VISIBLE SURFACE

DETECTION
The problem of Visibility – Occlusion.
Problem Definition:
Given a set of 3
3-D
D surfaces to be
projected onto a 2-D screen, obtain the
nearest surface corresponding to any point
on the screen.

Two types of methods used:

• Object-space methods (Continuous):

Compares parts of objects to each other
to determine which surfaces should be labeled
as visible (use of bounding boxes, and check
limits alongg each direction).
)
 Order the surfaces being drawn, such
that it p
provides the correct impression
p of
depth variations and positions.

Image Space methods (discrete):

Visibility is decided point by point at
each pixel position on the projection plane.
Screen resolution can be a limitation.

Hidden Surface – (a) Surface for rendering

or
(b)Line drawing
Coherence properties:
• Object Coherence – If one object is entirely
separate from another, do not compare.

• Face Coherence – smooth variations across a

face;; incrementally
y modify.
y

• Edge Coherence – Visibility changes if a

edge crosses behind a visible face.

• Implied edge coherence – Line of

intersection of a planar face penetrating
another, can be obtained from two points on
the intersection.
• Scanline coherence – Successive lines have
similar spans.

• Area Coherence – Span of adjacent group

of pixels is often covered by the same
visible face.

• Depth Coherence – Use difference equation

to estimate depths of nearby points on the
same surface.

• Frame Coherence – Pictures of two

successive frames of an animation
sequence are quite similar (small changes
in object
j and viewpoint).
p )
 Different Visible Surface
Detection Methods:

• Back-face Detection
• Depth (Z) buffer method
• Scan-line method
• Depth-sorting
ept so t g method
et od
• Area-subdivision method
• Octree methods
• A-buffer method
• BSP Trees
• Ray casting method

Visible
Vi ibl surface
f techniques
h i are 3D
versions of sorting algorithms
– basically
b i ll compare depth.
d th
 Back face Culling
g or removal
A Polygon (in 3D) is a back face if:
V N > 0.
V.N 0

Let V = (0,
(0 00, Vz) and N = Ai + Bj + Ck.
Ck

Then V
V.N
N = Vz.C;
C;

Let Vz be +ve (view along +ve Z-direction)

Z-direction),

Check the sign of C.

Condition of back face is thus:

sgn(C) > 0.
What happens if
V N = 0 ??
V.
N V

V
How to g
get normal vector (
(N)
) for a
3D surface, polygon ?
R Q

R=PXQ φ
P

-R

Order of vertices for calculation

of N – w.r.t
w r t the front side of the surface
surface.

Brighter Darker
Side side
Take
T k the
th order
d off vertices
ti to
t be
b
counter-clockwise for the brighter side.

Take the case of a cube:

Exterior faces of a cube:

 You can choose any two vectors (edges)
to obtain the normal to the surface.

Any risks in such a case, if we randomly

choose any two vectors?
Better Solution to obtain the direction
cosines of N:
n

Assume a =  (Yi −Y j )(Z i + Z j )

i =1
n vertices n
of the b =  ((Z i −Z j )(
)(X i + X j ) where j = i + 1
polygon: i =1

(Xi, Yi, Zi); n

i=1,2,…,n.
c =  (X i −X j )(Yi + Y j )
i =1

Compute: If i = n,
n j=1
 n
a =  (Yi −Y j )(Z i + Z j )
i =1
n
b =  (Z i −Z j )(X i + X j ) where j = i + 1
i =1
n
c =  (X i −X j )(Yi + Y j )
i =1

If i = n, j = 1
T k the
Take th expression
i off “c”:
“ ”
Can you relate that to one property of
the 2
2-DD polygon
pol gon in X-Y
X Y plane?
a, b and c also describe the projection
surface of the polygon on the Y-Z, Z-X and
X-Y planes (or along the X, Y and Z axis)
respectively.
i l
c = (X1 −X2)(Y1 +Y2) + (X2 −X3)(Y2 +Y3) + (X3 −X1)(Y3 +Y1)
= X1(Y2 −Y3)+X2(Y3 −Y1)+X3(Y1 −Y2)
If A is the total area of the polygon
polygon, the
projected areas are:

AYZ = a/2; AXZ = b/2; AXY = c/2;

The surface normal appears to be

pointing
i ti outwards.
t d Facing
F i towards
t d the
th viewer,
i
the edges of the polygon appears to be drawn
counter clockwise
counter-clockwise.
The p
polygon
yg is a back face,, if c > 0

If the view direction is along the –ve Z-

di
direction,
i the
h above
b condition
di i b
becomes:
c<0
D
Drawbacks
b k off b
back
k fface culling:
lli

•P
Partially
ti ll hidden
hidd faces
f cannott be
b determined
d t i d
by this method

• Not useful for ray tracing, or photometry/

radiosity.
radiosity

However, this is still

However
useful as a pre-processing
step as almost 50% of the
step,
surfaces are eliminated.
 Depth-buffer
Depth buffer
or
Z-buffer
ff method

COP
 Each (
(X,Y,Z)
, , )ppoint on a p
polygon
yg
surface, corresponds to the orthographic
projection point (X, Y) on the view plane.

At each point (X, Y) on the PP,

object depths are compared by using
the depth(Z) values.
Assume normalized coordinates:

(FCP) Zmax > Z > 0 (BCP), where Zmax = 1.

Two buffer areas are used:

(i) Depth (Z) buffer: To store the depth

values for each (X, Y) position, as
surfaces are processed.

(ii)
( ) Refresh Buffer: To store the intensity
y
value at each position (X, Y).
 Depth-buffer
Depth buffer
or
Z-buffer
ff method

COP
 Steps for Processing:
(a) Initialize depth(X,Y) = Z max
∀(X,
(X Y)
IB = background refresh(X,Y) = I B
intensity

(b) For each position on each polygon surface:

(i)
( ) Calculate depth
p Z for each p
position
(X, Y) on the polygon.

(ii) If Z < depth(X, Y) then

depth (X, Y) = Z;

refresh (X, Y) = Is (X, Y)

 Is is the projected intensity value of the
surface at position (X, Y), which has the
minimum value of Z Z, at the current stage of
iteration.
Equation of the surface :
AX + BY + CZ + D = 0;
Calculation of Z:
− AX − BY − D
Z=
C
Equation for Z at the next scan line:
Remember scanline/Polyfill algorithm?
U i
Using edge
d coherence,
h we gett for
f th
the nextt
scanline (Y+1):
X’ = X -1/m;
1/m;
 Y
Y-1

X X+1
A
Z X +1 = Z X − ;
C
A
+B
ZY + 1 = ZY + m ;
C
F
For a vertical
ti l edge:
d Z’ = Z + B/C;
B/C
 So to implement the algorithm,
algorithm three
constants are required for each surface:

A
( + B))
A m B
− , ,
C C C
Wh t iis th
What the special
i l condition,
diti C = 0?
Using the above three constants, we can
keep calculating the successive depth values
along
g and for successive scanlines. Similar
approaches can be used for curved surfaces:
Z = f(X,
( Y).)
 Z Buffer Algo:
Z-Buffer
for all (x,y)
depth(x y) = -∞ /* Watch this change */
depth(x,y)
refresh(x,y) = IB

for each polygon P

for each position (x,y) on polygon P
calculate depth z

if z > depth(x,y) then

p ( ,y) = z
1. depth(x,y)
2. refresh(x,y) = IP(x,y)

Lets take an example 

Z-values
Z l off th
the 4 4 9 8 7 6 5
coordinates: 4 4 8 7 7 6 5

4 4
4 4
Z-Buffer
Z Buffer
values
9 8 9 8
8 7 4 7 8 7 4
4 4 7 6 5 4
Z-values
Z l off th
the 4 4 9 8 7 6 5
coordinates: 4 4 8 7 7 6 5

9 8
7 8 7 4
7 6 5 4

Z-Buffer
values
IMAGES (after pseudo
pseudo-texture
texture rendering)
V
V
 SCAN-LINE
SCAN LINE Algorithms (for VSD)
Extension of 2-D Scanline (polyfill) algorithm.

Here we deal with a set of polygons.

Data structure used:
ET (Edge Table),
AET (Active Edge Table) and
PT (
(Polygon
yg Table).
)

Edge Table entries contain information

about edges of the polygon, bucket sorted
based on each edge’s smaller Y coordinate.

