This is an automated email from the git hooks/post-receive script.
sebastic-guest pushed a commit to branch upstream-master
in repository pktools.
commit d8c43962e6d4bd47ac3577c5facbc49d3aea8cd9
Author: Pieter Kempeneers <kempe...@gmail.com>
Date: Thu May 29 09:54:31 2014 +0200
ConfusionMatix: must find exact match in getClassIndex, pkfssvm and
pkfsann: confusion matrix global for class names
---
ChangeLog | 2 +
doc/mainpage.dox | 18 ++++---
src/algorithms/ConfusionMatrix.h | 2 +-
src/algorithms/Filter2d.cc | 8 +--
src/algorithms/Filter2d.h | 1 +
src/apps/pkdiff.cc | 36 ++++++-------
src/apps/pkfsann.cc | 106 +++++++++++++++++++++++++--------------
src/apps/pkfssvm.cc | 106 +++++++++++++++++++++++----------------
src/apps/pklas2img.cc | 3 +-
src/apps/pksvm.cc | 2 +-
10 files changed, 173 insertions(+), 111 deletions(-)
diff --git a/ChangeLog b/ChangeLog
index 73e020a..159086e 100755
--- a/ChangeLog
+++ b/ChangeLog
@@ -267,6 +267,8 @@ version 2.5.2
- configure script: GDAL>=1.10.0 is required for pkdiff
- programs ported to windows and GUI with Qt
removed underscore for QProcess in Windows
+ - algorithms
+ ConfusionMatrix.h: must find exact match in getClassIndex
- pkinfo
distinct long options
- pkclassify_svm -> pksvm
diff --git a/doc/mainpage.dox b/doc/mainpage.dox
index e2125a3..dba08e2 100644
--- a/doc/mainpage.dox
+++ b/doc/mainpage.dox
@@ -41,20 +41,26 @@ Please refer to pktools as: http://pktools.nongnu.org
(Pieter Kempeneers)
contact: kempenep at gmail.com
\section pktools_installation Installation
-Download the latest release from
http://download.savannah.gnu.org/releases/pktools/ (e.g., pktools-2.4.3.tar.gz)
+Download the latest release from
http://download.savannah.gnu.org/releases/pktools/ (pktools-latest.tar.gz)
Optionally: check the release md5sum to ensure originality.
\code
-md5sum pktools-2.5.2.tar.gz
+md5sum pktools-latest.tar.gz
\endcode
-The result should be identical to the content of the corresponding md5 file
+The result should be identical to the content of the corresponding md5 file
you can also download.
-Then unpack the release
+Ubuntu users can download and run the installation script install_pktools.sh:
+
+\code
+sudo bash install_pktools.sh
+\endcode
+
+For a manual installation, download and unpack the latest release
\code
-tar xzvf pktools-2.5.2.tar.gz
-cd pktools-2.5.2
+tar xzvf pktools-latest.tar.gz
+cd pktools-*
\endcode
Basic installation (please refer to INSTALL file for more advanced
configuration)
diff --git a/src/algorithms/ConfusionMatrix.h b/src/algorithms/ConfusionMatrix.h
index 4a2b22d..db11864 100644
--- a/src/algorithms/ConfusionMatrix.h
+++ b/src/algorithms/ConfusionMatrix.h
@@ -48,7 +48,7 @@ public:
int getClassIndex(std::string className) const {
int index=0;
for(index=0;index<m_classes.size();++index){
- if(m_classes[index].find(className)!=std::string::npos)
+ if(m_classes[index]==className)
break;
}
if(index>=m_classes.size())
diff --git a/src/algorithms/Filter2d.cc b/src/algorithms/Filter2d.cc
index e2e95eb..0fc643a 100644
--- a/src/algorithms/Filter2d.cc
+++ b/src/algorithms/Filter2d.cc
@@ -407,7 +407,7 @@ void filter2d::Filter2d::doit(const ImgReaderGdal& input,
ImgWriterGdal& output,
continue;
outBuffer[x/down]=0;
std::vector<double> windowBuffer;
- std::map<int,int> occurrence;
+ std::map<long int,int> occurrence;
int centre=dimX*(dimY-1)/2+(dimX-1)/2;
for(int j=-(dimY-1)/2;j<=dimY/2;++j){
for(int i=-(dimX-1)/2;i<=dimX/2;++i){
@@ -576,15 +576,15 @@ void filter2d::Filter2d::doit(const ImgReaderGdal& input,
ImgWriterGdal& output,
}
case(filter2d::countid):{
if(windowBuffer.size())
- outBuffer[x/down]+=occurrence.size();
+ outBuffer[x/down]=occurrence.size();
else
outBuffer[x/down]=(m_noDataValues.size())? m_noDataValues[0] : 0;
