REMOTE SENSING & GIS (18CV651)

MODULE 4

DATA MODELS

Vector data model

Vectors are graphical objects that have geometrical primitives such as points lines and polygons to
represent geographical entities in a computer graphics. Vectors have a precise direction length and
shape and can be defined by co-ordinate geometry.

The vector model is close to the traditional mapping approach with objects is represented as points
lines are areas. In a vector model the positions of points, lines and areas are precisely specified. The
position of each object is Defined by a co-ordinate pairs.

Representation of simple features:

The vector model is ideal to represent discrete entities. According to this concept discrete entities are
represented as points, lines and areas.

Point is simpler to input and analyse. Point feature is represented by a vector without length and
direction. The location of a point features such as a location of electric pole, postbox, tube wells can
be described by a single co-ordinate. Points are required to represent NTT course areas are negligible
or not important.

Lines defined by 2 points are used to represent features that are linear in nature. Roads, streams and
contour lines are the examples of line features. They can also be used to represent linear features that
do not have any physical existence such as in line showing an International border.

Areas are presented by a closed set of lines and are used to define features such as playgrounds,
buildings or administrative areas. These closed set up lines are referred 2 hours polygons or regions.
Polygons needs only points to input but the area perimeter and other geometric attributes may be
computed by the GIS software rather than manual input.

Regions similar to polygons but it may be containing a hole with in an area or region may contain
multiple polygons which are not adjacent. For example, private land lots scattered with a National
Forest should be subtracted from the forest to get the exact coverage in the area of the forest.

Vector data representation using points, lines and area is not always straight forward because it
depends on map scale, functions we wish to perform in our later analysis, on the criteria established

Mr. RAGHAVENDRA NAIK M D, Asst. Professor, JIT, DAVANGERE Page 1

by the government mapping agencies. For example, a city on a 1:100000 scale map may appear as a
point but the same city may appear as an area on a 1:24000 scale map.

Topology

Topology in GIS can be defined as a set of objects and object data that defines the relationship
between the objects. This is a conceptual representation of special features as points, lines and areas
in vector model. Topology refers to the relationship between spatial objects.

Important to vector data model are diagraphs (directed graphs), which include points and directed
lines. The directed lines are called arcs and the points where arcs meet or interest are called nodes.
Adjacency and incidence are two relationships that can be established between nodes and arcs in
diagraphs. If an arc joins two nodes, the nodes are said to be adjacent and incident with the arc.

In geometry a point is given by 2 D coordinates(x, y), while line, string(series of lines), and area are
given by a series of point coordinates, as shown in fig above. The topology however defines
additional structure as node, chain and polygon.

Mr. RAGHAVENDRA NAIK M D, Asst. Professor, JIT, DAVANGERE Page 2

Importance of topology

Topology has at least two advantages

1. Assurance of data quality. Topological relationships enables us to detect errors such as lines that
do not meet correctly or poly comes that are not closed properly. These kinds of errors must be
corrected to avoid incomplete features and to ensure Data integrity.

2. Topology can enhance GIS analysis.

3. Topology is necessary for certain special functions such as network routing through linear
networks.

Coverage and its data structure

A Coverage is a georelational data model that stores data, it contains both the special and attribute
data for Geographic features. Coverage you set up future classes to represent Geographic features.
Each feature class stores a set of points, lines polygons or annotation (text).

The coverage model incorporates the topological relationships into the structure of feature data. The
data structure of point coverage is simple. It contains feature identification numbers (IDs) and pair of
x, y coordinates.

The starting point of the arc is from node and end point is to node. The art node list sorts out the arc
node relationship.

Example: arc 2 has 12 as from node and 13 as the to node. The arc coordinate list shows the x-y
coordinates of the from node to node and other points (vertices) that makes up each arc.

Mr. RAGHAVENDRA NAIK M D, Asst. Professor, JIT, DAVANGERE Page 3

Arc 3 consists of from node at (2,9) to node (4,2) and two vertices at (2,6) and (4,4), therefore has
three line segment.

Figure shows the data structures of a polygon coverage. The polygon / arc list shows the relationship
between polygons and arcs.

Mr. RAGHAVENDRA NAIK M D, Asst. Professor, JIT, DAVANGERE Page 4

Examples: 1, 4, 6 connect to define polygon 101. Polygon 104 differs from the other polygons
because it is surrounded by polygon 102. To show that polygon 104 hole within polygon 102, the arc
list from polygon one or two contains a zero to separate the external and internal boundaries.
Polygon 104 is also an isolated polygon consisting of only arc 7. The left / right list in figure shows
the relation between arcs and their left and right polygons.

Example: arc 1 is a directed line from node 13 to node 11 and has polygon 100 as the polygon on the
left and polygon 101 as a polygon on the right. The arc co-ordinate list shows the nodes and vertices
that make up each node.

Shape file:

A shape file is a simple non topological format for storing the geometric location and attribute
information of the Geographic features. Geographic features in a shape file can be represented by
points, lines and polygons (areas). The workspace containing shape file may also contain dBase
tables, which can store additional attributes that can be joined to a shape file's features.

