@spec convertCorrespondencies(Evision.Mat.maybe_mat_in(), Evision.Mat.maybe_mat_in()) ::
  {Evision.Mat.t(), Evision.Mat.t()} | {:error, String.t()}

convertCorrespondencies

Positional Arguments

• cols: Evision.Mat.

  correspondence-position per line in line-bundle-space

• srcLocations: Evision.Mat.

  the source image location

Keyword Arguments

• mask: Evision.Mat.

  mask containing non-zero values for the elements to be retained

Return

• pts2d: Evision.Mat.t().

  2d points

• pts3d: Evision.Mat.t().

  3d points

Collect corresponding 2d and 3d points based on correspondencies and mask

Python prototype (for reference only):

convertCorrespondencies(cols, srcLocations[, pts2d[, pts3d[, mask]]]) -> pts2d, pts3d

@spec convertCorrespondencies(
  Evision.Mat.maybe_mat_in(),
  Evision.Mat.maybe_mat_in(),
  [{:mask, term()}] | nil
) :: {Evision.Mat.t(), Evision.Mat.t()} | {:error, String.t()}

convertCorrespondencies

Positional Arguments

• cols: Evision.Mat.

  correspondence-position per line in line-bundle-space

• srcLocations: Evision.Mat.

  the source image location

Keyword Arguments

• mask: Evision.Mat.

  mask containing non-zero values for the elements to be retained

Return

• pts2d: Evision.Mat.t().

  2d points

• pts3d: Evision.Mat.t().

  3d points

Collect corresponding 2d and 3d points based on correspondencies and mask

Python prototype (for reference only):

convertCorrespondencies(cols, srcLocations[, pts2d[, pts3d[, mask]]]) -> pts2d, pts3d

@spec drawCorrespondencies(Evision.Mat.maybe_mat_in(), Evision.Mat.maybe_mat_in()) ::
  Evision.Mat.t() | {:error, String.t()}

drawCorrespondencies

Positional Arguments

• cols: Evision.Mat.

  column coordinates in the line bundle

Keyword Arguments

• colors: Evision.Mat.

  colors for the markers. Defaults to white.

Return

• bundle: Evision.Mat.t().

  the lineBundle

Debug draw markers of matched correspondences onto a lineBundle

Python prototype (for reference only):

drawCorrespondencies(bundle, cols[, colors]) -> bundle

@spec drawCorrespondencies(
  Evision.Mat.maybe_mat_in(),
  Evision.Mat.maybe_mat_in(),
  [{:colors, term()}] | nil
) :: Evision.Mat.t() | {:error, String.t()}

drawCorrespondencies

Positional Arguments

• cols: Evision.Mat.

  column coordinates in the line bundle

Keyword Arguments

• colors: Evision.Mat.

  colors for the markers. Defaults to white.

Return

• bundle: Evision.Mat.t().

  the lineBundle

Debug draw markers of matched correspondences onto a lineBundle

Python prototype (for reference only):

drawCorrespondencies(bundle, cols[, colors]) -> bundle

@spec drawSearchLines(
  Evision.Mat.maybe_mat_in(),
  Evision.Mat.maybe_mat_in(),
  Evision.scalar()
) ::
  Evision.Mat.t() | {:error, String.t()}

drawSearchLines

Positional Arguments

• locations: Evision.Mat.

  the source locations of a line bundle

• color: Evision.scalar().

  the line color

Return

• img: Evision.Mat.t().

  the output image

Debug draw search lines onto an image

Python prototype (for reference only):

drawSearchLines(img, locations, color) -> img

@spec drawWireframe(
  Evision.Mat.maybe_mat_in(),
  Evision.Mat.maybe_mat_in(),
  Evision.Mat.maybe_mat_in(),
  Evision.scalar()
) :: Evision.Mat.t() | {:error, String.t()}

drawWireframe

Positional Arguments

• pts2d: Evision.Mat.

  the 2d points obtained by @ref projectPoints

• tris: Evision.Mat.

  triangle face connectivity

• color: Evision.scalar().

  line color

Keyword Arguments

• type: integer().

  line type. See @ref LineTypes.

