Raster data in detail

 Raster data is used in a GIS app when we want to display information that is continuous across an area and
cannot easily be divided into vector features.
 In previous image, Point, polyline and polygon features work well for representing some features on this
landscape, such as trees, roads and building footprints. However, we can’t map the grasslands and its variations
in colour and density of cover.
 Because when you give a the grass polygon
one value, they apply to the whole polygon
pixels, and a lot of information will be lost. so
vectors aren’t very good at representing
features that are not homogeneous (entirely the
same) all over. Because of simplification and
generalization of the value.

 Another approach you could take is to digitise

every small variation of grass colour and cover
as a separate polygon. The problem with that
approach is that it will take a huge amount of
work in order to create a good vector dataset

31 10 2021 | page. 1
  Using raster data is a solution to these problems. Many people use raster data as a backdrop to be used behind vector
layers in order to provide more meaning to the vector maps.

 Raster data is not only good for images that depict the real world surface (e.g. satellite images and aerial photographs),
they are also good for representing more abstract ideas (indirect information). For example, rasters can be used to show
rainfall trends over an area, or to depict the fire risk on a landscape. In these kinds of applications, each cell in the raster
represents a different value e.g. risk of fire on a scale of one to ten.

True colour raster images The calculated average minimum temperature

in the Western Cape for the month of March

31 10 2021 | page. 2
 Georeferencing
 Georeferencing is the process of defining exactly where on the earth’s surface an image or raster dataset
was created.

 This positional information is stored with the digital version of the aerial photo. When the GIS application
opens the photo, it uses the positional information to ensure that the photo appears in the correct place on
the map.

 Normally this positional information consists of a coordinate for the top left pixel in the image, the size of
each pixel in the X direction, the size of each pixel in the Y direction, and the amount (if any) by which the
image is rotated. With these few pieces of information, the GIS application can ensure that raster data are
displayed in the correct place.

 The georeferencing information for a raster is often provided in a small text file accompanying the raster.

31 10 2021 | page. 3
 Overview
 In the previous topics we have seen that vector features use geometry
(points, polylines and polygons) to represent the real world,
 However, raster data takes a different approach. Rasters are made up of a
matrix of pixels (also called cells), each containing a value that represents the
conditions for the area covered by that cell (see figure). In this topic we are
going to take a closer look at raster data, when it is useful and when it makes
more sense to use raster data.

 Each cell is covered with the coordinate system and has a location in the row
and column as well. According to these coordinate numbers, groups of pixels
(in millions) are saved on the hard disk. in a process similar to the coordinates
in point, polyline and polygon in the vector GIS

31 10 2021 | page. 4
 The computer stores these systems and deals with them, whether they have
dimensions (0, 1, 2, or more). Either by recording the true values (digital
numbers - DN) of the phenomenon or by reclassifying these values (Categorical
grid).

31 10 2021 | page. 5
31 10 2021 | page. 6
  In some cases, raster data can be computed. For example:
 an insurance company may take police crime incident reports and create a country wide raster map showing how high
the incidence of crime is likely to be in each area.
 Meteorologists (people who study weather patterns) might generate a province level raster showing average
temperature, rainfall and wind direction using data collected from weather stations (look previous figure).
In these cases, they will often use raster analysis techniques such as interpolation (which we describe in Topic Spatial
Analysis (Interpolation)).

 Sometimes raster data are created from vector data because the data owners want to share the data in an easy to use
format. For example, a company with road, rail, cadastral and other vector datasets may choose to generate a raster
version of these datasets so that employees can view these datasets in a web browser. This is normally only useful if the
attributes, that users need to be aware of, can be represented on the map with labels or symbology. If the user needs to
look at the attribute table for the data, providing it in raster format could be a bad choice because raster layers do not
usually have any attribute data associated with them.

31 10 2021 | page. 7
 Sources of Raster images
1)- Aerial photographs
 Spectral bands
 Thermal channels
 In aerial photography, an aeroplane flies
over an area with a camera mounted
underneath it. The photographs are then
imported into a computer and
georeferenced.

 photogrammetry and aerial photographs

have two basic types, photographic and
electromagnetic, which are the main
sources for raster maps in GIS.
 Photographic surveying (photogrammetry)
is the science and technology for obtaining quantitative and qualitative information about the phenomena of the Earth's
surface through photographic and electromagnetic images.

 These images are used in the production of maps and cadastral plans for various uses.

31 10 2021 | page. 8
 Elements aerial photos
The aerial photo contains elements that are useful to the GIS user in the
operations of guidance and image adjustments in his system, and these
elements are
1- Fiducial marks: to designate the main point on the image, and they
have different shapes on the four sides of the sound, or the midpoints of
the ribs, or both, and these points intersect at the location of the main
point of the image.

2- Altimeter: to indicate the height of the aircraft during the

photographing process.

