void StatisticalFlagger::SumToTop(Mask2DCPtr mask, int **sums, size_t width, size_t step, bool reverse)
{
if(reverse)
{
for(size_t y=width;y<mask->Height();++y)
{
int *column = sums[y];
for(size_t x=0;x<mask->Width();++x)
{
if(mask->Value(x, y - width/2))
column[x] += step;
}
}
} else {
for(size_t y=0;y<mask->Height() - width;++y)
{
int *column = sums[y];
for(size_t x=0;x<mask->Width();++x)
{
if(mask->Value(x, y + width/2))
column[x] += step;
}
}
}
}
void Morphology::calculateOpenings(Mask2DCPtr mask, int **values)
{
for(size_t y=0;y<mask->Height();++y)
{
size_t length = 0;
for(size_t x=0;x<mask->Width();++x)
{
if(mask->Value(x, y))
{
++length;
} else if(length > 0) {
for(size_t i=x-length;i<x;++i)
{
values[y][i] = length;
}
length = 0;
values[y][x] = 0;
} else {
values[y][x] = 0;
}
}
if(length > 0) {
for(size_t i=mask->Width()-length;i<mask->Width();++i)
{
values[y][i] = length;
}
}
}
for(size_t x=0;x<mask->Width();++x)
{
size_t length = 0;
for(size_t y=0;y<mask->Height();++y)
{
if(mask->Value(x, y))
{
++length;
} else if(length > 0) {
for(size_t i=y-length;i<y;++i)
{
if(values[i][x] < (int) length)
values[i][x] = -(int) length;
}
length = 0;
}
}
if(length > 0) {
for(size_t i=mask->Height()-length;i<mask->Height();++i)
{
if(values[i][x] < (int) length)
values[i][x] = -(int) length;
}
}
}
}
void Morphology::floodFill(Mask2DCPtr mask, SegmentedImagePtr output, const int *const *lengthWidthValues, size_t x, size_t y, size_t value)
{
std::stack<MorphologyPoint2D> points;
MorphologyPoint2D startPoint;
startPoint.x = x;
startPoint.y = y;
points.push(startPoint);
do {
MorphologyPoint2D p = points.top();
points.pop();
output->SetValue(p.x, p.y, value);
int z = lengthWidthValues[p.y][p.x];
if(p.x > 0 && output->Value(p.x-1, p.y) == 0 && mask->Value(p.x-1,p.y))
{
int zl = lengthWidthValues[p.y][p.x-1];
if((zl > 0 && z > 0) || (zl < 0 && z < 0))
{
MorphologyPoint2D newP;
newP.x = p.x-1; newP.y = p.y;
points.push(newP);
}
}
if(p.x < mask->Width()-1 && output->Value(p.x+1, p.y)==0 && mask->Value(p.x+1,p.y))
{
int zr = lengthWidthValues[p.y][p.x+1];
if((zr > 0 && z > 0) || (zr < 0 && z < 0))
{
MorphologyPoint2D newP;
newP.x = p.x+1; newP.y = p.y;
points.push(newP);
}
}
if(p.y > 0 && output->Value(p.x, p.y-1)==0 && mask->Value(p.x,p.y-1))
{
int zt = lengthWidthValues[p.y-1][p.x];
if((zt > 0 && z > 0) || (zt < 0 && z < 0))
{
MorphologyPoint2D newP;
newP.x = p.x; newP.y = p.y-1;
points.push(newP);
}
}
if(p.y < mask->Height()-1 && output->Value(p.x, p.y+1)==0 && mask->Value(p.x,p.y+1))
{
int zb = lengthWidthValues[p.y+1][p.x];
if((zb > 0 && z > 0) || (zb < 0 && z < 0))
{
MorphologyPoint2D newP;
newP.x = p.x; newP.y = p.y+1;
points.push(newP);
}
}
} while(points.size() != 0);
}
void Compress::Write(std::ofstream &stream, Image2DCPtr image, Mask2DCPtr mask)
{
const num_t
max = ThresholdTools::MaxValue(image, mask),
min = ThresholdTools::MinValue(image, mask),
mid = (min + max) / 2.0;
const num_t normalizeFactor = (num_t) ((2<<22) + ((2<<22)-1)) / (max - min);
