TEST(KMajority, Regression) {
std::vector<std::string> filenames;
filenames.push_back("brief_0.bin");
vlr::Mat descriptors(filenames);
cv::Mat centroids;
std::vector<int> labels;
double mytime = cv::getTickCount();
clustering::kmajority(16, 100, descriptors, centroids, labels);
mytime = ((double) cv::getTickCount() - mytime) / cv::getTickFrequency()
* 1000;
printf("Clustered [%d] points into [%d] clusters in [%lf] ms\n",
descriptors.rows, centroids.rows, mytime);
for (int j = 0; j < centroids.rows; ++j) {
printf(" Cluster %d:\n", j + 1);
FunctionUtils::printDescriptors(centroids.row(j));
}
for (int i = 0; i < descriptors.rows; ++i) {
printf(" Data point [%d] was assigned to cluster [%d]\n", i,
labels[i]);
}
EXPECT_FALSE(centroids.empty());
EXPECT_FALSE(labels.empty());
}
TEST(KMajority, Clustering) {
std::vector<std::string> filenames;
filenames.push_back("brief_0.bin");
vlr::Mat descriptors(filenames);
vlr::KMajorityParams params;
params["num.clusters"] = 10;
params["max.iterations"] = 10;
vlr::KMajority bofModel(descriptors, params);
bofModel.build();
EXPECT_FALSE(bofModel.getCentroids().empty());
EXPECT_TRUE(bofModel.getCentroids().rows == 10);
EXPECT_TRUE(
descriptors.rows >= 0
&& (size_t ) descriptors.rows
== bofModel.getClusterAssignments().size());
// Check all data has been assigned to same cluster
int cumRes = 0;
for (int k = 0; k < int(bofModel.getClusterCounts().size()); ++k) {
cumRes = cumRes + bofModel.getClusterCounts().at(k);
}
EXPECT_TRUE(cumRes == descriptors.rows);
}
void mitk::CoreExtActivator::StartInputDeviceModules()
{
IInputDeviceRegistry::Pointer inputDeviceRegistry(new mitk::InputDeviceRegistry());
berry::Platform::GetServiceRegistry().RegisterService(
mitk::CoreExtConstants::INPUTDEVICE_SERVICE,
inputDeviceRegistry);
// Gets the last setting of the preferences; if a device was selected,
// it will still be activated after a restart
berry::IPreferencesService::Pointer prefService = berry::Platform::GetServiceRegistry()
.GetServiceById<berry::IPreferencesService>(berry::IPreferencesService::ID);
berry::IPreferences::Pointer extPreferencesNode =
prefService->GetSystemPreferences()->Node(CoreExtConstants::INPUTDEVICE_PREFERENCES);
// Initializes the modules
std::vector<IInputDeviceDescriptor::Pointer> descriptors(inputDeviceRegistry->GetInputDevices());
for (std::vector<IInputDeviceDescriptor::Pointer>::const_iterator it = descriptors.begin();
it != descriptors.end(); ++it)
{
if (extPreferencesNode->GetBool((*it)->GetID(), false))
{
IInputDevice::Pointer temp = (*it)->CreateInputDevice();
temp->RegisterInputDevice();
}
}
} // end method StartInputDeviceModules
void AffineAdaptedFeature2D::operator()(InputArray _image, InputArray _mask,
vector<KeyPoint>& _keypoints,
OutputArray _descriptors,
bool useProvidedKeypoints) const
{
Mat image = _image.getMat();
Mat mask = _mask.getMat();
CV_Assert(useProvidedKeypoints == false);
CV_Assert(!affineTransformParams.empty());
vector<vector<KeyPoint> > keypoints(affineTransformParams.size());
vector<Mat> descriptors(affineTransformParams.size());
#pragma omp parallel for
for(size_t paramsIndex = 0; paramsIndex < affineTransformParams.size(); paramsIndex++)
{
const Vec2f& params = affineTransformParams[paramsIndex];
