CSE 167: Lecture 13: Bézier Curves. Jürgen P. Schulze, Ph.D. University of California, San Diego Fall Quarter 2011

Transcription

1 CSE 167: Introduction to Computer Graphics Lecture 13: Bézier Curves Jürgen P. Schulze, Ph.D. University of California, San Diego Fall Quarter 2011

2 Announcements Homework project #6 due Friday, Nov 18 at 1:30pm To be presented in lab 260 No grading this Friday, Nov 11 (Veterans Day) Last day for late grading project 5: Thursday, Nov 10 Final project presentations date+time confirmed: December 2 nd, 1-3pm, CSE

3 Lecture Overview Polynomial Curves cont d Bézier Curves Introduction Drawing Bézier curves Piecewise Bézier curves 3

4 Polynomial Curves Quadratic: (2 nd order) Cubic: (3 rd order) We usually define the curve for 0 t 1 4

5 Control Points Polynomial coefficients a, b, c, d can be interpreted as control points Remember: a, b, c, d have x,y,z components each Unfortunately, they do not intuitively describe the shape of the curve Goal: intuitive control points 5

6 Control Points How many control points? Two points define a line (1 st order) Three points define a quadratic curve (2 nd order) Four points define a cubic curve (3 rd order) k+1 points define a k-order curve Let s start with a line 6

7 First Order Curve Based on linear interpolation (LERP) Weighted average between two values Value could be a number, vector, color, Interpolate between points p 0 and p 1 with parameter t Defines a curve that is straight (first-order spline) t=0 corresponds to p 0 t=1 corresponds to p 1 t=0.5 corresponds to midpoint 7 p 0. 0<t<1 t=0. p 1 t=1 x(t) = Lerp( t, p 0, p 1 )= ( 1 t)p 0 + t p 1

8 Linear Interpolation Three equivalent ways to write it Expose different properties 1. Regroup for points p weighted sum of control points 2. Regroup for t polynomial in t 3. Matrix form 8

9 Weighted Average x(t) = (1 t)p 0 + (t)p 1 = B 0 (t) p 0 + B 1 (t)p 1, where B 0 (t) = 1 t and B 1 (t) = t Weights are a function of t Sum is always 1, for any value of t Also known as blending functions 9

10 Linear Polynomial x(t) = (p p 0 ) t + p vector point Curve is based at point p 0 Add the vector, scaled by t p 0..5(p 1 -p 0 ). p 1 -p 0. 10

11 Matrix Form Geometry matrix Geometric basis Polynomial basis In components 11

12 Tangent For a straight line, the tangent is constant Weighted average Polynomial Matrix form 12

13 Lecture Overview Polynomial Curves cont d Bézier Curves Introduction Drawing Bézier curves Piecewise Bézier curves 13

14 Bézier Curves Are a higher order extension of linear interpolation p 1 p 1 p 1 p 2 p 1 p 3 p 0 p 0 p 0 p 2 Linear Quadratic Cubic 14

15 Bézier Curves Give intuitive control over curve with control points Endpoints are interpolated, intermediate points are approximated Convex Hull property Variation-Diminishing property Many demo applets online, for example: Demo: ezier/bezier.html 15

16 Cubic Bézier Curve Most common case Defined by four control points: Two interpolated endpoints (points are on the curve) Two points control the tangents at the endpoints Points x on curve defined as function of parameter t p 0 x(t) p 1 p 2 16 p 3

17 Algorithmic Construction Algorithmic construction De Casteljau algorithm, developed at Citroen in 1959, named after its inventor Paul de Casteljau (pronounced Cast-all- Joe ) Developed independently from Bézier s work: Bézier created the formulation using blending functions, Casteljau devised the recursive interpolation algorithm 17

18 De Casteljau Algorithm A recursive series of linear interpolations Works for any order Bezier function, not only cubic Not very efficient to evaluate Other forms more commonly used But: Gives intuition about the geometry Useful for subdivision 18

19 De Casteljau Algorithm Given: Four control points A value of t (here t 0.25) p 1 p 0 p 2 p 3 19

20 De Casteljau Algorithm p 1 q 1 q 0 (t) = Lerp( t,p 0,p ) 1 p 0 q 0 ( ) q 1 (t) = Lerp t,p 1,p 2 p 2 q 2 (t) = Lerp( t,p 2,p ) 3 q 2 p 3 20

