static void Apply( GenericImage<P>& image, const ColorSaturationInstance& instance, bool useLUT = false )
{
if ( instance.Curve().IsIdentity() )
{
Console().WriteLn( "<end><cbr><* Identity *>" );
return;
}
size_type N = image.NumberOfPixels();
int numberOfThreads = Thread::NumberOfThreads( N, 16 );
size_type pixelsPerThread = N/numberOfThreads;
image.Status().Initialize( "Color saturation transformation, HSVL space", N );
ThreadData data( image, N );
if ( useLUT )
data.lut = MakeLUT( instance );
ReferenceArray<ColorSaturationThread<P> > threads;
for ( int i = 0, j = 1; i < numberOfThreads; ++i, ++j )
threads.Add( new ColorSaturationThread<P>( instance, data, image,
i*pixelsPerThread,
(j < numberOfThreads) ? j*pixelsPerThread : N ) );
AbstractImage::RunThreads( threads, data );
threads.Destroy();
image.Status() = data.status;
}
static void SuperPixelThreaded( Image& target, const GenericImage<P>& source, const DebayerInstance& instance )
{
int target_w = source.Width() >> 1;
int target_h = source.Height() >> 1;
target.AllocateData( target_w, target_h, 3, ColorSpace::RGB );
target.Status().Initialize( "SuperPixel debayering", target_h );
int numberOfThreads = Thread::NumberOfThreads( target_h, 1 );
int rowsPerThread = target_h/numberOfThreads;
AbstractImage::ThreadData data( target, target_h );
ReferenceArray<SuperPixelThread<P> > threads;
for ( int i = 0, j = 1; i < numberOfThreads; ++i, ++j )
threads.Add( new SuperPixelThread<P>( data, target, source, instance,
i*rowsPerThread,
(j < numberOfThreads) ? j*rowsPerThread : target_h ) );
AbstractImage::RunThreads( threads, data );
threads.Destroy();
target.Status() = data.status;
}
static void VNGThreaded( Image& target, const GenericImage<P>& source, const DebayerInstance& instance )
{
int target_w = source.Width();
int target_h = source.Height();
target.AllocateData( target_w, target_h, 3, ColorSpace::RGB );
target.Status().Initialize( "VNG debayering", target_h-4 );
int numberOfThreads = Thread::NumberOfThreads( target_h-4, 1 );
int rowsPerThread = (target_h - 4)/numberOfThreads;
AbstractImage::ThreadData data( target, target_h-4 );
ReferenceArray<VNGThread<P> > threads;
for ( int i = 0, j = 1; i < numberOfThreads; ++i, ++j )
threads.Add( new VNGThread<P>( data, target, source, instance,
i*rowsPerThread + 2,
(j < numberOfThreads) ? j*rowsPerThread + 2 : target_h-2 ) );
AbstractImage::RunThreads( threads, data );
threads.Destroy();
// copy top and bottom two rows from the adjecent ones
for ( int col = 0; col < target_w; col++ )
for ( int i = 0; i < 3; i++ )
{
target.Pixel( col, 0, i ) = target.Pixel( col, 1, i ) = target.Pixel( col, 2, i );
target.Pixel( col, target_h-1, i ) = target.Pixel( col, target_h-2, i ) = target.Pixel( col, target_h-3, i );
}
target.Status() = data.status;
}
void GenericImage::gen_normalMap(const char* filename)
{
GenericImage dst;
dst.width = width;
dst.height = height;
dst.components = 3;
dst.pixels = (unsigned char*)malloc(width * height * 3);
char* nmap = gen_normal();
//cast from char to unsigned char to save as an normalmap image
//and be able to look at the normalmap and make sense
int index=0;
unsigned char* p = (unsigned char*)dst.pixels;
for(int y = 0; y < height; y++)
{
for(int x = 0; x < width; x++)
{
p[3*index + 0] = (unsigned char)(nmap[3*index + 0] + 127);
p[3*index + 1] = (unsigned char)(nmap[3*index + 1] + 127);
p[3*index + 2] = (unsigned char)(nmap[3*index + 2] + 127);
//fprintf(stderr, "normal map[%d](%d, %d, %d)\n", index, nmap[3*index + 0], nmap[3*index + 1], nmap[3*index + 2]);
index++;
}
}
bool ret = dst.save(filename);
if(ret) printf("success! normalmap saved to image %s\n", filename);
else printf("error! normalmap save to image %s failed\n", filename);
}
static void ReadJP2KImage( GenericImage<P>& img, jas_stream_t* jp2Stream, jas_image_t* jp2Image )
{
int width = jas_image_cmptwidth( jp2Image, 0 );
int height = jas_image_cmptheight( jp2Image, 0 );
int numberOfChannels = jas_image_numcmpts( jp2Image );
jas_matrix_t* pixels = nullptr;
try
{
pixels = jas_matrix_create( 1, width );
if ( pixels == nullptr )
throw Error( "Memory allocation error reading JPEG2000 image" );
// Allocate pixel data
img.AllocateData( width, height, numberOfChannels,
(jas_clrspc_fam( jas_image_clrspc( jp2Image ) ) == JAS_CLRSPC_FAM_GRAY) ?