const float tilt = params[0];
const float phi = params[1];
if(tilt == 1.f) // tilt
{
detectAndComputeImpl(image, mask, keypoints[paramsIndex], descriptors[paramsIndex]);
}
else
{
Mat transformedImage, transformedMask;
Mat Ainv = affineSkew(image, mask, tilt, phi, transformedImage, transformedMask);
detectAndComputeImpl(transformedImage, transformedMask, keypoints[paramsIndex], descriptors[paramsIndex]);
// correct keypoints coordinates
CV_Assert(Ainv.type() == CV_32FC1);
const float* Ainv_ptr = Ainv.ptr<const float>();
for(size_t kpIndex = 0; kpIndex < keypoints[paramsIndex].size(); kpIndex++)
{
KeyPoint& kp = keypoints[paramsIndex][kpIndex];
float tx = Ainv_ptr[0] * kp.pt.x + Ainv_ptr[1] * kp.pt.y + Ainv_ptr[2];
float ty = Ainv_ptr[3] * kp.pt.x + Ainv_ptr[4] * kp.pt.y + Ainv_ptr[5];
kp.pt.x = tx;
kp.pt.y = ty;
}
}
}
// copy keypoints and descriptors to the output
_keypoints.clear();
Mat allDescriptors;
for(size_t paramsIndex = 0; paramsIndex < affineTransformParams.size(); paramsIndex++)
{
_keypoints.insert(_keypoints.end(), keypoints[paramsIndex].begin(), keypoints[paramsIndex].end());
allDescriptors.push_back(descriptors[paramsIndex]);
}
_descriptors.create(allDescriptors.size(), allDescriptors.type());
Mat _descriptorsMat = _descriptors.getMat();
allDescriptors.copyTo(_descriptorsMat);
}
int BundlerMatcher::extractSiftFeature(int fileIndex)
{
std::stringstream filepath;
filepath << mInputPath << mFilenames[fileIndex];
std::string tmp = filepath.str();
char* filename = &tmp[0];
bool extracted = true;
unsigned int imgId = 0;
ilGenImages(1, &imgId);
ilBindImage(imgId);
int nbFeatureFound = -1;
if(ilLoadImage(filename))
{
int w = ilGetInteger(IL_IMAGE_WIDTH);
int h = ilGetInteger(IL_IMAGE_HEIGHT);
int format = ilGetInteger(IL_IMAGE_FORMAT);
if (mSift->RunSIFT(w, h, ilGetData(), format, GL_UNSIGNED_BYTE))
{
int num = mSift->GetFeatureNum();
SiftKeyDescriptors descriptors(128*num);
SiftKeyPoints keys(num);
mSift->GetFeatureVector(&keys[0], &descriptors[0]);
//Save Feature in RAM
mFeatureInfos.push_back(FeatureInfo(w, h, keys, descriptors));
nbFeatureFound = num;
}
else
{
extracted = false;
}
}
else
{
extracted = false;
}
ilDeleteImages(1, &imgId);
if (!extracted)
{
std::cout << "Error while reading : " <<filename <<std::endl;
}
return nbFeatureFound;
}
std::vector<cv::Mat> SiftMatcher::compute(Keypoints& keypoints)
{
std::vector<cv::Mat> descriptors(keypoints.size());
cv::SiftDescriptorExtractor extractor;
for (size_t i = 0; i < keypoints.size(); ++i)
{
extractor.compute(m_images[i], keypoints[i], descriptors[i]);
}
return descriptors;
}
int
main(int argc, char** argv)
{
// Object for storing the point cloud.
pcl::PointCloud<pcl::PointXYZ>::Ptr cloud(new pcl::PointCloud<pcl::PointXYZ>);
// Object for storing the normals.
pcl::PointCloud<pcl::Normal>::Ptr normals(new pcl::PointCloud<pcl::Normal>);
// Object for storing the SHOT descriptors for each point.
pcl::PointCloud<pcl::SHOT352>::Ptr descriptors(new pcl::PointCloud<pcl::SHOT352>());
// Read a PCD file from disk.
if (pcl::io::loadPCDFile<pcl::PointXYZ>(argv[1], *cloud) != 0)
{
return -1;
}
// Note: you would usually perform downsampling now. It has been omitted here
// for simplicity, but be aware that computation can take a long time.
