1 / 54

Math Foundations of CG

Math Foundations of CG. Outline. Abstract spaces: objects and operations Field of real numbers R Vector space over R Euclidean spaces Affine spaces Affine combinations Convex combinations Frames Affine maps Euclidean spaces Read: angel, Appendices B and C, Ch 4.1-4.3. Geometric ADTs.

jmorey
Télécharger la présentation

Math Foundations of CG

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Math Foundations of CG Hofstra University

  2. Outline • Abstract spaces: objects and operations • Field of real numbers R • Vector space over R • Euclidean spaces • Affine spaces • Affine combinations • Convex combinations • Frames • Affine maps • Euclidean spaces • Read: angel, Appendices B and C, Ch 4.1-4.3 Hofstra University

  3. Geometric ADTs • Scalars, Points and Vectors are members of mathematical abstract sets • Abstract spaces for representing and manipulating these sets of objects • Field -- scalars • Linear Vector Space – vectors • Euclidean Space – add concept of distance • Affine Space – adds the point Hofstra University

  4. Field • Set, S, of scalars, with two operations • Addition, +, • Multiplication, ., satisfying the following • Axioms • Unique additive and multiplicative units, 0 (zero) and 1(one), s.t., • Every scalar has an additive inverse , s.t., • Every non-zero scalar, , has a multiplicative inverse , s.t., Hofstra University

  5. Field (cont) • Axioms (cont.) • Addition and multiplication are commutative operations • Addition and multiplication are associative • Multiplication is distributive over addition • Examples: • The field of the rational numbers, Q • The field of the real numbers, R • ADTs rational number and real number • Implementation: classes rational and double Hofstra University

  6. Linear Vector Spaces (defined over a field) • S is a field of scalars • The setVof objects called vectors , is a (linear) vector space defined over Sif there are two operations • Vector-vector addition, • Scalar-vector multiplication, satisfying the following • Axioms • Unique additive unit, the zero vector, • Every vector has additive inverse Hofstra University

  7. Vector Spaces (cont.) • Axioms (cont.) • Vector-vector addition is commutative and associative • Scalar-vector multiplication is distributive • Examples • Geometric vectors over R , i.e., directed line segments in 3D Hofstra University

  8. Vector Spaces (cont.) • Examples • n-tuples of real numbers • A vector is identified with an n-tuple Hofstra University

  9. Vector Spaces (cont.) • is a linear vector space over a field is alinear combination of , if • The non-zero vectors are linearly independent, if Hofstra University

  10. Vector Spaces (cont.) • is a linear vector space over a field • The vectors are linearly dependent, if they are not linearly independent • In which case one of them can be expressed as a non-trivial linear combination of the rest. (non-trivial means that not all coefficients are 0) • Any set of vectors that includes the zero vector is linearly dependent. • Basis : a maximal linear independent set of vectors, i.e., if one more vector is added to the set it becomes linearly dependent. • Any vector in the space is represented as unique linear combination of the basis vectors • Dimension: number of vectors in the basis Hofstra University

  11. Vector Spaces (cont.) • is a n-dimensional vector space over a field , and is a basis: • Every vector is represented uniquely as a linear combination of the basis, i.e., there exist unique scalars • representation (coordinates) of in the basis • Once a basis is selected, vectors are represented as n-toulpes of coordinates Hofstra University

  12. We are concerned with 3D vector space Represent w as linear combination of three linearly independent vectors, v1, v2, v3 v = 1v1 + 2v2 + 3v3 basis components Hofstra University

  13. Vector Spaces: Changes of Basis • How do we represent a vector if we change the basis? • Suppose the {v1,v2,v3} and {u1,u2,u3} are two bases. • Basis vector in second set can be represented in terms of the first basis • Given the representation of a vector in one basis, we can change to a representation of the same vector in the other basis by a linear transformation (i.e., matrix multiplication) Hofstra University

  14. Vector Spaces: Change of Basis • Vector Spaces: Change of basis is a linear operation. Hofstra University

  15. Vector Spaces: Change of Basis Notation: for a matrix,’ denote the transpose. Let in basis v the vector w is represented by a component column matrix a, and in u basis, by a component matrix b What is the relation between the two representations a and b? Hofstra University