ColorSpace::Gray : ColorSpace::RGB );
for ( int c = 0; c < numberOfChannels; ++c )
{
int n = jas_image_cmptprec( jp2Image, c );
bool s = jas_image_cmptsgnd( jp2Image, c ) != 0;
for ( int y = 0; y < height; ++y )
{
jas_image_readcmpt( jp2Image, c, 0, y, width, 1, pixels );
typename P::sample* f = img.ScanLine( y, c );
if ( n == 8 )
{
if ( s )
for ( int x = 0; x < width; ++x )
*f++ = P::ToSample( int8( jas_matrix_get( pixels, 0, x ) ) );
else
for ( int x = 0; x < width; ++x )
*f++ = P::ToSample( uint8( jas_matrix_get( pixels, 0, x ) ) );
}
else
{
if ( s )
for ( int x = 0; x < width; ++x )
*f++ = P::ToSample( int16( jas_matrix_get( pixels, 0, x ) ) );
else
for ( int x = 0; x < width; ++x )
*f++ = P::ToSample( uint16( jas_matrix_get( pixels, 0, x ) ) );
}
}
}
jas_matrix_destroy( pixels ), pixels = nullptr;
}
catch ( ... )
{
if ( pixels != nullptr )
jas_matrix_destroy( pixels );
throw;
}
}
static void CombineChannels( GenericImage<P>& img, int colorSpace, const String& baseId,
const Rect& r,
const GenericImage<P0>* src0, const GenericImage<P1>* src1, const GenericImage<P2>* src2 )
{
bool allChannels = src0 != 0 && src1 != 0 && src2 != 0;
typename P::sample* R = img.PixelData( 0 );
typename P::sample* G = img.PixelData( 1 );
typename P::sample* B = img.PixelData( 2 );
const RGBColorSystem& rgbws = img.RGBWorkingSpace();
for ( int y = r.y0; y < r.y1; ++y )
{
const typename P0::sample* data0 = (src0 != 0) ? src0->PixelAddress( r.x0, y ) : 0;
const typename P1::sample* data1 = (src1 != 0) ? src1->PixelAddress( r.x0, y ) : 0;
const typename P2::sample* data2 = (src2 != 0) ? src2->PixelAddress( r.x0, y ) : 0;
for ( int x = r.x0; x < r.x1; ++x, ++img.Status() )
{
if ( colorSpace == ColorSpaceId::RGB )
{
if ( data0 != 0 )
P0::FromSample( *R++, *data0++ );
if ( data1 != 0 )
P1::FromSample( *G++, *data1++ );
if ( data2 != 0 )
P2::FromSample( *B++, *data2++ );
}
else
{
RGBColorSystem::sample ch0, ch1, ch2;
RGBColorSystem::sample r, g, b;
if ( !allChannels )
{
P::FromSample( r, *R );
P::FromSample( g, *G );
P::FromSample( b, *B );
FromRGB( colorSpace, rgbws, ch0, ch1, ch2, r, g, b );
}
if ( data0 != 0 )
P0::FromSample( ch0, *data0++ );
if ( data1 != 0 )
P1::FromSample( ch1, *data1++ );
if ( data2 != 0 )
P2::FromSample( ch2, *data2++ );
ToRGB( colorSpace, rgbws, r, g, b, ch0, ch1, ch2 );
*R++ = P::ToSample( r );
*G++ = P::ToSample( g );
*B++ = P::ToSample( b );
}
}
}
}
static void WriteJPEGImage( const GenericImage<P>& image, JPEGWriter* writer )
{
if ( writer == 0 || !writer->IsOpen() )
throw Error( "JPEG format: Attempt to write an image before creating a file" );
StandardStatus status;
image.SetStatusCallback( &status );
image.SelectNominalChannels(); // JPEG doesn't support alpha channels
writer->WriteImage( image );
}
template <class P> static
void Rotate180( GenericImage<P>& image )
{
size_type N = image.NumberOfPixels();
int n = image.NumberOfChannels();
if ( image.Status().IsInitializationEnabled() )
image.Status().Initialize( "Rotate 180 degrees", n*N );
for ( int c = 0; c < n; ++c, image.Status() += N )
for ( int y0 = 0, y1 = image.Height()-1; y0 <= y1; ++y0, --y1 )
{
typename P::sample* f0 = image.ScanLine( y0, c );
typename P::sample* f1 = image.ScanLine( y1, c );
if ( y0 != y1 )
{
int x0 = 0, x1 = image.Width()-1;
while ( x0 < x1 )
{
pcl::Swap( f0[x0], f1[x1] );
pcl::Swap( f0[x1], f1[x0] );
++x0;
--x1;
}
if ( x0 == x1 )
pcl::Swap( f0[x0], f1[x0] );
}
else
for ( typename P::sample* f = f0, * g = f0+image.Width()-1; f < g; )
pcl::Swap( *f++, *g-- );
}
}
GenericImage<rgb3> to_small_gs( GenericImage<rgb3> const &input_image ) {
GenericImage<rgb3> image_output{input_image.width( ),
input_image.height( )};
std::transform( input_image.cbegin( ), input_image.cend( ),
image_output.begin( ), []( rgb3 const &rgb ) {
return static_cast<uint8_t>( rgb.too_float_gs( ) );
} );
return image_output;
}
static void ApplyInverseRealFourierTransform_1( GenericImage<P>& image, const DComplexImage& dft, bool parallel, int maxProcessors )
{
DImage tmp;
tmp.Status() = image.Status();
image.FreeData();
ApplyInverseRealFourierTransform_2( tmp, dft, parallel, maxProcessors );
image.SetStatusCallback( 0 );
image.Assign( tmp );
image.Status() = tmp.Status();
}
static void ReadJPEGImage( GenericImage<P>& image, JPEGReader* reader, int& readCount )
{
if ( reader == 0 || !reader->IsOpen() )
throw Error( "JPEG format: Attempt to read an image before opening a file" );
try
{
/*
* The readCount thing is a trick to allow reading the same JPEG image
* multiple times from the same format instance. Ugly but heck, it works.