// Estimate the normals.
pcl::NormalEstimation<pcl::PointXYZ, pcl::Normal> normalEstimation;
normalEstimation.setInputCloud(cloud);
normalEstimation.setRadiusSearch(10);
pcl::search::KdTree<pcl::PointXYZ>::Ptr kdtree(new pcl::search::KdTree<pcl::PointXYZ>);
normalEstimation.setSearchMethod(kdtree);
normalEstimation.compute(*normals);
// SHOT estimation object.
pcl::SHOTEstimation<pcl::PointXYZ, pcl::Normal, pcl::SHOT352> shot;
shot.setInputCloud(cloud);
shot.setInputNormals(normals);
// The radius that defines which of the keypoint's neighbors are described.
// If too large, there may be clutter, and if too small, not enough points may be found.
shot.setRadiusSearch(2);
shot.compute(*descriptors);
std::string ss_temp;
ss_temp.append(argv[1]);
std::size_t found = ss_temp.find_last_of("/\\");
ss_temp = ss_temp.substr(found+1);
std::string ss = "SHOT_";
ss.append(ss_temp);
std::cout << ss << '\n';
pcl::io::savePCDFileASCII (ss, *descriptors);
}
SamplesDesctiptors ParseDescriptors(const std::string &descriptorsFilename) {
std::ifstream descriptors(descriptorsFilename.c_str());
SamplesDesctiptors result;
while(!descriptors.eof()) {
std::string name;
std::getline(descriptors, name, '\t');
if(name == std::string()) {
break;
}
descriptors.get();
std::string gender;
std::getline(descriptors, gender, '\n');
result.push_back(std::make_pair(TRAIN_IMAGES + name,
gender == "Male" ? 0 : 1));
}
descriptors.close();
return result;
}
void PatternDetector::train_matcher(const Pattern& pattern)
{
// Store the pattern object
m_pattern = pattern;
// API of cv::DescriptorMatcher is somewhat tricky
// First we clear old train data:
m_matcher->clear();
// Then we add vector of descriptors (each descriptors matrix describe one image).
// This allows us to perform search across multiple images:
std::vector<cv::Mat> descriptors(1);
descriptors[0] = pattern.descriptors.clone();
m_matcher->add(descriptors);
// After adding train data perform actual train:
m_matcher->train();
}
pcl::PointCloud<DescriptorType>::Ptr PCLHighLevelCtl::computeFPFHDescriptors(float radius_search) {
pcl::PointCloud<DescriptorType>::Ptr descriptors(new pcl::PointCloud<DescriptorType>());
pcl::FPFHEstimationOMP<PointType, NormalType, DescriptorType> fpfh;
pcl::search::KdTree<PointType>::Ptr tree(new pcl::search::KdTree<PointType>);
if(!this->normals)
this->computeNormals();
if(!this->keypoints)
this->keypointsExtraction();
fpfh.setInputCloud(this->keypoints);
fpfh.setRadiusSearch(radius_search);
fpfh.setSearchSurface(this->cloud);
fpfh.setInputNormals(this->normals);
fpfh.setSearchMethod(tree);
fpfh.compute(*descriptors);
return descriptors;
}
std::vector<rafl::Descriptor_CPtr> ForestUtil::make_descriptors(const ORUtils::MemoryBlock<float>& featuresMB, size_t descriptorCount, size_t featureCount)
{
// Make sure that the features are available and up-to-date on the CPU.
featuresMB.UpdateHostFromDevice();
// Make the rafl feature descriptors.
const float *features = featuresMB.GetData(MEMORYDEVICE_CPU);
std::vector<rafl::Descriptor_CPtr> descriptors(descriptorCount);
#ifdef WITH_OPENMP
#pragma omp parallel for
#endif
for(int i = 0; i < static_cast<int>(descriptorCount); ++i)
{
// Copy the relevant features into a descriptor and add it.
const float *featuresForDescriptor = features + i * featureCount;
descriptors[i].reset(new rafl::Descriptor(featuresForDescriptor, featuresForDescriptor + featureCount));
}
return descriptors;
}
int
main(int argc, char** argv)
{
// Object for storing the point cloud.
pcl::PointCloud<pcl::PointXYZ>::Ptr cloud(new pcl::PointCloud<pcl::PointXYZ>);
// Object for storing the normals.
pcl::PointCloud<pcl::Normal>::Ptr normals(new pcl::PointCloud<pcl::Normal>);
// Object for storing the spin image for each point.
pcl::PointCloud<SpinImage>::Ptr descriptors(new pcl::PointCloud<SpinImage>());
// Read a PCD file from disk.
if (pcl::io::loadPCDFile<pcl::PointXYZ>(argv[1], *cloud) != 0)
{
cout << "Returning -1\n";
return -1;
}
// Note: you would usually perform downsampling now. It has been omitted here
// for simplicity, but be aware that computation can take a long time.