break;
}
case(filter2d::majority):{
if(occurrence.size()){
- std::map<int,int>::const_iterator maxit=occurrence.begin();
- for(std::map<int,int>::const_iterator
mit=occurrence.begin();mit!=occurrence.end();++mit){
+ std::map<long int,int>::const_iterator maxit=occurrence.begin();
+ for(std::map<long int,int>::const_iterator
mit=occurrence.begin();mit!=occurrence.end();++mit){
if(mit->second>maxit->second)
maxit=mit;
}
diff --git a/src/algorithms/Filter2d.h b/src/algorithms/Filter2d.h
index 30ef9da..dec5a32 100644
--- a/src/algorithms/Filter2d.h
+++ b/src/algorithms/Filter2d.h
@@ -165,6 +165,7 @@ private:
m_filterMap["scramble"]=filter2d::scramble;
m_filterMap["shift"]=filter2d::shift;
m_filterMap["linearfeature"]=filter2d::linearfeature;
+ m_filterMap["countid"]=filter2d::countid;
}
Vector2d<double> m_taps;
diff --git a/src/apps/pkdiff.cc b/src/apps/pkdiff.cc
index 84ff21c..f07d617 100644
--- a/src/apps/pkdiff.cc
+++ b/src/apps/pkdiff.cc
@@ -29,28 +29,28 @@ using namespace std;
int main(int argc, char *argv[])
{
- Optionpk<string> input_opt("i", "input", "Input image file.");
- Optionpk<string> reference_opt("ref", "reference", "Reference image file");
- Optionpk<string> layer_opt("ln", "ln", "layer name(s) in sample (leave empty
to select all)");
- Optionpk<string> output_opt("o", "output", "Output image file. Default is
empty: no output image, only report difference or identical.");
- Optionpk<string> ogrformat_opt("f", "f", "Output sample file
format","SQLite");
+ Optionpk<string> input_opt("i", "input", "Input raster dataset.");
+ Optionpk<string> reference_opt("ref", "reference", "Reference (raster or
vector) dataset");
+ Optionpk<string> layer_opt("ln", "ln", "layer name(s) in sample. Leave empty
to select all (for vector reference datasets only)");
+ Optionpk<string> output_opt("o", "output", "Output dataset (optional)");
+ Optionpk<string> ogrformat_opt("f", "f", "OGR format for output vector (for
vector reference datasets only)","SQLite");
Optionpk<string> mask_opt("m", "mask", "Mask image file. A single mask is
supported only, but several mask values can be used. See also msknodata option.
(default is empty)");
Optionpk<int> masknodata_opt("msknodata", "msknodata", "Mask value(s) where
image is invalid. Use negative value for valid data (example: use -t -1: if
only -1 is valid value)", 0);
- Optionpk<string> colorTable_opt("ct", "ct", "color table (file with 5
columns: id R G B ALFA (0: transparent, 255: solid)");
Optionpk<short> valueE_opt("\0", "correct", "Value for correct pixels", 0,2);
Optionpk<short> valueO_opt("\0", "omission", "Value for omission errors:
input label > reference label", 1,2);
Optionpk<short> valueC_opt("\0", "commission", "Value for commission errors:
input label < reference label", 2,1);
Optionpk<short> nodata_opt("nodata", "nodata", "No data value(s) in input or
reference dataset are ignored");
- Optionpk<short> band_opt("b", "band", "Input band", 0);
+ Optionpk<short> band_opt("b", "band", "Input raster band", 0);
Optionpk<bool> confusion_opt("cm", "confusion", "create confusion matrix (to
std out)", false);
- Optionpk<string> labelref_opt("lr", "lref", "attribute name of the reference
label in case reference is OGR vector file", "label");
- Optionpk<string> labelclass_opt("lc", "lclass", "attribute name of the
classified label in case output is OGR vector file", "class");
- Optionpk<short> boundary_opt("bnd", "boundary", "boundary for selecting the
sample", 1,1);
- Optionpk<bool> disc_opt("circ", "circular", "use circular disc kernel
boundary)", false,1);
- Optionpk<bool> homogeneous_opt("hom", "homogeneous", "only take homogeneous
regions into account", false,1);
- Optionpk<string> option_opt("co", "co", "Creation option for output file.
