CHAPTER 4

Lines and planes in dimension 3

Contents
4.1 Planes in E3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
4.1.1 Describing surfaces in E3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
4.1.2 Geometric description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
4.1.3 Parametric equations - local description . . . . . . . . . . . . . . . . . . . . . . 2
4.1.4 Cartesian equations - global description . . . . . . . . . . . . . . . . . . . . . . 3
4.1.5 Normal vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
4.1.6 Relative positions of two planes . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
4.1.7 Bundles of planes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
4.2 Lines in E3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4.2.1 Describing curves in E3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4.2.2 Geometric description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4.2.3 Parametric equations - local description . . . . . . . . . . . . . . . . . . . . . . 9
4.2.4 Cartesian equations - global description . . . . . . . . . . . . . . . . . . . . . . 10
4.2.5 Relative positions of two lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.3 Relative positions of a line and a plane . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4.4 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Unless otherwise stated, any coordinate system that we fix will be a right oriented orthonormal
coordinate system (reference frame). In this document Oxyz = (O, i, j, k) is a coordinate system of E3 .
Recall that V3 is the vector space of vectors which can be represented with points in E3 . Recall also
−−→
that we have a bijective map φO : E3 → V3 given by φO (P ) = OP .

1
Geometry - first year - IE [email protected]

4.1 Planes in E3
4.1.1 Describing surfaces in E3
There are essentially three ways in which we can describe a surface in E3 :
1. Via parametrizations, where, for two intervals I1 , I2 ⊆ R, you use a map ϕ : I1 × I2 → E3 to
specify for each pair (s, t) ∈ I1 × I2 a point in E3 . For example, the sphere of radius 1 centered at
the origin is given by φ : [0, 2π) × [0, π] → E3 with φ(s, t) = (cos(s) sin(t), sin(s) sin(t), cos(t)).

2. Via global equations, where you describe the surface as the set of all points whose coordinates
satisfy a given equation. For example, the sphere of radius 1 centered at the origin is the set
{P (x, y, z) ∈ E3 : x2 + y 2 + z2 = 1}.

3. Via geometric properties, where you describe the curve as the set of all points satisfying a
geometric property. For example, the sphere of radius 1 centered at the point O is the set
{P ∈ E3 : P is at distance 1 from O}.
Computationally you will most often use an approximation of the surface that you are interested in.
The simplest form of approximation is via triangles.

4.1.2 Geometric description

Recall from Lecture 1 that we can describe planes as sets of points S in E3 such that the set of vectors
which can be represented by points in S form a 2-dimensional vector subspace of V3 , i.e. for any
A∈S
−−→
φA (S) = { AB : B ∈ S} is a 2-dimensional vector subspace of V3 .

4.1.3 Parametric equations - local description

Since φA (S) is 2-dimensional, any basis will contain two vectors. Let v, w be a basis of the vector
−−→
space φA (S). Thus, for any two points A, B ∈ S, the vector AB is a linear combination of the basis
vectors, i.e. there exist unique scalars s, t ∈ R such that
−−→
AB = sv + tw.

2
Geometry - first year - IE [email protected]

Now, if you fix A and let B vary in the plane S then s and t vary in R. Since φA is a bijection, the
plane S can be described as
−−→
 
S = B ∈ E3 : AB = sv + tw for some s, t ∈ R .

In this description the point A ∈ S is arbitrary but fixed. We sometimes refer to it as the base point.
−−→
Now, the coordinates of a point B are the components of the vector OB and, rearranging the above
equation we have
−−→ −−−→
OB = OA + sv + tw. (4.1)
So, we can describe the plane S as the set of points B in E3 which satisfy Equation (4.1) for some
s, t ∈ R. This is called the vector equation of the plane S.
• The vector equation depends on the choice of the base point A.
• The vector equation depends on the choice of the vector v and w. In analogy with the case of
the line in E2 we may call such vectors direction vectors for the plane S.
• Hence a plane does not have a unique vector equation.
• The vector equation does not depend on the coordinate system. In the above description O can
be any point in E3 .
If we write Equation (4.1) in coordinates (relative to the coordinate system Oxyz) then we obtain
        


 x = xA + svx + twx x xA  vx  wx 
y = y + sv + tw or, in matrix form y  = yA  + s vy  + t wy  (4.2)
        

 A y y        
 z = z + sv + tw z z v w

A z z A z z

where A = A(xA , yA , zA ), v = v(vx , vy , vz ), w = v(wx , wy , wz ) and for different values s and t we obtain
different points (x, y, z) in the plane S. The three equations in the system (4.2) are called parametric
equations for the plane S.