// Estimate the normals.
pcl::NormalEstimation<pcl::PointXYZ, pcl::Normal> normalEstimation;
normalEstimation.setInputCloud(cloud);
normalEstimation.setRadiusSearch(0.03);
pcl::search::KdTree<pcl::PointXYZ>::Ptr kdtree(new pcl::search::KdTree<pcl::PointXYZ>);
normalEstimation.setSearchMethod(kdtree);
normalEstimation.compute(*normals);
// Spin image estimation object.
pcl::SpinImageEstimation<pcl::PointXYZ, pcl::Normal, SpinImage> si;
si.setInputCloud(cloud);
si.setInputNormals(normals);
// Radius of the support cylinder.
si.setRadiusSearch(0.02);
// Set the resolution of the spin image (the number of bins along one dimension).
// Note: you must change the output histogram size to reflect this.
si.setImageWidth(8);
si.compute(*descriptors);
}
void YawAngleEstimator::train()
{
printf("YawAngleEstimator:train\n");
vector<vector<KeyPoint>> kp(AngleNum,vector<KeyPoint>());
vector<Mat> descriptors(AngleNum,Mat());
featureExtract(YawTemplate, kp, descriptors, Feature);
if (useIndex)
{
//build Index with Lsh and Hamming distance
for (int i = 0; i < AngleNum; i++)
{
flann::Index tempIndex;
if (Feature == USE_SIFT)
{
tempIndex.build(descriptors[i], flann::KDTreeIndexParams(4), cvflann::FLANN_DIST_L2);
}
else
{
tempIndex.build(descriptors[i], flann::LshIndexParams(12, 20, 2), cvflann::FLANN_DIST_HAMMING);
}
YawIndex.push_back(tempIndex);
}
}
else
{
//build BFMathers
for (int i = 0; i < AngleNum; i++)
{
//record
fss<<"\nKeypoints number of template "<< i <<" is "<< descriptors[i].rows << endl;
BFMatcher tempMatcher;
vector<Mat> train_des(1, descriptors[i]);
tempMatcher.add(train_des);
tempMatcher.train();
matchers.push_back(tempMatcher);
}
}
}
TEST(KMajority, SaveLoad) {
std::vector<std::string> filenames;
filenames.push_back("brief_0.bin");
vlr::Mat descriptors(filenames);
vlr::KMajorityParams params;
params["num.clusters"] = 10;
params["max.iterations"] = 10;
params["nn.type"] = vlr::indexType::LINEAR;
vlr::KMajority bofModel(descriptors);
bofModel.build();
bofModel.save("test_vocab.yaml.gz");
vlr::KMajority bofModelLoaded;
bofModelLoaded.load("test_vocab.yaml.gz");
EXPECT_TRUE(
std::equal(bofModel.getCentroids().begin<uchar>(),
bofModel.getCentroids().end<uchar>(),
bofModelLoaded.getCentroids().begin<uchar>()));
}
folly::Future<int32_t> service_with_special_namesSvIf::future_descriptors() {
return apache::thrift::detail::si::future([&] { return descriptors(); });
}
int main(int argc, char* argv[])
{
//-- Main program parameters (default values)
double threshold = 0.03;
std::string output_histogram_image = "histogram_image.m";
std::string output_rsd_data = "rsd_data.m";
double rsd_normal_radius = 0.05;
double rsd_curvature_radius = 0.07;
double rsd_plane_threshold = 0.2;
//-- Show usage
if (pcl::console::find_switch(argc, argv, "-h") || pcl::console::find_switch(argc, argv, "--help"))
{
show_usage(argv[0]);
return 0;
}
//-- Check for parameters in arguments
if (pcl::console::find_switch(argc, argv, "-t"))
pcl::console::parse_argument(argc, argv, "-t", threshold);
else if (pcl::console::find_switch(argc, argv, "--threshold"))
pcl::console::parse_argument(argc, argv, "--threshold", threshold);
if (pcl::console::find_switch(argc, argv, "--histogram"))
pcl::console::parse_argument(argc, argv, "--histogram", output_histogram_image);