const uint32_t
width = image->Width(),
height = image->Height();
const char mode = 0;
stream.write(reinterpret_cast<const char*>(&max), sizeof(max));
stream.write(reinterpret_cast<const char*>(&min), sizeof(min));
stream.write(reinterpret_cast<const char*>(&width), sizeof(width));
stream.write(reinterpret_cast<const char*>(&height), sizeof(height));
stream.write(&mode, sizeof(mode));
for(unsigned y=0;y<height;++y)
{
for(unsigned x=0;x<width;++x)
{
if(!mask->Value(x, y))
{
int32_t value = (int32_t) round((image->Value(x, y) - mid) * normalizeFactor);
stream.write(reinterpret_cast<char*>(&value)+1, 3);
}
}
}
}
void Morphology::calculateVerticalCounts(Mask2DCPtr mask, int **values)
{
for(size_t x=0;x<mask->Width();++x)
{
size_t length = 0;
for(size_t y=0;y<mask->Height();++y)
{
if(mask->Value(x, y))
{
++length;
} else if(length > 0) {
for(size_t i=y-length;i<y;++i)
{
values[i][x] = length;
}
length = 0;
values[y][x] = 0;
} else {
values[y][x] = 0;
}
}
for(size_t i=mask->Height()-length;i<mask->Height();++i)
{
values[i][x] = length;
}
}
}
std::pair<TimeFrequencyData,TimeFrequencyMetaDataPtr> RSPReader::ReadSingleBeamlet(unsigned long timestepStart, unsigned long timestepEnd, unsigned beamletCount, unsigned beamletIndex)
{
std::pair<TimeFrequencyData,TimeFrequencyMetaDataPtr> data = ReadAllBeamlets(timestepStart, timestepEnd, beamletCount);
const unsigned width = timestepEnd - timestepStart;
Image2DPtr realX = Image2D::CreateZeroImagePtr(width, 1);
Image2DPtr imaginaryX = Image2D::CreateZeroImagePtr(width, 1);
Image2DPtr realY = Image2D::CreateZeroImagePtr(width, 1);
Image2DPtr imaginaryY = Image2D::CreateZeroImagePtr(width, 1);
Mask2DPtr mask = Mask2D::CreateUnsetMaskPtr(width, 1);
TimeFrequencyData allX = data.first.Make(Polarization::XX);
TimeFrequencyData allY = data.first.Make(Polarization::YY);
Image2DCPtr xr = allX.GetRealPart();
Image2DCPtr xi = allX.GetImaginaryPart();
Image2DCPtr yr = allY.GetRealPart();
Image2DCPtr yi = allY.GetImaginaryPart();
Mask2DCPtr maskWithBeamlets = data.first.GetSingleMask();
for(unsigned x=0;x<width;++x)
{
realX->SetValue(x, 0, xr->Value(x, beamletIndex));
imaginaryX->SetValue(x, 0, xi->Value(x, beamletIndex));
realY->SetValue(x, 0, yr->Value(x, beamletIndex));
imaginaryY->SetValue(x, 0, yi->Value(x, beamletIndex));
mask->SetValue(x, 0, maskWithBeamlets->Value(x, beamletIndex));
}
data.first = TimeFrequencyData(Polarization::XX, realX, imaginaryX, Polarization::YY, realY, imaginaryY);
data.first.SetGlobalMask(mask);
BandInfo band = data.second->Band();
band.channels[0] = data.second->Band().channels[beamletIndex];
band.channels.resize(1);
data.second->SetBand(band);
return data;
}
void Morphology::SegmentByMaxLength(Mask2DCPtr mask, SegmentedImagePtr output)
{
int **lengthWidthValues = new int*[mask->Height()];
for(size_t y=0;y<mask->Height();++y)
lengthWidthValues[y] = new int[mask->Width()];
calculateOpenings(mask, lengthWidthValues);
