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

added bitwise operations into gpu module

Browse Source
This commit is contained in:
Alexey Spizhevoy
2010-11-19 10:19:35 +00:00
parent 7df9aef99c
commit faf4d0bc74
4 changed files with 442 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

52
modules/gpu/include/opencv2/gpu/gpu.hpp
View File

@@ -474,33 +474,61 @@ namespace cv
//! computes magnitude of each (x(i), y(i)) vector
//! supports only floating-point source
CV_EXPORTS void magnitude(const GpuMat& x, const GpuMat& y, GpuMat& magnitude);
//! Acync version
//! Async version
CV_EXPORTS void magnitude(const GpuMat& x, const GpuMat& y, GpuMat& magnitude, const Stream& stream);
//! computes squared magnitude of each (x(i), y(i)) vector
//! supports only floating-point source
CV_EXPORTS void magnitudeSqr(const GpuMat& x, const GpuMat& y, GpuMat& magnitude);
//! Acync version
//! Async version
CV_EXPORTS void magnitudeSqr(const GpuMat& x, const GpuMat& y, GpuMat& magnitude, const Stream& stream);
//! computes angle (angle(i)) of each (x(i), y(i)) vector
//! supports only floating-point source
CV_EXPORTS void phase(const GpuMat& x, const GpuMat& y, GpuMat& angle, bool angleInDegrees = false);
//! Acync version
//! Async version
CV_EXPORTS void phase(const GpuMat& x, const GpuMat& y, GpuMat& angle, bool angleInDegrees, const Stream& stream);
//! converts Cartesian coordinates to polar
//! supports only floating-point source
CV_EXPORTS void cartToPolar(const GpuMat& x, const GpuMat& y, GpuMat& magnitude, GpuMat& angle, bool angleInDegrees = false);
//! Acync version
//! Async version
CV_EXPORTS void cartToPolar(const GpuMat& x, const GpuMat& y, GpuMat& magnitude, GpuMat& angle, bool angleInDegrees, const Stream& stream);
//! converts polar coordinates to Cartesian
//! supports only floating-point source
CV_EXPORTS void polarToCart(const GpuMat& magnitude, const GpuMat& angle, GpuMat& x, GpuMat& y, bool angleInDegrees = false);
//! Acync version
//! Async version
CV_EXPORTS void polarToCart(const GpuMat& magnitude, const GpuMat& angle, GpuMat& x, GpuMat& y, bool angleInDegrees, const Stream& stream);
//! Perfroms per-elements bit-wise inversion
CV_EXPORTS void bitwise_not(const GpuMat& src, GpuMat& dst);
//! Async version
CV_EXPORTS void bitwise_not(const GpuMat& src, GpuMat& dst, const Stream& stream);
//! Calculates per-element bit-wise disjunction of two arrays
CV_EXPORTS void bitwise_or(const GpuMat& src1, const GpuMat& src2, GpuMat& dst);
//! Async version
CV_EXPORTS void bitwise_or(const GpuMat& src1, const GpuMat& src2, GpuMat& dst, const Stream& stream);
//! Calculates per-element bit-wise conjunction of two arrays
CV_EXPORTS void bitwise_and(const GpuMat& src1, const GpuMat& src2, GpuMat& dst);
//! Async version
CV_EXPORTS void bitwise_and(const GpuMat& src1, const GpuMat& src2, GpuMat& dst, const Stream& stream);
//! Calculates per-element bit-wise "exclusive or" operation
CV_EXPORTS void bitwise_xor(const GpuMat& src1, const GpuMat& src2, GpuMat& dst);
//! Async version
CV_EXPORTS void bitwise_xor(const GpuMat& src1, const GpuMat& src2, GpuMat& dst, const Stream& stream);
//! Logical operators
CV_EXPORTS GpuMat operator ~ (const GpuMat& src);
CV_EXPORTS GpuMat operator | (const GpuMat& src1, const GpuMat& src2);
CV_EXPORTS GpuMat operator & (const GpuMat& src1, const GpuMat& src2);
CV_EXPORTS GpuMat operator ^ (const GpuMat& src1, const GpuMat& src2);
////////////////////////////// Image processing //////////////////////////////
//! DST[x,y] = SRC[xmap[x,y],ymap[x,y]] with bilinear interpolation.
@@ -523,7 +551,7 @@ namespace cv
//! Supported types of input disparity: CV_8U, CV_16S.
//! Output disparity has CV_8UC4 type in BGRA format (alpha = 255).
CV_EXPORTS void drawColorDisp(const GpuMat& src_disp, GpuMat& dst_disp, int ndisp);
//! Acync version
//! Async version
CV_EXPORTS void drawColorDisp(const GpuMat& src_disp, GpuMat& dst_disp, int ndisp, const Stream& stream);
//! Reprojects disparity image to 3D space.
@@ -532,12 +560,12 @@ namespace cv
//! Each element of this matrix will contain the 3D coordinates of the point (x,y,z,1), computed from the disparity map.
//! Q is the 4x4 perspective transformation matrix that can be obtained with cvStereoRectify.
CV_EXPORTS void reprojectImageTo3D(const GpuMat& disp, GpuMat& xyzw, const Mat& Q);
//! Acync version
//! Async version
CV_EXPORTS void reprojectImageTo3D(const GpuMat& disp, GpuMat& xyzw, const Mat& Q, const Stream& stream);
//! converts image from one color space to another
CV_EXPORTS void cvtColor(const GpuMat& src, GpuMat& dst, int code, int dcn = 0);
//! Acync version
//! Async version
CV_EXPORTS void cvtColor(const GpuMat& src, GpuMat& dst, int code, int dcn, const Stream& stream);
//! applies fixed threshold to the image.
@@ -793,7 +821,7 @@ namespace cv
//! Output disparity has CV_8U type.
void operator() ( const GpuMat& left, const GpuMat& right, GpuMat& disparity);
//! Acync version
//! Async version
void operator() ( const GpuMat& left, const GpuMat& right, GpuMat& disparity, const Stream & stream);
//! Some heuristics that tries to estmate
@@ -848,7 +876,7 @@ namespace cv
//! if disparity is empty output type will be CV_16S else output type will be disparity.type().
void operator()(const GpuMat& left, const GpuMat& right, GpuMat& disparity);
//! Acync version
//! Async version
void operator()(const GpuMat& left, const GpuMat& right, GpuMat& disparity, Stream& stream);
@@ -907,7 +935,7 @@ namespace cv
//! if disparity is empty output type will be CV_16S else output type will be disparity.type().
void operator()(const GpuMat& left, const GpuMat& right, GpuMat& disparity);
//! Acync version
//! Async version
void operator()(const GpuMat& left, const GpuMat& right, GpuMat& disparity, Stream& stream);
int ndisp;
@@ -963,7 +991,7 @@ namespace cv
//! disparity must have CV_8U or CV_16S type, image must have CV_8UC1 or CV_8UC3 type.
void operator()(const GpuMat& disparity, const GpuMat& image, GpuMat& dst);
//! Acync version
//! Async version
void operator()(const GpuMat& disparity, const GpuMat& image, GpuMat& dst, Stream& stream);
private:

128
modules/gpu/src/arithm.cpp
View File

@@ -81,6 +81,18 @@ void cv::gpu::cartToPolar(const GpuMat&, const GpuMat&, GpuMat&, GpuMat&, bool)
void cv::gpu::cartToPolar(const GpuMat&, const GpuMat&, GpuMat&, GpuMat&, bool, const Stream&) { throw_nogpu(); }
void cv::gpu::polarToCart(const GpuMat&, const GpuMat&, GpuMat&, GpuMat&, bool) { throw_nogpu(); }
void cv::gpu::polarToCart(const GpuMat&, const GpuMat&, GpuMat&, GpuMat&, bool, const Stream&) { throw_nogpu(); }
void cv::gpu::bitwise_not(const GpuMat&, GpuMat&) { throw_nogpu(); }
void cv::gpu::bitwise_not(const GpuMat&, GpuMat&, const Stream& stream) { throw_nogpu(); }
void cv::gpu::bitwise_or(const GpuMat&, const GpuMat&, GpuMat&) { throw_nogpu(); }
void cv::gpu::bitwise_or(const GpuMat&, const GpuMat&, GpuMat&, const Stream& stream) { throw_nogpu(); }
void cv::gpu::bitwise_and(const GpuMat&, const GpuMat&, GpuMat&) { throw_nogpu(); }
void cv::gpu::bitwise_and(const GpuMat&, const GpuMat&, GpuMat&, const Stream& stream) { throw_nogpu(); }
void cv::gpu::bitwise_xor(const GpuMat&, const GpuMat&, GpuMat&) { throw_nogpu(); }
void cv::gpu::bitwise_xor(const GpuMat&, const GpuMat&, GpuMat&, const Stream& stream) { throw_nogpu(); }
cv::gpu::GpuMat cv::gpu::operator ~ (const GpuMat&) { throw_nogpu(); return GpuMat(); }
cv::gpu::GpuMat cv::gpu::operator | (const GpuMat&, const GpuMat&) { throw_nogpu(); return GpuMat(); }
cv::gpu::GpuMat cv::gpu::operator & (const GpuMat&, const GpuMat&) { throw_nogpu(); return GpuMat(); }
cv::gpu::GpuMat cv::gpu::operator ^ (const GpuMat&, const GpuMat&) { throw_nogpu(); return GpuMat(); }
#else /* !defined (HAVE_CUDA) */
@@ -856,4 +868,120 @@ void cv::gpu::polarToCart(const GpuMat& magnitude, const GpuMat& angle, GpuMat&
::polarToCart_caller(magnitude, angle, x, y, angleInDegrees, StreamAccessor::getStream(stream));
}
//////////////////////////////////////////////////////////////////////////////
// Per-element bit-wise logical matrix operations
namespace cv { namespace gpu { namespace mathfunc
{
void bitwise_not_caller(const DevMem2D src, int elemSize, PtrStep dst, cudaStream_t stream);
void bitwise_or_caller(int cols, int rows, const PtrStep src1, const PtrStep src2, int elemSize, PtrStep dst, cudaStream_t stream);
void bitwise_and_caller(int cols, int rows, const PtrStep src1, const PtrStep src2, int elemSize, PtrStep dst, cudaStream_t stream);
void bitwise_xor_caller(int cols, int rows, const PtrStep src1, const PtrStep src2, int elemSize, PtrStep dst, cudaStream_t stream);
}}}
namespace
{
void bitwise_not_caller(const GpuMat& src, GpuMat& dst, cudaStream_t stream)
{
dst.create(src.size(), src.type());
mathfunc::bitwise_not_caller(src, src.elemSize(), dst, stream);
}
void bitwise_or_caller(const GpuMat& src1, const GpuMat& src2, GpuMat& dst, cudaStream_t stream)
{
CV_Assert(src1.size() == src2.size());
CV_Assert(src1.type() == src2.type());
dst.create(src1.size(), src1.type());
mathfunc::bitwise_or_caller(dst.cols, dst.rows, src1, src2, dst.elemSize(), dst, stream);
}
void bitwise_and_caller(const GpuMat& src1, const GpuMat& src2, GpuMat& dst, cudaStream_t stream)
{
CV_Assert(src1.size() == src2.size());
CV_Assert(src1.type() == src2.type());
dst.create(src1.size(), src1.type());
mathfunc::bitwise_and_caller(dst.cols, dst.rows, src1, src2, dst.elemSize(), dst, stream);
}
void bitwise_xor_caller(const GpuMat& src1, const GpuMat& src2, GpuMat& dst, cudaStream_t stream)
{
CV_Assert(src1.size() == src2.size());
CV_Assert(src1.type() == src2.type());
dst.create(src1.size(), src1.type());
mathfunc::bitwise_xor_caller(dst.cols, dst.rows, src1, src2, dst.elemSize(), dst, stream);
}
}
void cv::gpu::bitwise_not(const GpuMat& src, GpuMat& dst)
{
::bitwise_not_caller(src, dst, 0);
}
void cv::gpu::bitwise_not(const GpuMat& src, GpuMat& dst, const Stream& stream)
{
::bitwise_not_caller(src, dst, StreamAccessor::getStream(stream));
}
void cv::gpu::bitwise_or(const GpuMat& src1, const GpuMat& src2, GpuMat& dst)
{
::bitwise_or_caller(src1, src2, dst, 0);
}
void cv::gpu::bitwise_or(const GpuMat& src1, const GpuMat& src2, GpuMat& dst, const Stream& stream)
{
::bitwise_or_caller(src1, src2, dst, StreamAccessor::getStream(stream));
}
void cv::gpu::bitwise_and(const GpuMat& src1, const GpuMat& src2, GpuMat& dst)
{
::bitwise_and_caller(src1, src2, dst, 0);
}
void cv::gpu::bitwise_and(const GpuMat& src1, const GpuMat& src2, GpuMat& dst, const Stream& stream)
{
::bitwise_and_caller(src1, src2, dst, StreamAccessor::getStream(stream));
}
void cv::gpu::bitwise_xor(const GpuMat& src1, const GpuMat& src2, GpuMat& dst)
{
::bitwise_xor_caller(src1, src2, dst, 0);
}
void cv::gpu::bitwise_xor(const GpuMat& src1, const GpuMat& src2, GpuMat& dst, const Stream& stream)
{
::bitwise_xor_caller(src1, src2, dst, StreamAccessor::getStream(stream));
}
cv::gpu::GpuMat cv::gpu::operator ~ (const GpuMat& src)
{
GpuMat dst;
bitwise_not(src, dst);
return dst;
}
cv::gpu::GpuMat cv::gpu::operator | (const GpuMat& src1, const GpuMat& src2)
{
GpuMat dst;
bitwise_or(src1, src2, dst);
return dst;
}
cv::gpu::GpuMat cv::gpu::operator & (const GpuMat& src1, const GpuMat& src2)
{
GpuMat dst;
bitwise_and(src1, src2, dst);
return dst;
}
cv::gpu::GpuMat cv::gpu::operator ^ (const GpuMat& src1, const GpuMat& src2)
{
GpuMat dst;
bitwise_xor(src1, src2, dst);
return dst;
}
#endif /* !defined (HAVE_CUDA) */