if (pcl::console::find_switch(argc, argv, "-r"))
pcl::console::parse_argument(argc, argv, "-r", output_rsd_data);
else if (pcl::console::find_switch(argc, argv, "--rsd"))
pcl::console::parse_argument(argc, argv, "--rsd", output_rsd_data);
if (pcl::console::find_switch(argc, argv, "--rsd-params"))
pcl::console::parse_3x_arguments(argc, argv, "--rsd-params", rsd_normal_radius, rsd_curvature_radius,
rsd_plane_threshold);
//-- Get point cloud file from arguments
std::vector<int> filenames;
bool file_is_pcd = false;
filenames = pcl::console::parse_file_extension_argument(argc, argv, ".ply");
if (filenames.size() != 1)
{
filenames = pcl::console::parse_file_extension_argument(argc, argv, ".pcd");
if (filenames.size() != 1)
{
std::cerr << "No input file specified!" << std::endl;
show_usage(argv[0]);
return -1;
}
file_is_pcd = true;
}
//-- Print arguments to user
std::cout << "Selected arguments: " << std::endl
<< "\tRANSAC threshold: " << threshold << std::endl
#ifdef HISTOGRAM
<< "\tHistogram image: " << output_histogram_image << std::endl
#endif
#ifdef CURVATURE
<< "\tRSD:" << std::endl
<< "\t\tFile: " << output_rsd_data << std::endl
<< "\t\tNormal search radius: " << rsd_normal_radius << std::endl
<< "\t\tCurvature search radius: " << rsd_curvature_radius << std::endl
<< "\t\tPlane threshold: " << rsd_plane_threshold << std::endl
#endif
<< "\tInput file: " << argv[filenames[0]] << std::endl;
//-- Load point cloud data
pcl::PointCloud<pcl::PointXYZ>::Ptr source_cloud(new pcl::PointCloud<pcl::PointXYZ>);
if (file_is_pcd)
{
if (pcl::io::loadPCDFile(argv[filenames[0]], *source_cloud) < 0)
{
std::cout << "Error loading point cloud " << argv[filenames[0]] << std::endl << std::endl;
show_usage(argv[0]);
return -1;
}
}
else
{
if (pcl::io::loadPLYFile(argv[filenames[0]], *source_cloud) < 0)
{
std::cout << "Error loading point cloud " << argv[filenames[0]] << std::endl << std::endl;
show_usage(argv[0]);
return -1;
}
}
/********************************************************************************************
* Stuff goes on here
*********************************************************************************************/
//-- Initial pre-processing of the mesh
//------------------------------------------------------------------------------------
std::cout << "[+] Pre-processing the mesh..." << std::endl;
pcl::PointCloud<pcl::PointXYZ>::Ptr garment_points(new pcl::PointCloud<pcl::PointXYZ>);
MeshPreprocessor<pcl::PointXYZ> preprocessor;
preprocessor.setRANSACThresholdDistance(threshold);
preprocessor.setInputCloud(source_cloud);
preprocessor.process(*garment_points);
//-- Find bounding box (not really required)
pcl::MomentOfInertiaEstimation<pcl::PointXYZ> feature_extractor;
pcl::PointXYZ min_point_AABB, max_point_AABB;
pcl::PointXYZ min_point_OBB, max_point_OBB;
pcl::PointXYZ position_OBB;
Eigen::Matrix3f rotational_matrix_OBB;
feature_extractor.setInputCloud(garment_points);
feature_extractor.compute();
feature_extractor.getAABB(min_point_AABB, max_point_AABB);
feature_extractor.getOBB(min_point_OBB, max_point_OBB, position_OBB, rotational_matrix_OBB);
#ifdef CURVATURE
//-- Curvature stuff
//-----------------------------------------------------------------------------------
std::cout << "[+] Calculating curvature descriptors..." << std::endl;
// Object for storing the normals.