for(size_t y=0;y<mask->Height();++y)
{
for(size_t x=0;x<mask->Width();++x)
output->SetValue(x,y,0);
}
for(size_t y=0;y<mask->Height();++y)
{
for(size_t x=0;x<mask->Width();++x)
{
if(mask->Value(x, y) && output->Value(x,y) == 0)
{
floodFill(mask, output, lengthWidthValues, x, y, output->NewSegmentValue());
}
}
}
for(size_t y=0;y<mask->Height();++y)
delete[] lengthWidthValues[y];
delete[] lengthWidthValues;
}
void FrequencyFlagCountPlot::Add(class TimeFrequencyData &data, TimeFrequencyMetaDataCPtr meta)
{
_ignoreFirstChannel = _ignoreFirstChannel && data.ImageHeight() != 1;
size_t yStart = _ignoreFirstChannel ? 1 : 0;
for(size_t maskIndex=0;maskIndex<data.MaskCount();++maskIndex)
{
Mask2DCPtr mask = data.GetMask(maskIndex);
for(size_t y=yStart;y<mask->Height();++y)
{
double frequency = meta->Band().channels[y].frequencyHz;
size_t count = 0;
for(size_t x=0;x<mask->Width();++x)
{
if(mask->Value(x, y))
++count;
}
MapItem item = _counts[frequency];
item.count += count;
item.total += mask->Width();
_counts[frequency] = item;
}
}
}
void Morphology::floodFill(Mask2DCPtr mask, SegmentedImagePtr output, Mask2DPtr *matrices, size_t x, size_t y, size_t z, size_t value, int **hCounts, int **vCounts)
{
std::stack<MorphologyPoint3D> points;
MorphologyPoint3D startPoint;
startPoint.x = x;
startPoint.y = y;
startPoint.z = z;
points.push(startPoint);
do {
MorphologyPoint3D p = points.top();
points.pop();
if(mask->Value(p.x, p.y))
{
if(output->Value(p.x, p.y) == 0)
{
output->SetValue(p.x, p.y, value);
} else {
// now we need to decide whether to change this sample to the new segment or not
if(hCounts[p.y][p.x] < vCounts[p.y][p.x] && p.z == 2)
output->SetValue(p.x, p.y, value);
}
}
Mask2DPtr matrix = matrices[p.z];
matrix->SetValue(p.x, p.y, false);
if((p.z == 0 || p.z == 2) && matrices[1]->Value(p.x,p.y))
{
MorphologyPoint3D newP;
newP.x = p.x; newP.y = p.y; newP.z = 1;
points.push(newP);
}
if(p.x > 0 && matrix->Value(p.x-1,p.y))
{
MorphologyPoint3D newP;
newP.x = p.x-1; newP.y = p.y; newP.z = p.z;
points.push(newP);
}
if(p.x < mask->Width()-1 && matrix->Value(p.x+1,p.y))
{
MorphologyPoint3D newP;
newP.x = p.x+1; newP.y = p.y; newP.z = p.z; newP.z = p.z;
points.push(newP);
}
if(p.y > 0 && matrix->Value(p.x,p.y-1))
{
MorphologyPoint3D newP;
newP.x = p.x; newP.y = p.y-1; newP.z = p.z;
points.push(newP);
}
if(p.y < mask->Height()-1 && matrix->Value(p.x,p.y+1))
{
MorphologyPoint3D newP;
newP.x = p.x; newP.y = p.y+1; newP.z = p.z;
points.push(newP);
}
} while(points.size() != 0);
}
void StatisticalFlagger::MaskToInts(Mask2DCPtr mask, int **maskAsInt)
{
for(size_t y=0;y<mask->Height();++y)
{
int *column = maskAsInt[y];
for(size_t x=0;x<mask->Width();++x)
{
column[x] = mask->Value(x, y) ? 1 : 0;
}
}
}
bool StatisticalFlagger::SquareContainsFlag(Mask2DCPtr mask, size_t xLeft, size_t yTop, size_t xRight, size_t yBottom)
{
for(size_t y=yTop;y<=yBottom;++y)
{
for(size_t x=xLeft;x<=xRight;++x)
{
if(mask->Value(x, y))
return true;
}
}
return false;
}
