generic-library/opencv
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

added ensureSizeIsEnough into gpu module, updated reduction methods

Browse Source
This commit is contained in:
Alexey Spizhevoy 2011-01-18 12:36:01 +00:00
parent f3a2656808
commit cbb132ccb1
6 changed files with 151 additions and 115 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

20
doc/gpu_data_structures.tex
View File

@ -12,4 +12,24 @@ Creates continuous matrix in GPU memory.
\cvarg{m}{Destination matrix. Will be only reshaped if it has proper type and area (\texttt{rows} $\times$ \texttt{cols}).}
\end{description}
Also the following wrappers are available:
\cvdefCpp{GpuMat createContinuous(int rows, int cols, int type);\newline
void createContinuous(Size size, int type, GpuMat\& m);\newline
GpuMat createContinuous(Size size, int type);}
Matrix is called continuous if its elements are stored continuously, i.e. wuthout gaps in the end of each row.
\cvCppFunc{gpu::ensureSizeIsEnough}
Ensures that size of matrix is big enough and matrix has proper type. The function doesn't reallocate memory if matrix has proper attributes already.
\cvdefCpp{void ensureSizeIsEnough(int rows, int cols, int type, GpuMat\& m);}
\begin{description}
\cvarg{rows}{Minimum desired number of rows.}
\cvarg{cols}{Minimum desired number of cols.}
\cvarg{type}{Desired matrix type.}
\cvarg{m}{Destination matrix.}
\end{description}
Also the following wrapper is available:
\cvdefCpp{void ensureSizeIsEnough(Size size, int type, GpuMat\& m);}

6
modules/gpu/include/opencv2/gpu/gpu.hpp
View File

@ -252,9 +252,13 @@ namespace cv
#include "GpuMat_BetaDeprecated.hpp"
#endif
//! creates continuous GPU matrix
//! Creates continuous GPU matrix
CV_EXPORTS void createContinuous(int rows, int cols, int type, GpuMat& m);
//! Ensures that size of the given matrix is not less than (rows, cols) size
//! and matrix type is match specified one too
CV_EXPORTS void ensureSizeIsEnough(int rows, int cols, int type, GpuMat& m);
//////////////////////////////// CudaMem ////////////////////////////////
// CudaMem is limited cv::Mat with page locked memory allocation.
// Page locked memory is only needed for async and faster coping to GPU.

5
modules/gpu/include/opencv2/gpu/matrix_operations.hpp
View File

@ -364,6 +364,10 @@ inline GpuMat createContinuous(Size size, int type)
return m;
}
inline void ensureSizeIsEnough(Size size, int type, GpuMat& m)
{
ensureSizeIsEnough(size.height, size.width, type, m);
}
///////////////////////////////////////////////////////////////////////
@ -401,6 +405,7 @@ inline CudaMem::CudaMem(const Mat& m, int _alloc_type) : flags(0), rows(0), cols
inline CudaMem::~CudaMem()
{
release();
}
inline CudaMem& CudaMem::operator = (const CudaMem& m)

7
modules/gpu/src/matrix_operations.cpp
View File

@ -551,6 +551,13 @@ void cv::gpu::createContinuous(int rows, int cols, int type, GpuMat& m)
m = m.reshape(0, rows);
}
void cv::gpu::ensureSizeIsEnough(int rows, int cols, int type, GpuMat& m)
{
if (m.type() == type && m.rows >= rows && m.cols >= cols)
return;
m.create(rows, cols, type);
}
///////////////////////////////////////////////////////////////////////
//////////////////////////////// CudaMem //////////////////////////////

