OpenCV
4.1.0
Open Source Computer Vision
|
Classes | |
struct | NppStInterpolationState |
Enumerations | |
enum | NppStBorderType { nppStBorderNone = 0, nppStBorderClamp = 1, nppStBorderWrap = 2, nppStBorderMirror = 3 } |
enum | NppStInterpMode { nppStSupersample, nppStBicubic } |
enum NppStBorderType |
Border type
Filtering operations assume that each pixel has a neighborhood of pixels. The following structure describes possible ways to define non-existent pixels.
enum NppStInterpMode |
NCVStatus nppiStDecimate_32f_C1R | ( | Ncv32f * | d_src, |
Ncv32u | srcStep, | ||
Ncv32f * | d_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | srcRoi, | ||
Ncv32u | scale, | ||
NcvBool | readThruTexture | ||
) |
Downsamples (decimates) an image using the nearest neighbor algorithm. 32-bit float pixels, single channel.
NCVStatus nppiStDecimate_32f_C1R_host | ( | Ncv32f * | h_src, |
Ncv32u | srcStep, | ||
Ncv32f * | h_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | srcRoi, | ||
Ncv32u | scale | ||
) |
Downsamples (decimates) an image using the nearest neighbor algorithm. 32-bit float pixels, single channel. Host implementation.
NCVStatus nppiStDecimate_32s_C1R | ( | Ncv32s * | d_src, |
Ncv32u | srcStep, | ||
Ncv32s * | d_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | srcRoi, | ||
Ncv32u | scale, | ||
NcvBool | readThruTexture | ||
) |
Downsamples (decimates) an image using the nearest neighbor algorithm. 32-bit signed pixels, single channel.
NCVStatus nppiStDecimate_32s_C1R_host | ( | Ncv32s * | h_src, |
Ncv32u | srcStep, | ||
Ncv32s * | h_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | srcRoi, | ||
Ncv32u | scale | ||
) |
Downsamples (decimates) an image using the nearest neighbor algorithm. 32-bit signed pixels, single channel. Host implementation.
NCVStatus nppiStDecimate_32u_C1R | ( | Ncv32u * | d_src, |
Ncv32u | srcStep, | ||
Ncv32u * | d_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | srcRoi, | ||
Ncv32u | scale, | ||
NcvBool | readThruTexture | ||
) |
Downsamples (decimates) an image using the nearest neighbor algorithm. 32-bit unsigned pixels, single channel.
d_src | [IN] Source image pointer (CUDA device memory) |
srcStep | [IN] Source image line step |
d_dst | [OUT] Destination image pointer (CUDA device memory) |
dstStep | [IN] Destination image line step |
srcRoi | [IN] Region of interest in the source image |
scale | [IN] Downsampling scale factor (positive integer) |
readThruTexture | [IN] Performance hint to cache source in texture (true) or read directly (false) |
NCVStatus nppiStDecimate_32u_C1R_host | ( | Ncv32u * | h_src, |
Ncv32u | srcStep, | ||
Ncv32u * | h_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | srcRoi, | ||
Ncv32u | scale | ||
) |
Downsamples (decimates) an image using the nearest neighbor algorithm. 32-bit unsigned pixels, single channel. Host implementation.
h_src | [IN] Source image pointer (Host or pinned memory) |
srcStep | [IN] Source image line step |
h_dst | [OUT] Destination image pointer (Host or pinned memory) |
dstStep | [IN] Destination image line step |
srcRoi | [IN] Region of interest in the source image |
scale | [IN] Downsampling scale factor (positive integer) |
NCVStatus nppiStDecimate_64f_C1R | ( | Ncv64f * | d_src, |
Ncv32u | srcStep, | ||
Ncv64f * | d_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | srcRoi, | ||
Ncv32u | scale, | ||
NcvBool | readThruTexture | ||
) |
Downsamples (decimates) an image using the nearest neighbor algorithm. 64-bit float pixels, single channel.
NCVStatus nppiStDecimate_64f_C1R_host | ( | Ncv64f * | h_src, |
Ncv32u | srcStep, | ||
Ncv64f * | h_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | srcRoi, | ||
Ncv32u | scale | ||
) |
Downsamples (decimates) an image using the nearest neighbor algorithm. 64-bit float pixels, single channel. Host implementation.