Multiple options can be specified.");
+ Optionpk<string> labelref_opt("lr", "lref", "attribute name of the reference
label (for vector reference datasets only)", "label");
+ Optionpk<string> labelclass_opt("lc", "lclass", "attribute name of the
classified label (for vector reference datasets only)", "class");
+ Optionpk<short> boundary_opt("bnd", "boundary", "boundary for selecting the
sample (for vector reference datasets only)", 1,1);
+ Optionpk<bool> homogeneous_opt("hom", "homogeneous", "only take regions with
homogeneous boundary into account (for reference datasets only)", false,1);
+ Optionpk<bool> disc_opt("circ", "circular", "use circular boundary (for
vector reference datasets only)", false,1);
Optionpk<string> classname_opt("c", "class", "list of class names.");
Optionpk<short> classvalue_opt("r", "reclass", "list of class values (use
same order as in classname opt.");
+ Optionpk<string> colorTable_opt("ct", "ct", "color table in ascii format
having 5 columns: id R G B ALFA (0: transparent, 255: solid).");
+ Optionpk<string> option_opt("co", "co", "Creation option for output file.
Multiple options can be specified.");
Optionpk<short> verbose_opt("v", "verbose", "verbose", 0);
bool doProcess;//stop process when program was invoked with help option (-h
--help)
@@ -58,6 +58,7 @@ int main(int argc, char *argv[])
doProcess=input_opt.retrieveOption(argc,argv);
reference_opt.retrieveOption(argc,argv);
layer_opt.retrieveOption(argc,argv);
+ band_opt.retrieveOption(argc,argv);
confusion_opt.retrieveOption(argc,argv);
labelref_opt.retrieveOption(argc,argv);
classname_opt.retrieveOption(argc,argv);
@@ -68,16 +69,15 @@ int main(int argc, char *argv[])
output_opt.retrieveOption(argc,argv);
ogrformat_opt.retrieveOption(argc,argv);
labelclass_opt.retrieveOption(argc,argv);
- colorTable_opt.retrieveOption(argc,argv);
valueE_opt.retrieveOption(argc,argv);
valueO_opt.retrieveOption(argc,argv);
valueC_opt.retrieveOption(argc,argv);
- band_opt.retrieveOption(argc,argv);
- option_opt.retrieveOption(argc,argv);
- // class_opt.retrieveOption(argc,argv);
boundary_opt.retrieveOption(argc,argv);
- disc_opt.retrieveOption(argc,argv);
homogeneous_opt.retrieveOption(argc,argv);
+ disc_opt.retrieveOption(argc,argv);
+ colorTable_opt.retrieveOption(argc,argv);
+ option_opt.retrieveOption(argc,argv);
+ // class_opt.retrieveOption(argc,argv);
verbose_opt.retrieveOption(argc,argv);
}
catch(string predefinedString){
diff --git a/src/apps/pkfsann.cc b/src/apps/pkfsann.cc
index 94fbd3c..51d8b98 100644
--- a/src/apps/pkfsann.cc
+++ b/src/apps/pkfsann.cc
@@ -41,6 +41,7 @@ using namespace std;
//global parameters used in cost function getCost
+ConfusionMatrix cm;
map<string,short> classValueMap;
vector<std::string> nameVector;
vector<unsigned int> nctraining;
@@ -120,14 +121,14 @@ double getCost(const vector<Vector2d<float> >
&trainingFeatures)
vector<unsigned short> referenceVector;
vector<unsigned short> outputVector;
float rmse=0;
- ConfusionMatrix cm;
- //set names in confusion matrix using nameVector
- for(int iname=0;iname<nameVector.size();++iname){
- if(classValueMap.empty())
- cm.pushBackClassName(nameVector[iname]);
- else
if(cm.getClassIndex(type2string<short>(classValueMap[nameVector[iname]]))<0)
-
cm.pushBackClassName(type2string<short>(classValueMap[nameVector[iname]]));
- }
+ // ConfusionMatrix cm;
+ // //set names in confusion matrix using nameVector
+ // for(int iname=0;iname<nameVector.size();++iname){
+ // if(classValueMap.empty())
+ // cm.pushBackClassName(nameVector[iname]);
+ // else
if(cm.getClassIndex(type2string<short>(classValueMap[nameVector[iname]]))<0)
+ //
cm.pushBackClassName(type2string<short>(classValueMap[nameVector[iname]]));
+ // }
vector<Vector2d<float> > tmpFeatures(nclass);
for(int iclass=0;iclass<nclass;++iclass){
