AML710 CAD LECTURE 32

Solid Modeling Techniques

Constructive Solid Geometry

(CSG)
 Computational Solid Geometry
•  Primitive based: It is based on the notion that a physical
object can be divided into a set of primitives – basic
elements or shapes that can be combined in a certain
order following a set of rules ( Boolean operations) to
create the object.
•  Primitives themselves are considered valid CSG
models. Each primitives is bounded by closed and
orientable surfaces (half-spaces).
•  CSG model is fundamentally and topologically different
from a B-rep model in that it does not store faces, edges
and vertices explicitly. Instead it evaluates them
whenever needed by algorithms.
•  CSG representation is of considerable importance for
manufacturing
 Primitive based Vs Half-space based
CSG Schemes
There are two types of CSG schemes
•  Primitive Based CSG: It is based on bounded valid
solid primitives, r-sets. It is the most popular CSG
scheme.
•  Half space Based CSG :This scheme uses unbounded
half spaces (non r-sets). Bounded solid primitives are
considered composite half spaces and the boundaries
of these are the surfaces of the component half spaces.

- +

Half space based CSG

Primitive based CSG
 Data Structure of CSG
•  Like in B-rep, the database stores
topology and geometry.
•  The validity checking in CSG scheme
occurs indirectly. Each primitive that is
combined using a Boolean operations
(r-sets) to build the CSG model is
checked for its validity.
•  The common data structures used for
CSG are graphs and trees
 CSG Data Structure
Graph: A graph is defined as a set of nodes connected by a set of
branches or lines.

Path: Each node in a tree belongs to a path E.g. Path A to G is given

by (A,B,G) or (A,C,B,G) (in digraph )
Cycle: If starting and ending nodes are the same the path is called a
cycle E.g. Path (A,C,B,A) or (B,E,C,B)

A Root node A

B C B C

G F E D G F E D
Leaf nodes
Graph Digraph
 CSG Data Structure
Cyclic Graph: If a graph contains a cycle it is called cyclic otherwise it
is acyclic. Cycles are some times called loops or rings.
The digraph shown below is cyclic as it has two cycles.
Indegree: In a digraph, number of arrows entering a node
Outdegree: number of arrows going out of a node

Indegree = 3
B B C
Outdegree = 2

G F E D

Graph Digraph
 Tree and Binary Tree
Tree: A tree is an acyclic digraph in which a single node
called root node has a zero indegree and every other
nodes has an indegree of one.
Binary Tree: In a tree if the descendent of each node are in
order (from left to right) and each node except the leaf
node has two decedents (left and right), then the tree is
called a binary tree.
A A

B C B C

G F E D G F E D

Tree Binary Tree

 Binary Tree and Subtree
Subtree: Any binary tree can be thought of as joining together of 2
subtrees at any given node. A left subtree and a right subtree rooted
at two successor nodes. A perfect binary tree is one which has
(nL-nR)=0.
A A

C D
B C

B E
G F E D

Left subtree Right subtree F

G

nL = nR nL ≠ nR
Balanced binary tree Unbalanced binary tree
 Inverted Binary Tree for CSG Model
Inverted Binary Tree: If the direction of each connector
(arrow) is reversed then we get an inverted binary tree
wherein each node has one outdegree except the root
node.
Root node: A node with outdegree = 0. Any node that does
not have descendent.
Leaf node: Any node that does not have a predecessor or
indegree = 0.
Interior node: Any node with outdegree > 0 is an interior
node.
A

B C

Inverted Binary Tree G F E D

 Inverted Binary Tree as CSG Tree
CSG Tree: The inverted binary tree is very convenient to represent the
CSG operations. Here each of the leaf nodes is a valid solid
primitive. The intermediate nodes are the transition states of the
solid modeling (r-set) operations. The root node is the resulting solid
from the set operations. Here n primitives require (n-1) Boolean
operations to complete the construction of the object. The tree will
have (2n-1) nodes.

A DD
S=AUBUC
B C
S

A CUD D
AUB
B C
Primitives =4
Operations=3
A B C D
Nodes=2n-1=7
A B C D
CSG Tree
 Tree Traversal
Tree Traversal: Visiting the nodes in sequential or orderly
and efficient way. Many of the sorting and searching
algorithms need to do the tree traversal
Traversal methods: Two broad methods are:
1.  Depth first 2. Breadth first
The depth first is further divided into the following types
depending on the order in which the root node is visited
a) Preorder b) Inorder c)Postorder

1 2 3

2 3 1 3 1 2

Preorder Inorder Postorder

 Tree Traversal
Preorder: We have the following recursive algorithm
Algorithm
1
1.  Visit the root
2.  Traverse the left subtree in pre/o
3.  Traverse the right subtree in pre/o 2 9

