[0] << " " << (int) color[1] << " "<< (int) color[2] << " ]" << std::endl);
m_CustomMapper->SetChange(j, color);
}
// Create an output parameter to write the current output image
m_ParamOut->SetValue(m_CustomMapper->GetOutput());
m_ParamOut->InitializeWriters();
AddProcess(m_ParamOut->GetWriter(), GetParameterString("mapping.labelout"));
m_ParamOut->Write();
}
else
{
m_ColorMapper = ColorMapFilterType::New();
m_ColorMapper->SetInput(m_KMeansFilter->GetOutput());
m_ColorMapper->UseInputImageExtremaForScalingOn();
if (lut == "Relief")
{
ReliefColorMapFunctorType::Pointer reliefFunctor = ReliefColorMapFunctorType::New();
m_ColorMapper->SetColormap(reliefFunctor);
}
else
{
m_ColorMapper->SetColormap((ColorMapFilterType::ColormapEnumType) m_LutMap[lut]);
}
// reweight label into [0 255]
m_ColorMapper->GetColormap()->SetMinimumInputValue(0.0);
m_ColorMapper->GetColormap()->SetMaximumInputValue(255.0);
m_ParamOut->SetValue(m_ColorMapper->GetOutput());
m_ParamOut->InitializeWriters();
AddProcess(m_ParamOut->GetWriter(), GetParameterString("mapping.labelout"));
m_ParamOut->Write();
}
}
SetParameterOutputImage("out", m_KMeansFilter->GetOutput());
}
// KMeans filter
KMeansFilterType::Pointer m_KMeansFilter;
FloatVectorImageType::Pointer m_InImage;
OutputImageParameter::Pointer m_ParamOut;
KMeansFilterType::Pointer m_KMeansFilter;
FloatVectorImageType::Pointer m_InImage;
ChangeLabelFilterType::Pointer m_CustomMapper;
ColorMapFilterType::Pointer m_ColorMapper;
std::map<std::string, unsigned int> m_LutMap;
};
}
}
......
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment