Voronoi Diagram and Delaunay Triangulation
Chih-Hung Liu
October 21
Outline
1 Voronoi Diagrams and Delaunay Triangulations Properties and Duality
2 3D geometric transformation
Voronoi Diagram
Given a setSofnpoint sites, Voronoi DiagramV(S)is a planar subdivision
1 Each region contains exactly one sitep∈Sand is denoted byVR(p,S).
2 For each pointx ∈VR(p,S),pis its closest site inS. VR(p,S)is the locus of points closer topthan any other site.
Voronoi Diagram
Given a setSofnpoint sites, Voronoi DiagramV(S)is a planar subdivision
1 Each region contains exactly one sitep∈Sand is denoted byVR(p,S).
2 For each pointx ∈VR(p,S),pis its closest site inS. VR(p,S)is the locus of points closer topthan any other site.
Voronoi Diagram
Given a setSofnpoint sites, Voronoi DiagramV(S)is a planar subdivision
1 Each region contains exactly one sitep∈Sand is denoted byVR(p,S).
2 For each pointx ∈VR(p,S),pis its closest site inS. VR(p,S)is the locus of points closer topthan any other site.
Voronoi Diagram
Given a setSofnpoint sites, Voronoi DiagramV(S)is a planar subdivision
1 Each region contains exactly one sitep∈Sand is denoted byVR(p,S).
2 For each pointx ∈VR(p,S),pis its closest site inS. VR(p,S)is the locus of points closer topthan any other site.
p
Voronoi Diagram
Given a setSofnpoint sites, Voronoi DiagramV(S)is a planar subdivision
1 Each region contains exactly one sitep∈Sand is denoted byVR(p,S).
2 For each pointx ∈VR(p,S),pis its closest site inS. VR(p,S)is the locus of points closer topthan any other site.
p
Voronoi Diagram
Given a setSofnpoint sites, Voronoi DiagramV(S)is a planar subdivision
1 Each region contains exactly one sitep∈Sand is denoted byVR(p,S).
2 For each pointx ∈VR(p,S),pis its closest site inS.
VR(p,S)is the locus of points closer topthan any other site.
p
Voronoi Diagram
Given a setSofnpoint sites, Voronoi DiagramV(S)is a planar subdivision
1 Each region contains exactly one sitep∈Sand is denoted byVR(p,S).
2 For each pointx ∈VR(p,S),pis its closest site inS.
VR(p,S)is the locus of points closer topthan any other site.
p x
Voronoi Region
BisectorB(p,q)={x ∈R2|d(x,p) =d(x,q)}.
D(p,q)={x ∈R2|d(x,p)<d(x,q)}.
Two half-planesD(p,q)andD(q,p)separated byB(p,q).
VR(p,S) = \
q∈S,q6=p
D(p,q).
p
B(p, q)
Voronoi Region
BisectorB(p,q)={x ∈R2|d(x,p) =d(x,q)}.
D(p,q)={x ∈R2|d(x,p)<d(x,q)}.
Two half-planesD(p,q)andD(q,p)separated byB(p,q).
VR(p,S) = \
q∈S,q6=p
D(p,q).
p
q
B(p, q) D(p, q)
D(q, p)
Voronoi Region
BisectorB(p,q)={x ∈R2|d(x,p) =d(x,q)}.
D(p,q)={x ∈R2|d(x,p)<d(x,q)}.
Two half-planesD(p,q)andD(q,p)separated byB(p,q).
VR(p,S) = \
q∈S,q6=p
D(p,q).
p q
Voronoi Region
BisectorB(p,q)={x ∈R2|d(x,p) =d(x,q)}.
D(p,q)={x ∈R2|d(x,p)<d(x,q)}.
Two half-planesD(p,q)andD(q,p)separated byB(p,q).
VR(p,S) = \
q∈S,q6=p
D(p,q).
p
q
D(p, q) D(q, p)
Voronoi Region
BisectorB(p,q)={x ∈R2|d(x,p) =d(x,q)}.
D(p,q)={x ∈R2|d(x,p)<d(x,q)}.
