300 likes | 410 Vues
Footpoint distance as a measure of distance computation between curves and surfaces. Bharath Ram Sundar *, Abhijit Chunduru *, Rajat Tiwari *, Ashish Gupta^ and Ramanathan Muthuganapathy *. *Department of Engineering Design Indian Institute of Technology Madras Chennai, India
E N D
Footpoint distance as a measure of distance computation between curves and surfaces Bharath Ram Sundar*, AbhijitChunduru*, RajatTiwari*, AshishGupta^ and RamanathanMuthuganapathy*. *Department of Engineering Design Indian Institute of Technology Madras Chennai, India ^Renishaw, Pune India (Formerly worked in India Science lab, General Motors, India)
Overview • Introduction • Statement • Curve-curve case • Distance query- Surface-Surface • Distance query- Curve-Surface
Introduction • Most of CAD design requirements can be modeled as geometric queries, such as distance to edge, planarity, gap, interference and parallelism. • Typically done in discrete domain, thus there is need to solve in continuous domain. • Should be scalable efficiently for a larger domain.
Motivation • Commercial CAD packages offer elementary computations, difficult to scale and generally discretely computed.
Disadvantages with discrete computation • Approximation- Queries made are approximate as faceted models are approximate representation of geometry. • Computational complexity-Computational expense increases with densely faceted model. • Result remapping- Mapping back to original geometry further adds to approximation.
Problem Statement • Surface-Surface • Given two freeform surfaces, compute regions on each surface, such that, for any point (P) in a region on one surface there lies a corresponding point (P’) on the other surface at a distance less than a threshold value.
Curve-Surface • Given a freeform curve and a set of freeform surfaces, compute segments of the curve where the minimum distance between the curve and any of the surfaces is more than a threshold value.
Existing Distance functions • Typical minimum distance computation is performed. • Hausdorff distance.
Contributions • To the best of our knowledge, no work seems to exist that compute corresponding patches of curves/surfaces satisfying above or below a certain distance value, which is the focus of this work. Our major contributions are: • Footpoint distance measure has been proposed as a measure for distance computation. • Established points of correspondence through footpoints was explored in the case of curve-curve case and found to be an useful tool. • Corresponding surface patches for the surface surface case are identified using footpoint distance. Alpha shape has been used to detect boundaries including island regions. • A lower-envelope based approach has been proposed and demonstrated for the distance query between a curve and a set of surfaces.
Curve-curve • It is desired to find the exact distance bounds for curves C1(t) and C2(r) that correspond. • Let d1 be Minimum of the antipodal distances.Let d2be the subsequent minima. • Distance for the shown segments of is bound by the distances d1 and d2.
Surface-Surface Let S1(u1, v1) and S2(u2, v2) be the surfaces and D1(u1, v1, u2, v2) be the distance function. The basic partial differential equations for extremum are • Symbolic representation of bisector surface is possible for curve-curve case. • Such a representation for the bisector of a pair of surfaces and subsequently for D1(u1, v1, u2, v2) has not been shown to be possible yet. • Using antipodal points as the start looks infeasible and this motivated us to directly work on the footpoint distance, given a query distance.
Surface-Surface Distance query • Footpoint distance and α-shape • Solving distance query • Boundary detection using α-shape • Boundary identification for islands
Boundary detection using α-shape Patches in the form of point sets on both surfaces for Dq =0.8
Boundary detection using α-shape Points Sets in parametric space Dq =0.8
α-shape α-shape for S1 α-shape for S2 α-shapes for points in parametric space for Dq =0.8
Boundary identification for islands Island boundaries identified in parametric space. Regions on the surfaces for the identified boundaries.
Curve and set of free-form Surfaces • We initially find all the bisector points B between a curve C(t) and a surface S = S(u,v), which can be identified by solving the following equations
Footpoints Lines joining minimum distance footpoints.
Min footpoint distance function • For a point on the curve, there are several footpoints on the surface. • We take the minimum distance footpoint (MinF)..
Curve-Surface Lower envelope • Given a distance query valuDq, Lower envelope technique is then computed about Dq .
Conclusions • Algorithms for computing distance between curves and surfaces that satisfy a distance input value has been proposed and implemented. • Footpoint distance has been shown to be an appropriate distance measure for the intended problems. • Implementation Results have been provided.