tmpFeatures[iclass].resize(trainingFeatures[iclass].size(),nFeatures);
@@ -272,11 +273,6 @@ int main(int argc, char *argv[])
int nband=0;
int startBand=2;//first two bands represent X and Y pos
- vector<double> offset;
- vector<double> scale;
- vector< Vector2d<float> > trainingPixels;//[class][sample][band]
- vector< Vector2d<float> > testPixels;//[class][sample][band]
-
// if(priors_opt.size()>1){//priors from argument list
// priors.resize(priors_opt.size());
// double normPrior=0;
@@ -300,11 +296,16 @@ int main(int argc, char *argv[])
classValueMap[classname_opt[iclass]]=classvalue_opt[iclass];
}
//----------------------------------- Training
-------------------------------
+ vector<double> offset;
+ vector<double> scale;
+ vector< Vector2d<float> > trainingPixels;//[class][sample][band]
+ vector< Vector2d<float> > testPixels;//[class][sample][band]
+ map<string,Vector2d<float> > trainingMap;
+ map<string,Vector2d<float> > testMap;
vector<string> fields;
+
//organize training data
trainingPixels.clear();
- map<string,Vector2d<float> > trainingMap;
- map<string,Vector2d<float> > testMap;
if(verbose_opt[0]>=1)
std::cout << "reading imageVector file " << training_opt[0] << std::endl;
try{
@@ -339,6 +340,12 @@ int main(int argc, char *argv[])
cerr << error << std::endl;
exit(1);
}
+ catch(std::exception& e){
+ std::cerr << "Error: ";
+ std::cerr << e.what() << std::endl;
+ std::cerr << CPLGetLastErrorMsg() << std::endl;
+ exit(1);
+ }
catch(...){
cerr << "error catched" << std::endl;
exit(1);
@@ -354,23 +361,22 @@ int main(int argc, char *argv[])
std::cout << "training pixels: " << std::endl;
map<string,Vector2d<float> >::iterator mapit=trainingMap.begin();
while(mapit!=trainingMap.end()){
- if(classValueMap.size()){
- //check if name in training is covered by classname_opt (values can not
be 0)
- if(classValueMap[mapit->first]>0){
- if(verbose_opt[0])
- std::cout << mapit->first << " -> " << classValueMap[mapit->first] <<
std::endl;
- }
- else{
- std::cerr << "Error: names in classname option are not complete, please
check names in training vector and make sure classvalue is > 0" << std::endl;
- exit(1);
- }
- }
+ // if(classValueMap.size()){
+ // //check if name in training is covered by classname_opt (values can
not be 0)
+ // if(classValueMap[mapit->first]>0){
+ // if(verbose_opt[0])
+ // std::cout << mapit->first << " -> " <<
classValueMap[mapit->first] << std::endl;
+ // }
+ // else{
+ // std::cerr << "Error: names in classname option are not
complete, please check names in training vector and make sure classvalue is >
0" << std::endl;
+ // exit(1);
+ // }
+ // }
//delete small classes
if((mapit->second).size()<minSize_opt[0]){
trainingMap.erase(mapit);
continue;
}
- nameVector.push_back(mapit->first);
trainingPixels.push_back(mapit->second);
if(verbose_opt[0]>1)
std::cout << mapit->first << ": " << (mapit->second).size() << "
samples" << std::endl;
@@ -383,16 +389,6 @@ int main(int argc, char *argv[])
mapit=testMap.begin();
while(mapit!=testMap.end()){
- if(classValueMap.size()){
- //check if name in test is covered by classname_opt (values can not be 0)
- if(classValueMap[mapit->first]>0){
- ;//ok, no need to print to std::cout
- }
- else{
- std::cerr << "Error: names in classname option are not complete, please
check names in test vector and make sure classvalue is > 0" << std::endl;
- exit(1);
- }
- }
//no need to delete small classes for test sample
testPixels.push_back(mapit->second);
if(verbose_opt[0]>1)
@@ -480,6 +476,42 @@ int main(int argc, char *argv[])
// std::cout << std::endl;
}
+ //new
+ mapit=trainingMap.begin();
+ bool doSort=true;
+ try{
+ while(mapit!=trainingMap.end()){
+ nameVector.push_back(mapit->first);
+ if(classValueMap.size()){
+ //check if name in training is covered by classname_opt (values can not