Two half-planesD(p,q)andD(q,p)separated byB(p,q).
VR(p,S) = \
q∈S,q6=p
D(p,q).
p
Voronoi Region
BisectorB(p,q)={x ∈R2|d(x,p) =d(x,q)}.
D(p,q)={x ∈R2|d(x,p)<d(x,q)}.
Two half-planesD(p,q)andD(q,p)separated byB(p,q).
VR(p,S) = \
q∈S,q6=p
D(p,q).
p
Voronoi Edge and Vertex
Voronoi Edge
Common intersection between twoadjacentVoronoi regionsVR(p,S)andVR(q,S)
A piece ofB(p,q) Voronoi Vertex
Common intersection among more than two Voronoi regionsVR(p,S),VR(q,S),VR(r,S), and so on.
Voronoi Edge and Vertex
Voronoi Edge
Common intersection between twoadjacentVoronoi regionsVR(p,S)andVR(q,S)
A piece ofB(p,q)
Voronoi Vertex
Common intersection among more than two Voronoi regionsVR(p,S),VR(q,S),VR(r,S), and so on.
p
q
Voronoi Edge and Vertex
Voronoi Edge
Common intersection between twoadjacentVoronoi regionsVR(p,S)andVR(q,S)
A piece ofB(p,q)
Voronoi Vertex
Common intersection among more than two Voronoi regionsVR(p,S),VR(q,S),VR(r,S), and so on.
p q
Voronoi Edge and Vertex
Voronoi Edge
Common intersection between twoadjacentVoronoi regionsVR(p,S)andVR(q,S)
A piece ofB(p,q) Voronoi Vertex
Common intersection among more than two Voronoi regionsVR(p,S),VR(q,S),VR(r,S), and so on.
p
q r
Growing Circle
Grow a circle from a pointx on the plane
Hit one sitep∈S→x belongs toVR(p,S)
Hit two sitesp,q∈S→x belongs to the Voronoi edge betweenVR(p,S)andVR(q,S)
Hit more than two sitesp,q,r, . . .∈S→x is the Voronoi vertex amongVR(p,S),VR(q,S),VR(r,S),. . .
Growing Circle
Grow a circle from a pointx on the plane
Hit one sitep∈S→x belongs toVR(p,S)
Hit two sitesp,q∈S→x belongs to the Voronoi edge betweenVR(p,S)andVR(q,S)
Hit more than two sitesp,q,r, . . .∈S→x is the Voronoi vertex amongVR(p,S),VR(q,S),VR(r,S),. . .
x
Growing Circle
Grow a circle from a pointx on the plane Hit one sitep∈S→x belongs toVR(p,S)
Hit two sitesp,q∈S→x belongs to the Voronoi edge betweenVR(p,S)andVR(q,S)
Hit more than two sitesp,q,r, . . .∈S→x is the Voronoi vertex amongVR(p,S),VR(q,S),VR(r,S),. . .
p
Growing Circle
Grow a circle from a pointx on the plane Hit one sitep∈S→x belongs toVR(p,S)
Hit two sitesp,q∈S→x belongs to the Voronoi edge betweenVR(p,S)andVR(q,S)
Hit more than two sitesp,q,r, . . .∈S→x is the Voronoi vertex amongVR(p,S),VR(q,S),VR(r,S),. . .
x p
Growing Circle
Grow a circle from a pointx on the plane Hit one sitep∈S→x belongs toVR(p,S)
Hit two sitesp,q∈S→x belongs to the Voronoi edge betweenVR(p,S)andVR(q,S)
Hit more than two sitesp,q,r, . . .∈S→x is the Voronoi vertex amongVR(p,S),VR(q,S),VR(r,S),. . .