21 De Casteljau Algorithm q 1 q r 0 0 r 1 ( ) r 0 (t) = Lerp t,q 0 (t),q 1 (t) r 1 (t) = Lerp( t,q 1 (t),q 2 (t)) q 2 21

22 De Casteljau Algorithm r 0 x r 1 x(t) = Lerp ( t,r 0 (t),r 1 (t)) 22

23 De Casteljau Algorithm p 1 p 0 x p 2 Applets Demo: p 3

24 Recursive Linear Interpolation x = Lerp t,r 0,r 1 ( ) r = Lerp ( t,q,q ) ( ) r 1 = Lerp t,q 1,q 2 q 0 = Lerp( t,p 0,p ) 1 q 1 = Lerp( t,p 1,p ) 2 q 2 = Lerp( t,p 2,p ) 3 p 0 p 1 p 2 p 3 p 1 q 0 r 0 p 2 x q 1 r 1 p 3 q 2 p 4 24

25 Expand the LERPs q 0 (t) = Lerp( t,p 0,p 1 )= ( 1 t)p 0 + tp 1 q 1 (t) = Lerp( t,p 1,p 2 )= ( 1 t)p 1 + tp 2 q 2 (t) = Lerp( t,p 2,p 3 )= ( 1 t)p 2 + tp 3 r 0 (t) = Lerp ( t,q 0 (t),q 1 (t) )= 1 t r 1 (t) = Lerp( t,q 1 (t),q 2 (t))= 1 t ( ) ( )p 0 + tp 1 )+ t ( 1 t )p 1 + tp 2 ( )p 1 + tp 2 )+ t ( 1 t)p 2 + tp 3 ( ) 1 t ( ) 1 t x(t) = Lerp t,r 0 (t),r 1 (t) = ( 1 t) ( 1 t) ( 1 t)p 0 + tp 1 )+ t ( 1 t)p 1 + tp 2 +t ( 1 t) ( 1 t)p 1 + tp 2 )+ t ( 1 t)p 2 + tp 3 25 ( ) ( ) ( ) ( ) ( ) ( )

26 Weighted Average of Control Points Regroup for p: x(t) = 1 t ( )p 0 + tp 1 )+ t ( 1 t)p 1 + tp 2 ( )p 1 + tp 2 )+ t( 1 t)p 2 + tp 3 ) ( ) ( 1 t) 1 t +t 1 t ( ) ( ) ( ) 1 t ( ) x(t) = 1 t ( ) 3 p 0 + 3( 1 t) 2 tp t ( )t 2 p 2 + t 3 p 3 B 0 (t ) B (t ) x(t) = t 3 + 3t 2 3t + 1 p 0 + 3t 3 6t 2 + 3t ( ) ( ) + 3t 3 + 3t p + t 3 2 { p 3 B 2 (t ) ( ) B 3 (t ) ( ) p 1 26

27 Cubic Bernstein Polynomials x(t) = B 0 ( t)p 0 + B 1 ( t)p 1 + B 2 ( t)p 2 + B 3 ( t)p 3 The cubic Bernstein polynomials : B 0 ( t)= t 3 + 3t 2 3t + 1 B 1 ( t)= 3t 3 6t 2 + 3t B 2 ( t)= 3t 3 + 3t 2 ( t)= t 3 B 3 B i (t) = 1 Weights B i (t) add up to 1 for any t 27

28 General Bernstein Polynomials B 1 0 B 1 1 ( t)= t + 1 B 2 0 ( t)= t 2 2t + 1 B 3 0 ( t)= t 3 + 3t 2 3t + 1 ( t)= t B 2 1 ( t)= 2t 2 + 2t B 3 1 ( t)= 3t 3 6t 2 + 3t B 2 2 ( t)= t 2 B 3 2 ( t)= 3t 3 + 3t 2 B 3 3 ( t)= t 3 28 B i n ( t)= n i ( 1 t)n i t B n i ( t) = 1 n ( ) i i = n! i! ( n i)! n! = factorial of n (n+1)! = n! x (n+1)

29 General Bézier Curves nth-order Bernstein polynomials form nth-order Bézier curves B i n ( t)= n ( 1 t)n i t i n i=0 x( t)= B n i ( t)p i ( ) i 29

30 Bézier Curve Properties Overview: Convex Hull property Variation Diminishing property Affine Invariance 30

31 Definitions Convex hull of a set of points: Polyhedral volume created such that all lines connecting any two points lie completely inside it (or on its boundary) Convex combination of a set of points: Weighted average of the points, where all weights between 0 and 1, sum up to 1 Any convex combination always of a set of points lies within the convex hull 31