be 0)
+ if(classValueMap[mapit->first]>0){
+
if(cm.getClassIndex(type2string<short>(classValueMap[mapit->first]))<0){
+
cm.pushBackClassName(type2string<short>(classValueMap[mapit->first]),doSort);
+ }
+ }
+ else{
+ std::cerr << "Error: names in classname option are not complete,
please check names in training vector and make sure classvalue is > 0" <<
std::endl;
+ exit(1);
+ }
+ }
+ else
+ cm.pushBackClassName(mapit->first,doSort);
+ ++mapit;
+ }
+ }
+ catch(BadConversion conversionString){
+ std::cerr << "Error: did you provide class pairs names (-c) and integer
values (-r) for each class in training vector?" << std::endl;
+ exit(1);
+ }
+ if(classname_opt.empty()){
+ //std::cerr << "Warning: no class name and value pair provided for all "
<< nclass << " classes, using string2type<int> instead!" << std::endl;
+ for(int iclass=0;iclass<nclass;++iclass){
+ if(verbose_opt[0])
+ std::cout << iclass << " " << cm.getClass(iclass) << " -> " <<
string2type<short>(cm.getClass(iclass)) << std::endl;
+
classValueMap[cm.getClass(iclass)]=string2type<short>(cm.getClass(iclass));
+ }
+ }
+
//Calculate features of trainig set
nctraining.resize(nclass);
nctest.resize(nclass);
diff --git a/src/apps/pkfssvm.cc b/src/apps/pkfssvm.cc
index a314c76..5dfccf6 100644
--- a/src/apps/pkfssvm.cc
+++ b/src/apps/pkfssvm.cc
@@ -44,6 +44,7 @@ using namespace std;
#define Malloc(type,n) (type *)malloc((n)*sizeof(type))
//global parameters used in cost function getCost
+ConfusionMatrix cm;
map<string,short> classValueMap;
vector<std::string> nameVector;
vector<unsigned int> nctraining;
@@ -51,9 +52,9 @@ vector<unsigned int> nctest;
Optionpk<std::string> svm_type_opt("svmt", "svmtype", "type of SVM (C_SVC,
nu_SVC,one_class, epsilon_SVR, nu_SVR)","C_SVC");
Optionpk<std::string> kernel_type_opt("kt", "kerneltype", "type of kernel
function (linear,polynomial,radial,sigmoid) ","radial");
Optionpk<unsigned short> kernel_degree_opt("kd", "kd", "degree in kernel
function",3);
-Optionpk<float> gamma_opt("g", "gamma", "gamma in kernel function",0);
+Optionpk<float> gamma_opt("g", "gamma", "gamma in kernel function",1.0);
Optionpk<float> coef0_opt("c0", "coef0", "coef0 in kernel function",0);
-Optionpk<float> ccost_opt("cc", "ccost", "the parameter C of C-SVC,
epsilon-SVR, and nu-SVR",1);
+Optionpk<float> ccost_opt("cc", "ccost", "the parameter C of C-SVC,
epsilon-SVR, and nu-SVR",1000);
Optionpk<float> nu_opt("nu", "nu", "the parameter nu of nu-SVC, one-class SVM,
and nu-SVR",0.5);
Optionpk<float> epsilon_loss_opt("eloss", "eloss", "the epsilon in loss
function of epsilon-SVR",0.1);
Optionpk<int> cache_opt("cache", "cache", "cache memory size in MB",100);
@@ -148,14 +149,6 @@ double getCost(const vector<Vector2d<float> >
&trainingFeatures)
if(verbose_opt[0]>2)
std::cout << "SVM is now trained" << std::endl;
- ConfusionMatrix cm;
- //set names in confusion matrix using nameVector
- for(int iname=0;iname<nameVector.size();++iname){
- if(classValueMap.empty())
- cm.pushBackClassName(nameVector[iname]);
- else
if(cm.getClassIndex(type2string<short>(classValueMap[nameVector[iname]]))<0)
-
cm.pushBackClassName(type2string<short>(classValueMap[nameVector[iname]]));
- }
if(cv_opt[0]>1){
double *target = Malloc(double,prob.l);
svm_cross_validation(&prob,¶m,cv_opt[0],target);
@@ -172,6 +165,7 @@ double getCost(const vector<Vector2d<float> >
&trainingFeatures)
}
else{
struct svm_node *x_test;
+ vector<double> result(nclass);
x_test = Malloc(struct svm_node,(nFeatures+1));
for(int iclass=0;iclass<nclass;++iclass){
for(int isample=0;isample<nctest[iclass];++isample){
@@ -182,7 +176,8 @@ double getCost(const vector<Vector2d<float> >
&trainingFeatures)
x_test[nFeatures].index=-1;
double predict_label=0;
//todo: make distinction between svm_predict and
svm_predict_probability?