3 8 10
11
Preorder
Traversal

4 7 12 13

5 6 14 15
 Tree Traversal
Reverse Preorder: We have the following recursive algorithm
Algorithm
15
1.  Traverse the right subtree in rp/o
2.  Traverse the left subtree in rp/o
3.  Visit the root 14 7

13 8 6
5
Reverse Preorder
Traversal

12 9 4 3

11 10 2 1
 Tree Traversal
Inorder: We have the following recursive algorithm
Algorithm
8
1.  Traverse the left subtree in inorder
2.  Visit the root
3.  Traverse the right subtree 6 10

4 7 9
12
Inorder
Traversal

2 5 11 14

1 3 13 15
 Tree Traversal
Reverse Inorder: We have the following recursive algorithm
Algorithm
8
1.  Traverse the right subtree
2.  Visit the root
3.  Traverse the left subtree 10 6
in reverse inorder

12 9 7
4
Reverse Inorder
Traversal

14 11 5 2

15 13 3 1
 Tree Traversal
Postorder: We have the following recursive algorithm
Algorithm
15
1.  Traverse the left subtree
in postorder
2.  Traverse the right subtree
7 14
3.  Visit the root

5 6 8
13
Postorder Traversal

3 4 9 12

1 2 10 11
 Tree Traversal
Reverse Postorder: We have the following recursive algorithm
Algorithm
1
1.  Visit the root
2.  Traverse the right subtree in
reverse postorder
9 2
3.  Traverse the left subtree in
reverse postorder

11 10 8
3
Reverse Postorder
Traversal

13 12 7 4

15 14 6 5
 Basic Elements
Primitives: Bounded solid primitives are the basic building blocks of
CSG. These (parametric) solids have two sets of geometric data:
1.  Configuration Parameters (size information)
2.  Rigid motion Parameters (orientation information)
Y Y
L

R
P H X
B D Z
X
Y Y
Z BLOCK CYLINDER

WEDGE SPHERE

H P R
X X
D Z
Z B
 Building Operations
The main building operations are regularised set operatoins
like union (U*), intersection (∩*) and difference (-*).
Hence the CSG models are known as set-theoretic,
Boolean or combinatorial models.
In contrast to Euler operations, the Boolean operations are
not based on any equation or law. They are based on the
set theory and the closure property. These operations
are considered higher-level operations than Euler
operations.
Some implementations of solid modelers provide derived
types of operations like ASSEMBLE and GLUE
 Main algorithms in CSG Operations
1.  Edge / Solid intersection algorithm
2.  Computing set membership classification
a) Divide and conquer: It is like ray tracing. Instead of a ray an edge is
used as a reference
b) Neighborhood: It deals with in, on and out decisions
When a point is in the interior of solid face then it is called face
neighborhood
Edge neighborhood occurs when the point lies on the solid edge
When a point is a vertex, vertex neighborhood occurs. This is a
complex case because the point is shared between three solid
faces.

P P
P
 Summary of a CSG algorithm
The following steps describe a general CSG algorithm based on divide
and conquer approach:
1.  Generate a sufficient number of t-faces, set of faces of participating primitives, say A
and B.
2.  Classify self edges of A w.r.t A including neighborhood.
3.  Classify self edges of A w.r.t B using D & C paradigm. If A or B is not primitive then
this step is followed recursively.
4.  Combine the classifications in step 2 and 3 via Boolean operations.
5.  Regularize the ‘on’ segment that result from step 4 discarding the segments that
belong to only one face of S.
6.  Store the final ‘on’ segments that result from step 5 as part of the boundary of S.
Steps 2 to 6 is performed for each of t-edge of a given t-face of A.
7.  Utilize the surface/surface intersection to find cross edges that result from intersecting
faces of B (one at a time) with the same t-face mentioned in step 6.
8.  Classify each cross edge w.r.t S by repeating steps 2 to 4 with the next self edge of A.
9.  Repeat steps 5 and 6 for each cross edge
10.  Repeat steps 2 to 9 for each t-face of A.
11.  Repeat steps 2 to 6 for each t-face of B.
 A CSG Example
Create the CSG model of the following solid S.
y d S

x
c d AU*B-*C
b B
z
A c + -
a AU*B C
Geometry of the primitives
BLOCK A: xL = a − d , y L = d , z L = c, PA ( x, y, z ) = PA (d ,0,−c)
xL = d , y L = b, z L = c, PB ( x, y, z ) = PB (0,0,−c) A B
BLOCK B:
CYLINDER C: R = R, H = d , PC ( x, y, z ) = PC (d + a / 2, d ,−c / 2)

E M(E,B) M(E,A)
A U* ∩* -*
Classify E on A M(E,A)
E out A E on S
E out B M(E,B) E E on B
B E in A E on S E on S
E on S U*
Combine Null ∩* Classify Combine
E on S -*