32 Convex Hull Property A Bézier curve is a convex combination of the control points (by definition, see Bernstein polynomials) Bézier curve is always inside the convex hull Makes curve predictable Allows culling, intersection testing, adaptive tessellation Demo: p 1 p 3 p 0 p 2 32

33 Variation Diminishing Property If the curve is in a plane, this means no straight line intersects a Bézier curve more times than it intersects the curve's control polyline Curve is not more wiggly than control polyline 33

34 Affine Invariance Transforming Bézier curves Two ways to transform: Transform the control points, then compute resulting spline points Compute spline points then transform them Either way, we get the same points Curve is defined via affine combination of points Invariant under affine transformations Convex hull property remains true 34

35 Cubic Polynomial Form Start with Bernstein form: x(t) = ( t 3 + 3t 2 3t + 1)p 0 + ( 3t 3 6t 2 + 3t)p 1 + ( 3t 3 + 3t 2 )p 2 + ( t 3 )p 3 Regroup into coefficients of t : x(t) = ( p 0 + 3p 1 3p 2 + p 3 )t 3 + ( 3p 0 6p 1 + 3p 2 )t 2 + ( 3p 0 + 3p 1 )t + ( p 0 )1 x(t) = at 3 + bt 2 + ct + d ( ) ( ) ( ) ( ) a = p 0 + 3p 1 3p 2 + p 3 b = 3p 0 6p 1 + 3p 2 c = 3p 0 + 3p 1 d = p 0 Good for fast evaluation Precompute constant coefficients (a,b,c,d) Not much geometric intuition 35

36 Cubic Matrix Form x(t) = a r t 3 r b c r t 2 d t 1 r a = p 0 + 3p 1 3p 2 + p 3 r b = ( 3p 0 6p 1 + 3p 2 ) r c = ( 3p 0 + 3p 1 ) d = p 0 ( ) ( ) t x(t) = [ p 0 p 1 p 2 p 3 ] t t { T G Bez B Bez Other types of cubic splines use different basis matrices B Bez 36

37 Cubic Matrix Form In 3D: 3 parallel equations for x, y and z: 37 [ ] x x (t) = p 0 x p 1x p 2 x p 3x x y (t) = p 0 y p 1y p 2 y p 3y x z (t) = p 0z p 1z p 2z p 3z t t t t t t t t t

38 Matrix Form Bundle into a single matrix x(t) = t 3 p 0 x p 1x p 2 x p 3x p 0 y p 1y p 2 y p 3y t t p 0z p 1z p 2z p 3z x(t) = G Bez B Bez T x(t) = C T Efficient evaluation Pre-compute C Take advantage of existing 4x4 matrix hardware support 38

39 Lecture Overview Polynomial Curves cont d Bézier Curves Introduction Drawing Bézier curves Piecewise Bézier curves 39

40 Drawing Bézier Curves Draw line segments or individual pixels Approximate the curve as a series of line segments (tessellation) Uniform sampling Adaptive sampling Recursive subdivision 40

41 Uniform Sampling Approximate curve with N straight segments N chosen in advance Evaluate x i = x( t i ) where t i = i N for i = 0, 1,K, N Connect the points with lines Too few points? Poor approximation Curve is faceted Too many points? x i = a r i3 N + b r i2 3 N + c r i 2 N + d Slow to draw too many line segments Segments may draw on top of each other x(t) x 0 x 1 x 2 x 3 x 4 41

42 Adaptive Sampling Use only as many line segments as you need Fewer segments where curve is mostly flat More segments where curve bends Segments never smaller than a pixel x(t) 42

43 Recursive Subdivision Any cubic curve segment can be expressed as a Bézier curve Any piece of a cubic curve is itself a cubic curve Therefore: Any Bézier curve can be broken down into smaller Bézier curves 43

44 De Casteljau Subdivision p 1 p 0 q r 0 0 r x 1 p 2 De Casteljau construction points are the control points of two Bézier sub-segments q 2 p 3 44

45 Adaptive Subdivision Algorithm Use De Casteljau construction to split Bézier segment For each half If flat enough : draw line segment Else: recurse Curve is flat enough if hull is flat enough Test how far the approximating control points are from a straight segment If less than one pixel, the hull is flat enough 45

46 Drawing Bézier Curves With OpenGL Indirect OpenGL support for drawing curves: Define evaluator map (glmap) Draw line strip by evaluating map (glevalcoord) Optimize by pre-computing coordinate grid (glmapgrid and glevalmesh) More details about OpenGL implementation: /opengl_nurbs.pdf 46