NCVStatus nppiStDecimate_64s_C1R | ( | Ncv64s * | d_src, |
Ncv32u | srcStep, | ||
Ncv64s * | d_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | srcRoi, | ||
Ncv32u | scale, | ||
NcvBool | readThruTexture | ||
) |
Downsamples (decimates) an image using the nearest neighbor algorithm. 64-bit signed pixels, single channel.
NCVStatus nppiStDecimate_64s_C1R_host | ( | Ncv64s * | h_src, |
Ncv32u | srcStep, | ||
Ncv64s * | h_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | srcRoi, | ||
Ncv32u | scale | ||
) |
Downsamples (decimates) an image using the nearest neighbor algorithm. 64-bit signed pixels, single channel. Host implementation.
NCVStatus nppiStDecimate_64u_C1R | ( | Ncv64u * | d_src, |
Ncv32u | srcStep, | ||
Ncv64u * | d_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | srcRoi, | ||
Ncv32u | scale, | ||
NcvBool | readThruTexture | ||
) |
Downsamples (decimates) an image using the nearest neighbor algorithm. 64-bit unsigned pixels, single channel.
NCVStatus nppiStDecimate_64u_C1R_host | ( | Ncv64u * | h_src, |
Ncv32u | srcStep, | ||
Ncv64u * | h_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | srcRoi, | ||
Ncv32u | scale | ||
) |
Downsamples (decimates) an image using the nearest neighbor algorithm. 64-bit unsigned pixels, single channel. Host implementation.
NCVStatus nppiStFilterColumnBorder_32f_C1R | ( | const Ncv32f * | pSrc, |
NcvSize32u | srcSize, | ||
Ncv32u | nSrcStep, | ||
Ncv32f * | pDst, | ||
NcvSize32u | dstSize, | ||
Ncv32u | nDstStep, | ||
NcvRect32u | oROI, | ||
NppStBorderType | borderType, | ||
const Ncv32f * | pKernel, | ||
Ncv32s | nKernelSize, | ||
Ncv32s | nAnchor, | ||
Ncv32f | multiplier | ||
) |
Column linear filter. 32-bit floating point image, single channel
Apply vertical linear filter
pSrc | [IN] Source image pointer (CUDA device memory) |
srcSize | [IN] Source image size |
nSrcStep | [IN] Source image line step |
pDst | [OUT] Destination image pointer (CUDA device memory) |
dstSize | [OUT] Destination image size |
nDstStep | [IN] |
oROI | [IN] Region of interest in the source image |
borderType | [IN] Type of border |
pKernel | [IN] Pointer to column kernel values (CUDA device memory) |
nKernelSize | [IN] Size of the kernel in pixels |
nAnchor | [IN] The kernel column alignment with respect to the position of the input pixel |
multiplier | [IN] Value by which the computed result is multiplied |
NCVStatus nppiStFilterRowBorder_32f_C1R | ( | const Ncv32f * | pSrc, |
NcvSize32u | srcSize, | ||
Ncv32u | nSrcStep, | ||
Ncv32f * | pDst, | ||
NcvSize32u | dstSize, | ||
Ncv32u | nDstStep, | ||
NcvRect32u | oROI, | ||
NppStBorderType | borderType, | ||
const Ncv32f * | pKernel, | ||
Ncv32s | nKernelSize, | ||
Ncv32s | nAnchor, | ||
Ncv32f | multiplier | ||
) |
Row linear filter. 32-bit floating point image, single channel
Apply horizontal linear filter
pSrc | [IN] Source image pointer (CUDA device memory) |
srcSize | [IN] Source image size |
nSrcStep | [IN] Source image line step |
pDst | [OUT] Destination image pointer (CUDA device memory) |
dstSize | [OUT] Destination image size |
nDstStep | |
oROI | [IN] Region of interest in the source image |
borderType | [IN] Type of border |
pKernel | [IN] Pointer to row kernel values (CUDA device memory) |
nKernelSize | [IN] Size of the kernel in pixels |
nAnchor | [IN] The kernel row alignment with respect to the position of the input pixel |
multiplier | [IN] Value by which the computed result is multiplied |
NCVStatus nppiStGetInterpolationBufferSize | ( | NcvSize32u | srcSize, |
Ncv32u | nStep, | ||
Ncv32u * | hpSize | ||
) |
Size of a buffer required for interpolation.