• cullBackface: bool.

  enable back-face culling based on CCW order

Return

• img: Evision.Mat.t().

  the output image

Draw a wireframe of a triangle mesh

Python prototype (for reference only):

drawWireframe(img, pts2d, tris, color[, type[, cullBackface]]) -> img

@spec drawWireframe(
  Evision.Mat.maybe_mat_in(),
  Evision.Mat.maybe_mat_in(),
  Evision.Mat.maybe_mat_in(),
  Evision.scalar(),
  [cullBackface: term(), type: term()] | nil
) :: Evision.Mat.t() | {:error, String.t()}

drawWireframe

Positional Arguments

• pts2d: Evision.Mat.

  the 2d points obtained by @ref projectPoints

• tris: Evision.Mat.

  triangle face connectivity

• color: Evision.scalar().

  line color

Keyword Arguments

• type: integer().

  line type. See @ref LineTypes.

• cullBackface: bool.

  enable back-face culling based on CCW order

Return

• img: Evision.Mat.t().

  the output image

Draw a wireframe of a triangle mesh

Python prototype (for reference only):

drawWireframe(img, pts2d, tris, color[, type[, cullBackface]]) -> img

@spec extractControlPoints(
  integer(),
  integer(),
  Evision.Mat.maybe_mat_in(),
  Evision.Mat.maybe_mat_in(),
  Evision.Mat.maybe_mat_in(),
  Evision.Mat.maybe_mat_in(),
  {number(), number()},
  Evision.Mat.maybe_mat_in()
) :: {Evision.Mat.t(), Evision.Mat.t()} | {:error, String.t()}

extractControlPoints

Positional Arguments

• num: integer().

  number of control points

• len: integer().

  search radius (used to restrict the ROI)

• pts3d: Evision.Mat.

  the 3D points of the mesh

• rvec: Evision.Mat.

  rotation between mesh and camera

• tvec: Evision.Mat.

  translation between mesh and camera

• k: Evision.Mat.

  camera intrinsic

• imsize: Size.

  size of the video frame

• tris: Evision.Mat.

  triangle face connectivity

Return

• ctl2d: Evision.Mat.t().

  the 2D locations of the control points

• ctl3d: Evision.Mat.t().

  matching 3D points of the mesh

Extract control points from the projected silhouette of a mesh see @cite drummond2002real Sec 2.1, Step b

Python prototype (for reference only):

extractControlPoints(num, len, pts3d, rvec, tvec, K, imsize, tris[, ctl2d[, ctl3d]]) -> ctl2d, ctl3d

@spec extractControlPoints(
  integer(),
  integer(),
  Evision.Mat.maybe_mat_in(),
  Evision.Mat.maybe_mat_in(),
  Evision.Mat.maybe_mat_in(),
  Evision.Mat.maybe_mat_in(),
  {number(), number()},
  Evision.Mat.maybe_mat_in(),
  [{atom(), term()}, ...] | nil
) :: {Evision.Mat.t(), Evision.Mat.t()} | {:error, String.t()}

extractControlPoints

Positional Arguments

• num: integer().

  number of control points

• len: integer().

  search radius (used to restrict the ROI)

• pts3d: Evision.Mat.

  the 3D points of the mesh

• rvec: Evision.Mat.

  rotation between mesh and camera

• tvec: Evision.Mat.

  translation between mesh and camera

• k: Evision.Mat.

  camera intrinsic

• imsize: Size.

  size of the video frame

• tris: Evision.Mat.

  triangle face connectivity

Return

• ctl2d: Evision.Mat.t().

  the 2D locations of the control points

• ctl3d: Evision.Mat.t().

  matching 3D points of the mesh

Extract control points from the projected silhouette of a mesh see @cite drummond2002real Sec 2.1, Step b

Python prototype (for reference only):

extractControlPoints(num, len, pts3d, rvec, tvec, K, imsize, tris[, ctl2d[, ctl3d]]) -> ctl2d, ctl3d

@spec extractLineBundle(
  integer(),
  Evision.Mat.maybe_mat_in(),
  Evision.Mat.maybe_mat_in()
) ::
  {Evision.Mat.t(), Evision.Mat.t()} | {:error, String.t()}

extractLineBundle

Positional Arguments

• len: integer().

  the search radius. The bundle will have 2*len + 1 columns.

• ctl2d: Evision.Mat.

  the search lines will be centered at this points and orthogonal to the contour defined by them. The bundle will have as many rows.