*/
if ( readCount )
{
String filePath = reader->Path();
reader = new JPEGReader;
reader->Open( filePath );
}
StandardStatus status;
image.SetStatusCallback( &status );
reader->ReadImage( image );
if ( readCount )
delete reader;
++readCount;
}
catch ( ... )
{
if ( readCount )
delete reader;
throw;
}
}
static void Apply( GenericImage<P>& image, const PhotometricSuperflatInstance& instance )
{
PolynomialSurface* S;
if ( !File::Exists( instance.starDatabasePath ) )
throw Error( "No such file: " + instance.starDatabasePath );
S = new PolynomialSurface( instance.starDatabasePath, image.Width(), image.Height() );
//S->PrintCatalog();
S->PrintCatalogSummary();
S->PlotXYKeyedToRelativeFlux(false);
String eqn = S->ComputeBestFitModel(instance.fitDegree);
S->PlotXYKeyedToRelativeFlux(true);
S->ShowBestFitModelImage();
delete(S);
};
static void Apply( GenericImage<P>& image, const AnnotationInstance& instance )
{
int relPosX = 0, relPosY = 0;
Bitmap annotationBmp = AnnotationRenderer::CreateAnnotationBitmap( instance, relPosX, relPosY, false );
// blend bitmap to the image
image.Blend( annotationBmp, Point( instance.annotationPositionX - relPosX,
instance.annotationPositionY - relPosY ) );
}
template <class P1, class P2> static
void ApplyFilter_2( GenericImage<P1>& image, const GenericImage<P2>& sharp,
float amount, float threshold, float deringing,
float rangeLow, float rangeHigh, pcl_bool disableExtension, int c, pcl_bool highPass )
{
float rangeWidth = 1 + rangeHigh + rangeLow;
bool isRange = rangeWidth + 1 != 1;
StandardStatus callback;
StatusMonitor monitor;
monitor.SetCallback( &callback );
monitor.Initialize( "<end><cbr>Larson-Sekanina filter", image.NumberOfPixels() );
for ( int x = 0; x < image.Width(); ++x )
for ( int y = 0; y < image.Height(); ++y, ++monitor )
{
double f1, f2;
P1::FromSample( f1, image.Pixel( x, y, c ) );
P2::FromSample( f2, sharp.Pixel( x, y ) );
Apply_PixelValues( f1, f2, threshold, deringing, amount, highPass );
if ( disableExtension )
image.Pixel( x, y, c ) = P1::ToSample( f1 );
else
{
if ( isRange )
f1 = (f1 + rangeLow)/rangeWidth;
image.Pixel( x, y, c ) = P1::ToSample( pcl::Range( f1, 0.0, 1.0 ) );
}
}
if ( disableExtension )
Console().WarningLn( "<end><cbr>*** Warning: Dynamic range extension has been disabled - check pixel values!" );
}
template <class P> static
void Apply( GenericImage<P>& image, const HistogramTransformation& H )
{
if ( image.IsEmptySelection() )
return;
image.SetUnique();
Rect r = image.SelectedRectangle();
int h = r.Height();
int numberOfThreads = H.IsParallelProcessingEnabled() ? Min( H.MaxProcessors(), pcl::Thread::NumberOfThreads( h, 1 ) ) : 1;
int rowsPerThread = h/numberOfThreads;
H.UpdateFlags();
size_type N = image.NumberOfSelectedSamples();
if ( image.Status().IsInitializationEnabled() )
image.Status().Initialize( "Histogram transformation", N );
ThreadData<P> data( image, H, N );
PArray<Thread<P> > threads;
for ( int i = 0, j = 1; i < numberOfThreads; ++i, ++j )
threads.Add( new Thread<P>( data, i*rowsPerThread, (j < numberOfThreads) ? j*rowsPerThread : h ) );
AbstractImage::RunThreads( threads, data );
threads.Destroy();
image.Status() = data.status;
}
template <class P> static
void Apply( GenericImage<P>& image, const typename P::sample* lut )
{
if ( image.IsEmptySelection() )
return;
Rect r = image.SelectedRectangle();
if ( image.Status().IsInitializationEnabled() )
image.Status().Initialize( "LUT-based histogram transformation", image.NumberOfSelectedSamples() );
for ( int c = image.FirstSelectedChannel(), w = r.Width(); c <= image.LastSelectedChannel(); ++c )
for ( int y = r.y0; y < r.y1; ++y )
for ( typename P::sample* p = image.ScanLine( y, c ) + r.x0, * pw = p + w; p < pw; ++p )
*p = lut[*p];
image.Status() += image.NumberOfSelectedSamples();
}
void Normalize( GenericImage<P>& image )
{
image.Status().Initialize( "Normalizing sample values", 2*image.NumberOfNominalSamples() );
image.Status().DisableInitialization();
image.SelectNominalChannels();
image.Truncate( -instance.rangeLow, 1 + instance.rangeHigh ); // N*n
image.Normalize(); // N*n
image.Status().EnableInitialization();
}
static void Apply( GenericImage<P>& image, const FFTConvolution& F )
{
Rect r = image.SelectedRectangle();
if ( F.m_h.IsNull() )
if ( !F.m_filter.IsNull() )
F.m_h = Initialize( *F.m_filter, r.Width(), r.Height(), F.IsParallelProcessingEnabled(), F.MaxProcessors() );
else
F.m_h = Initialize( F.m_image, r.Width(), r.Height(), F.IsParallelProcessingEnabled(), F.MaxProcessors() );
Convolve( image, *F.m_h, F.IsParallelProcessingEnabled(), F.MaxProcessors() );
}