101
modules/gpu/src/cuda/mathfunc.cu
View File

@@ -238,4 +238,105 @@ namespace cv { namespace gpu { namespace mathfunc
{
compare_ne<float, float>(src1, src2, dst);
}
//////////////////////////////////////////////////////////////////////////////
// Per-element bit-wise logical matrix operations
__global__ void bitwise_not_kernel(int cols, int rows, const PtrStep src, PtrStep dst)
{
const int x = blockDim.x * blockIdx.x + threadIdx.x;
const int y = blockDim.y * blockIdx.y + threadIdx.y;
if (x < cols && y < rows)
{
dst.ptr(y)[x] = ~src.ptr(y)[x];
}
}
void bitwise_not_caller(const DevMem2D src, int elemSize, PtrStep dst, cudaStream_t stream)
{
dim3 threads(16, 16, 1);
dim3 grid(divUp(src.cols * elemSize, threads.x), divUp(src.rows, threads.y), 1);
bitwise_not_kernel<<<grid, threads, 0, stream>>>(src.cols * elemSize, src.rows, src, dst);
if (stream == 0)
cudaSafeCall(cudaThreadSynchronize());
}
__global__ void bitwise_or_kernel(int cols, int rows, const PtrStep src1, const PtrStep src2, PtrStep dst)
{
const int x = blockDim.x * blockIdx.x + threadIdx.x;
const int y = blockDim.y * blockIdx.y + threadIdx.y;
if (x < cols && y < rows)
{
dst.ptr(y)[x] = src1.ptr(y)[x] | src2.ptr(y)[x];
}
}
void bitwise_or_caller(int cols, int rows, const PtrStep src1, const PtrStep src2, int elemSize, PtrStep dst, cudaStream_t stream)
{
dim3 threads(16, 16, 1);
dim3 grid(divUp(cols * elemSize, threads.x), divUp(rows, threads.y), 1);
bitwise_or_kernel<<<grid, threads, 0, stream>>>(cols * elemSize, rows, src1, src2, dst);
if (stream == 0)
cudaSafeCall(cudaThreadSynchronize());
}
__global__ void bitwise_and_kernel(int cols, int rows, const PtrStep src1, const PtrStep src2, PtrStep dst)
{
const int x = blockDim.x * blockIdx.x + threadIdx.x;
const int y = blockDim.y * blockIdx.y + threadIdx.y;
if (x < cols && y < rows)
{
dst.ptr(y)[x] = src1.ptr(y)[x] & src2.ptr(y)[x];
}
}
void bitwise_and_caller(int cols, int rows, const PtrStep src1, const PtrStep src2, int elemSize, PtrStep dst, cudaStream_t stream)
{
dim3 threads(16, 16, 1);
dim3 grid(divUp(cols * elemSize, threads.x), divUp(rows, threads.y), 1);
bitwise_and_kernel<<<grid, threads, 0, stream>>>(cols * elemSize, rows, src1, src2, dst);
if (stream == 0)
cudaSafeCall(cudaThreadSynchronize());
}
__global__ void bitwise_xor_kernel(int cols, int rows, const PtrStep src1, const PtrStep src2, PtrStep dst)
{
const int x = blockDim.x * blockIdx.x + threadIdx.x;
const int y = blockDim.y * blockIdx.y + threadIdx.y;
if (x < cols && y < rows)
{
dst.ptr(y)[x] = src1.ptr(y)[x] ^ src2.ptr(y)[x];
}
}
void bitwise_xor_caller(int cols, int rows, const PtrStep src1, const PtrStep src2, int elemSize, PtrStep dst, cudaStream_t stream)
{
dim3 threads(16, 16, 1);
dim3 grid(divUp(cols * elemSize, threads.x), divUp(rows, threads.y), 1);
bitwise_xor_kernel<<<grid, threads, 0, stream>>>(cols * elemSize, rows, src1, src2, dst);
if (stream == 0)
cudaSafeCall(cudaThreadSynchronize());
}
}}}