Growing Circle
Grow a circle from a pointx on the plane Hit one sitep∈S→x belongs toVR(p,S)
Hit two sitesp,q∈S→x belongs to the Voronoi edge betweenVR(p,S)andVR(q,S)
Hit more than two sitesp,q,r, . . .∈S→x is the Voronoi vertex amongVR(p,S),VR(q,S),VR(r,S),. . .
x p
q
Growing Circle
Grow a circle from a pointx on the plane Hit one sitep∈S→x belongs toVR(p,S)
Hit two sitesp,q∈S→x belongs to the Voronoi edge betweenVR(p,S)andVR(q,S)
Hit more than two sitesp,q,r, . . .∈S→x is the Voronoi vertex amongVR(p,S),VR(q,S),VR(r,S),. . .
p x
Growing Circle
Grow a circle from a pointx on the plane Hit one sitep∈S→x belongs toVR(p,S)
Hit two sitesp,q∈S→x belongs to the Voronoi edge betweenVR(p,S)andVR(q,S)
Hit more than two sitesp,q,r, . . .∈S→x is the Voronoi vertex amongVR(p,S),VR(q,S),VR(r,S),. . .
x
Growing Circle
Grow a circle from a pointx on the plane Hit one sitep∈S→x belongs toVR(p,S)
Hit two sitesp,q∈S→x belongs to the Voronoi edge betweenVR(p,S)andVR(q,S)
Hit more than two sitesp,q,r, . . .∈S→x is the Voronoi vertex amongVR(p,S),VR(q,S),VR(r,S),. . .
x p
r
Growing Circle
Grow a circle from a pointx on the plane Hit one sitep∈S→x belongs toVR(p,S)
Hit two sitesp,q∈S→x belongs to the Voronoi edge betweenVR(p,S)andVR(q,S)
Hit more than two sitesp,q,r, . . .∈S→x is the Voronoi vertex amongVR(p,S),VR(q,S),VR(r,S),. . .
x p
r
q
Wavefront Model (Growth Model)
Grow circles from∀p∈S at unit speed
x ∈R2is first hit by a circle fromp→x belongs toVR(p,S) x ∈R2is first hit by two circles frompandq→x belongs to a Voronoi edge betweenVR(p,S)andVR(q,S) x ∈R2is first hit by three circles fromp,q, andr →x is a Voronoi vertex amongVR(p,S),VR(q,S)andVR(r,S)
Wavefront Model (Growth Model)
Grow circles from∀p∈S at unit speed
x ∈R2is first hit by a circle fromp→x belongs toVR(p,S) x ∈R2is first hit by two circles frompandq→x belongs to a Voronoi edge betweenVR(p,S)andVR(q,S) x ∈R2is first hit by three circles fromp,q, andr →x is a Voronoi vertex amongVR(p,S),VR(q,S)andVR(r,S)
x p
Wavefront Model (Growth Model)
Grow circles from∀p∈S at unit speed
x ∈R2is first hit by a circle fromp→x belongs toVR(p,S)
x ∈R2is first hit by two circles frompandq→x belongs to a Voronoi edge betweenVR(p,S)andVR(q,S) x ∈R2is first hit by three circles fromp,q, andr →x is a Voronoi vertex amongVR(p,S),VR(q,S)andVR(r,S)
Wavefront Model (Growth Model)
Grow circles from∀p∈S at unit speed
x ∈R2is first hit by a circle fromp→x belongs toVR(p,S)
x ∈R2is first hit by two circles frompandq→x belongs to a Voronoi edge betweenVR(p,S)andVR(q,S) x ∈R2is first hit by three circles fromp,q, andr →x is a Voronoi vertex amongVR(p,S),VR(q,S)andVR(r,S)
x p
Wavefront Model (Growth Model)
Grow circles from∀p∈S at unit speed
x ∈R2is first hit by a circle fromp→x belongs toVR(p,S) x ∈R2is first hit by two circles frompandq→x belongs to a Voronoi edge betweenVR(p,S)andVR(q,S)
x ∈R2is first hit by three circles fromp,q, andr →x is a Voronoi vertex amongVR(p,S),VR(q,S)andVR(r,S)
p q x
Wavefront Model (Growth Model)
Grow circles from∀p∈S at unit speed
x ∈R2is first hit by a circle fromp→x belongs toVR(p,S) x ∈R2is first hit by two circles frompandq→x belongs to a Voronoi edge betweenVR(p,S)andVR(q,S)
