Curves¶

Implementation of various curve utilities

splipy.utils.curve.curve_length_parametrization(pts, normalize=False)[source]¶

Calculate knots corresponding to a curvelength parametrization of a set of points.

Parameters:	pts (numpy.array) – A set of points normalize (bool) – Whether to normalize the parametrization
Returns:	The parametrization
Return type:	[float]

splipy.utils.curve.get_curve_points(curve)[source]¶

Evaluate the curve in all its knots.

Parameters:	curve (splipy.Curve) – The curve
Returns:	The curve points
Type:	numpy.array

Images¶

Implementation of image based mesh generation.

splipy.utils.image.get_corners(X, L=50, R=30, D=15)[source]¶

Detects corners of traced outlines using the SAM04 algorithm.

The outline is assumed to constitute a discrete closed curve where each point is included just once. Increasing D and R will give the same number of corners or fewer corners.

Parameters:	X (numpy.array) – A traced outline forming a discrete closed curve (size n × 2) L (float) – Controls the scale at which corners are measured. R (float) – Controls how close corners can appear. D (float) – The lower bound for the corner metric. Corner candidates with lower metric than this are rejected.
Returns:	The indices of X that consitute corner points
Return type:	numpy.array

splipy.utils.image.image_convex_surface(filename)[source]¶

Generate a single B-spline surface corresponding to convex black domain of a black/white mask image. The algorithm traces the boundary and searches for 4 natural corner points. It will then generate 4 boundary curves which will be used to create the surface by Coons Patch.

Parameters:	filename (str) – Name of image file to read
Returns:	B-spline surface
Return type:	splipy.Surface

splipy.utils.image.image_curves(filename)[source]¶

Generate B-spline curves corresponding to the edges in a black/white mask image.

Parameters:	filename (str) – Name of image file to read
Returns:	All curves generated by tracing the black/white edges
Return type:	[splipy.Curve]

splipy.utils.image.image_height(filename, N=[30, 30], p=[4, 4])[source]¶

Generate a B-spline surface approximation given by the heightmap in a grayscale image.

Parameters:	filename (str) – Name of image file to read N ((int,int)) – Number of controlpoints in u-direction p ((int,int)) – Polynomial order (degree+1)
Returns:	Normalized (all values between 0 and 1) heightmap approximation
Return type:	splipy.Surface

Refinement¶

Implementation of various refinement schemes.

splipy.utils.refinement.edge_refine(obj, S, n[, direction=0])[source]¶

Refine an object towards both edges in a direction, by sampling an arctan function.

Parameters:	obj (splipy.SplineObject) – The object to refine S (float) – The slope of the arctan function. n (int) – The number of knots to insert direction – The direction to refine in

splipy.utils.refinement.geometric_refine(obj, alpha, n[, direction=0][, reverse=False])[source]¶

Refine a spline object by making a geometric distribution of element sizes.

Parameters:	obj (splipy.SplineObject) – The object to refine alpha (float) – The length ratio between two sequential knot segments n (int) – The number of knots to insert direction – The direction to refine in reverse (bool) – Set to True to refine towards the other end

splipy.utils.refinement.subdivide(objs, n)[source]¶

Subdivide a list of objects by splitting them up along existing knot lines. The resulting partition will roughly the same number of elements on all pieces. By splitting along n lines, we generate n + 1 new blocks.

The number of subdivisions can be a list of integers: one for each direction, or a single integer for uniform sudivision.

Parameters:	objs – Objects to split n (int or [int]) – Number of subdivisions to perform
Returns:	New objects
Return type:	[splipy.SplineObject]

Smoothing¶

Implementation of various smoothing operations on a per-controlpoint level.

splipy.utils.smooth.smooth(obj)[source]¶

Smooth an object by setting the interior control points to the average of itself and all neighbours (e.g. 9 for surfaces, 27 for volumes). The edges are kept unchanged, and any rational weights are kept unchanged.

Parameters:	obj (splipy.SplineObject) – The object to smooth