Requires several such buffers. See
srcSize | [IN] Frame size (both frames must be of the same size) |
nStep | [IN] Frame line step |
hpSize | [OUT] Where to store computed size (host memory) |
NCVStatus nppiStIntegral_32f32f_C1R | ( | Ncv32f * | d_src, |
Ncv32u | srcStep, | ||
Ncv32f * | d_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | roiSize, | ||
Ncv8u * | pBuffer, | ||
Ncv32u | bufSize, | ||
cudaDeviceProp & | devProp | ||
) |
Creates an integral image representation for the input image
NCVStatus nppiStIntegral_32f32f_C1R_host | ( | Ncv32f * | h_src, |
Ncv32u | srcStep, | ||
Ncv32f * | h_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | roiSize | ||
) |
Creates an integral image representation for the input image. Host implementation
NCVStatus nppiStIntegral_8u32u_C1R | ( | Ncv8u * | d_src, |
Ncv32u | srcStep, | ||
Ncv32u * | d_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | roiSize, | ||
Ncv8u * | pBuffer, | ||
Ncv32u | bufSize, | ||
cudaDeviceProp & | devProp | ||
) |
Creates an integral image representation for the input image
d_src | [IN] Source image pointer (CUDA device memory) |
srcStep | [IN] Source image line step |
d_dst | [OUT] Destination integral image pointer (CUDA device memory) |
dstStep | [IN] Destination image line step |
roiSize | [IN] Region of interest of the source image |
pBuffer | [IN] Pointer to the pre-allocated temporary buffer (CUDA device memory) |
bufSize | [IN] Size of the pBuffer in bytes |
devProp | [IN] CUDA device properties structure, containing texture alignment information |
NCVStatus nppiStIntegral_8u32u_C1R_host | ( | Ncv8u * | h_src, |
Ncv32u | srcStep, | ||
Ncv32u * | h_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | roiSize | ||
) |
Creates an integral image representation for the input image. Host implementation
h_src | [IN] Source image pointer (Host or pinned memory) |
srcStep | [IN] Source image line step |
h_dst | [OUT] Destination integral image pointer (Host or pinned memory) |
dstStep | [IN] Destination image line step |
roiSize | [IN] Region of interest of the source image |
NCVStatus nppiStIntegralGetSize_32f32f | ( | NcvSize32u | roiSize, |
Ncv32u * | pBufsize, | ||
cudaDeviceProp & | devProp | ||
) |
Calculates the size of the temporary buffer for integral image creation
NCVStatus nppiStIntegralGetSize_8u32u | ( | NcvSize32u | roiSize, |
Ncv32u * | pBufsize, | ||
cudaDeviceProp & | devProp | ||
) |
Calculates the size of the temporary buffer for integral image creation
roiSize | [IN] Size of the input image |
pBufsize | [OUT] Pointer to host variable that returns the size of the temporary buffer (in bytes) |
devProp | [IN] CUDA device properties structure, containing texture alignment information |
NCVStatus nppiStInterpolateFrames | ( | const NppStInterpolationState * | pState | ) |
Interpolate frames (images) using provided optical flow (displacement field). 32-bit floating point images, single channel
pState | [IN] structure containing all required parameters (host memory) |
NCVStatus nppiStRectStdDev_32f_C1R | ( | Ncv32u * | d_sum, |
Ncv32u | sumStep, | ||
Ncv64u * | d_sqsum, | ||
Ncv32u | sqsumStep, | ||
Ncv32f * | d_norm, | ||
Ncv32u | normStep, | ||
NcvSize32u | roi, | ||
NcvRect32u | rect, | ||
Ncv32f | scaleArea, | ||
NcvBool | readThruTexture | ||
) |
Computes standard deviation for each rectangular region of the input image using integral images.