4.1.4 Cartesian equations - global description

As in the case of the line in E2 , it is possible to eliminate the parameters s, t in (4.2) to obtain
vx vy x − xA z − zA
! ! ! !
vx vz x − xA y − yA
− − = − − . (4.3)
wx wz wx wy wx wy wx wz

3
Geometry - first year - IE [email protected]

We will not give this equation a name, because it is a bit much to keep in mind, and one has to make
sense of what happens when the denominators are zero. The whole point here, is that it is a linear
equation in x, y and z and that it can be obtained by eliminating the parameters in (4.2).
There is an easier way of describing S with a linear equation. For this you can interpret (4.2) as
−−→
saying that the vector AB is linearly dependent on the vectors v and w. With this in mind, the point
B(x, y, z) lies in the plane S if and only if
 
x − xA y − yA z − zA  −−→
 vx vy vz  = 0 ⇔ [ AB , v, w] = 0 (4.4)
wx wy wz 
 

Recall that this was the coplanarity condition which we deduced when we discussed the box product.
−−→
It says that the volume of the parallelepiped spanned by the vectors AB , v, w is zero. We may refer
to this expression as the zero-volume equation of the plane S.

Proposition 4.1. Every plane in E3 can be described with a linear equation in three variables

ax + by + cz + d = 0 (4.5)

relative to a fixed coordinate system and any linear equation in two variables describes a plane rela-
tive to a fixed coordinate system if the constants a, b, c are not all zero.

• Equation (4.5) is called the Cartesian equation of the plane it describes.

• Notice that there are infinitely many Cartesian equations describing the same plane, since you
can multiply one equation by a non-zero constant.

• In a fixed coordinate system equations of a plane are the same up to multiplication by a non-
zero scalar.

If a plane is given with the Cartesian equation (4.5), you can rearrange it in the form

x y z d d d
+ + =1 where A=− ,B = − . and C=− .
A B C a b c
In this form we have the equation of the plane where we can read off the intersection points with the
coordinate axes since the line intersects Ox in (A, 0, 0), it intersects Oy in (0, B, 0) and it intersects Oz
in (0, 0, C).

4.1.5 Normal vectors

Normal vectors play an important role both in abstract geometry and in applications. Here we con-
sider normal vectors of planes in E3 .
The idea is again simple: we discussed in Lecture 2 that if we fix a vector n ∈ V3 then n⊥ is a
vector subspace of V3 and V3 = ⟨n⟩ ⊕ n⊥ . The vector subspace ⟨n⟩ is 1-dimensional (it is spanned by

4
Geometry - first year - IE [email protected]

one vector) and dim(V3 ) = 3. Therefore n⊥ must have dimension 2. So, returning to the description
of a plane as a set of points of the form
−−→
 
S = B ∈ E3 : AB = sv + tw for some s, t ∈ R ,

we may choose n perpendicular to both v and w. In other words, we may choose n such that v, w are
a basis of n⊥ . Then, we have
−−→ −−→
   
S = B ∈ E3 : n ⊥ AB = B ∈ E3 : n · AB = 0 ,

where A is some point in S. In other words, the plane S can be described as the set of point B ∈ E3
such that
−−→ −−→ −−−→
n · AB = 0 or, equivalently n · ( OB − OA ) = 0.

The vector n is perpendicular to the plane S and any non-zero vector with this property is called a
−−−→
normal vector for the plane S. If we look at this equation in coordinates, then OA = (xA , yA , zA ) = A ∈ S
and with n = n(nx , ny , nz ) we have
nx (x − xA ) + ny (y − yA ) + nz (z − zA ) = 0 (4.6)
where (x, y, z) is an arbitrary point in the plane S. This is a linear equation for S and since all other
equations of S (in the coordinate system Oxy) are obtained from (4.6) by multiplying with a non-zero
constant, we see that the coefficients of x, y and z in a Cartesian equation of S can be interpreted as
the components of a normal vector.
If a plane is given via parametric equations, with v and w as direction vectors. How can we obtain
a normal vector? You can transform the parametric equations in a Cartesian equation and read off
the coefficients of x, y, z or you recall that v × w is a vector which is perpendicular to both v and w,
thus, it is a normal vector for the plane.