static void ApplyInverseRealFourierTransform( GenericImage<P>& image, const ImageVariant& dft, bool parallel, int maxProcessors )
{
if ( !dft || dft->IsEmpty() )
{
image.FreeData();
return;
}
switch ( dft.BitsPerSample() )
{
case 32: ApplyInverseRealFourierTransform_1( image, static_cast<const FComplexImage&>( *dft ), parallel, maxProcessors ); break;
case 64: ApplyInverseRealFourierTransform_1( image, static_cast<const DComplexImage&>( *dft ), parallel, maxProcessors ); break;
}
}
template <class P> static
void HorizontalMirror( GenericImage<P>& image )
{
size_type N = image.NumberOfPixels();
int n = image.NumberOfChannels();
if ( image.Status().IsInitializationEnabled() )
image.Status().Initialize( "Horizontal mirror", n*N );
for ( int c = 0; c < n; ++c, image.Status() += N )
for ( int y = 0; y < image.Height(); ++y )
for ( typename P::sample* f = image.ScanLine( y, c ),
* g = f + image.Width()-1; f < g; )
{
pcl::Swap( *f++, *g-- );
}
}
GenericImage<rgb3>
FilterDAWGS::filter( GenericImage<rgb3> const &input_image ) {
std::vector<uint32_t> const keys = [&]( ) {
std::vector<uint32_t> v{};
v.resize( input_image.size( ) );
daw::algorithm::parallel::transform(
input_image.begin( ), input_image.end( ), v.begin( ),
[]( auto rgb ) { return daw::imaging::FilterDAWGS::too_gs( rgb ); } );
daw::algorithm::parallel::sort( v.begin( ), v.end( ) );
v.erase( std::unique( v.begin( ), v.end( ) ), v.end( ) );
return v;
}( );
// If we must compress as there isn't room for number of grayscale items
if( keys.size( ) <= 256 ) {
std::cerr << "Already a grayscale image or has enough room for all "
"possible values and no compression needed:"
<< keys.size( ) << std::endl;
return impl::to_small_gs( input_image );
}
std::array<uint32_t, 256> bins = [&keys]( ) {
std::array<uint32_t, 256> a{};
auto const inc = static_cast<float>( keys.size( ) ) / 256.0f;
for( size_t n=0; n<255; ++n ) {
a[n] = keys[static_cast<size_t>(static_cast<float>(n)*inc)];
}
a[255] = keys.back( );
return a;
}( );
GenericImage<rgb3> output_image{input_image.width( ),
input_image.height( )};
daw::algorithm::parallel::transform(
input_image.cbegin( ), input_image.cend( ), output_image.begin( ),
[&bins]( auto rgb ) -> uint8_t {
auto const val = FilterDAWGS::too_gs( rgb );
for( uint8_t n=0; n<static_cast<uint8_t>( bins.size( ) ); ++n ) {
if( bins[n] >= val ) {
return n;
}
}
std::abort( );
} );
return output_image;
}
static void Apply( GenericImage<P>& image, const CurvesTransformationInstance& instance, bool useLUT = false )
{
int numberOfCurves = 0;
if ( !instance[CurveIndex::RGBK].IsIdentity() )
numberOfCurves = image.NumberOfNominalChannels();
if ( image.IsColor() )
{
for ( int c = 0; c < image.NumberOfNominalChannels(); ++c )
if ( !instance[c].IsIdentity() )
++numberOfCurves;
if ( !instance[CurveIndex::L].IsIdentity() || !instance[CurveIndex::a].IsIdentity() || !instance[CurveIndex::b].IsIdentity() || !instance[CurveIndex::c].IsIdentity() )
++numberOfCurves;
if ( !instance[CurveIndex::H].IsIdentity() || !instance[CurveIndex::S].IsIdentity() )
++numberOfCurves;
}
if ( image.HasAlphaChannels() && !instance[CurveIndex::A].IsIdentity() )
++numberOfCurves;
if ( numberOfCurves == 0 )
{
Console().WriteLn( "<end><cbr><* Identity *>" );
return;
}
size_type N = image.NumberOfPixels();
int numberOfThreads = Thread::NumberOfThreads( N, 256 );
size_type pixelsPerThread = N/numberOfThreads;
image.Status().Initialize( "Curves transformation", numberOfCurves*N );
ThreadData data( image, numberOfCurves*N );
if ( useLUT )
data.lut.Generate( image, instance );
ReferenceArray<CurvesThread<P> > threads;
for ( int i = 0, j = 1; i < numberOfThreads; ++i, ++j )
threads.Add( new CurvesThread<P>( instance, data, image,
i*pixelsPerThread,
(j < numberOfThreads) ? j*pixelsPerThread : N ) );
AbstractImage::RunThreads( threads, data );
threads.Destroy();
image.Status() = data.status;
}
template <class P> static
void VerticalMirror( GenericImage<P>& image )
{
size_type N = image.NumberOfPixels();
int n = image.NumberOfChannels();
if ( image.Status().IsInitializationEnabled() )
image.Status().Initialize( "Vertical mirror", n*N );
for ( int c = 0; c < n; ++c, image.Status() += N )
for ( int y0 = 0, y1 = image.Height()-1; y0 < y1; ++y0, --y1 )
for ( typename P::sample* f0 = image.ScanLine( y0, c ),
* f1 = image.ScanLine( y1, c ),
* fw = f0 + image.Width();
f0 < fw; )
{
pcl::Swap( *f0++, *f1++ );
}
}
template <class P, class S> static
void Apply( GenericImage<P>& image, S*,
const ICCProfileTransformation& T,
ICCProfileTransformation::transformation_handle transformation )
{
if ( image.IsEmptySelection() || T.Profiles().IsEmpty() )