x ∈R2is first hit by three circles fromp,q, andr →x is a Voronoi vertex amongVR(p,S),VR(q,S)andVR(r,S)
p q x
Wavefront Model (Growth Model)
Grow circles from∀p∈S at unit speed
x ∈R2is first hit by a circle fromp→x belongs toVR(p,S) x ∈R2is first hit by two circles frompandq→x belongs to a Voronoi edge betweenVR(p,S)andVR(q,S) x ∈R2is first hit by three circles fromp,q, andr →x is a Voronoi vertex amongVR(p,S),VR(q,S)andVR(r,S)
p q
r x
Wavefront Model (Growth Model)
Grow circles from∀p∈S at unit speed
x ∈R2is first hit by a circle fromp→x belongs toVR(p,S) x ∈R2is first hit by two circles frompandq→x belongs to a Voronoi edge betweenVR(p,S)andVR(q,S) x ∈R2is first hit by three circles fromp,q, andr →x is a Voronoi vertex amongVR(p,S),VR(q,S)andVR(r,S)
p q
r x
Unbounded Region
VR(p,S)isunboundedif and only ifp is a vertex of the convex hull ofS.
Select a pointcin the convex hull Shoot a ray−→
cpfromc top For any pointx ∈−→
cp\cp,x belongs toVR(p,S)
−→
cpextends to the infinity.
IfSis in convex position,V(S)is a tree.
An unbounded Voronoi edge corresponds to a hull edge.
p
Unbounded Region
VR(p,S)isunboundedif and only ifp is a vertex of the convex hull ofS.
Select a pointcin the convex hull
Shoot a ray−→
cpfromc top For any pointx ∈−→
cp\cp,x belongs toVR(p,S)
−→
cpextends to the infinity.
IfSis in convex position,V(S)is a tree.
An unbounded Voronoi edge corresponds to a hull edge.
c p
Unbounded Region
VR(p,S)isunboundedif and only ifp is a vertex of the convex hull ofS.
Select a pointcin the convex hull Shoot a ray−→
cpfromc top
For any pointx ∈−→
cp\cp,x belongs toVR(p,S)
−→
cpextends to the infinity.
IfSis in convex position,V(S)is a tree.
An unbounded Voronoi edge corresponds to a hull edge.
p
Unbounded Region
VR(p,S)isunboundedif and only ifp is a vertex of the convex hull ofS.
Select a pointcin the convex hull Shoot a ray−→
cpfromc top For any pointx ∈−→
cp\cp,x belongs toVR(p,S)
−→
cpextends to the infinity.
IfSis in convex position,V(S)is a tree.
An unbounded Voronoi edge corresponds to a hull edge.
c p
x
Unbounded Region
VR(p,S)isunboundedif and only ifp is a vertex of the convex hull ofS.
Select a pointcin the convex hull Shoot a ray−→
cpfromc top For any pointx ∈−→
cp\cp,x belongs toVR(p,S)
−→
cpextends to the infinity.
IfSis in convex position,V(S)is a tree.
An unbounded Voronoi edge corresponds to a hull edge.
p x
Unbounded Region
VR(p,S)isunboundedif and only ifp is a vertex of the convex hull ofS.
Select a pointcin the convex hull Shoot a ray−→
cpfromc top For any pointx ∈−→
cp\cp,x belongs toVR(p,S)
−→
cpextends to the infinity.
IfSis in convex position,V(S)is a tree.
An unbounded Voronoi edge corresponds to a hull edge.
c p
x
Unbounded Region
VR(p,S)isunboundedif and only ifp is a vertex of the convex hull ofS.
Select a pointcin the convex hull Shoot a ray−→
cpfromc top For any pointx ∈−→
cp\cp,x belongs toVR(p,S)
−→
cpextends to the infinity.
IfSis in convex position,V(S)is a tree.
An unbounded Voronoi edge corresponds to a hull edge.
Unbounded Region
VR(p,S)isunboundedif and only ifp is a vertex of the convex hull ofS.