- predict_label = svm_predict(svm,x_test);
+ assert(svm_check_probability_model(svm));
+ predict_label = svm_predict_probability(svm,x_test,&(result[0]));
string refClassName=nameVector[iclass];
string className=nameVector[static_cast<short>(predict_label)];
if(classValueMap.size())
@@ -212,6 +207,7 @@ double getCost(const vector<Vector2d<float> >
&trainingFeatures)
free(prob.x);
free(x_space);
svm_free_and_destroy_model(&(svm));
+
return(cm.kappa());
}
@@ -307,11 +303,6 @@ int main(int argc, char *argv[])
int nband=0;
int startBand=2;//first two bands represent X and Y pos
- vector<double> offset;
- vector<double> scale;
- vector< Vector2d<float> > trainingPixels;//[class][sample][band]
- vector< Vector2d<float> > testPixels;//[class][sample][band]
-
// if(priors_opt.size()>1){//priors from argument list
// priors.resize(priors_opt.size());
// double normPrior=0;
@@ -329,6 +320,8 @@ int main(int argc, char *argv[])
std::sort(band_opt.begin(),band_opt.end());
// map<string,short> classValueMap;//global variable for now (due to getCost)
+ // vector<std::string> nameVector;//global variable for now (due to getCost)
+
if(classname_opt.size()){
assert(classname_opt.size()==classvalue_opt.size());
for(int iclass=0;iclass<classname_opt.size();++iclass)
@@ -336,13 +329,18 @@ int main(int argc, char *argv[])
}
//----------------------------------- Training
-------------------------------
- struct svm_problem prob;
- vector<string> fields;
- //organize training data
- trainingPixels.clear();
- testPixels.clear();
+ vector<double> offset;
+ vector<double> scale;
+ vector< Vector2d<float> > trainingPixels;//[class][sample][band]
+ vector< Vector2d<float> > testPixels;//[class][sample][band]
map<string,Vector2d<float> > trainingMap;
map<string,Vector2d<float> > testMap;
+ vector<string> fields;
+
+ struct svm_problem prob;
+ //organize training data
+ // trainingPixels.clear();
+ // testPixels.clear();
if(verbose_opt[0]>=1)
std::cout << "reading training file " << training_opt[0] << std::endl;
try{
@@ -377,6 +375,12 @@ int main(int argc, char *argv[])
cerr << error << std::endl;
exit(1);
}
+ catch(std::exception& e){
+ std::cerr << "Error: ";
+ std::cerr << e.what() << std::endl;
+ std::cerr << CPLGetLastErrorMsg() << std::endl;
+ exit(1);
+ }
catch(...){
cerr << "error catched" << std::endl;
exit(1);
@@ -391,23 +395,11 @@ int main(int argc, char *argv[])
std::cout << "training pixels: " << std::endl;
map<string,Vector2d<float> >::iterator mapit=trainingMap.begin();
while(mapit!=trainingMap.end()){
- if(classValueMap.size()){
- //check if name in training is covered by classname_opt (values can not
be 0)
- if(classValueMap[mapit->first]>0){
- if(verbose_opt[0])
- std::cout << mapit->first << " -> " << classValueMap[mapit->first] <<
std::endl;
- }
- else{
- std::cerr << "Error: names in classname option are not complete, please
check names in training vector and make sure classvalue is > 0" << std::endl;
- exit(1);
- }
- }
//delete small classes
if((mapit->second).size()<minSize_opt[0]){
trainingMap.erase(mapit);
continue;
}
- nameVector.push_back(mapit->first);
trainingPixels.push_back(mapit->second);
if(verbose_opt[0]>1)
std::cout << mapit->first << ": " << (mapit->second).size() << "
samples" << std::endl;
@@ -420,16 +412,6 @@ int main(int argc, char *argv[])
mapit=testMap.begin();
while(mapit!=testMap.end()){
- if(classValueMap.size()){
- //check if name in test is covered by classname_opt (values can not be 0)
- if(classValueMap[mapit->first]>0){
- ;//ok, no need to print to std::cout
- }
- else{
- std::cerr << "Error: names in classname option are not complete, please