d_sum | [IN] Integral image pointer (CUDA device memory) |
sumStep | [IN] Integral image line step |
d_sqsum | [IN] Squared integral image pointer (CUDA device memory) |
sqsumStep | [IN] Squared integral image line step |
d_norm | [OUT] Stddev image pointer (CUDA device memory). Each pixel contains stddev of a rect with top-left corner at the original location in the image |
normStep | [IN] Stddev image line step |
roi | [IN] Region of interest in the source image |
rect | [IN] Rectangular region to calculate stddev over |
scaleArea | [IN] Multiplication factor to account decimated scale |
readThruTexture | [IN] Performance hint to cache source in texture (true) or read directly (false) |
NCVStatus nppiStRectStdDev_32f_C1R_host | ( | Ncv32u * | h_sum, |
Ncv32u | sumStep, | ||
Ncv64u * | h_sqsum, | ||
Ncv32u | sqsumStep, | ||
Ncv32f * | h_norm, | ||
Ncv32u | normStep, | ||
NcvSize32u | roi, | ||
NcvRect32u | rect, | ||
Ncv32f | scaleArea | ||
) |
Computes standard deviation for each rectangular region of the input image using integral images. Host implementation
h_sum | [IN] Integral image pointer (Host or pinned memory) |
sumStep | [IN] Integral image line step |
h_sqsum | [IN] Squared integral image pointer (Host or pinned memory) |
sqsumStep | [IN] Squared integral image line step |
h_norm | [OUT] Stddev image pointer (Host or pinned memory). Each pixel contains stddev of a rect with top-left corner at the original location in the image |
normStep | [IN] Stddev image line step |
roi | [IN] Region of interest in the source image |
rect | [IN] Rectangular region to calculate stddev over |
scaleArea | [IN] Multiplication factor to account decimated scale |
NCVStatus nppiStResize_32f_C1R | ( | const Ncv32f * | pSrc, |
NcvSize32u | srcSize, | ||
Ncv32u | nSrcStep, | ||
NcvRect32u | srcROI, | ||
Ncv32f * | pDst, | ||
NcvSize32u | dstSize, | ||
Ncv32u | nDstStep, | ||
NcvRect32u | dstROI, | ||
Ncv32f | xFactor, | ||
Ncv32f | yFactor, | ||
NppStInterpMode | interpolation | ||
) |
Resize. 32-bit floating point image, single channel
Resizes image using specified filter (interpolation type)
pSrc | [IN] Source image pointer (CUDA device memory) |
srcSize | [IN] Source image size |
nSrcStep | [IN] Source image line step |
srcROI | [IN] Source image region of interest |
pDst | [OUT] Destination image pointer (CUDA device memory) |
dstSize | [IN] Destination image size |
nDstStep | [IN] Destination image line step |
dstROI | [IN] Destination image region of interest |
xFactor | [IN] Row scale factor |
yFactor | [IN] Column scale factor |
interpolation | [IN] Interpolation type |
NCVStatus nppiStSqrIntegral_8u64u_C1R | ( | Ncv8u * | d_src, |
Ncv32u | srcStep, | ||
Ncv64u * | d_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | roiSize, | ||
Ncv8u * | pBuffer, | ||
Ncv32u | bufSize, | ||
cudaDeviceProp & | devProp | ||
) |
Creates a squared integral image representation for the input image
d_src | [IN] Source image pointer (CUDA device memory) |
srcStep | [IN] Source image line step |
d_dst | [OUT] Destination squared integral image pointer (CUDA device memory) |
dstStep | [IN] Destination image line step |
roiSize | [IN] Region of interest of the source image |
pBuffer | [IN] Pointer to the pre-allocated temporary buffer (CUDA device memory) |
bufSize | [IN] Size of the pBuffer in bytes |
devProp | [IN] CUDA device properties structure, containing texture alignment information |
NCVStatus nppiStSqrIntegral_8u64u_C1R_host | ( | Ncv8u * | h_src, |
Ncv32u | srcStep, | ||
Ncv64u * | h_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | roiSize | ||
) |
Creates a squared integral image representation for the input image. Host implementation