return;
if ( image.ColorSpace() != ColorSpace::RGB && image.ColorSpace() != ColorSpace::Gray )
throw Error( String().Format( "Unsupported color space %X in ICC color transformation.", image.ColorSpace() ) );
image.EnsureUnique();
Rect r = image.SelectedRectangle();
int h = r.Height();
int numberOfThreads = T.IsParallelProcessingEnabled() ? Min( T.MaxProcessors(), pcl::Thread::NumberOfThreads( h, 1 ) ) : 1;
int rowsPerThread = h/numberOfThreads;
size_type N = image.NumberOfSelectedPixels();
if ( image.Status().IsInitializationEnabled() )
image.Status().Initialize( "In-place ICC color profile transformation", N );
ThreadData<P> data( image, T, transformation, N );
ReferenceArray<Thread<P,S> > threads;
for ( int i = 0, j = 1; i < numberOfThreads; ++i, ++j )
threads.Add( new Thread<P,S>( data,
i*rowsPerThread,
(j < numberOfThreads) ? j*rowsPerThread : h ) );
AbstractImage::RunThreads( threads, data );
threads.Destroy();
image.Status() = data.status;
}
template <class P> static
void RealMinMax( const GenericImage<P>& image, const Rect& r, int c, double& min, double& max, int y0, int y1 )
{
int w = r.Width();
min = max = image( r.LeftTop(), c );
for ( int y = r.y0+y0, y01 = r.y0+y1; y < y01; ++y )
{
const typename P::sample* f = image.ScanLine( y, c ) + r.x0;
const typename P::sample* fw = f + w;
do
{
if ( *f < min )
min = *f;
else if ( max < *f )
max = *f;
}
while ( ++f < fw );
}
}
static void LoadAndTransformImage( GenericImage<P>& transform, const GenericImage<P1>& image, bool parallel, int maxProcessors )
{
Rect r = image.SelectedRectangle();
if ( !r.IsRect() )
return;
int w = FFTC::OptimizedLength( r.Width() );
int h = FFTC::OptimizedLength( r.Height() );
int dw2 = (w - r.Width()) >> 1;
int dh2 = (h - r.Height()) >> 1;
transform.AllocateData( w, h, image.NumberOfSelectedChannels(),
(image.NumberOfSelectedChannels() < 3 || image.FirstSelectedChannel() != 0) ?
ColorSpace::Gray : image.ColorSpace() );
transform.Zero().Move( image, Point( dw2, dh2 ) );
ApplyInPlaceFourierTransform( transform, FFTDirection::Forward, parallel, maxProcessors );
}
template <class P> inline
static void Apply( GenericImage<P>& image, const IntegerResample& Z )
{
int width = image.Width();
int w0 = width;
int height = image.Height();
int h0 = height;
Z.GetNewSizes( width, height );
if ( width == w0 && height == h0 )
return;
if ( width == 0 || height == 0 )
{
image.FreeData();
return;
}
image.EnsureUnique();
typename P::sample* f = 0;
typename P::sample** f0 = 0;
int n = image.NumberOfChannels();
size_type N = image.NumberOfPixels();
typename GenericImage<P>::color_space cs0 = image.ColorSpace();
StatusMonitor status = image.Status();
int z = pcl::Abs( Z.ZoomFactor() );
int z2 = z*z;
int n2 = z2 >> 1;
try
{
if ( status.IsInitializationEnabled() )
{
String info = (Z.ZoomFactor() > 0) ? "Upsampling" : "Downsampling";
info.AppendFormat( " %d:%d, %dx%d",
(Z.ZoomFactor() > 0) ? z : 1, (Z.ZoomFactor() > 0) ? 1 : z, width, height );
if ( Z.ZoomFactor() < 0 )
{
info += ", ";
switch ( Z.DownsampleMode() )
{
default:
case IntegerDownsampleMode::Average: info += "average"; break;
case IntegerDownsampleMode::Median: info += "median"; break;
case IntegerDownsampleMode::Maximum: info += "maximum"; break;
case IntegerDownsampleMode::Minimum: info += "minimum"; break;
}
}
status.Initialize( info, n*N );
}
GenericVector<typename P::sample> fm;
if ( Z.ZoomFactor() < 0 && Z.DownsampleMode() == IntegerDownsampleMode::Median )
fm = GenericVector<typename P::sample>( z2 );
f0 = image.ReleaseData();
for ( int c = 0; c < n; ++c, status += N )
{
f = image.Allocator().AllocatePixels( width, height );
if ( Z.ZoomFactor() > 0 )
{
const typename P::sample* f0c = f0[c];
for ( int y = 0; y < h0; ++y )
{
int yz = y*z;
for ( int x = 0; x < w0; ++x )
{
int xz = x*z;
typename P::sample v = *f0c++;
for ( int i = 0; i < z; ++i )
{
typename P::sample* fi = f + (size_type( yz + i )*width + xz);
for ( int j = 0; j < z; ++j )
*fi++ = v;
}
}
}
}
else
{
typename P::sample* fz = f;
for ( int y = 0; y < height; ++y )
{
const typename P::sample* fy = f0[c] + size_type( y )*z*w0;
for ( int x = 0; x < width; ++x )
{
const typename P::sample* fyx = fy + x*z;
switch ( Z.DownsampleMode() )
{
default:
case IntegerDownsampleMode::Average:
{
double s = 0;
for ( int i = 0; i < z; ++i, fyx += w0 )
for ( int j = 0; j < z; ++j )
s += fyx[j];
*fz++ = typename P::sample( P::IsFloatSample() ? s/z2 : Round( s/z2 ) );
}
break;
case IntegerDownsampleMode::Median:
{
typename P::sample* fmi = *fm;
for ( int i = 0; i < z; ++i, fyx += w0 )
for ( int j = 0; j < z; ++j )
*fmi++ = fyx[j];
*fz++ = (z & 1) ?