40
modules/gpu/src/matrix_reductions.cpp
View File

@ -159,7 +159,7 @@ Scalar cv::gpu::sum(const GpuMat& src, GpuMat& buf)
Size bufSize;
sum::get_buf_size_required(src.cols, src.rows, src.channels(), bufSize.width, bufSize.height);
buf.create(bufSize, CV_8U);
ensureSizeIsEnough(bufSize, CV_8U, buf);
Caller caller = callers[hasAtomicsSupport(getDevice())][src.depth()];
if (!caller) CV_Error(CV_StsBadArg, "sum: unsupported type");
@ -192,7 +192,7 @@ Scalar cv::gpu::sqrSum(const GpuMat& src, GpuMat& buf)
Size bufSize;
sum::get_buf_size_required(src.cols, src.rows, src.channels(), bufSize.width, bufSize.height);
buf.create(bufSize, CV_8U);
ensureSizeIsEnough(bufSize, CV_8U, buf);
Caller caller = callers[hasAtomicsSupport(getDevice())][src.depth()];
if (!caller) CV_Error(CV_StsBadArg, "sqrSum: unsupported type");
@ -265,7 +265,7 @@ void cv::gpu::minMax(const GpuMat& src, double* minVal, double* maxVal, const Gp
Size bufSize;
get_buf_size_required(src.cols, src.rows, src.elemSize(), bufSize.width, bufSize.height);
buf.create(bufSize, CV_8U);
ensureSizeIsEnough(bufSize, CV_8U, buf);
if (mask.empty())
{
@ -292,31 +292,31 @@ namespace cv { namespace gpu { namespace mathfunc { namespace minmaxloc {
template <typename T>
void min_max_loc_caller(const DevMem2D src, double* minval, double* maxval,
int minloc[2], int maxloc[2], PtrStep valbuf, PtrStep locbuf);
int minloc[2], int maxloc[2], PtrStep valBuf, PtrStep locBuf);
template <typename T>
void min_max_loc_mask_caller(const DevMem2D src, const PtrStep mask, double* minval, double* maxval,
int minloc[2], int maxloc[2], PtrStep valbuf, PtrStep locbuf);
int minloc[2], int maxloc[2], PtrStep valBuf, PtrStep locBuf);
template <typename T>
void min_max_loc_multipass_caller(const DevMem2D src, double* minval, double* maxval,
int minloc[2], int maxloc[2], PtrStep valbuf, PtrStep locbuf);
int minloc[2], int maxloc[2], PtrStep valBuf, PtrStep locBuf);
template <typename T>
void min_max_loc_mask_multipass_caller(const DevMem2D src, const PtrStep mask, double* minval, double* maxval,
int minloc[2], int maxloc[2], PtrStep valbuf, PtrStep locbuf);
int minloc[2], int maxloc[2], PtrStep valBuf, PtrStep locBuf);
}}}}
void cv::gpu::minMaxLoc(const GpuMat& src, double* minVal, double* maxVal, Point* minLoc, Point* maxLoc, const GpuMat& mask)
{
GpuMat valbuf, locbuf;
minMaxLoc(src, minVal, maxVal, minLoc, maxLoc, mask, valbuf, locbuf);
GpuMat valBuf, locBuf;
minMaxLoc(src, minVal, maxVal, minLoc, maxLoc, mask, valBuf, locBuf);
}
void cv::gpu::minMaxLoc(const GpuMat& src, double* minVal, double* maxVal, Point* minLoc, Point* maxLoc,
const GpuMat& mask, GpuMat& valbuf, GpuMat& locbuf)
const GpuMat& mask, GpuMat& valBuf, GpuMat& locBuf)
{
using namespace mathfunc::minmaxloc;
@ -348,23 +348,23 @@ void cv::gpu::minMaxLoc(const GpuMat& src, double* minVal, double* maxVal, Point
int minLoc_[2];
int maxLoc_[2];
Size valbuf_size, locbuf_size;
get_buf_size_required(src.cols, src.rows, src.elemSize(), valbuf_size.width,
valbuf_size.height, locbuf_size.width, locbuf_size.height);
valbuf.create(valbuf_size, CV_8U);
locbuf.create(locbuf_size, CV_8U);
Size valBufSize, locBufSize;
get_buf_size_required(src.cols, src.rows, src.elemSize(), valBufSize.width,
valBufSize.height, locBufSize.width, locBufSize.height);
ensureSizeIsEnough(valBufSize, CV_8U, valBuf);
ensureSizeIsEnough(locBufSize, CV_8U, locBuf);
if (mask.empty())
{
Caller caller = callers[hasAtomicsSupport(getDevice())][src.type()];
if (!caller) CV_Error(CV_StsBadArg, "minMaxLoc: unsupported type");
caller(src, minVal, maxVal, minLoc_, maxLoc_, valbuf, locbuf);
caller(src, minVal, maxVal, minLoc_, maxLoc_, valBuf, locBuf);
}
else
{
MaskedCaller caller = masked_callers[hasAtomicsSupport(getDevice())][src.type()];
if (!caller) CV_Error(CV_StsBadArg, "minMaxLoc: unsupported type");
caller(src, mask, minVal, maxVal, minLoc_, maxLoc_, valbuf, locbuf);
caller(src, mask, minVal, maxVal, minLoc_, maxLoc_, valBuf, locBuf);
}
if (minLoc) { minLoc->x = minLoc_[0]; minLoc->y = minLoc_[1]; }
@ -411,9 +411,9 @@ int cv::gpu::countNonZero(const GpuMat& src, GpuMat& buf)
CV_Assert(src.channels() == 1);
CV_Assert(src.type() != CV_64F || hasNativeDoubleSupport(getDevice()));
Size buf_size;
get_buf_size_required(src.cols, src.rows, buf_size.width, buf_size.height);
buf.create(buf_size, CV_8U);
Size bufSize;
get_buf_size_required(src.cols, src.rows, bufSize.width, bufSize.height);
ensureSizeIsEnough(bufSize, CV_8U, buf);
Caller caller = callers[hasAtomicsSupport(getDevice())][src.type()];
if (!caller) CV_Error(CV_StsBadArg, "countNonZero: unsupported type");