void UVImager::Image(const TimeFrequencyData &data, TimeFrequencyMetaDataCPtr metaData, unsigned frequencyIndex)
{
if(_uvReal == 0)
Empty();
Image2DCPtr
real = data.GetRealPart(),
imaginary = data.GetImaginaryPart();
Mask2DCPtr
flags = data.GetSingleMask();
for(unsigned i=0;i<data.ImageWidth();++i) {
switch(_imageKind) {
case Homogeneous:
if(flags->Value(i, frequencyIndex)==0.0L) {
num_t
vr = real->Value(i, frequencyIndex),
vi = imaginary->Value(i, frequencyIndex);
if(std::isfinite(vr) && std::isfinite(vi))
{
num_t u,v;
GetUVPosition(u, v, i, frequencyIndex, metaData);
SetUVValue(u, v, vr, vi, 1.0);
SetUVValue(-u, -v, vr, -vi, 1.0);
}
}
break;
case Flagging:
if((flags->Value(i, frequencyIndex)!=0.0L && !_invertFlagging) ||
(flags->Value(i, frequencyIndex)==0.0L && _invertFlagging)) {
num_t u,v;
GetUVPosition(u, v, i, frequencyIndex, metaData);
SetUVValue(u, v, 1, 0, 1.0);
SetUVValue(-u, -v, 1, 0, 1.0);
}
break;
}
}
}
void ImagerAction::Perform(ArtifactSet &artifacts, ProgressListener &progress)
{
boost::mutex::scoped_lock lock(_imagerMutex);
UVImager *imager = artifacts.Imager();
if(imager == 0)
throw BadUsageException("No imager available to create image.");
TimeFrequencyData &data = artifacts.ContaminatedData();
TimeFrequencyMetaDataCPtr metaData = artifacts.MetaData();
if(data.PolarisationCount() > 1)
{
TimeFrequencyData *tmp = data.CreateTFData(StokesIPolarisation);
data = *tmp;
delete tmp;
}
bool btPlaneImager = true;
if(btPlaneImager)
{
typedef double ImagerNumeric;
BaselineTimePlaneImager<ImagerNumeric> btImager;
BandInfo band = metaData->Band();
Image2DCPtr
inputReal = data.GetRealPart(),
inputImag = data.GetImaginaryPart();
Mask2DCPtr mask = data.GetSingleMask();
size_t width = inputReal->Width();
for(size_t t=0;t!=width;++t)
{
UVW uvw = metaData->UVW()[t];
size_t channelCount = inputReal->Height();
std::vector<std::complex<ImagerNumeric> > data(channelCount);
for(size_t ch=0;ch!=channelCount;++ch) {
if(mask->Value(t, ch))
data[ch] = std::complex<ImagerNumeric>(0.0, 0.0);
else
data[ch] = std::complex<ImagerNumeric>(inputReal->Value(t, ch), inputImag->Value(t, ch));
}
btImager.Image(uvw.u, uvw.v, uvw.w, band.channels[0].frequencyHz, band.channels[1].frequencyHz-band.channels[0].frequencyHz, channelCount, &(data[0]), imager->FTReal());
}
} else {
progress.OnStartTask(*this, 0, 1, "Imaging baseline");
for(size_t y=0;y<data.ImageHeight();++y)
{
imager->Image(data, metaData, y);
progress.OnProgress(*this, y, data.ImageHeight());
}
progress.OnEndTask(*this);
}
}
void Morphology::calculateOpenings(Mask2DCPtr mask, Mask2DPtr *values, int **hCounts, int **vCounts)
{
//const int zThreshold = 5;
for(size_t y=0;y<mask->Height();++y)
{
for(size_t x=0;x<mask->Width();++x)
{
bool v = mask->Value(x, y);
values[0]->SetValue(x, y, v && (hCounts[y][x] > vCounts[y][x]));
values[1]->SetValue(x, y, v && false);
//values[1]->SetValue(x, y, v && (abs(hCounts[y][x] - vCounts[y][x]) < zThreshold));
values[2]->SetValue(x, y, v && (hCounts[y][x] <= vCounts[y][x]));