*Select( *fm, fm.At( z2 ), n2 ) :
P::FloatToSample( 0.5*(double( *Select( *fm, fm.At( z2 ), n2 ) ) +
double( *Select( *fm, fm.At( z2 ), n2-1 ) )) );
}
break;
case IntegerDownsampleMode::Maximum:
{
*fz = P::MinSampleValue();
for ( int i = 0; i < z; ++i, fyx += w0 )
for ( int j = 0; j < z; ++j )
if ( fyx[j] > *fz )
*fz = fyx[j];
++fz;
}
break;
case IntegerDownsampleMode::Minimum:
{
*fz = P::MaxSampleValue();
for ( int i = 0; i < z; ++i, fyx += w0 )
for ( int j = 0; j < z; ++j )
if ( fyx[j] < *fz )
*fz = fyx[j];
++fz;
}
break;
}
}
}
}
image.Allocator().Deallocate( f0[c] );
f0[c] = f;
f = 0;
}
image.ImportData( f0, width, height, n, cs0 ).Status() = status;
}
catch ( ... )
{
if ( f != 0 )
image.Allocator().Deallocate( f );
if ( f0 != 0 )
{
for ( int c = 0; c < n; ++c )
if ( f0[c] != 0 )
image.Allocator().Deallocate( f0[c] );
image.Allocator().Deallocate( f0 );
}
image.FreeData();
throw;
}
}
template <class P> static
void Apply( const GenericImage<P>& image, MultiscaleMedianTransform& T )
{
InitializeStructures();
bool statusInitialized = false;
StatusMonitor& status = (StatusMonitor&)image.Status();
try
{
if ( status.IsInitializationEnabled() )
{
status.Initialize( String( T.m_medianWaveletTransform ? "Median-wavelet" : "Multiscale median" ) + " transform",
image.NumberOfSelectedSamples()*T.m_numberOfLayers*(T.m_medianWaveletTransform ? 2 : 1) );
status.DisableInitialization();
statusInitialized = true;
}
GenericImage<P> cj0( image );
cj0.Status().Clear();
for ( int j = 1, j0 = 0; ; ++j, ++j0 )
{
GenericImage<P> cj( cj0 );
cj.Status() = status;
MedianFilterLayer( cj, T.FilterSize( j0 ), T.m_parallel, T.m_maxProcessors );
if ( T.m_medianWaveletTransform )
{
GenericImage<P> w0( cj0 );
GenericImage<P> d0( cj0 );
d0 -= cj;
for ( int c = 0; c < d0.NumberOfChannels(); ++c )
{
w0.SelectChannel( c );
d0.SelectChannel( c );
cj.SelectChannel( c );
double t = T.m_medianWaveletThreshold*d0.MAD( d0.Median() )/0.6745;
for ( typename GenericImage<P>::sample_iterator iw( w0 ), id( d0 ), ic( cj ); iw; ++iw, ++id, ++ic )
if ( Abs( *id ) > t )
*iw = *ic;
}
w0.ResetSelections();
cj.ResetSelections();
w0.Status() = cj.Status();
LinearFilterLayer( w0, T.FilterSize( j0 ), T.m_parallel, T.m_maxProcessors );
cj = w0;
}
status = cj.Status();
cj.Status().Clear();
if ( T.m_layerEnabled[j0] )
{
cj0 -= cj;
T.m_transform[j0] = Image( cj0 );
}
if ( j == T.m_numberOfLayers )
{
if ( T.m_layerEnabled[j] )
T.m_transform[j] = Image( cj );
break;
}
cj0 = cj;
}
if ( statusInitialized )
status.EnableInitialization();
}
catch ( ... )
{
T.DestroyLayers();
if ( statusInitialized )
status.EnableInitialization();
throw;
}
}
template <class P> static
void Compute( const GenericImage<P>& image, ImageStatistics::Data& data, bool /*parallel*/, int /*maxProcessors*/ )
{
data.AssignStatisticalData( ImageStatistics::Data() );
size_type N = image.NumberOfSelectedPixels();
if ( N == 0 )
return;
if ( image.Status().IsInitializationEnabled() )
image.Status().Initialize( "Computing image statistics", N );
size_type NS = N/8;
size_type NN = N - 7*NS;
Rect rect = image.SelectedRectangle();
int channel = image.SelectedChannel();
data.minimum = data.maximum = 0;
data.minPos = data.maxPos = Point( 0 );
if ( data.rejectLow || data.rejectHigh )
{
// Rejection bounds in the native range
double s0 = 0, s1 = 0;
if ( data.rejectLow )
s0 = data.low * P::MaxSampleValue();
if ( data.rejectHigh )
s1 = data.high * P::MaxSampleValue();
Array<double> v;
v.Reserve( N ); // clearly, optimize for speed
typename GenericImage<P>::const_roi_sample_iterator i( image, rect, channel );
P::FromSample( data.minimum, *i );
data.maximum = data.minimum;
if ( data.noExtremes )
{
if ( data.rejectLow )
{
if ( data.rejectHigh )
{
for ( ; i; ++i )
if ( *i > s0 )
if ( *i < s1 )
{
double f; P::FromSample( f, *i );