4.1.6 Relative positions of two planes

• [Intersections] Assume that we have two planes
π1 : a1 x + b1 y + c1 z + d1 = 0 and π2 : a2 x + b2 y + c2 z + d2 = 0.
In order to determine if they intersect or not one has to discuss the system:
(
π1 : a1 x + b1 y + c1 z + d1 = 0
(4.7)
π2 : a2 x + b2 y + c2 z + d2 = 0
Recall Lecture 12 of your Algebra course last semester. Let A be the matrix of the system and
à the extended matrix of the system.

5
Geometry - first year - IE [email protected]

– two planes either intersect in a line, the coordinates of the points on the line will be solu-
tions to (4.7), this happens if the rank of A and the rank of à equals 2; or
– they don’t intersect and (4.7) doesn’t have solutions, in which case the planes are parallel,
this happens if the rank of A is strictly less than the rank of Ã; or
– the solution to system (4.7) depends on two parameters in which case π1 = π2 , this hap-
pens if the rank of A and the rank of à are equal to 1.

Notice that (a1 , b1 , c1 ) is a normal vector for π1 and (a2 , b2 , c2 ) is a normal vector for π2 . These are
the rows of the matrix A, so the last two cases above occur if these two vectors are proportional,
since the rank of A equals 1.

• [Angles]

For planes we have the notion of dihedral angle. Two planes π1 and π2 define two dihedral
angles which are supplementary.
These two angles can also be described with normal vectors: if n1 and n2 are normal vectors
for π1 and π2 respectively, then the two angles between π1 and π2 are congruent to ∡(n1 , n2 )
and ∡(−n1 , n2 ). So, if these vectors are known we may calculate
n1 · n2
cos ∡(n1 , n2 ) = . (4.8)
∥n1 ∥ · ∥n2 ∥

Clearly, if the two planes are parallel, then the two angles are 0° and 180°.

• [Distances] We first consider the distance from a point to a plane

Proposition 4.2. Suppose you have a plane π : ax + by + cz + d = 0 and a point P (xP , yP , zP ) in E3 .

The distance from P to π is
|axP + byP + czP + d|
d(P , π) = √ . (4.9)
a2 + b 2 + c 2

6
Geometry - first year - IE [email protected]

Now, considering the distances between two planes π1 and π2 we have the following cases:

– if π1 and π2 intersect then the distance between them is zero d(π1 , π2 ) = 0.

– if π1 and π2 do not intersect, i.e. if they are parallel, then

d(π1 , π2 ) = d(P1 , π2 ) = d(π1 , P2 )

for any point P1 ∈ π1 and any point P2 ∈ π2 .

4.1.7 Bundles of planes

Definition. Let ℓ ⊆ E3 be a line. The set Πℓ of all planes in E3 containing ℓ is called a bundle of planes
and ℓ is called the axis of the bundle Πℓ .

Proposition 4.3. If π1 : a1 x + b1 y + c1 z + d1 = 0 and π2 : a2 x + b2 y + c2 z + d2 = 0 are two distinct planes

in the bundle Πℓ , then Πℓ consists of planes having equations of the form

πλ,µ : λ(a1 x + b1 y + c1 z + d1 ) + µ(a2 x + b2 y + c2 z + d2 ) = 0.

where λ, µ ∈ R not both zero.

• Bundles of planes are useful in practice when a line ℓ is given as the intersection of two planes
and one wants to find the equation of a plane containing ℓ and satisfying some other conditions.
For example, the condition that it contains some point P which does not belong to ℓ or that it
is parallel to a given line.

