Bézier curves

Manopt.BezierSegmentType
BezierSegment

A type to capture a Bezier segment. With $n$ points, a Beziér segment of degree $n-1$ is stored. On the Euclidean manifold, this yields a polynomial of degree $n-1$.

This type is mainly used to encapsulate the points within a composite Bezier curve, which consist of an AbstractVector of BezierSegments where each of the points might be a nested array on a PowerManifold already.

Not that this can also be used to represent tangent vectors on the control points of a segment.

See also: de_casteljau.

Constructor

BezierSegment(pts::AbstractVector)

Given an abstract vector of pts generate the corresponding Bézier segment.

source
Manopt.de_casteljauMethod
de_casteljau(M::AbstractManifold, b::BezierSegment NTuple{N,P}) -> Function

return the Bézier curve $β(⋅;b_0,…,b_n): [0,1] → \mathcal M$ defined by the control points $b_0,…,b_n∈\mathcal M$, $n∈\mathbb N$, as a BezierSegment. This function implements de Casteljau's algorithm[Casteljau1959][Casteljau1963] gneralized to manifolds[PopielNoakes2007]: Let $γ_{a,b}(t)$ denote the shortest geodesic connecting $a,b∈\mathcal M$. Then the curve is defined by the recursion

\[\begin{aligned} β(t;b_0,b_1) &= \gamma_{b_0,b_1}(t)\\ β(t;b_0,…,b_n) &= \gamma_{β(t;b_0,…,b_{n-1}), β(t;b_1,…,b_n)}(t), \end{aligned}\]

and P is the type of a point on the Manifold M.

de_casteljau(M::AbstractManifold, B::AbstractVector{<:BezierSegment}) -> Function

Given a vector of Bézier segments, i.e. a vector of control points $B=\bigl( (b_{0,0},…,b_{n_0,0}),…,(b_{0,m},… b_{n_m,m}) \bigr)$, where the different segments might be of different degree(s) $n_0,…,n_m$. The resulting composite Bézier curve $c_B:[0,m] → \mathcal M$ consists of $m$ segments which are Bézier curves.

\[c_B(t) := \begin{cases} β(t; b_{0,0},…,b_{n_0,0}) & \text{ if } t ∈[0,1]\\ β(t-i; b_{0,i},…,b_{n_i,i}) & \text{ if } t∈(i,i+1], \quad i∈\{1,…,m-1\}. \end{cases}\]
de_casteljau(M::AbstractManifold, b::BezierSegment, t::Real)
de_casteljau(M::AbstractManifold, B::AbstractVector{<:BezierSegment}, t::Real)
de_casteljau(M::AbstractManifold, b::BezierSegment, T::AbstractVector) -> AbstractVector
de_casteljau(
    M::AbstractManifold,
    B::AbstractVector{<:BezierSegment},
    T::AbstractVector
) -> AbstractVector

Evaluate the Bézier curve at time t or at times t in T.

source
Manopt.get_bezier_degreeMethod
get_bezier_degree(M::AbstractManifold, b::BezierSegment)

return the degree of the Bézier curve represented by the tuple b of control points on the manifold M, i.e. the number of points minus 1.

source
Manopt.get_bezier_degreesMethod
get_bezier_degrees(M::AbstractManifold, B::AbstractVector{<:BezierSegment})

return the degrees of the components of a composite Bézier curve represented by tuples in B containing points on the manifold M.

source
Manopt.get_bezier_inner_pointsMethod
get_bezier_inner_points(M::AbstractManifold, B::AbstractVector{<:BezierSegment} )
get_bezier_inner_points(M::AbstractManifold, b::BezierSegment)

returns the inner (i.e. despite start and end) points of the segments of the composite Bézier curve specified by the control points B. For a single segment b, its inner points are returned

source
Manopt.get_bezier_junction_tangent_vectorsMethod
get_bezier_junction_tangent_vectors(M::AbstractManifold, B::AbstractVector{<:BezierSegment})
get_bezier_junction_tangent_vectors(M::AbstractManifold, b::BezierSegment)

returns the tangent vectors at start and end points of the composite Bézier curve pointing from a junction point to the first and last inner control points for each segment of the composite Bezier curve specified by the control points B, either a vector of segments of controlpoints.