v.Append( f );
}
}
else
{
for ( ; i; ++i )
if ( *i > s0 )
{
double f; P::FromSample( f, *i );
v.Append( f );
}
}
}
else
{
for ( ; i; ++i )
if ( *i < s1 )
{
double f; P::FromSample( f, *i );
v.Append( f );
}
}
}
else // !data.noExtremes
{
bool extremesSeen = false;
if ( data.rejectLow )
{
if ( data.rejectHigh )
{
for ( int y = rect.y0; y < rect.y1; ++y )
for ( int x = rect.x0; x < rect.x1; ++x, ++i )
if ( *i > s0 )
if ( *i < s1 )
{
double f; P::FromSample( f, *i );
v.Append( f );
if ( extremesSeen )
{
if ( f < data.minimum )
{
data.minimum = f;
data.minPos.x = x;
data.minPos.y = y;
}
else if ( f > data.maximum )
{
data.maximum = f;
data.maxPos.x = x;
data.maxPos.y = y;
}
}
else
{
data.minimum = data.maximum = f;
data.minPos.x = data.maxPos.x = x;
data.minPos.y = data.maxPos.y = y;
extremesSeen = true;
}
}
}
else
{
for ( int y = rect.y0; y < rect.y1; ++y )
for ( int x = rect.x0; x < rect.x1; ++x, ++i )
if ( *i > s0 )
{
double f; P::FromSample( f, *i );
v.Append( f );
if ( extremesSeen )
{
if ( f < data.minimum )
{
data.minimum = f;
data.minPos.x = x;
data.minPos.y = y;
}
else if ( f > data.maximum )
{
data.maximum = f;
data.maxPos.x = x;
data.maxPos.y = y;
}
}
else
{
data.minimum = data.maximum = f;
data.minPos.x = data.maxPos.x = x;
data.minPos.y = data.maxPos.y = y;
extremesSeen = true;
}
}
}
}
else
{
for ( int y = rect.y0; y < rect.y1; ++y )
for ( int x = rect.x0; x < rect.x1; ++x, ++i )
if ( *i < s1 )
{
double f; P::FromSample( f, *i );
v.Append( f );
if ( extremesSeen )
{
if ( f < data.minimum )
{
data.minimum = f;
data.minPos.x = x;
data.minPos.y = y;
}
else if ( f > data.maximum )
{
data.maximum = f;
data.maxPos.x = x;
data.maxPos.y = y;
}
}
else
{
data.minimum = data.maximum = f;
data.minPos.x = data.maxPos.x = x;
data.minPos.y = data.maxPos.y = y;
extremesSeen = true;
}
}
}
}
data.count = v.Length();
if ( !data.noSumOfSquares )
data.sumOfSquares = pcl::SumOfSquares( v.Begin(), v.End() );
image.Status() += NS;
if ( !data.noMean )
{
data.mean = pcl::Mean( v.Begin(), v.End() );
image.Status() += NS;
if ( !data.noVariance )
{
data.variance = pcl::Variance( v.Begin(), v.End(), data.mean );
data.stdDev = Sqrt( data.variance );
}
image.Status() += NS;
}
else
{
image.Status() += 2*NS;
}
if ( !data.noMedian )
{
data.median = pcl::Median( v.Begin(), v.End() );
image.Status() += NS;
if ( !data.noAvgDev )
data.avgDev = pcl::AvgDev( v.Begin(), v.End(), data.median );
image.Status() += NS;
if ( !data.noMAD )
{
data.MAD = pcl::MAD( v.Begin(), v.End(), data.median );
if ( !data.noBWMV )
data.bwmv = pcl::BiweightMidvariance( v.Begin(), v.End(), data.median, 1.4826*data.MAD );
}
if ( !data.noPBMV )
data.pbmv = pcl::BendMidvariance( v.Begin(), v.End(), data.median, 0.2 );
image.Status() += NS;
}
else
{
image.Status() += 3*NS;
}
if ( !data.noSn )
data.Sn = pcl::Sn( v.Begin(), v.End() );
image.Status() += NS;
if ( !data.noQn )
data.Qn = pcl::Qn( v.Begin(), v.End() );
image.Status() += NN;
}
else
{
data.count = N;
DMatrix V( image, rect, channel );
if ( !data.noExtremes )
{
double* i = V.Begin();
data.minimum = data.maximum = *i;
for ( int y = rect.y0; y < rect.y1; ++y )
for ( int x = rect.x0; x < rect.x1; ++x, ++i )
if ( *i < data.minimum )
{
data.minimum = *i;
data.minPos.x = x;
data.minPos.y = y;
}
else if ( *i > data.maximum )
{
data.maximum = *i;
data.maxPos.x = x;
data.maxPos.y = y;
}
}
if ( !data.noSumOfSquares )
data.sumOfSquares = pcl::SumOfSquares( V.Begin(), V.End() );
image.Status() += NS;
if ( !data.noMean )
{
data.mean = pcl::Mean( V.Begin(), V.End() );
image.Status() += NS;
if ( !data.noVariance )
{
data.variance = pcl::Variance( V.Begin(), V.End(), data.mean );
data.stdDev = Sqrt( data.variance );
}
image.Status() += NS;
}
else
{
image.Status() += 2*NS;
}
if ( !data.noMedian )
{
data.median = pcl::Median( V.Begin(), V.End() );