7
Geometry - first year - IE [email protected]

• There is redundancy in the two parameters λ, µ, meaning that there are not two independent
parameters here. If λ , 0 then one can divide the equation of πλ,µ by λ to obtain
π1,t : (a1 x + b1 y + c1 z + d1 ) + t(a2 x + b2 y + c2 z + d2 ) = 0.
µ
where λ = t ∈ R. So π1, µ and πλ,µ are in fact the same planes.
λ

• A reduced bundle is the set of all planes Πℓ with axis ℓ from which we remove one plane, i.e. it is
Πℓ \ {π2 } for some π2 ∈ Πℓ . With the above notation and discussion, it is the set

π1,t : (a1 x + b1 y + c1 z + d1 ) + t(a2 x + b2 y + c2 z + d2 ) = 0 : t ∈ R .
The fact that we use one parameter instead of two, to describe almost all planes containing ℓ,
greatly simplifies calculations.

Definition. Let W ⊆ V3 be a vector subspace of dimension 2. The set ΠW of all planes in A hav-
ing associated vector subspace W is called an improper bundle of planes, and W is called the vector
subspace associated to the bundle ΠW .

• The connection between bundles of planes and improper bundles of planes is best understood
through projective geometry, where we can think about the improper bundle of planes as the
set of all planes intersecting in a line at infinity.

4.2 Lines in E3
4.2.1 Describing curves in E3
There are essentially three ways in which we can describe a curve in E3 :
1. Via parametrizations, where, for an interval I ⊆ R, you use a map ϕ : I → E3 to specify for
each t ∈ I a point in E3 . For example, a circle of radius 1 centered at the origin is given by
φ : [0, 2π) → E3 with φ(s, t) = (cos(t), sin(t), 0).
2. Via global equations, where you describe the curve as the set of all points whose coordinates
satisfy two given equations. For example, the above circle can be described as the set {P (x, y, z) ∈
E3 : x2 + y 2 = 1 and z = 0}.
3. Via geometric properties, where you describe the curve as the set of all points satisfying some
geometric properties. For example, the above circle can be described as the set of points {P ∈
E3 : P is at distance 1 from O and P lies in the plane Oxy}.

4.2.2 Geometric description

Recall from Lecture 1 that we can describe lines in E3 as being sets of points S such that the set of
vectors which can be represented by points in S form a 1-dimensional vector subspace of V3 , i.e. for
any A ∈ S
−−→
φA (S) = { AB : B ∈ S} is a 1-dimensional vector subspace of V3 .

8
Geometry - first year - IE [email protected]

4.2.3 Parametric equations - local description

−−→
If S is a line then for any two distinct points A, B in S the vector AB is called a direction vector of S.
Since φA (S) is 1-dimensional, all direction vectors are linearly dependent and v is a direction vector
−−→
for S if and only if it is linearly dependent on AB . So, for any direction vector v of S there is a scalar
t ∈ R such that
−−→
AB = tv.
Now, if you fix A and let B vary on the line then t varies in R. Since φA is a bijection, the line S can
be described as
−−→
 
S = B ∈ E3 : AB = tv for some t ∈ R .

In this description the point A ∈ S is arbitrary but fixed. We sometimes refer to it as the base point.
−−→
Now, the coordinates of a point B are the components of the vector OB and, rearranging the above
equation we have
−−→ −−−→
OB = OA + tv. (4.10)
So, we can describe the line S as being the set of points B in E3 which satisfy Equation (4.10) for some
t ∈ R. This is called the vector equation of the line S.

• The vector equation depends on the choice of the base point A.

• The vector equation depends on the choice of the direction vector v.

• Hence, a line does not have a unique vector equation.

• The vector equation does not depend on the coordinate system. In the above description O can
be any point in E3 .

If we write Equation (4.10) in coordinates (relative to the coordinate system Oxyz) then we obtain
      


 x = xA + tvx x xA  vx 
y = y + tv or, in matrix form y  = yA  + t vy  (4.11)
      

 A y      
 z = z + tv z z v

A z A z

where A = A(xA , yA , zA ), v = v(vx , vy , vz ) and for different values t we obtain different points (x, y, z) on
the line. The three equations in the system (4.11) are called parametric equations for the line S.