h_src | [IN] Source image pointer (Host or pinned memory) |
srcStep | [IN] Source image line step |
h_dst | [OUT] Destination squared integral image pointer (Host or pinned memory) |
dstStep | [IN] Destination image line step |
roiSize | [IN] Region of interest of the source image |
NCVStatus nppiStSqrIntegralGetSize_8u64u | ( | NcvSize32u | roiSize, |
Ncv32u * | pBufsize, | ||
cudaDeviceProp & | devProp | ||
) |
Calculates the size of the temporary buffer for squared integral image creation
roiSize | [IN] Size of the input image |
pBufsize | [OUT] Pointer to host variable that returns the size of the temporary buffer (in bytes) |
devProp | [IN] CUDA device properties structure, containing texture alignment information |
NCVStatus nppiStTranspose_128_C1R | ( | void * | d_src, |
Ncv32u | srcStep, | ||
void * | d_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | srcRoi | ||
) |
Transposes an image. 128-bit pixels of any type, single channel
NCVStatus nppiStTranspose_128_C1R_host | ( | void * | d_src, |
Ncv32u | srcStep, | ||
void * | d_dst, | ||
Ncv32u | dstStep, | ||
NcvSize32u | srcRoi | ||
) |
Transposes an image. 128-bit pixels of any type, single channel. Host implementation
NCVStatus nppiStTranspose_32f_C1R | ( | Ncv32f * | d_src, |
Ncv32u | srcStride, | ||
Ncv32f * | d_dst, | ||
Ncv32u | dstStride, | ||
NcvSize32u | srcRoi | ||
) |
Transposes an image. 32-bit float pixels, single channel
NCVStatus nppiStTranspose_32f_C1R_host | ( | Ncv32f * | h_src, |
Ncv32u | srcStride, | ||
Ncv32f * | h_dst, | ||
Ncv32u | dstStride, | ||
NcvSize32u | srcRoi | ||
) |
Transposes an image. 32-bit float pixels, single channel. Host implementation
NCVStatus nppiStTranspose_32s_C1R | ( | Ncv32s * | d_src, |
Ncv32u | srcStride, | ||
Ncv32s * | d_dst, | ||
Ncv32u | dstStride, | ||
NcvSize32u | srcRoi | ||
) |
Transposes an image. 32-bit signed pixels, single channel
NCVStatus nppiStTranspose_32s_C1R_host | ( | Ncv32s * | h_src, |
Ncv32u | srcStride, | ||
Ncv32s * | h_dst, | ||
Ncv32u | dstStride, | ||
NcvSize32u | srcRoi | ||
) |
Transposes an image. 32-bit signed pixels, single channel. Host implementation
NCVStatus nppiStTranspose_32u_C1R | ( | Ncv32u * | d_src, |
Ncv32u | srcStride, | ||
Ncv32u * | d_dst, | ||
Ncv32u | dstStride, | ||
NcvSize32u | srcRoi | ||
) |
Transposes an image. 32-bit unsigned pixels, single channel
d_src | [IN] Source image pointer (CUDA device memory) |
srcStride | [IN] Source image line step |
d_dst | [OUT] Destination image pointer (CUDA device memory) |
dstStride | [IN] Destination image line step |
srcRoi | [IN] Region of interest of the source image |
NCVStatus nppiStTranspose_32u_C1R_host | ( | Ncv32u * | h_src, |
Ncv32u | srcStride, | ||
Ncv32u * | h_dst, | ||
Ncv32u | dstStride, | ||
NcvSize32u | srcRoi | ||
) |
Transposes an image. 32-bit unsigned pixels, single channel. Host implementation
h_src | [IN] Source image pointer (Host or pinned memory) |
srcStride | [IN] Source image line step |
h_dst | [OUT] Destination image pointer (Host or pinned memory) |
dstStride | [IN] Destination image line step |
srcRoi | [IN] Region of interest of the source image |
NCVStatus nppiStTranspose_64f_C1R | ( | Ncv64f * | d_src, |
Ncv32u | srcStride, | ||
Ncv64f * | d_dst, | ||
Ncv32u | dstStride, | ||
NcvSize32u | srcRoi | ||
) |
Transposes an image. 64-bit float pixels, single channel
NCVStatus nppiStTranspose_64f_C1R_host | ( | Ncv64f * | h_src, |
Ncv32u | srcStride, | ||
Ncv64f * | h_dst, | ||
Ncv32u | dstStride, | ||
NcvSize32u | srcRoi | ||
) |
Transposes an image. 64-bit float pixels, single channel. Host implementation