Entries within a bucket are ordered by

i
increasing
i X-coordinate
X di t off their
th i endpoint.
d i t
 Structure of each entry in ET:
• X-Coordn. of the end point with the smaller
Y-Coordn.
• Y-Coordn. of the edge’s other end point
• ΔX = 1/m
• Polygon ID X Ymax ΔX ID

Structure of each entry in PT:

• Coefficients of the plane equations
• Shading or color information of the polygon
• Flag (IN/OUT), initialized to ‘false’

Plane Shading
ID IN/OUT
Coeffs.
Coe s Info.
Take Adjacent (not successive) pairs of intersections to fill;
FILL (within pair) if POLY_FLAG is set to IN.

B E

L1
L2
L3
D
C
L4
F
L5

L3
ABC IN; ABC IN; ABC OUT;
DEF OUT DEF IN. DEF IN.

Compare Z values in this region

to find the visible Z value.
 B E

L1
L2
L3
D
C
L4
F
L5

AET Contents
Scan Line Entries
L5 AB CA
L4 AB CA FD EF
L3, L2 AB DE BC EF
L1 AB BC DE EF
Three non-intersecting
non intersecting polygons
pol gons

J B I
E

D C
F
A
G H

GHIJ is behind ABC and DEF.

 So, when scanline leaves edge BC, it is
still inside polygons DEF and GHIJ. If
polygons do not intersect, depth calculations
and comparisons
p between GHIJ and DEF can
be avoided.

Thus depth computations are

unnecessary when the scanline leaves an
obscured polygon. It is required only when it
leaves an obscuring polygon.

Additional treatment is necessary for

i
intersecting
i polygons.
l
 Depth-sorting
ept so t g oro Painter’s
a te s Algorithm
go t
Paint the polygons in the frame buffer
in order of decreasing distance from the
viewpoint.
Broad Steps:
• S
Surfaces
f are sorted
t d iin iincreasing
i order
d
of DEPTH.
• Resolve ambiguities when polygons
overlap (in DEPTH), splitting polygons
if necessary.
• Surfaces are scan converted in order
order,
starting with the surface of greatest
DEPTH.
DEPTH
Principle:

Each layer of paint (polygon surface of

an object) covers up the previous layers
while
hil drawing.
d i

ZMax1
S1
ZMin1
ZMax2
S2
ZMin2
X
Z
Since,, ZMin1 > ZMax2 no overlapp occurs.
S1 is first scan converted, and then S2.
This g
goes on, as long g as no overlap occurs.

ZMax1 S1
ZMin1
ZMax2 S2
ZMin2 V
X
Z
If depth overlap occurs, additional
comparisons are necessary to reorder the
surfaces.
The following set of tests are used to
ensure that no re-orderingg of surfaces is
necessary:
 Four(4) Tests in Painter’s algorithm

1.The boundary rectangles in the X-Y plane

for the two surfaces do not overlap.

2.Surface S is completely behind the

overlapping surface relative to the viewing
position.

3.The overlapping surface is completely in

f
front
t off S relative
l ti to
t the
th viewing
i i position.
iti

4.The
4 The projections of the ttwo
ossurfaces
rfaces onto
the view plane do not overlap.
T
Tests must be
b performed
f d iin order,
d as specified.
ifi d

Condition to RE-ORDER the surfaces:

If any one of the 4 tests is TRUE, we

proceed to the next overlapping surface. i.e.
If all overlapping surfaces pass at least one of
the tests, none of them is behind S.
No re-ordering is required, and then S is
scan converted.

RE ORDERing is required if all the four

RE-ORDERing
tests fail.
 TEST #1:
The boundary rectangles in the X-Y
plane for the two surfaces do not overlap.
overlap

We have depth overlap

overlap, but no overlap
in X-direction.
Hence, Test #1 is passed, scan convert
S2 and then S1.