  16. Vector Spaces: Change of Basis Example Given a basis v, we want to change to a new basis u old new new old Let w has representation a in the old basis, v, in old Then the representation b of w in the new basis u is in new Hofstra University

  17. Vector Spaces • ADT n-D vector • ADT vector implementation: class vector • Internally work with n-tuples of real numbers • Use matrix algebra in the implementation of methods Hofstra University

  18. Euclidean Space • We add the notion of a distance and angle to a vector space by means of inner (dot) product. • is an Euclidean space, if it is vector space with dot (scalar,inner) product , i.e., for vectors u and u v is a real number, such that • Axioms Hofstra University

  19. Euclidean Space (cont.) • The length of a vector • The norm of a vector • Two non-zero vectors u and v are orthogonal if u v= 0 • The angle between two vectors is given by • Unit vector: a vector of length 1 • Normalizing a vector: Hofstra University

  20. Euclidean Space (cont.) • Orthonormal basis: a basis consisting of unit vectors which are mutually orthogonal • Projections: • If v is unit vector, the projection of u on v is u.v |u|cos = u v/|v| is the orthogonal projection of u onto v Hofstra University

  21. Euclidean Space (cont.) • Cross Product(in 3D) of two vectors u and v is a vector n = u v, • n is orthogonal to v and u, • the triple (u,v,n) is right-handed, • The length Hofstra University

  22. Euclidean Spaces (cont) • Example: orthonormal basis, and • Dot product: • Cross product: Hofstra University

  23. Euclidean Space (cont.) • We can construct orthonormal basis in 3D by using the dot and cross product • Given vector u, • Set • Calculate s.t. • Calculate • The basis is orthonormal Hofstra University

  24. Affine Spaces • Given a vector space A, an affine spaceA over the vector space has two types of objects: • points, P,Q,… • and vectors, u,v,… and is defined by the following axioms • All axioms of the vector space • Operations relating points and vectors • Point-point subtraction gives unique vector, • Point-vector addition gives unique point, Hofstra University

  25. Affine Spaces (cont) • Operations relating points and vectors • Subtraction of two points yields a vector:v = P – Q • Point-vector addition yields a point:P = v + Q All operations: • point-point subtraction, • point-vector addition, • vector-vector addition, • scalar-vector multiplication Hofstra University

  26. Affine Spaces (cont) • Axioms: • Two points define unique vector, • Point and vector define unique point, • head-to-tail axiom: given points P,R, for any other point Q, • If O is an arbitrary point, Hofstra University

  27. Line: parametric equation • A line, defied by a point P0 and a vector dconsists of all points P obtained by P () = P0 + d where  varies over all scalars. • P () is a point for any value of  • The equation is called the parametric form of the line • For non-negative values, we get a ray emanating from P0 in the direction of d Hofstra University

  28. Plane: parametric equation • A plane defined by a pointP0 and two non collinear vectors (I.e. two linearly independent vectors) u and v, consists of all points T(,): T(,) = P0 +  u +  v  v T(,)  u u P0 v Hofstra University

  29. Affine Spaces (cont) • All operations: • point-point subtraction, • point-vector addition, • vector-vector addition, • scalar-vector multiplication • Point-point addition is not defined, but addition-like combinations of points are well-defined. Hofstra University

  30. Affine Combinations of Two Points • Given two points Q and R, and two scalars where the affine combination of Q and R with coefficients is a point P denoted by • All affine combinations of two points generate the line through that points. Hofstra University

  31. Affine Combinations of Three Points • Given three points P, Q, and R, and three scalars where the affine combination of the three pointswithcoefficients is a point T, denoted • The point T is defined by • All affine combinations of three non collinear points generate the plane through that points. T R Q P Hofstra University

  32. Affine Combinations of n Points • Given an affine space A, a point P is an affine combination of , iff, there exist scalars • The affine combination is denoted by • If the vectors are coplanar, what is the set of all affine combinations of the n points? such that P = 1P1 + 2P2 + … + nPn Hofstra University

  33. Convexity • Convex set– a set in which a line segment connecting any two pints of the set is entirely in the set. • For 0   the affine combinations of points Q and R is the line segment connecting Q and R • This line segment is convex • The midpoint, =0.5 • Give the affine combination representing a point dividing the line Segment in ratio m:n, starting from Q Hofstra University