9
Geometry - first year - IE [email protected]

4.2.4 Cartesian equations - global description

It is possible to eliminate the parameter t in (4.11) in order to obtain
x − xA y − yA z − zA
= = . (4.12)
vx vy vz
We refer to the Equations (4.12) as the symmetric equations of the line S. It could happen that vx , vy
or vz are zero. In that case, translate back to the parametric equations to understand what happens.
• For lines in E3 there is no notion of ‘slope’.
• We have just described a line with two linear equations (relative to the coordinate system Oxyz).
Proposition 4.4. Every line in E3 can be described with two linear equations in three variables
(
a1 x + b1 y + c1 z + d1 = 0
(4.13)
a2 x + b2 y + c2 z + d2 = 0
relative to a fixed coordinate system and any compatible system of two linear equations of rank 2 in
three variable describes a line relative to a fixed coordinate system.
• The Equations (4.5) are called Cartesian equations of the line it describes.
• The Equations (4.5) describe a line as an intersection of two planes.

4.2.5 Relative positions of two lines

• [Intersections] Assume we have two lines
( (
a1 x + b1 y + c1 z + d1 = 0 a3 x + b3 y + c3 z + d3 = 0
ℓ1 : and ℓ2 : .
a2 x + b2 y + c2 z + d2 = 0 a4 x + b4 y + c4 z + d4 = 0
One way to determine if they intersect is to discuss the system:



 a1 x + b1 y + c1 z + d1 = 0
 a2 x + b2 y + c2 z + d2 = 0


. (4.14)
a3 x + b3 y + c3 z + d3 = 0





 a x+b y +c z+d = 0
4 4 4 4

10
Geometry - first year - IE [email protected]

As you did in Lecture 12 of your Algebra course last semester. However, it is easier to discuss
the relative positions of lines in E3 via their parametric equations:
 


 x = x1 + tvx 

 x = x2 + tux
ℓ1 :  y = y + tv und ℓ : y = y2 + tuy
 
 1 y 2  
 z = z + tv
  z = z + tu

1 z 2 z

We have the following cases

– If the direction vectors v and u are proportional then the two lines are parallel.
– If they are parallel and have a point in common then the two lines are equal.
– If they are not parallel then they are coplanar (they lie in the same plane) if
 
x1 − x2 y1 − y2 z1 − z2 
 vx vy vz  = 0.
ux uy uz 
 

in which case they intersect in exactly one point.

– If they are not parallel and they don’t intersect then we say that the two lines ℓ1 and ℓ2 are
skew relative to each other.

• [Angles]

Two lines define two angles: if v1 is a direction vector for ℓ1 and if v2 is a direction vector for ℓ2
then the two angles described by ℓ1 and ℓ2 are ∡(v1 , v2 ) and ∡(−v1 , v2 ). They are supplementary
angles so if you can measure one of them you know the other one. Recall from Lecture 2 that
v1 · v2
cos ∡(v1 , v2 ) = . (4.15)
∥v1 ∥ · ∥v2 ∥
Clearly, if the two lines are parallel, then the two angles are 0° and 180°.

11
Geometry - first year - IE [email protected]

• [Distances] We first consider the distance from a point to a lines

Proposition 4.5. Suppose you have a line



 x = xA + tvx
P (xP , yP , zP ) ∈ E3 .

ℓ1 :  y = yA + tvy and a point


 z = z + tv

A z

The distance from P to ℓ is

−−→
∥ P A × v∥
d(P , ℓ) = .
∥v∥

Now, considering the distances between two lines ℓ1 and ℓ2 . We have the following cases:

– if ℓ1 and ℓ2 intersect then the distance between them is zero d(ℓ1 , ℓ2 ) = 0.

– if ℓ1 and ℓ2 do not intersect and they are parallel, then

d(ℓ1 , ℓ2 ) = d(P1 , ℓ2 ) = d(ℓ1 , P2 )

for any point P1 ∈ ℓ1 and any point P2 ∈ ℓ2 .

– if ℓ1 and ℓ2 are skew relative to each other, there is a unique plane π1 containing ℓ1 which
is parallel to ℓ2 and there is a unique plane π2 which contains ℓ2 and is parallel to ℓ1 , thus

d(ℓ1 , ℓ2 ) = d(π1 , P2 ) = d(P1 , π2 )

for any point P1 ∈ ℓ1 and any point P2 ∈ ℓ2 .

