potok/opencv
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2026-07-31 08:13:04 +04:00
Code Packages Releases Wiki Activity

used new device layer for cv::cuda::LUT

Browse Source
This commit is contained in:
Vladislav Vinogradov
2013-10-01 15:24:17 +04:00
parent 06f07944c8
commit 23cc31e041
3 changed files with 210 additions and 216 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/cudaarithm/src/core.cpp
-213
View File
@@ -130,217 +130,4 @@ void cv::cuda::flip(InputArray _src, OutputArray _dst, int flipCode, Stream& str
funcs[src.depth()][src.channels() - 1](src, dst, flipCode, StreamAccessor::getStream(stream));
}
////////////////////////////////////////////////////////////////////////
// LUT
#if (CUDA_VERSION >= 5000)
namespace
{
class LookUpTableImpl : public LookUpTable
{
public:
LookUpTableImpl(InputArray lut);
void transform(InputArray src, OutputArray dst, Stream& stream = Stream::Null());
private:
int lut_cn;
int nValues3[3];
const Npp32s* pValues3[3];
const Npp32s* pLevels3[3];
GpuMat d_pLevels;
GpuMat d_nppLut;
GpuMat d_nppLut3[3];
};
LookUpTableImpl::LookUpTableImpl(InputArray _lut)
{
nValues3[0] = nValues3[1] = nValues3[2] = 256;
Npp32s pLevels[256];
for (int i = 0; i < 256; ++i)
pLevels[i] = i;
d_pLevels.upload(Mat(1, 256, CV_32S, pLevels));
pLevels3[0] = pLevels3[1] = pLevels3[2] = d_pLevels.ptr<Npp32s>();
GpuMat lut;
if (_lut.kind() == _InputArray::GPU_MAT)
{
lut = _lut.getGpuMat();
}
else
{
Mat hLut = _lut.getMat();
CV_Assert( hLut.total() == 256 && hLut.isContinuous() );
lut.upload(Mat(1, 256, hLut.type(), hLut.data));
}
lut_cn = lut.channels();
CV_Assert( lut.depth() == CV_8U );
CV_Assert( lut.rows == 1 && lut.cols == 256 );
lut.convertTo(d_nppLut, CV_32S);
if (lut_cn == 1)
{
pValues3[0] = pValues3[1] = pValues3[2] = d_nppLut.ptr<Npp32s>();
}
else
{
cuda::split(d_nppLut, d_nppLut3);
pValues3[0] = d_nppLut3[0].ptr<Npp32s>();
pValues3[1] = d_nppLut3[1].ptr<Npp32s>();
pValues3[2] = d_nppLut3[2].ptr<Npp32s>();
}
}
void LookUpTableImpl::transform(InputArray _src, OutputArray _dst, Stream& _stream)
{
GpuMat src = _src.getGpuMat();
const int cn = src.channels();
CV_Assert( src.type() == CV_8UC1 || src.type() == CV_8UC3 );
CV_Assert( lut_cn == 1 || lut_cn == cn );
_dst.create(src.size(), src.type());
GpuMat dst = _dst.getGpuMat();
cudaStream_t stream = StreamAccessor::getStream(_stream);
NppStreamHandler h(stream);
NppiSize sz;
sz.height = src.rows;
sz.width = src.cols;
if (src.type() == CV_8UC1)
{
nppSafeCall( nppiLUT_Linear_8u_C1R(src.ptr<Npp8u>(), static_cast<int>(src.step),
dst.ptr<Npp8u>(), static_cast<int>(dst.step), sz, d_nppLut.ptr<Npp32s>(), d_pLevels.ptr<Npp32s>(), 256) );
}
else
{
nppSafeCall( nppiLUT_Linear_8u_C3R(src.ptr<Npp8u>(), static_cast<int>(src.step),
dst.ptr<Npp8u>(), static_cast<int>(dst.step), sz, pValues3, pLevels3, nValues3) );
}
if (stream == 0)
cudaSafeCall( cudaDeviceSynchronize() );
}
}
#else // (CUDA_VERSION >= 5000)
namespace
{
class LookUpTableImpl : public LookUpTable
{
public:
LookUpTableImpl(InputArray lut);
void transform(InputArray src, OutputArray dst, Stream& stream = Stream::Null());
private:
int lut_cn;
Npp32s pLevels[256];
int nValues3[3];
const Npp32s* pValues3[3];
const Npp32s* pLevels3[3];
Mat nppLut;
Mat nppLut3[3];
};
LookUpTableImpl::LookUpTableImpl(InputArray _lut)
{
nValues3[0] = nValues3[1] = nValues3[2] = 256;
for (int i = 0; i < 256; ++i)
pLevels[i] = i;
pLevels3[0] = pLevels3[1] = pLevels3[2] = pLevels;
Mat lut;
if (_lut.kind() == _InputArray::GPU_MAT)
{
lut = Mat(_lut.getGpuMat());
}
else
{
Mat hLut = _lut.getMat();
CV_Assert( hLut.total() == 256 && hLut.isContinuous() );
lut = hLut;
}
lut_cn = lut.channels();
CV_Assert( lut.depth() == CV_8U );
CV_Assert( lut.rows == 1 && lut.cols == 256 );
lut.convertTo(nppLut, CV_32S);
if (lut_cn == 1)
{
pValues3[0] = pValues3[1] = pValues3[2] = nppLut.ptr<Npp32s>();
}
else
{
cv::split(nppLut, nppLut3);
pValues3[0] = nppLut3[0].ptr<Npp32s>();
pValues3[1] = nppLut3[1].ptr<Npp32s>();
pValues3[2] = nppLut3[2].ptr<Npp32s>();
}
}
void LookUpTableImpl::transform(InputArray _src, OutputArray _dst, Stream& _stream)
{
GpuMat src = _src.getGpuMat();
const int cn = src.channels();
CV_Assert( src.type() == CV_8UC1 || src.type() == CV_8UC3 );
CV_Assert( lut_cn == 1 || lut_cn == cn );
_dst.create(src.size(), src.type());
GpuMat dst = _dst.getGpuMat();
cudaStream_t stream = StreamAccessor::getStream(_stream);
NppStreamHandler h(stream);
NppiSize sz;
sz.height = src.rows;
sz.width = src.cols;
if (src.type() == CV_8UC1)
{
nppSafeCall( nppiLUT_Linear_8u_C1R(src.ptr<Npp8u>(), static_cast<int>(src.step),
dst.ptr<Npp8u>(), static_cast<int>(dst.step), sz, nppLut.ptr<Npp32s>(), pLevels, 256) );
}
else
{
nppSafeCall( nppiLUT_Linear_8u_C3R(src.ptr<Npp8u>(), static_cast<int>(src.step),
dst.ptr<Npp8u>(), static_cast<int>(dst.step), sz, pValues3, pLevels3, nValues3) );
}
if (stream == 0)
cudaSafeCall( cudaDeviceSynchronize() );
}
}
#endif // (CUDA_VERSION >= 5000)
Ptr<LookUpTable> cv::cuda::createLookUpTable(InputArray lut)
{
return makePtr<LookUpTableImpl>(lut);
}
#endif /* !defined (HAVE_CUDA) */