  34. Convex (affine) combinations • Convex combinations: affine combinations with positive coefficients,P = 1P1 + 2P2 + … + nPn • 1 + 2 +…+ n = 1 • i  0, i = 1,2,…,n • Convex hull of a set of points is the set of all convex combination of this points. • In particular, for any two points of the set, the line segment connecting the points is in the convex hull, thus the convex hull is a convex set. • In fact, the convex hull it is the smallest convex set that contains the original points. Hofstra University

  35. Geometric ADTs: Convexity • The convex hull could be thought of as the set of points that we form by stretching a tight-fitting surface over the given set of points – shrink wrapping the points. • It is the smallest convex object that includes the set of points Hofstra University

  36. Convex Polygons • A convex polygon is completely specified by the set of its vertices • A convex polygon: the convex hull of the vertices • Given equilateral triangle give the representation of the center of the mass Hofstra University

  37. A normal to a plane • Normaln to a plane : unit vector orthogonal to the plane • If we are given the parametric equation of the plane T(,) = P0 +  u +  v, n = u  v /| u  v| • Given a polygon, write the outward normal • Given a point P0 and a vector n , there is unique plane that goes through P0 and has normal n: it consists of all points P satisfying the normal equation of the plane Hofstra University

  38. Affine Spaces: Frames • Frame: a basis at fixed origin • Select a point O (origin) and a basis (coordinate vectors) • Any vector u can be represented as uniquely as a linear combination of the basis vectors • Any point P can be represented uniquely as • Thus, we have affine coordinates for points and for vectors • Given a frame, points and vectors can be represented uniquely by their affine coordinates Hofstra University

  39. Affine Speces: Frames (cont) • If we change frames the coordinates change. • The change of basis in a vector space is a linear transformation (represented as matrix multiplication) • The change of frame in an affine space is NOT linear transformation • We extend the affine coordiantes, by adding one more dimension. The new coordinates are called homogeneous coordinates. • The change of frame in homogeneous coordinates is a linear transformation (i.e represented as matrix multiplication) Hofstra University

  40. Affine coordinates in 3D Given a frame (P0, v1, v2, v3), a vector w and a point P can be represented uniquely by: The affine coordinates (representations) of the vector and point are Hofstra University

  41. Homogeneous Coordinates • Use four dimensional column matrices to represent both points and vectors in homogeneous coordinates • The first three components are the affine coordinates • To maintain a distinction between points and vectors we use the fourth component: for a vector it is 0 and for a point it is 1 Hofstra University

  42. From affine to Homogeneous Coordinates • Affine coordinate equations and representations: • We agree that • The homogeneous-coordinate equations and representations: affine-coordinate representations homogeneous-coordinate representation of the point and the vector Hofstra University

  43. Homogeneous Coordinates Point Vector Frame We can write the coordinate equations in matrix form. For example, We carry out operations on points and vectors using their homogeneous-coordinate representation and ordinary matrix algebra Hofstra University

  44. Homogeneous Coordinates Change of Frame two frames change of frames matrix Change of frames is a linear transformation in homogeneous coordinates. All affine transformations can be represented as matrix multiplications in homogeneous coordinates. Hofstra University

  45. Taking a Picture Hofstra University

  46. Frames In OpenGL Hofstra University

  47. OpenGL: Initial Camera Position • Primitives are subject to transformation according to model-view matrix. • Objects are modeled independently from the location of the camera. • OpenGL places the camera at the origin of the world frame pointing in the negative z direction. • The model-view matrix is the identity iff the camera frame and world frame are the same. Hofstra University

  48. Default Position Object and camera at the origin Hofstra University

  49. Separation of the two frames glMatrixMode(GL_MODELVIEW); glLoadIdentity( ); glTranslatef( 0.0, 0.0,-d); Hofstra University

  50. Frames In OpenGL • We use two frames: the camera frame and the world frame • We regard the camera frame as fixed • The model-view matrix, M, is the change of frame matrix, from world-frame coordinates to camera-frame coordinates. • Could be thought as translating the camera frame at a distance d, in direction of negative z-axis, -k, at the new location, centered at P. Hofstra University

More Related