RASTER & VECTOR DATA

REPRESENTATION

Harikrishna M
Department of Civil Engineering
National Institute of Technology Calicut

National Institute of Technology Calicut

RASTER DATA REPRESENTATION

National Institute of Technology Calicut 2

 INTRODUCTION
 Raster model
 Field-based model of geographic data representation

 Best employed for continuous geographic data

 Commonly known as tessellation model

 Basic spatial data unit: Predefined area of space for which

attribute data are explicitly recorded
 Regular or Irregular tessellation

 Regular tessellation : triangular, square and hexagonal

 Regular square or rectangular tessellation: Raster data model

 Compatible with different types of hardware devices

 Compatible with concepts and methods of bit-mapped images

 Compatible with grid-oriented coordinate systems

National Institute of Technology Calicut 3
Nature & Characteristics of Raster Data

National Institute of Technology Calicut 4

 Nature & Characteristics of Raster Data
 Subdividing a geographic space into grid cells
 Linear dimensions of each cell define the spatial resolution
of the data
 Size of the smallest object in the geographic space to be
represented: Minimum Mapping Unit (MMU)
 Grid size should be less than half the size of MMU
 Each grid cell must contain a value
 Value indicates the quantity or characteristic of the spatial
object or phenomenon that is found at that location of the
cell
 Grid cell values can be directly for computation or indirectly
as code numbers references to an associated table 5
National Institute of Technology Calicut
 Nature & Characteristics of Raster Data
 In raster database, values pertaining to different
characteristics in the same cell location are stored in
separate files (map layers)
 Raster data processing involves use of multiple raster
files
 Positions of spatial objects or phenomena in raster
model are represented only to the nearest cell
 Representation does not always correspond to the
spatial object or phenomena in the real world

National Institute of Technology Calicut 6

 Nature & Characteristics of Raster Data
 Individual cells that make up the real world object are
the entities
 Identities of individual spatial objects are lost in raster
data model
 Raster data are stored as a linear array of attribute
values
 Location of each cell is implicitly defined by its row and
column numbers
 Location of the cells can be computed when the data are
used for display and analysis
National Institute of Technology Calicut 7
 Nature & Characteristics of Raster Data
 To translate a linear array storage to a two dimensional
display, information should be stored in the header section
of the data file
 File Header contains information on
 Number of bits used to represent the value in each cell
 Number of rows and columns
 Type of image
 Legend
 Coordinate transformation
 File formats for raster data files vary depending on the
algorithm used for data compression
National Institute of Technology Calicut 8
 Principles of Raster Data Compression
 Single raster data files contain several million grid cells
 Black and white line map measuring 50cm× 50cm when
scanned at a resolution of 25 micrometers (~ 1000dpi) will
produce 400 million pixels
 Actual size of file depends on bit depth: number of bits used
to represent the value of the pixel
 Therefore, data compression is an important feature of
digital representation of raster data
 Assumptions in raster data file compression
 Cells representing areas of same entity type have identical values
 Patterns of values tend to be spatially clumped
National Institute of Technology Calicut 9
 Principles of Raster Data Compression
 A number of algorithms are used to handle adjacent cells of
identical values
 Run Length Coding
 Adjacent cells along a row with the same value are treated as a group called as a
“run”
 Instead of repeatedly storing the same value for each cell, the value is stored
once, together with the number of cells that make the run
 Simple to understand and easy to implement
 Disadvantage : Does not give a good compression ratio
 More efficient compression technology is wavelet compression
 Multi-resolution Seamless Image Database (MrSID) gives
compression ratio between 15:1 to 20:1 for 8-bit gray scale images
and 30:1 to 40:1 for 24-bit colour images
National Institute of Technology Calicut 10
 Principles of Raster Data Compression

Original Raster Data 10A2B3A8B1C4B2C3B3C

Start encoding
A A A A A A A A A A A A

A A A A B B A A A A B B

A A A B B B A A A B B B

B B B B B C B B B B B C

B B B B C C B B B B C C

B B B C C C B B B C C C

Total number of values = 18

Total number of values = 36
National Institute of Technology Calicut 11
 Quadtree Data Representation
 Hierarchical tessellation model that uses grid cells of
variable sizes
 Geographic space is divided by the process of recursive
decomposition
 Instead of dividing the entire geographic space into grid
cells of the same size, quadtree model uses finer
subdivision of areas when finer detail occurs
 Decomposition continues till a predefined maximum
number of iterations is reached, which determines the
minimum cell size that can be represented