• img: Evision.Mat.

  the image to read the pixel intensities values from

Return

• bundle: Evision.Mat.t().

  line bundle image with size ctl2d.rows() x (2 * len + 1) and the same type as @p img

• srcLocations: Evision.Mat.t().

  the source pixel locations of @p bundle in @p img as CV_16SC2

Extract the line bundle from an image

Python prototype (for reference only):

extractLineBundle(len, ctl2d, img[, bundle[, srcLocations]]) -> bundle, srcLocations

@spec extractLineBundle(
  integer(),
  Evision.Mat.maybe_mat_in(),
  Evision.Mat.maybe_mat_in(),
  [{atom(), term()}, ...] | nil
) :: {Evision.Mat.t(), Evision.Mat.t()} | {:error, String.t()}

extractLineBundle

Positional Arguments

• len: integer().

  the search radius. The bundle will have 2*len + 1 columns.

• ctl2d: Evision.Mat.

  the search lines will be centered at this points and orthogonal to the contour defined by them. The bundle will have as many rows.

• img: Evision.Mat.

  the image to read the pixel intensities values from

Return

• bundle: Evision.Mat.t().

  line bundle image with size ctl2d.rows() x (2 * len + 1) and the same type as @p img

• srcLocations: Evision.Mat.t().

  the source pixel locations of @p bundle in @p img as CV_16SC2

Extract the line bundle from an image

Python prototype (for reference only):

extractLineBundle(len, ctl2d, img[, bundle[, srcLocations]]) -> bundle, srcLocations

@spec findCorrespondencies(Keyword.t()) :: any() | {:error, String.t()}

@spec findCorrespondencies(Evision.Mat.maybe_mat_in()) ::
  {Evision.Mat.t(), Evision.Mat.t()} | {:error, String.t()}

findCorrespondencies

Positional Arguments

• bundle: Evision.Mat.

  the line bundle

Return

• cols: Evision.Mat.t().

  correspondence-position per line in line-bundle-space

• response: Evision.Mat.t().

  the sobel response for the selected point

Find corresponding image locations by searching for a maximal sobel edge along the search line (a single row in the bundle)

Python prototype (for reference only):

findCorrespondencies(bundle[, cols[, response]]) -> cols, response

@spec findCorrespondencies(Evision.Mat.maybe_mat_in(), [{atom(), term()}, ...] | nil) ::
  {Evision.Mat.t(), Evision.Mat.t()} | {:error, String.t()}

findCorrespondencies

Positional Arguments

• bundle: Evision.Mat.

  the line bundle

Return

• cols: Evision.Mat.t().

  correspondence-position per line in line-bundle-space

• response: Evision.Mat.t().

  the sobel response for the selected point

Find corresponding image locations by searching for a maximal sobel edge along the search line (a single row in the bundle)

Python prototype (for reference only):

findCorrespondencies(bundle[, cols[, response]]) -> cols, response

@spec rapid(
  Evision.Mat.maybe_mat_in(),
  integer(),
  integer(),
  Evision.Mat.maybe_mat_in(),
  Evision.Mat.maybe_mat_in(),
  Evision.Mat.maybe_mat_in(),
  Evision.Mat.maybe_mat_in(),
  Evision.Mat.maybe_mat_in()
) ::
  {number(), Evision.Mat.t(), Evision.Mat.t(), number()} | {:error, String.t()}

rapid

Positional Arguments

• img: Evision.Mat.

  the video frame

• num: integer().

  number of search lines

• len: integer().

  search line radius

• pts3d: Evision.Mat.

  the 3D points of the mesh

• tris: Evision.Mat.

  triangle face connectivity

• k: Evision.Mat.

  camera matrix

Return

• retval: float

• rvec: Evision.Mat.t().

  rotation between mesh and camera. Input values are used as an initial solution.

• tvec: Evision.Mat.t().

  translation between mesh and camera. Input values are used as an initial solution.

• rmsd: double*.

  the 2d reprojection difference

High level function to execute a single rapid @cite harris1990rapid iteration

1. @ref extractControlPoints
2. @ref extractLineBundle
3. @ref findCorrespondencies
4. @ref convertCorrespondencies
5. @ref solvePnPRefineLM @return ratio of search lines that could be extracted and matched

Python prototype (for reference only):

rapid(img, num, len, pts3d, tris, K, rvec, tvec) -> retval, rvec, tvec, rmsd