}
}
}
void Mask2D::EnlargeHorizontallyAndSet(Mask2DCPtr smallMask, int factor)
{
for(size_t x=0;x<smallMask->Width();++x)
{
size_t binSize = factor;
if(binSize + x*factor > _width)
binSize = _width - x*factor;
for(size_t y=0;y<_height;++y)
{
for(size_t binX=0;binX<binSize;++binX)
{
size_t curX = x*factor + binX;
SetValue(curX, y, smallMask->Value(x, y));
}
}
}
}
void HighPassFilter::setFlaggedValuesToZeroAndMakeWeights(const Image2DCPtr &inputImage, const Image2DPtr &outputImage, const Mask2DCPtr &inputMask, const Image2DPtr &weightsOutput)
{
const size_t width = inputImage->Width();
for(size_t y=0;y<inputImage->Height();++y)
{
for(size_t x=0;x<width;++x)
{
if(inputMask->Value(x, y) || !isfinite(inputImage->Value(x, y)))
{
outputImage->SetValue(x, y, 0.0);
weightsOutput->SetValue(x, y, 0.0);
} else {
outputImage->SetValue(x, y, inputImage->Value(x, y));
weightsOutput->SetValue(x, y, 1.0);
}
}
}
}
void Mask2D::EnlargeVerticallyAndSet(Mask2DCPtr smallMask, int factor)
{
for(size_t y=0;y<smallMask->Height();++y)
{
size_t binSize = factor;
if(binSize + y*factor > _height)
binSize = _height - y*factor;
for(size_t x=0;x<_width;++x)
{
for(size_t binY=0;binY<binSize;++binY)
{
size_t curY = y*factor + binY;
SetValue(x, curY, smallMask->Value(x, y));
}
}
}
}
void Compress::WriteSubtractFrequencies(std::ofstream &stream, Image2DCPtr image, Mask2DCPtr mask)
{
const num_t
max = ThresholdTools::MaxValue(image, mask),
min = ThresholdTools::MinValue(image, mask);
const num_t normalizeFactor = (num_t) ((2<<22) + ((2<<22)-1)) / (max - min);
//const num_t normalizeFactor = 256.0;
const uint32_t
width = image->Width(),
height = image->Height();
const char mode = 1;
stream.write(reinterpret_cast<const char*>(&max), sizeof(max));
stream.write(reinterpret_cast<const char*>(&min), sizeof(min));
stream.write(reinterpret_cast<const char*>(&width), sizeof(width));
stream.write(reinterpret_cast<const char*>(&height), sizeof(height));
stream.write(&mode, sizeof(mode));
std::vector<int32_t> basis(width);
for(size_t x=0;x<width;++x)
{
SampleRowPtr row = SampleRow::CreateFromColumn(image, x);
basis[x] = (int32_t) round(row->Median() * normalizeFactor);
}
stream.write(reinterpret_cast<char*>(&basis[0]), sizeof(basis));
for(unsigned y=0;y<height;++y)
{
for(unsigned x=0;x<width;++x)
{
if(!mask->Value(x, y))
{
int32_t value = (int32_t) (round(image->Value(x, y) * normalizeFactor) - basis[x]);
stream.write(reinterpret_cast<char*>(&value)+1, 3);
}
}
}
}
void FrequencyPowerPlot::Add(class TimeFrequencyData &data, TimeFrequencyMetaDataCPtr meta)
{
Image2DCPtr image = data.GetSingleImage();
Mask2DCPtr mask = data.GetSingleMask();
for(size_t y=0;y<image->Height();++y)
{
double frequency = meta->Band().channels[y].frequencyHz;
size_t count = 0;
long double value = 0.0L;
for(size_t x=0;x<image->Width();++x)
{
if(!mask->Value(x, y))
{
++count;
value += image->Value(x, y);
}
}
MapItem &item = _values[frequency];
item.total += value;