Select a pointcin the convex hull Shoot a ray−→
cpfromc top For any pointx ∈−→
cp\cp,x belongs toVR(p,S)
−→
cpextends to the infinity.
IfSis in convex position,V(S)is a tree.
An unbounded Voronoi edge corresponds to a hull edge.
Voronoi Diagram (Mathematic Definition)
Voronoi DiagramV(S) V(S) =R2\([
p∈S
VR(p,S)) = [
p∈S
∂VR(p,S)
∂VR(p,S)is the boundary ofVR(p,S)
∂VR(p,S)6⊂VR(p,S)
V(S)is the union of all the Voronoi edges Voronoi EdgeebetweenVR(p,S)andVR(q,S)
e=∂VR(p,S)∩∂VR(q,S)
Voronoi Vertexv amongVR(p,S),VR(q,S), andVR(r,S)
Complexity of V (S)
Theorem
V(S)hasO(n)edges and vertices. The average number of edges of a Voronoi region is less than6.
Add a large curveΓ
Γonly passes through unbounded edges ofV(S) Cut unbounded pieces outsideΓ
One additional face and several edges and vertices.
Complexity of V (S)
Theorem
V(S)hasO(n)edges and vertices. The average number of edges of a Voronoi region is less than6.
Euler’s Polyhedron Formula:v −e+f =1+c v: # of vertices,e: # of edges,f: # of faces, andc: # number of connected components.
An edge hastwoendpoints, and a vertex is incident to at leastthreeedges.
3v ≤2e→v ≤2e/3 f =n+1andc=1
v =1+c+e−f =e+1−n≤2e/3→e≤3n−3 e=v+f−1−c=v+n−1≥3v/2→v ≤2n−2
Triangulation
Definition
Given a setS of points on the plane, atriangulationis maximal collection ofnon-crossingline segments amongS.
Crossing (pq) p
q
Triangulation
Definition
Given a setS of points on the plane, atriangulationis maximal collection ofnon-crossingline segments amongS.
p
q
Triangulation
Definition
Given a setS of points on the plane, atriangulationis maximal collection ofnon-crossingline segments amongS.
Triangulation
Delaunay Edge
Definition
An edgepq is calledDelaunayif there exists a circle passing throughpandqand containingnoother point in its interior.
p
q
Delaunay Edge
Definition
An edgepq is calledDelaunayif there exists a circle passing throughpandqand containingnoother point in its interior.
pqisDelaunay p
q
Delaunay Edge
Definition
An edgepq is calledDelaunayif there exists a circle passing throughpandqand containingnoother point in its interior.
p q
Delaunay Triangulation
Definition
ADelaunay Triangulationis a triangulation whose edges are allDelaunay.
For each face, there exists a circle passing all its vertices and containing no other point.
Delaunay Triangulation
Definition
ADelaunay Triangulationis a triangulation whose edges are allDelaunay.
For each face, there exists a circle passing all its vertices and containing no other point.
Delaunay Triangulation
Definition
ADelaunay Triangulationis a triangulation whose edges are allDelaunay.
For each face, there exists a circle passing all its vertices and containing no other point.
Delaunay Triangulation
Definition
ADelaunay Triangulationis a triangulation whose edges are allDelaunay.
For each face, there exists a circle passing all its vertices and containing no other point.
General Position Assumption
1 No more thantwopoint sites arecolinear
V(S)isconnected
2 No more thanthreepoint sites arecocircular (At mostthreepoints are on the same circle)
degreeof each Voronoi vertex is exactly3.
Each face of the Delaunay triangulation is atriangle.
There is auniqueDelaunay triangulation.
General Position Assumption
1 No more thantwopoint sites arecolinear V(S)isconnected
2 No more thanthreepoint sites arecocircular (At mostthreepoints are on the same circle)
degreeof each Voronoi vertex is exactly3.
Each face of the Delaunay triangulation is atriangle.
There is auniqueDelaunay triangulation.
General Position Assumption
1 No more thantwopoint sites arecolinear V(S)isconnected
2 No more thanthreepoint sites arecocircular (At mostthreepoints are on the same circle)
degreeof each Voronoi vertex is exactly3.