47 Lecture Overview Polynomial Curves cont d Bézier Curves Introduction Drawing Bézier curves Piecewise Bézier curves 47

48 More Control Points Cubic Bézier curve limited to 4 control points Cubic curve can only have one inflection (point where curve changes direction of bending) Need more control points for more complex curves k-1 order Bézier curve with k control points Hard to control and hard to work with Intermediate points don t have obvious effect on shape Changing any control point changes the whole curve Want local support: each control point only influences nearby portion of curve 48

49 Piecewise Curves Sequence of simple (low-order) curves, end-to-end Known as a piecewise polynomial curve Sequence of line segments Piecewise linear curve Sequence of cubic curve segments Piecewise cubic curve (here piecewise Bézier) 49

50 Mathematical Continuity C 0 continuity: Curve segments are connected C 1 continuity: C 0 & 1st-order derivatives agree at joints Tangents/normals are C 0 continuous Important for smooth shading C 2 continuity: C 1 & 2nd-order derivatives agree at joints Tangents/normals are C 1 continuous Important for high quality reflections

51 Global Parameterization Given N curve segments x 0 (t), x 1 (t),, x N-1 (t) Each is parameterized for t from 0 to 1 Define a piecewise curve Global parameter u from 0 to N x 0 (u), 0 u 1 x(u) = x 1 (u 1), 1 u 2 M M x N 1 (u ( N 1)), N 1 u N x(u) = x i (u i), where i = u (and x(n) = x N 1 (1)) Alternate: solution u also goes from 0 to 1 x(u) = x i (Nu i), where i = Nu 51

52 Piecewise-Linear Curve Given N+1 points p 0, p 1,, p N Define curve x(u) = Lerp(u i,p i,p i+1 ), i u i + 1 = (1 u + i)p i + (u i)p i+1, i = u p 1 x(1.5) x(2.9) p 3 p p 4 p 6 2 x(5.25) p p 5 0 N+1 points define N linear segments x(i)=p i C 0 continuous by construction C 1 at p i when p i -p i-1 = p i+1 -p i 52

53 Piecewise Bézier curve Given 3N + 1 points p 0,p 1,K,p 3N Define N Bézier segments: x 0 (t) = B 0 (t)p 0 + B 1 (t)p 1 + B 2 (t)p 2 + B 3 (t)p 3 x 1 (t) = B 0 (t)p 3 + B 1 (t)p 4 + B 2 (t)p 5 + B 3 (t)p 6 M x N 1 (t) = B 0 (t)p 3N 3 + B 1 (t)p 3N 2 + B 2 (t)p 3N 1 + B 3 (t)p 3N p 7 p 8 p 6 x 2 (t) p 9 p 0 p p 2 1 x 0 (t) p 3 x 1 (t) x 3 (t) p 10 p 11 p 12 p 4 p 5 53

54 Piecewise Bézier Curve Parameter in 0<=u<=3N x(u) = x 0 ( 1 u), 0 u 3 3 x 1 ( 1 u 1), 3 u 6 3 M M x N 1 ( 1 3 u (N 1)), 3N 3 u 3N 1 x(u) = x i ( u i 3 ), where i = 1 u 3 x(8.75) x 0 (t) x 1 (t) x 2 (t) x 3 (t) u=0 x(3.5) u=12 54

55 Piecewise Bézier Curve 3N+1 points define N Bézier segments x(3i)=p 3i C 0 continuous by construction C 1 continuous at p 3i when p 3i - p 3i-1 = p 3i+1 - p 3i C 2 is harder to achieve p 4 p 1 p 2 P 3 p 6 p 1 p 2 P 3 p 6 p 4 p 5 p 0 p 5 p 0 C 1 discontinuous C 1 continuous 55

56 Piecewise Bézier Curves Used often in 2D drawing programs Inconveniences Must have 4 or 7 or 10 or 13 or (1 plus a multiple of 3) control points Some points interpolate, others approximate Need to impose constraints on control points to obtain C 1 continuity C 2 continuity more difficult Solutions User interface using Bézier handles Generalization to B-splines or NURBS 56

57 Bézier Handles Segment end points (interpolating) presented as curve control points Midpoints (approximating points) presented as handles Can have option to enforce C 1 continuity Adobe Illustrator 57

58 Next Lecture Parametric surfaces 58