item.count += count;
}
}
Image2DPtr Compress::Read(std::ifstream &stream, Image2DPtr image, Mask2DCPtr mask)
{
num_t max = 0.0, min = 0.0;
size_t width = 0, height = 0;
char mode = 0;
stream.read(reinterpret_cast<char*>(&max), sizeof(max));
stream.read(reinterpret_cast<char*>(&min), sizeof(min));
stream.read(reinterpret_cast<char*>(&width), sizeof(width));
stream.read(reinterpret_cast<char*>(&height), sizeof(height));
stream.read(&mode, sizeof(mode));
num_t normalizeFactor = (max - min) / (num_t) ((2<<22) + ((2<<22)-1));
for(unsigned y=0;y<height;++y)
{
for(unsigned x=0;x<width;++x)
{
if(!mask->Value(x, y))
{
int32_t value;
stream.read(reinterpret_cast<char*>(&value), 3);
value >>= 8;
image->SetValue(x, y, value / normalizeFactor + min);
}
}
}
void TimeFrequencyImager::WriteNewFlagsPart(Mask2DCPtr newXX, Mask2DCPtr newXY, Mask2DCPtr newYX, Mask2DCPtr newYY, int antenna1, int antenna2, int spectralWindow, size_t timeOffset, size_t timeEnd, size_t leftBorder, size_t rightBorder)
{
initializePolarizations();
checkPolarizations();
size_t frequencyCount = _measurementSet->FrequencyCount();
std::map<double,size_t> observationTimes;
setObservationTimes(*_measurementSet, observationTimes);
casa::Table *table = _measurementSet->OpenTable(true);
casa::ScalarColumn<int> antenna1Column(*table, "ANTENNA1");
casa::ScalarColumn<int> antenna2Column(*table, "ANTENNA2");
casa::ScalarColumn<int> windowColumn(*table, "DATA_DESC_ID");
casa::ScalarColumn<double> timeColumn(*table, "TIME");
casa::ArrayColumn<bool> flagColumn(*table, "FLAG");
ScalarColumnIterator<int> antenna1Iter = ScalarColumnIterator<int>::First(antenna1Column);
ScalarColumnIterator<int> antenna2Iter = ScalarColumnIterator<int>::First(antenna2Column);
ScalarColumnIterator<int> windowIter = ScalarColumnIterator<int>::First(windowColumn);
ScalarColumnIterator<double> timeIter = ScalarColumnIterator<double>::First(timeColumn);
ArrayColumnIterator<bool> flagIter = ArrayColumnIterator<bool>::First(flagColumn);
if(frequencyCount != newXX->Height())
{
std::cerr << "The frequency count in the measurement set (" << frequencyCount << ") does not match the image!" << std::endl;
}
if(timeEnd - timeOffset != newXX->Width())
{
std::cerr << "The number of time scans to write in the measurement set (" << (timeEnd - timeOffset) << ") does not match the image (" << newXX->Width() << ") !" << std::endl;
}
size_t rowsWritten = 0;
for(size_t i=0;i<table->nrow();++i) {
if((*antenna1Iter) == (int) antenna1 &&
(*antenna2Iter) == (int) antenna2 &&
(*windowIter) == (int) spectralWindow)
{
double time = *timeIter;
size_t timeIndex = observationTimes.find(time)->second;
if(timeIndex >= timeOffset + leftBorder && timeIndex < timeEnd - rightBorder)
{
casa::Array<bool> flag = *flagIter;
casa::Array<bool>::iterator j = flag.begin();
for(size_t f=0;f<(size_t) frequencyCount;++f) {
if(_stokesIIndex >= 0)
{
if(_readStokesIDirectly) *j = newXX->Value(timeIndex - timeOffset, f);
++j;
}
if(_xxIndex >= 0)
