OpenCV中OpenCL模块函数

It currently develop and test on GPU devices only. This includes both discrete GPUs(NVidia,AMD), as well as integrated chips(AMD APU and intel HD devices).

The ocl module can be found under the “modules”directory. In “modules/ocl/src” you can find the source code for the cpp class that wrap around the direct kernel invocation. The kernels themselves can be found in “modules/ocl/src/kernels.” Samples can be found under “samples/ocl.”Accuracy tests can be found in “modules/ocl/test,”and performance tests under “module/ocl/perf.”

If a function support 4-channel operator, it should support 3-channel operator as well, because All the 3-channel matrix(i.e. RGB image) are represented by 4-channel matrix in oclMat. It means 3-channel image have 4-channel space with the last channel unused.

 

1.      getDevice：returns the list of devices；

2.      setDevice：sets adevice and initializes it；

3.      setBinpath：if you call this function and set a valid path, the OCL module will save the complied kernel to the address in the first time and reload the binary since that, it can save compilation time at the runtime；

4.      getoclContext：returns the pointer to the opencl context；

5.      getoclCommandQueue：returns the pointer to the opencl command queue；

6.      class::oclMat：OpenCV C++1-D or 2-D dense array class, the oclMat is the mirror of Mat with the extension of ocl feature；

7.      oclMat::convertTo：the method converts source pixel values to the target datatype, saturate cast is applied in the end to avoid possible overflows；

8.      oclMat::copyTo：copies the matrix to another one；

9.      oclMat::setTo：sets all or some of the array elements to the specified value；

10.  absdiff：computes per-element absolute difference between two arrays or between array and ascalar；

11.  add：computes per-element addition between two arrays or between array and a scalar；

12.  subtract：computes per-element subtract between two arrays or between array and a scalar；

13.  multiply：computes per-element multiply between two arrays or between array and a scalar；

14.  divide：computes per-element divide between two arrays or between array and a scalar；

15.  bitwise_and：computesper-element bitwise_and between two arrays or between array and a scalar；

16.  bitwise_or：computes per-element bitwise_or between two arrays or between array and a scalar；

17.  bitwise_xor：computes per-element bitwise_xor between two arrays or between array and a scalar；

18.  bitwise_not：the function bitwise not compute per-element bit-wise inversion of the source array；

19.  cartToPolar：calculates the magnitude and angle of 2d vectors；

20.  polarToCart：calculates the Cartesian coordinates of each 2D vector represented by the correspondingelements of magnitude and angle；

21.  compare：performs per-element comparison of two arrays or an array and scalar value；

22.  exp：calculates the exponent of every element of the input array；

23.  log：calculates the log of every element of the input array；

24.  LUT：performs a look-up table transform of an array；

25.  magnitude：calculates magnitude of 2D vectors formed from the corresponding element of x and y arrays；

26.  flip：flips the array in one of three different ways(row and column indices are 0-based)；

27.  meanStdDev：compute the mean and the standard deviation M of array elements, independently for eachchannel, and return it via the output parameters；

28.  merge：composes a multi-channel array from several single-channel arrays；

29.  split：split multi-channel array into separate single-channel arrays；

30.  norm：calculates absolute array norm, absolute difference norm, or relative difference norm；

31.  phase：computes the rotation angle of each 2D vector that is formed from the corresponding elementsof x and y；

32.  pow：raises every element of the input array to p；

33.  transpose：transposes a matrix；

34.  dft(cv::dft)：performs a forward or inverse discrete Fourier transform(1D or 2D) of the floating pointmatrix；

35.  gemm(cv::gemm)：performs generalized matrix multiplication；

36.  countNonZero：returns the number of non-zero elements in src；

37.  minMax：finds global minimum and maximum in a whole array or sub-array；

38.  minMaxLoc：find minimum and maximum element values and their positions；

39.  sum：returns the sum of matrix elements for each channel；

40.  sqrSum：returns the squared sum of matrix elements for each channel；

41.  Sobel：computesthe first x- or y- spatial image derivative using Sobel operator；

42.  Scharr：computes the first x- or y- spatial image derivative using Scharr operator；

43.  GaussianBlur：convolves the source image with the specified Gaussian kernel；

44.  boxFilter：smoothes image using box filter；

45.  Laplacian：calculates the Laplacian of the source image by adding up the second x and y derivativescalculated using the Sobel operator；

46.  convolue：convolves an image with the kernel；

47.  bilateralFilter：applies bilateral filter to the image；

48.  copyMakeBorder：forms a border around the image；

49.  dilate：dilatesthe source image using the specified structuring element that determines theshape of a pixel neighborhood over which the maximum is taken；

50.  erode：erodes the source image using the specified structuring element that determines the shapeof a pixel neighborhood over which the minimum is taken；