Each face of the Delaunay triangulation is atriangle. There is auniqueDelaunay triangulation.
General Position Assumption
1 No more thantwopoint sites arecolinear V(S)isconnected
2 No more thanthreepoint sites arecocircular (At mostthreepoints are on the same circle) degreeof each Voronoi vertex is exactly3.
Each face of the Delaunay triangulation is atriangle. There is auniqueDelaunay triangulation.
General Position Assumption
1 No more thantwopoint sites arecolinear V(S)isconnected
2 No more thanthreepoint sites arecocircular (At mostthreepoints are on the same circle) degreeof each Voronoi vertex is exactly3.
Each face of the Delaunay triangulation is atriangle.
There is auniqueDelaunay triangulation.
General Position Assumption
1 No more thantwopoint sites arecolinear V(S)isconnected
2 No more thanthreepoint sites arecocircular (At mostthreepoints are on the same circle) degreeof each Voronoi vertex is exactly3.
Each face of the Delaunay triangulation is atriangle.
There is auniqueDelaunay triangulation.
Duality
Theorem
Under the general position assumption, the Delaunay triangulation isa dual graphof the Voronoi diagram.
A sitep↔a Voronoi regionVR(p,S)
A Delaunay edgepq↔a Voronoi edge betweenVR(p,S) andVR(q,S)
A Delaunay triangle∆pqr ↔a Voronoi vertex among VR(p,S),VR(q,S)andVR(r,S)
Duality
Theorem
Under the general position assumption, the Delaunay triangulation isa dual graphof the Voronoi diagram.
A sitep↔a Voronoi regionVR(p,S)
A Delaunay edgepq↔a Voronoi edge betweenVR(p,S) andVR(q,S)
A Delaunay triangle∆pqr ↔a Voronoi vertex among VR(p,S),VR(q,S)andVR(r,S)
p
q
p
x x
Duality
Theorem
Under the general position assumption, the Delaunay triangulation isa dual graphof the Voronoi diagram.
A sitep↔a Voronoi regionVR(p,S)
A Delaunay edgepq↔a Voronoi edge betweenVR(p,S) andVR(q,S)
A Delaunay triangle∆pqr ↔a Voronoi vertex among VR(p,S),VR(q,S)andVR(r,S)
p
q q
r
p
r
x x
Geometric Transformation from 2D to 3D
A paraboloidP ={(x1,x2,x3)|x12+x22=x3}in 3D For a pointx = (x1,x2)in 2D,x0= (x1,x2,x12+x22)is its lifted image in 3D
x0←vertical projection fromx toP For a setAof points in 2D, its lifted image A0={x0= (x1,x2,x12+x22)|x = (x1,x2)∈A}
E Z
C’
p p’
q q’
r r’
x3=x12+x22
Circle in 2D ↔ Planar Curve in P
Lemma
LetCbe a circle in the plane. ThenC0 is a planar curve on the paraboloidP
C is given byr2= (x1−c1)2+ (x2−c2)2 r2=x12+x22−2x1c1−2x2c2+c12+c22 C0 satisfiesx12+x22=x3
Substitutingx12+x22byx3, we obtain a planeE x3−2x1c1−2x2c2+c12+c22−r2=0 C0 =P∩E
Intersection betweenE andP is a planar curve
Lower Convex Hull
S0 onP →S0in convex position
Each point ofS0 is a vertex ofconv(S0)
Lower convex hulllconv(S0)ofS0is the part ofconv(S0) visible fromx3=−∞
Duality between DT(S ) and lconv (S
0) (1)
Theorem
The Delaunay triangulationDT(S)equals to the vertical projection onto thex1x2-planeof the lower convex hull ofS0