S2
If we have X
S1
overlap, check
f
for th
the rest.
t
Z X1 X2 X3 X4 X
TEST #2:
Surface S is completely behind the
overlapping surface relative to the viewing
position.
S1

S2

Z X

Fig. 1. S1 is completely behind/inside the

overlapping surface S2
TEST #3:
The overlapping surface is completely in
f
front
t off S relative
l ti to
t the
th viewing
i i position.
iti

S1 is not completely behind S2

S2.
So, Test #2 fails.
S1
Fig. 2. Overlapping S2
surface
f S2 iis completely
l t l
front/outside S1.
Z X

In Fig. 2, S2 is in front of S1, but S1 is

not completely inside S2 – Test #2 it not TRUE
or FAILS, although Test #3 is TRUE.
 How to check these conditions?

i) Set the plane equation of S2, such that the

surface S2 is towards the viewing position.

ii) Substitute the coordinates of all vertices of

S1 into the plane equation of S2 and check
for the sign.
sign

iii) If all vertices of S1 are inside S2,

S2 then
S1 is behind S2. (Fig. 1).
iiv)
) If all
ll vertices
i off S1 are outside
id S2,
S2 S1 iis
in front of S2.
TEST #4:
Th projections
The j ti off the
th two
t surfaces
f onto
t
the view plane do not overlap.

S2
S1

Z X
 Four(4) Tests in Painter’s algorithm
- revisited
1.The boundary rectangles in the X-Y plane
for the two surfaces do not overlap.
overlap

2.Surface
2 Surface S is completely behind the
overlapping surface relative to the viewing
position.
position

3.The
3 The overlapping surface is completely in
front of S relative to the viewing position.

4.The projections of the two surfaces onto

the view plane do not overlap.
T
Tests must be
b performed
f d iin order,
d as specified.
ifi d

Condition to RE-ORDER the surfaces:

If any one of the 4 tests is TRUE, we

proceed to the next overlapping surface. i.e.
If all overlapping surfaces pass at least one of
the tests, none of them is behind S.
No re-ordering is required, and then S is
scan converted.

RE ORDERing is required if all the four

RE-ORDERing
tests fail.
Case Studies – examples of

Painter’s Algorithm
 Case Study - I

S2

S1

X
Z
Initial
I iti l order:
d
S1 -> S2

Change the order:

S2 ->
> S1
 C
Case S
Study
d - II
S3
Initial Order: S1 S2
S1 -> S2 -> S3
S3.

S1 -> S2
S2, X
Z
Test 1 passed.

Check S1, S3. All tests fail.

Interchange. S3 -> S2 -> S1.

Check S2 and S3. All tests fail again.

Correct Order: S2 -> S3 -> S1.

Process off checking
P h ki th
the order,
d
results in an infinite loop.

Avoided by setting a FLAG for a

surface that has been re re-ordered
ordered
or shuffled. If it has to be altered
again split it into two parts.
again, parts
 0 0 0
Y Y 0
-0.5 -0.5
-0.5
05 -0.5
05

S1 S2
S2 S1

0.5 0.5 0
0.5 0.5
0 0 0

X Fig. B X
Fig A
Fig.

What happens in these cases ?

 Area sub-division method

Examine and divide if necessary

y

• Works in image-space
image space

• Area Coherence is exploited

• Divide-and-conquer
q strategy.
gy

WARNOCK’s
WARNOCK s
Algorithm
 Possible relationships of the area of interest
(rectangular,
g AOI) and the projection
j of the
polygon:
1.Surrounding
1 Surrounding polygons are completely
overlapping the area of interest.

2. Intersecting polygons intersect the area.

3. Contained polygons are completely inside

the area.

4. Disjoint
j polygons
p yg area completely
p y outside
the area.
 Four cases of Polygon
yg p position w.r.t rectangular
g AOI.

Intersecting

S
Surrounding
di

Contained

Disjoint
 Treat interior part of the intersection
and contained as equivalent. Disjoint is
irrelevant The decisions about division of an
irrelevant.
area is based on:

1.All polygons are disjoint from the area.

Display
p y background
g color.

2.Only
y one intersectinggppart or interior
(contained): Fill the area with background
color and then scan-convert the polygon.