51.  morphologyEx：a wrapper for erode and dilate；

52.  pyrDown(cv::pyrDown)：smoothes an image and downsamples it；

53.  pyrUp(cv::pyrUp)：upsamples an image and then smoothes it；

54.  columnSum：computes a vertical(column) sum；

55.  blendLinear：performs linear blending of two images；

56.  cornerHarris：calculate Harris corner；

57.  cornerMinEigenVal：calculate MinEigenVal；

58.  calcHist：calculates histogram of one or more arrays；

59.  remap：transforms the source image using the specified map；

60.  resize：resizes an image；

61.  warpAffine： transforms the source image using the specified matrix；

62.  warpPerspective：applies a perspective transformation to an image；

63.  cvtColor：converts image from one color space to another；

64.  threshold：applies fixed-level thresholding to a single-channel array；

65.  buildWarpPlaneMaps：builds plane warping maps；

66.  buildWarpCylindricalMaps：builds cylindrical warping maps；

67.  buildWarpSphericalMaps：builds spherical warping maps；

68.  buildWarpPerspectiveMaps(ocl::warpPerspective)：builds transformation maps for perspective transformation；

69.  buildWarpAffineMaps(ocl::warpAffine)：builds transformation maps for affine transformation；

70.  class::PyrLKOpticalFlow(cv::calcOpticalFlowPyrLK)：class used for calculating an optical flow；

71.  PyrLKOpticalFlow::sparse：calculate an optical flow for a sparse feature set；

72.  PyrLKOpticalFlow::dense：calculate dense optical flow；

73.  PyrLKOpticalFlow::releaseMemory：releases inner buffers memory；

74.  interpolateFrames：interpolateframes(images) using provided optical flow(displacement field)；

75.  class::OclCascadeClassifier：cascade classifier class used for object detection；

76.  OclCascadeClassifier::oclHaarDetectObjects：returns the detected objects by a list rectangles；

77.  struct::MatchTemplateBuf：class providing memory buffers for matchTemplate function, plus it allows to adjustsome specific parameters；

78.  matchTemplate(cv::matchTemplate)：computes aproximity map for a raster template and an image where the template is searchedfor；

79.  Canny(cv::Canny)：finds edgesin an image using the Canny algorithm；

80.  class::BruteForceMatcher_OCL_base(cv::DescriptorMatcher,cv::BFMatcher)：Brute-force descriptor matcher, for each descriptor in the firstset, this matcher finds the closest descriptor in the second set by trying eachone；

81.  BruteForceMatcher_OCL_base::match：finds the best match for each descriptor from a query set with traindescriptors；

82.  BruteForceMatcher_OCL_base::makeGpuCollection：performs a GPU collection of traindescriptors and masks in a suitable format for the matchCollection function；

83.  BruteForceMatcher_OCL_base::matchDownload：downloads matrices obtained viamatchSingle or matchCollection to vector with DMatch；

84.  BruteForceMatcher_OCL_base::matchConvert：converts matrices obtained viamatchSingle or matchCollection to vector with DMatch；

85.  BruteForceMatcher_OCL_base::knnMatch：finds the k best matches for eachdescriptor from a query set with train descriptors；

86.  BruteForceMatcher_OCL_base::knnMatchDownload：downloads matrices obtained viaknnMatchSingle or knnMatch2Collection to vector with DMatch；

87.  BruteForceMatcher_OCL_base::knnMatchConvert：converts matrices obtained viaknnMatchSingle or knnMatch2Collection to CPU vector with DMatch；

88.  BruteForceMatcher_OCL_base::radiusMatch：for each query descriptor, finds thebest matches with a distance less than a given threshold；

89.  BruteForceMatcher_OCL_base::radiusMatchDownload：downloads matrices obtained viaradiusMatchSingle or radiusMatchCollection to vector with DMatch；

90.  BruteForceMatcher_OCL_base::radiusMatchConvert：converts matrices obtained viaradiusMatchSingle or radiusMatchCollection to vector with DMatch；

91.  struct::HOGDescriptor：the class implements Histogram ofOriented Gradients object detector；

92.  HOGDescriptor::getDescriptorSize：returns the number of coefficientsrequired for the classification；

93.  HOGDescriptor::getBlockHistogramSize：returns the block histogram size；

94.  HOGDescriptor::setSVMDetector：sets coefficients for the linear SVMclassifier；

95.  HOGDescriptor::getDefaultPeopleDetector：returns coefficients of the classifiertrained for people detection(for default window size)；

96.  HOGDescriptor::getPeopleDetector48x96：returns coefficients of the classifiertrained for people detection(for 48x96 windows)；

97.  HOGDescriptor::getPeopleDetector64x128：returns coefficients of the classifiertrained for people detection(for 64x128 windows)；

98.  HOGDescriptor::detect：performs object detection without amulti-scale window；

99.  HOGDescriptor::detectMultiScale：performs object detection with amulti-scale window；

100.  HOGDescriptor::getDescriptors：returns block descriptors computed forthe whole image；