image.Status() += NS;
if ( !data.noAvgDev )
data.avgDev = pcl::AvgDev( V.Begin(), V.End(), data.median );
image.Status() += NS;
if ( !data.noMAD )
{
data.MAD = pcl::MAD( V.Begin(), V.End(), data.median );
if ( !data.noBWMV )
data.bwmv = pcl::BiweightMidvariance( V.Begin(), V.End(), data.median, 1.4826*data.MAD );
}
if ( !data.noPBMV )
data.pbmv = pcl::BendMidvariance( V.Begin(), V.End(), data.median, 0.2 );
image.Status() += NS;
}
else
{
image.Status() += 3*NS;
}
if ( !data.noSn )
data.Sn = pcl::Sn( V.Begin(), V.End() );
image.Status() += NS;
if ( !data.noQn )
data.Qn = pcl::Qn( V.Begin(), V.End() );
image.Status() += NN;
}
}
static void Apply( GenericImage<P>& img, const View& view, const FluxCalibrationInstance& instance )
{
FITSKeywordArray inputKeywords;
view.Window().GetKeywords( inputKeywords );
if ( KeywordExists( inputKeywords, "FLXMIN" ) ||
KeywordExists( inputKeywords, "FLXRANGE" ) ||
KeywordExists( inputKeywords, "FLX2DN" ) )
{
throw Error( "Already calibrated image" );
}
if ( img.IsColor() )
throw Error( "Can't calibrate a color image" );
float Wc = instance.p_wavelength.GetValue( inputKeywords );
float Tr = Max( 1.0F, instance.p_transmissivity.GetValue( inputKeywords ) );
float Delta = instance.p_filterWidth.GetValue( inputKeywords );
float Ap = instance.p_aperture.GetValue( inputKeywords ) / 10; // mm -> cm
float Cobs = Max( 0.0F, instance.p_centralObstruction.GetValue( inputKeywords ) ) / 10; // mm -> cm
float ExpT = instance.p_exposureTime.GetValue( inputKeywords );
float AtmE = Max( 0.0F, instance.p_atmosphericExtinction.GetValue( inputKeywords ) );
float G = Max( 1.0F, instance.p_sensorGain.GetValue( inputKeywords ) );
float QEff = Max( 1.0F, instance.p_quantumEfficiency.GetValue( inputKeywords ) );
if ( Wc <= 0 )
throw Error( "Invalid filter wavelength" );
if ( Tr <= 0 || Tr > 1 )
throw Error( "Invalid filter transmissivity" );
if ( Delta <= 0 )
throw Error( "Invalid filter width" );
if ( Ap <= 0 )
throw Error( "Invalid aperture" );
if ( Cobs < 0 || Cobs >= Ap )
throw Error( "Invalid central obstruction area" );
if ( ExpT <= 0 )
throw Error( "Invalid exposure time" );
if ( AtmE < 0 || AtmE >= 1 )
throw Error( "Invalid atmospheric extinction" );
if ( G <= 0 )
throw Error( "Invalid sensor gain" );
if ( QEff <= 0 || QEff > 1 )
throw Error( "Invalid quantum efficiency" );
FITSKeywordArray keywords;
float pedestal = 0;
bool foundPedestal = false;
for ( FITSKeywordArray::const_iterator i = inputKeywords.Begin(); i != inputKeywords.End(); ++i )
if ( i->name == "PEDESTAL" )
{
if ( i->value.TryToFloat( pedestal ) )
foundPedestal = true;
pedestal /= 65535; // 2^16-1 maximum value of a 16bit CCD.
}
else
keywords.Add( *i );
if ( foundPedestal )
Console().NoteLn( "<end><cbr><br>* FluxCalibration: PEDESTAL keyword found: " + view.FullId() );
// double F = Wc * inv_ch * (1 - Tr) * Delta * Ap * Cobs * ExpT * AtmE * G * QEff;
double F = Wc * inv_ch * (1 - AtmE) * Delta * ( Const<double>::pi() / 4 * ( Ap*Ap - Cobs*Cobs ) ) * ExpT * Tr * G * QEff;
size_type N = img.NumberOfPixels();
typename P::sample* f = img.PixelData( 0 );
const typename P::sample* fN = f + N;
double flxMin = DBL_MAX;
double flxMax = 0;
for ( ; f < fN; ++f, ++img.Status() )
{
double I; P::FromSample( I, *f );
I = (I - pedestal)/F;
*f = P::ToSample( I );
if ( I < flxMin )
flxMin = I;
if ( I > flxMax )
flxMax = I;
}
img.Rescale();
keywords.Add( FITSHeaderKeyword( "FLXMIN",
IsoString().Format( "%.8e", flxMin ),
"" ) );
keywords.Add( FITSHeaderKeyword( "FLXRANGE",
IsoString().Format( "%.8e", flxMax - flxMin ),
"FLXRANGE*pixel_value + FLXMIN = erg/cm^2/s/nm" ) );
keywords.Add( FITSHeaderKeyword( "FLX2DN",
IsoString().Format( "%.8e", F*65535 ),
"(FLXRANGE*pixel_value + FLXMIN)*FLX2DN = DN" ) );
view.Window().SetKeywords( keywords );
}