pcl::PointCloud<pcl::Normal>::Ptr normals(new pcl::PointCloud<pcl::Normal>);
// Object for storing the RSD descriptors for each point.
pcl::PointCloud<pcl::PrincipalRadiiRSD>::Ptr descriptors(new pcl::PointCloud<pcl::PrincipalRadiiRSD>());
// Note: you would usually perform downsampling now. It has been omitted here
// for simplicity, but be aware that computation can take a long time.
// Estimate the normals.
pcl::NormalEstimationOMP<pcl::PointXYZ, pcl::Normal> normalEstimation;
normalEstimation.setInputCloud(garment_points);
normalEstimation.setRadiusSearch(rsd_normal_radius);
pcl::search::KdTree<pcl::PointXYZ>::Ptr kdtree(new pcl::search::KdTree<pcl::PointXYZ>);
normalEstimation.setSearchMethod(kdtree);
normalEstimation.setViewPoint(0, 0 , 2);
normalEstimation.compute(*normals);
// RSD estimation object.
pcl::RSDEstimation<pcl::PointXYZ, pcl::Normal, pcl::PrincipalRadiiRSD> rsd;
rsd.setInputCloud(garment_points);
rsd.setInputNormals(normals);
rsd.setSearchMethod(kdtree);
// Search radius, to look for neighbors. Note: the value given here has to be
// larger than the radius used to estimate the normals.
rsd.setRadiusSearch(rsd_curvature_radius);
// Plane radius. Any radius larger than this is considered infinite (a plane).
rsd.setPlaneRadius(rsd_plane_threshold);
// Do we want to save the full distance-angle histograms?
rsd.setSaveHistograms(false);
rsd.compute(*descriptors);
//-- Save to mat file
std::ofstream rsd_file(output_rsd_data.c_str());
for (int i = 0; i < garment_points->points.size(); i++)
{
rsd_file << garment_points->points[i].x << " "
<< garment_points->points[i].y << " "
<< garment_points->points[i].z << " "
<< descriptors->points[i].r_min << " "
<< descriptors->points[i].r_max << "\n";
}
rsd_file.close();
#endif
#ifdef HISTOGRAM
//-- Obtain histogram image
//-----------------------------------------------------------------------------------
std::cout << "[+] Calculating histogram image..." << std::endl;
HistogramImageCreator<pcl::PointXYZ> histogram_image_creator;
histogram_image_creator.setInputPointCloud(garment_points);
histogram_image_creator.setResolution(1024);
histogram_image_creator.setUpsampling(true);
histogram_image_creator.compute();
Eigen::MatrixXi image = histogram_image_creator.getDepthImageAsMatrix();
//-- Temporal fix to get image (through file)
std::ofstream file(output_histogram_image.c_str());
file << image;
file.close();
#endif
return 0;
}
int SiftGPU::DetectAndGenerateDesc()
{
printf("\n ----------- DetectAndGenerateDesc inside --------------- \n");
float prelim_contrastThreshold = 0.5 * contrastThreshold / intvls;
struct detection_data* ddata;
int o, i, r, c;
int num=0; // Number of keypoins detected
int numRemoved=0; // The number of key points rejected because they failed a test
int numberExtrema = 0;
int number = 0;
int intvlsSum = intvls + 3;
int OffsetAct = 0;
int OffsetNext = 0;
int OffsetPrev = 0;
Keys keysArray[SIFT_MAX_NUMBER_KEYS];
/*for (int j = 0 ; j < SIFT_MAX_NUMBER_KEYS ; j++)
{
keysArray[j].x = 0.0;
keysArray[j].y = 0.0;
keysArray[j].intvl = 0.0;
keysArray[j].octv = 0.0;
keysArray[j].subintvl = 0.0;