source
Manopt.get_bezier_junctionsFunction
get_bezier_junctions(M::AbstractManifold, B::AbstractVector{<:BezierSegment})
get_bezier_junctions(M::AbstractManifold, b::BezierSegment)

returns the start and end point(s) of the segments of the composite Bézier curve specified by the control points B. For just one segment b, its start and end points are returned.

source
Manopt.get_bezier_pointsFunction
get_bezier_points(
    M::AbstractManifold,
    B::AbstractVector{<:BezierSegment},
    reduce::Symbol=:default
)
get_bezier_points(M::AbstractManifold, b::BezierSegment, reduce::Symbol=:default)

returns the control points of the segments of the composite Bézier curve specified by the control points B, either a vector of segments of controlpoints or a.

This method reduces the points depending on the optional reduce symbol

  • :default – no reduction is performed
  • :continuous – for a continuous function, the junction points are doubled at $b_{0,i}=b_{n_{i-1},i-1}$, so only $b_{0,i}$ is in the vector.
  • :differentiable – for a differentiable function additionally $\log_{b_{0,i}}b_{1,i} = -\log_{b_{n_{i-1},i-1}}b_{n_{i-1}-1,i-1}$ holds. hence $b_{n_{i-1}-1,i-1}$ is ommited.

If only one segment is given, all points of b – i.e. b.pts is returned.

source
Manopt.get_bezier_segmentsMethod
get_bezier_segments(M::AbstractManifold, c::AbstractArray{P}, d[, s::Symbol=:default])

returns the array of BezierSegments B of a composite Bézier curve reconstructed from an array c of points on the manifold M and an array of degrees d.

There are a few (reduced) representations that can get extended; see also get_bezier_points. For ease of the following, let $c=(c_1,…,c_k)$ and $d=(d_1,…,d_m)$, where $m$ denotes the number of components the composite Bézier curve consists of. Then

  • :default$k = m + \sum_{i=1}^m d_i$ since each component requires one point more than its degree. The points are then ordered in tuples, i.e.
    \[B = \bigl[ [c_1,…,c_{d_1+1}], (c_{d_1+2},…,c_{d_1+d_2+2}],…, [c_{k-m+1+d_m},…,c_{k}] \bigr]\]
  • :continuous$k = 1+ \sum_{i=1}{m} d_i$, since for a continuous curve start and end point of successive components are the same, so the very first start point and the end points are stored.
    \[B = \bigl[ [c_1,…,c_{d_1+1}], [c_{d_1+1},…,c_{d_1+d_2+1}],…, [c_{k-1+d_m},…,b_{k}) \bigr]\]
  • :differentiable – for a differentiable function additionally to the last explanation, also the second point of any segment was not stored except for the first segment. Hence $k = 2 - m + \sum_{i=1}{m} d_i$ and at a junction point $b_n$ with its given prior point $c_{n-1}$, i.e. this is the last inner point of a segment, the first inner point in the next segment the junction is computed as $b = \exp_{c_n}(-\log_{c_n} c_{n-1})$ such that the assumed differentiability holds
source

Literature

  • Casteljau1959

    de Casteljau, P.: Outillage methodes calcul, Enveloppe Soleau 40.040 (1959), Institute National de la Propriété Industrielle, Paris.

  • Casteljau1963

    de Casteljau, P.: Courbes et surfaces à pôles, Microfiche P 4147-1, André Citroën Automobile SA, Paris, (1963).

  • PopielNoakes2007

    Popiel, T. and Noakes, L.: Bézier curves and $C^2$ interpolation in Riemannian manifolds. Journal of Approximation Theory (2007), 148(2), pp. 111–127.- doi: 10.1016/j.jat.2007.03.002.