check names in test vector and make sure classvalue is > 0" << std::endl;
- exit(1);
- }
- }
//no need to delete small classes for test sample
testPixels.push_back(mapit->second);
if(verbose_opt[0]>1)
@@ -518,6 +500,44 @@ int main(int argc, char *argv[])
// std::cout << std::endl;
}
+
+ //new
+ // map<string,Vector2d<float> >::iterator mapit=trainingMap.begin();
+ mapit=trainingMap.begin();
+ bool doSort=true;
+ try{
+ while(mapit!=trainingMap.end()){
+ nameVector.push_back(mapit->first);
+ if(classValueMap.size()){
+ //check if name in training is covered by classname_opt (values can not
be 0)
+ if(classValueMap[mapit->first]>0){
+
if(cm.getClassIndex(type2string<short>(classValueMap[mapit->first]))<0){
+
cm.pushBackClassName(type2string<short>(classValueMap[mapit->first]),doSort);
+ }
+ }
+ else{
+ std::cerr << "Error: names in classname option are not complete,
please check names in training vector and make sure classvalue is > 0" <<
std::endl;
+ exit(1);
+ }
+ }
+ else
+ cm.pushBackClassName(mapit->first,doSort);
+ ++mapit;
+ }
+ }
+ catch(BadConversion conversionString){
+ std::cerr << "Error: did you provide class pairs names (-c) and integer
values (-r) for each class in training vector?" << std::endl;
+ exit(1);
+ }
+ if(classname_opt.empty()){
+ //std::cerr << "Warning: no class name and value pair provided for all "
<< nclass << " classes, using string2type<int> instead!" << std::endl;
+ for(int iclass=0;iclass<nclass;++iclass){
+ if(verbose_opt[0])
+ std::cout << iclass << " " << cm.getClass(iclass) << " -> " <<
string2type<short>(cm.getClass(iclass)) << std::endl;
+
classValueMap[cm.getClass(iclass)]=string2type<short>(cm.getClass(iclass));
+ }
+ }
+
//Calculate features of training (and test) set
nctraining.resize(nclass);
nctest.resize(nclass);
diff --git a/src/apps/pklas2img.cc b/src/apps/pklas2img.cc
index dad94b0..1f50a53 100644
--- a/src/apps/pklas2img.cc
+++ b/src/apps/pklas2img.cc
@@ -146,10 +146,11 @@ int main(int argc,char **argv) {
lasReader.open(input_opt[iinput]);
}
catch(string errorString){
- cout << errorString << endl;
+ cerr << errorString << endl;
exit(1);
}
catch(...){
+ cerr << "Error opening input " << input_opt[iinput] << endl;
exit(2);
}
nPoints=lasReader.getPointCount();
diff --git a/src/apps/pksvm.cc b/src/apps/pksvm.cc
index fa0918a..23a2e0c 100644
--- a/src/apps/pksvm.cc
+++ b/src/apps/pksvm.cc
@@ -86,7 +86,7 @@ int main(int argc, char *argv[])
Optionpk<string> output_opt("o", "output", "output classification image");
Optionpk<string> oformat_opt("of", "oformat", "Output image format (see
also gdal_translate). Empty string: inherit from input image");
Optionpk<string> option_opt("co", "co", "Creation option for output file.
Multiple options can be specified.");
- Optionpk<string> colorTable_opt("ct", "ct", "colour table in ascii format
having 5 columns: id R G B ALFA (0: transparent, 255: solid)");
+ Optionpk<string> colorTable_opt("ct", "ct", "color table in ascii format
having 5 columns: id R G B ALFA (0: transparent, 255: solid)");
Optionpk<string> prob_opt("prob", "prob", "probability image.");
Optionpk<string> entropy_opt("entropy", "entropy", "entropy image (measure
for uncertainty of classifier output","",2);
Optionpk<string> active_opt("active", "active", "ogr output for active
training sample.","",2);
--
Alioth's /usr/local/bin/git-commit-notice on
/srv/git.debian.org/git/pkg-grass/pktools.git
_______________________________________________
Pkg-grass-devel mailing list
Pkg-grass-devel@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-grass-devel