keysArray[j].scx = 0.0;
keysArray[j].scy = 0.0;
keysArray[j].ori = 0.0;
}*/
for( o = 0; o < octvs; o++ )
{
for( i = 0; i < intvlsSum; i++ )
{
OffsetNext += sizeOfImages[o];
if( i > 0 && i <= intvls )
{
num = 0;
detectExtremaGPU->Process(subtractGPU->cmBufPyramid, gaussFilterGPU->cmBufPyramid, imageWidthInPyramid[o], imageHeightInPyramid[o], OffsetPrev, OffsetAct, OffsetNext, &num, prelim_contrastThreshold, i, o, keysArray);
number = num;
struct detection_data* ddata;
cv::Mat descriptors(number,128,CV_32F,0.0);
//descriptors =
for(int ik = 0; ik < number ; ik++)
{
cv::KeyPoint* key = new cv::KeyPoint(keysArray[ik].scx, keysArray[ik].scy, 1);
key->octave = keysArray[ik].octv;
key->angle = keysArray[ik].ori;
keypoints.push_back(*key);
for(int i = 0; i < 128 ; i++ )
{
descriptors.at<float>(ik,i) = keysArray[ik].desc[i];
}
}
}
OffsetPrev = OffsetAct;
OffsetAct += sizeOfImages[o];
}
}
return 0;
}
void xyzrgb2xyzsift::compute() {
CLOG(LTRACE) << "xyzrgb2xyzsift::compute";
pcl::PointCloud<pcl::PointXYZRGB>::Ptr cloud = in_cloud_xyzrgb.read();
cloud->is_dense = true ;
cv::Mat rgbimg(cloud->height, cloud->width, CV_8UC3, cv::Scalar::all(0)) ;
for (register int row = 0; row < cloud->height; row++)
for (register int col = 0; col < cloud->width ; col++) {
pcl::PointXYZRGB& pt = cloud->at(col, row) ;
cv::Vec3b &rgbvec = rgbimg.at<cv::Vec3b>(row, col) ;
rgbvec[2] = pt.r ; //Creating image for all pixels: normal and NaN
rgbvec[1] = pt.g ;
rgbvec[0] = pt.b ;
}
std::vector<cv::KeyPoint> keypoints;
cv::Mat descriptors;
try {
//-- Step 1: Detect the keypoints.
cv::SiftFeatureDetector detector;
detector.detect(rgbimg, keypoints);
//-- Step 2: Calculate descriptors (feature vectors).
cv::SiftDescriptorExtractor extractor;
extractor.compute( rgbimg, keypoints, descriptors);
} catch (...) {
LOG(LERROR) << "sdasdas\n";
}
pcl::PointCloud<PointXYZSIFT>::Ptr cloud_xyzsift(new pcl::PointCloud<PointXYZSIFT>);
//Create output point cloud (this time - unorganized)
cloud_xyzsift->height = 1 ;
cloud_xyzsift->width = keypoints.size() ;
cloud_xyzsift->is_dense = false ;
cloud_xyzsift->points.resize(cloud_xyzsift->height * cloud_xyzsift->width) ;
//Convert SIFT keypoints/descriptors into PointXYZRGBSIFT cloud
pcl::PointCloud<PointXYZSIFT>::iterator pt_iter = cloud_xyzsift->begin();
for (register int i = 0; i < keypoints.size() ; i++) {
//std::cout << " " << keypoints[i].pt.x << " " << keypoints[i].pt.y << std::endl ;
int keyx = std::min(std::max(((unsigned int) round(keypoints[i].pt.x)), 0u), cloud->width) ;
int keyy = std::min(std::max(((unsigned int) round(keypoints[i].pt.y)), 0u), cloud->height) ;
//std::cout << keyx << " " << keyy << std::endl ;
cv::Mat desc = descriptors(cv::Range(i,i + 1), cv::Range::all()) ;
//std::cout << "descriptor " << desc << std::endl ;
//Make XYZSIFT point
PointXYZSIFT& pt_out = *pt_iter++ ;
pcl::PointXYZRGB& pt_in = cloud->at(keyx, keyy) ;
pt_out.x = pt_in.x; pt_out.y = pt_in.y; pt_out.z = pt_in.z ;
for(int j=0; j<descriptors.cols;j++){
pt_out.descriptor[j] = descriptors.row(i).at<float>(j);
}
}
out_cloud_xyzsift.write(cloud_xyzsift);
}