3.Single surrounding polygon, and no

intersecting or contained polygon. Use
color of surrounding polygon to shade the
area.
4. More than one polygon intersect, contained
in and surrounding the area. But the
surrounding polygon is in front of all other
polygons.
l Then
Th fill the
h area with
i h the
h color
l
of the surrounding polygon.
Step 4 is implemented by comparing the
Z-coordinates of the p
planes of all p
polygons
yg
(involved) at the four corners of the area
If all four tests fail
fail, divide the
rectangular area into four equal parts.

Recursively apply this logic, till you reach

the lowest minimum area or maximum
resolution – pixel size.
Use nearest surface in that case to
paint/shade.
 X Decision – Step 4
X

Z
Z
Intersection with:
Contained Polygon Surrounding
Polygon
Intersecting Polygon
 Four cases of Polygon
yg p position w.r.t rectangular
g AOI.

Intersecting

S
Surrounding
di

Contained

Disjoint
Case Studies – examples of

Area Sub-division method

(i) (ii)

(iii) (iv)
( )
BSP (Binary Space Partition) trees
 P1
P2 C

A
D
P1

B BP F B
FP
FP
BP P2F P2B

F B F B
BSP (Binary Space
A C B D
Partition) trees
Salient feat
features:
es

• Identify surfaces that are inside/front and

outside/back w.r.t. the partitioning plane at
each step of the space division, relative to
the viewing direction.

• Start with any plane and find one set of

objects
bj t behind
b hi d andd the
th restt in
i the
th front.
f t

• In the tree
tree, the objects are represented as
terminal nodes with front objects as left
branches and back objects as right
branches.
• BSP tree’s root is a polygon selected from
those to be displayed.

• One polygon each from the root polygon’s

front and back half plane, becomes its front
and back children

• The
h algorithm
l i h terminates
i when
h each
h node
d
contains only a single polygon.

• Intersection and sorting at object space/

precision
i i
 3
5a 5 2 4
2 5b
5b 5a 1
1 3
4

5a 5
2
3 5b
1 3
4
1 4
2 5b
5a
 5

5a 5 4
2
3
5b
1 3 1
4
2

5
2
3
1 3
2 4 4

5a 1 5b
 5a
5
2
V
5b 3
1 3
4 2 5b

1 5a 4
Also see:
Other applications:
potential visibility sets;
• Robot navigation
KD trees, R+-trees;
• Collision Detecion
Bounding Volume Hierarchy (BVH)
• GIS
and
Shadow Volume BSP Tree (SVBSP) • Image Registration
 Display
p y list of a BSP tree
• BSP tree may be traversed in a “modified in-
order tree
tree” walk to yield a correctly priority-
ordered polygon list for an arbitrary
viewpoint.
viewpoint
• If the viewer is in the root polygon’s front
half space
space, the algorithm must first display:
1.All polygons in the root’s rear half-
space.
space
2.Then the root.
3.And finally all polygons in the front
half-space.
• Use back face culling
• Each of the root’s children is recursively
processed.
processed
 Disadvantages
g

– more polygon splitting may occur than

in Painter's algorithm

– appropriate partitioning hyperplane

selection is quite complicated and
difficult

Finding Optimal BSP:

Finding an optimal root node is by testing a small

number o
of candidates.

The node that results in the smallest number of

nodes
d iin th
the BSP ttree should
h ld bbe chosen
h
VSD - Ray Tracing

COP
 Figure for illustrating
VSD - Ray
R Tracing
T i

In reality, possibilities with a light ray:

absorption,
o p o , reflection,
o , refraction,
o , scattering,
g, fluorescence
o or
o
chromatic aberration (error in lens).
Remember? For Z-buffer:
Each (X,Y,Z) point on a polygon surface,
corresponds to the orthographic projection
point (X, Y) on the view plane.
At each point (X, Y) on the PP, object
depths are compared by using the depth(Z)
values.
l

For Ray-tracing:
• Shoot ray from eye point through pixel (x,y)
into scene

• Intersect with all surfaces, find first one the

ray
y hits
 F Ray-tracing
For R t i (cont’d):
( t’d)

• Li
Line off Si
Sight
ht off each
h pixel
i l iis iintersected
t t d
with all surfaces

• Shade that point to compute the color of

pixel (x,y)
(x y)