Although the shape file treats a point as a pair of x, y coordinates a line as a series of points and a
polygon has a series of lines. No files describe the special relationship between these geometric
objects. The geometry of Shape file is stored into basic files: the .shp file stores the feture geometry,
and .shx file maintains the spatial index of the features geometry.

To main advantages of shape files are

1. They can display more rapidly on the computer monitor then topology based data

2. They can be used across different software packages.

Shape files can be converted to coverages and vice versa. The conversion froma shape file to a
coverage required the building of topological relationships and the removal of duplicate arcs. The
conversion from a coverage to a shapefile is simpler. But if a coverage as topological error such as
lines not joined properly the errors can be lead to problems of missing features in the shape file.

Relational database

In a relational database all data is stored and accessed via relations. Relations that store data are
called base relations and implementations are called tables. Other relations do not store data but are
computed by applying relational operations to store relations. These relations are sometimes called
derived relations.

Mr. RAGHAVENDRA NAIK M D, Asst. Professor, JIT, DAVANGERE Page 5

A relational database is a database that group data using common attributes found in the data set. The
resulting clumps organized data are much easier for people to understand.

Example: a dataset containing all the real estate transactions in a town can be grouped by the year the
transaction occurred, or it can be grouped by the sale price of the transaction, all it can be grouped by
the buyer’s last name and so on.

Such a grouping uses the relational model. In such a database is called a relational database.

The software used to do this grouping is called a relational database management system. The term
relational database often referred to this type of software.

Relational database are currently the predominant choice in storing financial records, manufacturing
and logistical information, personal data and much more.

Raster data model

Raster organizes special features in a regularly spaced grid of cells as pixels. Raster data represent
points with single cells, lines with sequence of neighboring cells and areas collection of continuous
cells.

As in a traditional raster image, the x-axis is the column indicator and the y axis is arrow indicator.
Not only Geographic data but other picturing kal data can also be represented and stored in a
computer as raster.

Each cell in the raster contains a single value. The raster data model uses a regular grid to cover the
space and the value in each cell to respond to the characteristics of a special phenomenon at the cell
location. Conceptually, the variation of special phenomenon is reflected by the changes in the cell
value.

Mr. RAGHAVENDRA NAIK M D, Asst. Professor, JIT, DAVANGERE Page 6

The raster cell size is an important factor. Small cells improve data quality because they can provide
more details. As cell size increases, data definition decreases or blurs. Cell size in a raster file is
referred to as resolution.

Elements of raster data model

1. Cell value: each cell in a raster carries a value which represents the characteristics of a special
phenomenon at the location denoted by its row and column depending on the coding of its cell
values, a raster can be either an integer or a floating point raster.

An integer value as no decimal digits, whereas a floating point value does. A land cover raster May
use 1 for urban land use, 2 for forest land, 3 for water body and so on. Floating point cell values
represent continuous numeric data. For example, a precipitation raster may have precipitation values
of 20.15, 12.23 and so on. We can assess the cell values of an integer raster through a value attribute
table. But a floating point raster usually does not have a value attribute table because of its
potentially large number of records.

2. Cell size: the same size determines the resolution of the raster data model. A cell size of 10 m
means that each cell measures 10m X 10m = 100 square metre. A Cell size of 30m means that each
cell measures 900 square metre. Therefore a 10m raster as a finer resolution than a 30m raster.

3. Raster bands:

A raster may have single band or multiple bands. Each cell in a multi band raster is associated with
more than one cell value. Example for multiband raster in a satellite image which may have 5, 7 or
more bands at each cell location. Each cell in a single band raster as only one all value. An example
of a single band raster is an elevation raster which has 1 elevation value at each cell location.

Special reference:

Raster data must have the special reference information so that they can align especially with other
datasets in a GIS. For example to superimpose an elevation raster on a vector based soil layer we
must first make sure that both datasets are based on the same co-ordinate system.

Types raster data:

Satellite imagery Remotely sensed satellite data are familiar to GIS users. The spatial resolution of a
satellite image relates to the ground pixel size. Example, special resolution of 30 m means each pixel
image satellite image corresponding ground pixel of 900 square meter. The pixel value also called
the brightness value represents light energy reflected or emitted from the Earth surface. The
Mr. RAGHAVENDRA NAIK M D, Asst. Professor, JIT, DAVANGERE Page 7
REMOTE SENSING & GIS (18CV651)

measurement of light energy is based on spectral bands. Panchromatic images are comprised of
single spectral bands where as multispectral images comprised of multiple bands.

Digital elevation models (DEM): It consists of an array of uniformly spaced elevation data. A DEM
is a point based but it can be easily being converted to raster data by placing a elevation point at the
centre of a cell. Most GIS uses in the US use DEMs from the USGS. USGS DEMs include 7.5
minute DEM 30 minute DEM 1 DEGREE DEM and Alaska DEM (7.5 minute DEMS provide
elevation data at a spacing of 30 m 10 m).