NCVStatus nppiStTranspose_64s_C1R | ( | Ncv64s * | d_src, |
Ncv32u | srcStride, | ||
Ncv64s * | d_dst, | ||
Ncv32u | dstStride, | ||
NcvSize32u | srcRoi | ||
) |
Transposes an image. 64-bit signed pixels, single channel
NCVStatus nppiStTranspose_64s_C1R_host | ( | Ncv64s * | h_src, |
Ncv32u | srcStride, | ||
Ncv64s * | h_dst, | ||
Ncv32u | dstStride, | ||
NcvSize32u | srcRoi | ||
) |
Transposes an image. 64-bit signed pixels, single channel. Host implementation
NCVStatus nppiStTranspose_64u_C1R | ( | Ncv64u * | d_src, |
Ncv32u | srcStride, | ||
Ncv64u * | d_dst, | ||
Ncv32u | dstStride, | ||
NcvSize32u | srcRoi | ||
) |
Transposes an image. 64-bit unsigned pixels, single channel
NCVStatus nppiStTranspose_64u_C1R_host | ( | Ncv64u * | h_src, |
Ncv32u | srcStride, | ||
Ncv64u * | h_dst, | ||
Ncv32u | dstStride, | ||
NcvSize32u | srcRoi | ||
) |
Transposes an image. 64-bit unsigned pixels, single channel. Host implementation
NCVStatus nppiStVectorWarp_PSF1x1_32f_C1 | ( | const Ncv32f * | pSrc, |
NcvSize32u | srcSize, | ||
Ncv32u | nSrcStep, | ||
const Ncv32f * | pU, | ||
const Ncv32f * | pV, | ||
Ncv32u | nVFStep, | ||
Ncv32f | timeScale, | ||
Ncv32f * | pDst | ||
) |
Warp image using provided 2D vector field and 1x1 point spread function. 32-bit floating point image, single channel
During warping pixels from the source image may fall between pixels of the destination image. PSF (point spread function) describes how the source image pixel affects pixels of the destination. For 1x1 PSF only single pixel with the largest intersection is affected (similar to nearest interpolation).
Destination image size and line step must be the same as the source image size and line step
pSrc | [IN] Source image pointer (CUDA device memory) |
srcSize | [IN] Source image size |
nSrcStep | [IN] Source image line step |
pU | [IN] Pointer to horizontal displacement field (CUDA device memory) |
pV | [IN] Pointer to vertical displacement field (CUDA device memory) |
nVFStep | [IN] Displacement field line step |
timeScale | [IN] Value by which displacement field will be scaled for warping |
pDst | [OUT] Destination image pointer (CUDA device memory) |
NCVStatus nppiStVectorWarp_PSF2x2_32f_C1 | ( | const Ncv32f * | pSrc, |
NcvSize32u | srcSize, | ||
Ncv32u | nSrcStep, | ||
const Ncv32f * | pU, | ||
const Ncv32f * | pV, | ||
Ncv32u | nVFStep, | ||
Ncv32f * | pBuffer, | ||
Ncv32f | timeScale, | ||
Ncv32f * | pDst | ||
) |
Warp image using provided 2D vector field and 2x2 point spread function. 32-bit floating point image, single channel
During warping pixels from the source image may fall between pixels of the destination image. PSF (point spread function) describes how the source image pixel affects pixels of the destination. For 2x2 PSF all four intersected pixels will be affected.
Destination image size and line step must be the same as the source image size and line step
pSrc | [IN] Source image pointer (CUDA device memory) |
srcSize | [IN] Source image size |
nSrcStep | [IN] Source image line step |
pU | [IN] Pointer to horizontal displacement field (CUDA device memory) |
pV | [IN] Pointer to vertical displacement field (CUDA device memory) |
nVFStep | [IN] Displacement field line step |
pBuffer | |
timeScale | [IN] Value by which displacement field will be scaled for warping |
pDst | [OUT] Destination image pointer (CUDA device memory) |
NCVStatus nppiStVectorWarpGetBufferSize | ( | NcvSize32u | srcSize, |
Ncv32u | nSrcStep, | ||
Ncv32u * | hpSize | ||
) |
Size of buffer required for vector image warping.
srcSize | [IN] Source image size |
nSrcStep | [IN] Source image line step |
hpSize | [OUT] Where to store computed size (host memory) |