• Consider only the closest surface for shading

• Based on optics of image formation

formation, paths of
light rays are traced

• Light rays are traced backward

 For Ray-tracing
Ray tracing (cont’d):
(cont d):

• Suitable for complex curved surfaces

• Computationally expensive

• Need
d off an efficient
ffi i ray-surface
f iintersection
i
technique

• Almost all visual effects can be generated

g

• Can be parallelized

• Has aliasing problems

 Mathematically, the intersection
problem is of finding the roots of an equation:

Surface – RAY = 0;

Ray Equation: r(t) = t (P – C)

Eqns. for
LINE: x = x0 +tΔx ; y = y0 +tΔy ; z = z0 +tΔz ;
SPHERE: (x − a) + ( y −b) + (z −c) = r
2 2 2 2
Substitution
gives us:
( x0 + tΔx) − 2a( x0 + tΔx) + a
2 2

+ ( y0 + tΔy) − 2b( y0 + tΔy) + b

2 2

+ (z0 + tΔz) − 2c(z0 + tΔz) + c = r

2 2 2

Collecting terms
gives us:
(Δx + Δy + Δz )t +
2 2 2 2

2t[Δx( x0 − a) + Δy( y0 − b) + Δz(z0 − c)]

+ ( x0 − a) + ( y0 − b) + (z0 − c) − r = 0
2 2 2 2
 The Equation is Quadratic, with
coefficients from the constants of sphere and
ray equations.

Cases:
• No real roots - Surface and ray do
not intersect

• One real root - Ray tangentially

grazes the surface

• Two
T reall roots
t - From b
F both
th
intersections, get
th one with
the ith the
th
smallest value of t.
What is the shade, at that point of
i t
intersection?
ti ?

Sphere has center (a, b, c).

The surface normal at any point of

intersection is:
xp − a y p − b z p − c
[ , , ]
r r r
What about intersection with other (regular) non-
linear surfaces ??

• Gaussian
G i
• Ellipsoid
• Directional sinusoid
• Torus
Eqns. for
LINE : x = x0 + tΔx ; y = y0 + tΔy ; z = z0 + tΔz ;
PLANE : A + By
Ax B + Cz
C +D=0
Substitution gives us:
( Ax0 + By0 + Cz0 + D)
t =−
( AΔx + BΔy + CΔz)
• Denominator should not be zero;

• Find if intersection is within the polygon, by

projecting onto a suitable coordinate plane;

• Overall processing is ray-wise, not

polygon-wise.
l i
Read about:

• Stochastic/Distributed Ray Tracing (Monte Carlo on Graphics)

• Global Illumination

• Radiosity
R di it

• The
Th Metropolis
M t li light
li ht transport
t t (MLT) - SIGGRAPH 1997

• Bi-directional
Bi directional Ray Tracing
 Rendering with global illumination: Light is reflected by
surfaces, and colored light transfers from one surface to
another Color from the red wall and green wall (latter not
another.
visible) reflects onto other surfaces in the scene. Also
notable is the caustic projected onto the red wall from
li ht passing
light i through
th h the
th glass
l sphere.
h
Comparison
p of VSD (
(HSR)
) techniques
q

Algorithms/
Memory Speed
Methods
Two Depth
Z B ff
Z-Buffer
arrays complexity
Apriori
One
Painter’s sorting helps
array
y
speed-up
d
Object ( p
O(#pixels,
Ray casting data #surfaces or
base objects)
Scanline, Area
- Slowest
sub-division
sub division
Algorithms Issues in
Remarks
/Methods Implementation
Scan conversion,,
Z B ff
Z-Buffer C
Commonly
l used
d
Hardware
Splitting and
sorting the
Painter’s Scan conversion
major
bottleneck
Excellent
Spatial data
Ray
y casting
g structures helpp for CSG,
speedup h d
shadows,
transparency
Cannot be
Scanline, Area
Hard generalized for non-
sub-division
sub division
polygonal models.
VSD - “Visible Surface Detection”
i also
is l called:
ll d

Hidden Surface Elimination (HSE); Also: HSR, OC.

So, what is HLE:

Hidden Line Elimination.

Problem:
Take any VSD algorithm and
g
convert to an HLE algorithm
End of lectures on

VISIBLE SURFACE
DETECTION