National Institute of Technology Calicut 12

Quadtree Data Representation

13
National Institute of Technology Calicut
 Quadtree Data Representation
 For a typical topographic map of 50cm × 50cm, it takes 11
iterations to reach a resolution of 0.5mm and 13 iterations to
reach a resolution of 0.12mm
 No explicit storage of coordinates is necessary for the
quadtree model
 Position of individual cells can be found using cell
identification number
 Hierarchical numbering system of cell identification allows
location of each cell to be computed relative to map origin
 Descriptive attributes associated with cells are stored as
feature codes of individual cells
National Institute of Technology Calicut 14
 Quadtree Data Representation
 Advantages
 Data-storage and search techniques are well researched and
understood
 Compatible with Cartesian coordinate system for cartographic
applications
 Recursive sub-division of geographic space facilitates physically
distributed storage, economic use of memory and expedites
browsing operations
 Allows variable spatial resolution to be represented in accordance
with degree of complexity of geographic surface
 Disadvantages
 Difficult compared to simple raster model
 Relatively complex process to
National Institute of Technology Calicut 15
 Quadtree Data Representation
 Disadvantages
 Difficult compared to simple raster model
 Relatively complex process to modify quadtree indices and tables

 Best suited for

 Areas where data is relatively homogeneous
 Applications that require high-performance spatial search of the
database

National Institute of Technology Calicut 16

VECTOR DATA REPRESENTATION

National Institute of Technology Calicut 17

 INTRODUCTION
 Object-based approach
 Best suited to represent discrete objects
 Spatial objects are identified individually and represented
mathematically as coordinates
 More complex than a raster data model
 More difficult to implement
 All vector models are built on 2 concepts
 Decomposition of spatial objects into basic graphical elements
 Use of topology (spatial relationships) to represent spatial objects in
addition to geometry (coordinates)

National Institute of Technology Calicut 18

 Nature & Characteristics of Vector Data
 Represented using coordinates

 Spatial objects are represented by one of the 3 basic

graphical elements: points, lines, polygons

 When graphical elements representing an individual

identifiable real-world feature are logically grouped together,
a graphical entity is formed

 Graphical elements are “dumb” graphics and graphical

entities are “intelligent” graphics
National Institute of Technology Calicut 19
Nature & Characteristics of Vector Data

National Institute of Technology Calicut 20

Nature & Characteristics of Vector Data
Spaghetti

National Institute of Technology Calicut 21

 Spaghetti Data Model
 Any polygons that lie adjacent to each other must be
made up of their own lines, or stands of spaghetti

 In other words, each polygon must be uniquely defined

by its own set of X, Y coordinate pairs, even if the
adjacent polygons share the exact same boundary
information

 Data that are collected but not structured are said to be

in the spaghetti data model
National Institute of Technology Calicut 22
 Spaghetti Data Model
 Vector data obtained by map digitisation are said to be
in this data model as they are not structured
 Spaghetti data model stores graphical elements, but not
graphical entities, defined by strings of coordinates
 Spatial relationships are not explicitly encoded within the
spaghetti model
 Considerable redundancy in this data model, as
boundaries between adjacent polygons are stored twice
 Vector data in this model should be properly structured
for use in GIS
National Institute of Technology Calicut 23
 Topological Data Model
 Structured data built on the concept of topology
 Among several topological data models, the commonly
used is the arc-node data model
 “Arc”: line segment & “Node”: end points of line segment
 Arc-node model also stores graphical elements, but also
stores explicitly the spatial relations between graphical
elements and relationships between arcs and their
respective nodes
 Stored relationships allow graphical entities to be
constructed from basic graphical elements
National Institute of Technology Calicut 24
 Vector Data Model
 Vector data should be properly linked to descriptive data
in geographic databases
 This is achieved using unique feature identifiers (FID)
assigned to individual spatial objects
 By using common FIDs, graphical and descriptive
elements of vector data are correctly cross-referenced
during database creation and spatial data processing
 Usually automates process, but linkage to descriptive
data is normally a manual process