More concretely, if the two lines are

 


 x = x1 + tvx 

 x = x2 + tux
ℓ1 :  y = y1 + tvy and ℓ2 :  y = y2 + tuy
 
 
 z = z + tv
  z = z + tu

1 z 2 z

then  
x − x1 y − y1 z − z1 
π1 :  vx vy vz  = 0
ux uy uz 
 

12
Geometry - first year - IE [email protected]

Hence, by (4.9),    
 x2 − x1 y2 − y1 z2 − z1  
  vx vy vz  
ux uy uz  
  
d(ℓ1 , ℓ2 ) = d(π1 , P2 ) =  
q  .
 a2x + a2y + a2z 
 
 

where a is the normal vector of the plane π1 corresponding to the above equation, i.e.
     
 vy vz  vz vx   vx vy 
ax =  , ay =   and az =  .
uy uz  uz ux  ux uy 

• [Common perpendicular line of two skew lines] With the notation in the previous paragraph
consider the line   

 x − x1 y − y1 z − z1 
π1′ :  vx

vy vz  = 0,





ax ay az 

  
ℓ:
  
x − x y − y z − z 
2 2 2



 π′ :  u

u u  = 0.


 2   x y z

  ax ay az 


viewed as the intersection of the two planes π1′ and π2′ . It is a calculation to check that

– ℓ ⊥ ℓ1 and ℓ ⊥ ℓ2 , and that

– ℓ ∩ ℓ1 , ∅ and ℓ ∩ ℓ2 , ∅.

4.3 Relative positions of a line and a plane

• [Intersection] Consider the plane

π : ax + by + cz + d = 0

and the line 



 x = x0 + tvx
ℓ: y = y0 + tvy


 z = z + tv .

0 z

In order to see if they intersect, we check which points in ℓ satisfy the equation of π:

a(x0 + tvx ) + b(y0 + tvy ) + c(z0 + tvz ) + d = 0 ⇔ (avx + bvy + cvz )t + ax0 + by0 + cz0 + d = 0. (4.16)

Notice that n(a, b, c) is a normal vector for π. The possibilities are

– n · v = 0 in which case the line is parallel to the plane.

– If ℓ ∥ π and Equation (4.16) is not satisfied, then the line lies outside the plane π.

13
Geometry - first year - IE [email protected]

– If ℓ ∥ π and Equation (4.16) is satisfied, then any t ∈ R is a solution. In this case the line
lies in the plane π.
– If ℓ ∦ π then Equation (4.16) has the unique solution

ax0 + by0 + cz0 + d

t0 = − .
avx + bvy + cvz

Hence, the point corresponding to the parammeter t0 is the intersection point ℓ ∩ π.

• [Angle] Let n be a normal vector for the plane π and v be a direction vector for the line ℓ. If
n · v < 0 then the angle ∡(n, v) is obtuse, so, replace n by −n or v by −v so that the angle ∡(n, v)
is acute and n · v > 0.
The line ℓ and the plane π determine two angles: 90° ± ∡(n, v) which can be calculated as usual
with the scalar product if n and v are known.

• [Distance] For the distances between a plane π and a line ℓ we have the following cases:

– if π and ℓ intersect, then the distance between them is zero d(ℓ, π) = 0.

– if π and ℓ do not intersect, i.e. if they are parallel, then

d(ℓ, π) = d(P , π)

for any point P ∈ ℓ.

14
Geometry - first year - IE [email protected]

4.4 Applications
• [Triangulation] In computer graphics it is common to describe a surface by a triangulation, i.e.
as a union of triangles:

• [Light] When modeling light on a surface one uses the normal vectors to the surface as de-
scribed in Lecture 2. If the surface is given by a triangulation, the luminosity of each triangle
can be calculated with the normal vector corresponding to that triangle, more precisely, with
the normal vector to the plane determined by the corresponding triangle. Normal vectors are
generally pre-computed, and are part of the ‘model’.

• [Ray tracing]

Ray tracing algorithms consider rays passing through the camera and each pixel of the screen.
These rays are intersected with the objects in the scene in order to decide what color to choose
for the pixels. For this, one intersects the rays with the planes of the triangles. After finding
the intersection points P one has to decide if P is inside or outside the triangle.

15