Global DEMs: DEMs different resolutions are now available on the Global scale. SRTM DEMs are
available for land areas outside the United States with the resolution of 90 M at the equator.

Digital orthophotos: A digital orthophoto quad(DOQ) digitized image prepared from an aerial
photographs or other remotely sensed data, in which the displacement caused by camera 10 and
Terrain relief has been removed. A digital orthophoto is georeferenced and can be registered with
topographic and other maps.

Bi level scanned files: It is scanned image containing values 1 or 0. Bi level scanned files are
usually made for the purpose of digitizing. They are scanned from paper that contains boundaries of
soils, parcels and other features. Bi level scanned files are converted to vector based features.

Digital Raster Graphics (DRGs): It is a scanned image of USGS topographic map. USGS scans the
7.5 minute topographic map thus providing a DRG with a ground resolution of 2.4m. The USGS
uses up to 13 colours on each 7.5 minute DRG

Graphic files: Maps, photographs and images can be stored as digital graphic files. Many popular
graphic files are in raster format, such as TIFF( tagged image file format), GIF( Graphics interchange
format) and JPEG( joint photographic experts group).

GIS software - specific raster data: GIS packages use raster data that are imported from DEMs,
satellite images, graphic files and vector data.

Raster data structure:

Raster data structure refers to the storage of raster data. The three common data structures are,

Cell by cell encoding: This method provides the simplest raster data structure. Raster is stored as
matrix, and its cell values are written into a file by row and column.

Mr. RAGHAVENDRA NAIK M D, Asst. Professor, JIT, DAVANGERE Page 8

Run length encoding: Run length encoding records the cell values by row and by group. Hey Guru
refers to adjacent cells with the same cell value. Figure shows the run length encoding of a polygon.
For each row, the starting cell and the ending self denote the length of the group (run) that falls
within the polygon.

Quad Tree: instead of walking along one row at a time, Quad tree uses recussive decomposition to
divide a raster into a hierarchy of quadrants. Recussive decomposition refers to the process of
continuous subdivision until every quadrant in a quad tree contains only one cell value.

Figure shows a raster with the polygon in gray and a quad tree that stores the features. The quad tree
contains nodes and branches. A node represents a quadrant, a node can be a non leaf node or leaf
node. In non leaf node represents a quadrant that has different cell values. A non leaf node is
therefore a branch point, meaning that quadrant is subjected to subdivision. A leaf node, on the other
hand, represents a quadrant that has the same cell value. A leaf node is therefore an end point which
can be coded with the value of homogeneous quadrant (gray or white).

Mr. RAGHAVENDRA NAIK M D, Asst. Professor, JIT, DAVANGERE Page 9

For example NW quadrant with spatial index 0 has 3 leaf node and one non leaf node which can be
further subdivided. The non leaf node again subdivided into single leaf node with gray and white
shade. Similarly all quadrants are subdivided into nodes and branches.

Data conversion:

Data conversion is a standard functionality in a GIS package. The conversion of vector data to raster
data is called rasterization and the conversion of raster data to vector data is called vectorisation.

The simplex of the two conversion methods rasterization involves basic steps

Mr. RAGHAVENDRA NAIK M D, Asst. Professor, JIT, DAVANGERE Page 10

1. First step sets of a raster with the specified service size to convert the area extent of the vector data
and initially assigns all cell values as zero's.

2. Second step Changes the values of those cells that corresponds to points, lines or polygon
boundaries. The cell value is set to 1 for a point, the line's value for a line and the polygons value for
polygon boundary.

3. Third step fills the interior of the polygon outline with a polygon value comes from rasterization I
usually related to the design of computer algorithm size of raster cell and boundary complexity.

Vectorisation involved three basic elements, line thinning line extraction and topological
reconstruction. Line in the vector data model has length but no width. Raster lines in a standard file
ok but usually occupy several pixels in width. Raster lines must be thinned, ideally to a 1 cell width,
for vectorisation. Line extraction is a process of determining where individual lines begin and end.

Advantages and disadvantages of raster and vector:

Raster model advantages:

 Simple data structure

 Easy and efficient overlaying

 Compatible with remote sensing imagery

 High spatial variability is efficiently represented

 Efficient to the represent continuous data

Disadvantages

 Large file size

 All the objects a series of pixels, no identity for discrete objects other than point / pixels

 Difficult to build topological relationship

 Inefficient projection transformations

 Lots of information when using large cells

 Difficult to edit

Mr. RAGHAVENDRA NAIK M D, Asst. Professor, JIT, DAVANGERE Page 11

Vector model advantages

 Smaller file size

 Individual identity for discrete objects like line polygon etc.

 Efficient for topological relationship

 Accurate map output

 Easy to edit

Disadvantages

 Complex data structure

 Difficult overlay operation

 High special variability is insufficiently represented

 Not compatible with remote sensing imagery

 Not appropriate to represent continuous data

Mr. RAGHAVENDRA NAIK M D, Asst. Professor, JIT, DAVANGERE Page 12