p,q,r ∈S.C: circumcircle ofp,q,r C0 lies on a planeE defined byp0,q0,r0 a pointx insideC ↔lifted imagex0belowE
E Z
C’
C p p’
q q’
r r’
x3=x12+x22
Duality between DT(S ) and lconv (S
0) (2)
Theorem
The Delaunay triangulationDT(S)equals to the vertical projection onto thex1x2-planeof the lower convex hull ofS0
p,q,r defines a triangle ofDT(S)
↔no point ofSinCdefined byp,q,r
↔no point ofS0 belowE defined byp0,q0,r0
↔p0,q0,r0 defines a facet oflconv(S0)
Computing a convex hull in 3D takesO(nlogn)time
Another Viewpoint of paraboloid
For eachs= (s1,s2)∈S, a paraboloid
Ps={(x1,x2,x3)|x3= (x1−s1)2+ (x2−s2)2}
For eachx = (σ1, σ2)inx1x2plane, vertical distance fromx toPsisd(x,s)2
Opaqueand of pairwisedifferentcolors Looking fromx3=−∞upward→V(S)
Vertical fromx upward first hitsPs →x ∈VR(p,S) Ps∩Pt →B(s,t)
Lower envelope ofS
s∈SPs→V(S)
Wavefront model revisited
Ps ={(x1,x2,x3)|x3=f((x1−s1)2+ (x2−s2)2)}
f is astrictly increasing function Lower envelope ofS
s∈SPs→V(S)
f(x) =√ x =p
(x1−s1)2+ (x2−s2)2
Cones of slope45◦ with apices at sitess∈S
Expanding circlesCsfrom sitess∈Sat equalunitspeed timet = radiusr
r2= (x1−s1)2+ (x2−s2)2 x3=p
(x1−s1)2+ (x2−s2)2=radius=time
x first hit byCs↔upward vertical projection fromx first hit Ps
Wavefront model revisited
Ps ={(x1,x2,x3)|x3=f((x1−s1)2+ (x2−s2)2)}
f is astrictly increasing function Lower envelope ofS
s∈SPs→V(S) f(x) =√
x =p
(x1−s1)2+ (x2−s2)2
Cones of slope45◦with apices at sitess∈S
Expanding circlesCsfrom sitess∈Sat equalunitspeed timet = radiusr
r2= (x1−s1)2+ (x2−s2)2 x3=p
(x1−s1)2+ (x2−s2)2=radius=time
x first hit byCs↔upward vertical projection fromx first hit Ps
Wavefront model revisited
Ps ={(x1,x2,x3)|x3=f((x1−s1)2+ (x2−s2)2)}
f is astrictly increasing function Lower envelope ofS
s∈SPs→V(S) f(x) =√
x =p
(x1−s1)2+ (x2−s2)2
Cones of slope45◦with apices at sitess∈S
Expanding circlesCsfrom sitess∈Sat equalunitspeed timet = radiusr
r2= (x1−s1)2+ (x2−s2)2 x3=p
(x1−s1)2+ (x2−s2)2=radius=time
x first hit byCs↔upward vertical projection fromx first hit Ps
Wavefront model revisited
Ps ={(x1,x2,x3)|x3=f((x1−s1)2+ (x2−s2)2)}
f is astrictly increasing function Lower envelope ofS
s∈SPs→V(S) f(x) =√
x =p
(x1−s1)2+ (x2−s2)2
Cones of slope45◦with apices at sitess∈S
Expanding circlesCsfrom sitess∈Sat equalunitspeed timet = radiusr
r2= (x1−s1)2+ (x2−s2)2 x3=p
(x1−s1)2+ (x2−s2)2=radius=time
x first hit byCs↔upward vertical projection fromx first hit Ps
Wavefront model revisited
Ps ={(x1,x2,x3)|x3=f((x1−s1)2+ (x2−s2)2)}
f is astrictly increasing function Lower envelope ofS
s∈SPs→V(S) f(x) =√
x =p
(x1−s1)2+ (x2−s2)2
Cones of slope45◦with apices at sitess∈S
Expanding circlesCsfrom sitess∈Sat equalunitspeed timet = radiusr
r2= (x1−s1)2+ (x2−s2)2 x3=p
(x1−s1)2+ (x2−s2)2=radius=time
x first hit byCs↔upward vertical projection fromx first hit P