{
if(_readXX) *j = newXX->Value(timeIndex - timeOffset, f);
++j;
}
if(_xyIndex >= 0)
{
if(_readXY) *j = newXY->Value(timeIndex - timeOffset, f);
++j;
}
if(_yxIndex >= 0)
{
if(_readYX) *j = newYX->Value(timeIndex - timeOffset, f);
++j;
}
if(_yyIndex >= 0)
{
if(_readYY) *j = newYY->Value(timeIndex - timeOffset, f);
++j;
}
}
flagIter.Set(flag);
++rowsWritten;
}
}
++antenna1Iter;
++antenna2Iter;
++timeIter;
++windowIter;
++flagIter;
}
AOLogger::Debug << "Rows written: " << rowsWritten << '\n';
delete table;
}
std::pair<TimeFrequencyData,TimeFrequencyMetaDataPtr> RSPReader::ReadChannelBeamlet(unsigned long timestepStart, unsigned long timestepEnd, unsigned beamletCount, unsigned beamletIndex)
{
const unsigned width = timestepEnd - timestepStart;
std::pair<TimeFrequencyData,TimeFrequencyMetaDataPtr> data = ReadSingleBeamlet(timestepStart*(unsigned long) 256, timestepEnd*(unsigned long) 256, beamletCount, beamletIndex);
TimeFrequencyData allX = data.first.Make(Polarization::XX);
TimeFrequencyData allY = data.first.Make(Polarization::YY);
Image2DCPtr xr = allX.GetRealPart();
Image2DCPtr xi = allX.GetImaginaryPart();
Image2DCPtr yr = allY.GetRealPart();
Image2DCPtr yi = allY.GetImaginaryPart();
Mask2DCPtr mask = data.first.GetSingleMask();
Image2DPtr
outXR = Image2D::CreateUnsetImagePtr(width, 256),
outXI = Image2D::CreateUnsetImagePtr(width, 256),
outYR = Image2D::CreateUnsetImagePtr(width, 256),
outYI = Image2D::CreateUnsetImagePtr(width, 256);
Mask2DPtr
outMask = Mask2D::CreateUnsetMaskPtr(width, 256);
std::vector<double> observationTimes;
for(unsigned long timestep = 0;timestep < timestepEnd-timestepStart;++timestep)
{
unsigned long timestepIndex = timestep * 256;
SampleRow
realX = SampleRow::MakeFromRow(xr.get(), timestepIndex, 256, 0),
imaginaryX = SampleRow::MakeFromRow(xi.get(), timestepIndex, 256, 0),
realY = SampleRow::MakeFromRow(yr.get(), timestepIndex, 256, 0),
imaginaryY = SampleRow::MakeFromRow(yi.get(), timestepIndex, 256, 0);
FFTTools::FFT(realX, imaginaryX);
FFTTools::FFT(realY, imaginaryY);
realX.SetVerticalImageValues(outXR.get(), timestep);
imaginaryX.SetVerticalImageValues(outXI.get(), timestep);
realY.SetVerticalImageValues(outYR.get(), timestep);
imaginaryY.SetVerticalImageValues(outYI.get(), timestep);
observationTimes.push_back(data.second->ObservationTimes()[timestepIndex + 256/2]);
size_t validValues = 0;
for(unsigned y=0;y<256;++y)
{
if(!mask->Value(timestepIndex + y, 0))
++validValues;
}
for(unsigned y=0;y<256;++y)
{
outMask->SetValue(timestep, y , validValues == 0);
}
}
data.first = TimeFrequencyData(Polarization::XX, outXR, outXI, Polarization::YY, outYR, outYI);
data.first.SetGlobalMask(outMask);
BandInfo band = data.second->Band();
band.channels.clear();
for(unsigned i=0;i<256;++i)
{
ChannelInfo channel;
channel.frequencyHz = i+1;
channel.frequencyIndex = i;
band.channels.push_back(channel);
}
data.second->SetBand(band);
data.second->SetObservationTimes(observationTimes);
return data;
}