National Institute of Technology Calicut 25

Topological Data Model

26

National Institute of Technology Calicut

 Vector Data Model
 Representation of vector data is governed by scale of
input data
 Possibility of representing vector data differently at
different scales is associated with 2 important concepts
 Cartographic generalisation: whereby line and area objects
are represented by more coordinates at a larger scale than at
a smaller scale
 Cartographic symbolisation: whereby vector data are
represented by different symbols that serve to visually
distinguish them from one another when the data is displayed

National Institute of Technology Calicut 27

 Vector Data Model
 Vector data is stored as integers or floating-point
numbers
 To avoid rounding errors that occur during data
processing, most software products store data by using
double-precision, floating point numbers
 However, this does not ensure precise representation of
spatial objects as
 Precision of data storage does not always mean accurate
description of data
 Boundaries of spatial objects are fuzzy rather than exact
entities
National Institute of Technology Calicut 28
 Concept of a Topological Map
 Topology
 Branch of mathematics that studies those properties of
geometric figures that are unchanged when the shape of a
figure is twisted, stretched, shrunk or distorted without
breaking
 Field of geometry concerned with spatial relationships rather
than with rigid coordinates

 Topological relationships for spatial objects was

proposed by Corbett (1979)

National Institute of Technology Calicut 29

 Concept of a Topological Map
 Topological relationships include 3 basic elements
 Adjacency
 Containment
 Connectivity

National Institute of Technology Calicut 30

 Concept of a Topological Map
 Topological map
 Map that contains explicit topological information on top of the
geographical information expressed in coordinates
 All spatial entities are decomposed and represented in forms
of 3 basic graphical elements
 Point entity is identified by a unique feature identifier
 Line entity is identified by the line itself, as well as, its nodes
 Topological relationships among linear entities are computed
and stored by using the identifier of the nodes
 Polygon entities are formed by using line entities and their
respective nodes
National Institute of Technology Calicut 31
 Concept of a Topological Map
 Topological map
 Once formed, polygon entities are individually identified by a
unique identification number
 Topological relationship between polygon entities are
computed and stored by using adjacency information stored
with the line entities
 Adjacency information includes nodes of the line and
identifiers of the polygons on the left and right of the line
 Process of computing topological relationships: Topology
building

National Institute of Technology Calicut 32

Concept of a Topological Map

33
Concept of a Topological Map

34
Types of Topology

35
National Institute of Technology Calicut
Uses of Topological Relationships in GIS
 Functional areas where used

 Data Input and Representation

 Spatial Search

 Construction of complex spatial objects from basic graphical

elements

 Integrity checks in database creation

National Institute of Technology Calicut 36
 Data Input & Representation
 Manual creation of topological map is next to impossible
 Using topological relationships, the method of spaghetti
digitising can be adopted for graphical data input
 Arc-based digitising procedure wherein no particular
sequence in digitising needs to be followed
 Storing adjacency information in the form of identifiers
removes the need to duplicate line data
 When data are plotted to show all the polygons, lines
are plotted and not polygons

National Institute of Technology Calicut 37

Data Input & Representation

38
Spatial Search by Topological Relationships
 Tedious to find manually the data for parcels surrounding a
particular parcel in a land title information system
 Process requires first using an index map to obtain parcel
identification number (PID) of all neighbouring parcels and
retrieving data from appropriate records
 In GIS, adjacency information can be used to identify all
polygons that share a common boundary and apply their
PIDs to retrieve descriptive data from database
 Connectivity information can be used for spatial search
using line data

National Institute of Technology Calicut 39

Spatial Search by Topological Relationships

National Institute of Technology Calicut 40

Construction of Complex Spatial Objects
 Complex spatial objects are represented as complex
polygons in geographic databases
 2 types of complex polygons
 Those containing one or more holes (islands or enclaves)
 Those made up of two or more polygons that are not
physically connected
 Polygons can be constructed from vector lines using
topological information or a common identifier

National Institute of Technology Calicut 41

Construction of Complex Spatial Objects

National Institute of Technology Calicut 42

Integrity Checks in Database Creation
 Graphical data must not contain any of the topological
errors

43
Integrity Checks in Database Creation
 During topology building process, the computer will
identify topological errors and flag them automatically

 Data input operator can check the errors and decide

how they must be corrected

 Corrections of some errors can be done using the

concept of tolerance or interactive manual editing

44
THANK YOU

National Institute of Technology Calicut 45