3- Hour: Determining the time during which the shooting process took
place, and how to process the image from the shadows

4- Inclinometer: to know the inclination of the main axis of the camera

lens on the plane during the photographing process.
5- Basic dimension (focal length of the camera lens): It is possible to calculate the appropriate drawing scale for the printed
aerial image.
6- Other information: In addition to this, other information should be printed on top of the image, such as the serial number
of the image, the number of the camera, the date of shooting and the airline number.
31 10 2021 | page. 9
 Use of aerial photographs in GIS
Aerial images can be added in different ways in the GIS. Where as the most
important methods of using aerial images in the GIS are the following:

1- In areas where no maps are available. Where the images are collected in a
unified, scaled-up image called a Planimetric Orthophoto map that can be used
instead of real maps, but the maps are distinguished from them by the
presence of names for the phenomena of the Earth's surface.

- An orthophoto, orthophotograph, orthoimage or orthoimage is an aerial

photograph or satellite imagery geometrically corrected ("orthorectified")
such that the scale is uniform: the photo or image follows a given map
projection.

2- As an accessory for maps. It is possible to take advantage of the modified

planar aerial photographs for details that are difficult to identify using
charts and maps, with the exception of steep slopes or covered with dense
forests, the features of the surface of the earth can easily be distinguished
from roads, buildings, etc. Therefore, a GIS user can use aerial photographs
31 10to
2021 supplement
| page. 10 the missing information on the maps.
 Sources of Raster images
2)- Satellite Images (Remote Sensing).
 Spectral bands
 Thermal channels
 Satellite imagery is created when satellites
orbiting the earth point special digital cameras
towards the earth and then take an image of the
area on earth they are passing over.
 Once the image has been taken it is sent back to
earth using radio signals to special receiving
stations such as the one shown in the figure
below. The process of capturing raster data from
an aeroplane or satellite is called remote
sensing.
 Many consider satellite images as an extension of the (latest)
aerial photos, but satellite photos are distinguished from
aerial photos in that they contain a lot of spectral
information as a result of their design.
 The derivation of spectral information (the spectrum of
electromagnetic waves) from space images is the subject of the
31 10 2021science
| page. 11 of analysing digital images for remote sensing.
 Sources of Raster images
2)- Satellite Images (Remote Sensing).
 Spectral bands
 Thermal channels

The CSIR Satellite Applications Center at Hartebeeshoek near Johannesburg. Special antennae track satellites as they
pass overhead and download images using radio waves

31 10 2021 | page. 12
 Electromagnetic Spectrum
 The ratios of reflected, absorbed, or
emitted energy will vary depending on the
different features of the Earth.
 These differences allow us to distinguish
between the different features of the types
of surface phenomena such as water, barren
land or vegetation, so that the reflected
electromagnetic radiation will give a
different spectral signature.

 Spectral signature curves indicate the

amount of energy that is reflected or
emitted from the Earth's surface, which is
recorded by a remote sensing instrument.

 At a certain wavelength, if the reflectance

is high, this means that the reflecting
object is light in colour, and vice versa.
We notice from the above figure that the
vegetation will reflect more radiation at
infrared wavelengths.
31 10 2021 | page. 13
 Types of satellite images
Classification according to its spectral structure Satellite images
can collect data in one or more regions of the electromagnetic
spectrum. Accordingly, satellite images are divided into:
1- Panchromatic image: An image that contains information in one
region of the electromagnetic spectrum.

2- Multispectral Imagery: Images that contain information in a number

of ranges ranging from three to nine ranges.

3- Hyperspectral Imagery: Images that contain information in a number

of ranges ranging from ten to several hundred ranges.

4- Radar & Microwave Imagery: This image represents one of the types
of monochrome images, but it differs from it in that it records data
in the electromagnetic spectrum, both radar and microwave.

31 10 2021 | page. 14
 Satellite images – Landsat 8 CHARACTERISTICS
Launch Date • February 11, 2013
Vehicle • Atlas-V rocket
Site • Vandenberg Air Force Base, California
Spacecraft
• 3.14 terabit solid-state data recorder
• Power provided by a single 9 x 0.4 meter solar array and
one 125 Ampere-Hour (AHr), Nickel-Hydrogen (NiH2) battery
• Weight: 2,071 kg (4,566 lbs) Length: 3 m (9.8 ft)
• Diameter: 2.4 m (7.9 ft)

Communications
• Direct Downlink with Solid State Recorders (SSR)
• Data rate: 384 Mbps on X-band frequency; 260.92 Mbps on S-band frequency

Orbit
• Worldwide Reference System-2 (WRS-2) path/row system
• Sun-synchronous orbit at an altitude of 705 km (438 mi)
• 233 orbit cycle; covers the entire globe every 16 days (except for the highest polar latitudes)
• Inclined 98.2° (slightly retrograde)
• Circles the Earth every 98.9 minutes
• Equatorial crossing time: 10:00 a.m. +/- 15 minutes

31 10 2021 | page. 15
 Satellite images - Landsat 8 Bands
Sensor Band number Band name Wavelength (μm) Resolution (m) Band Applications

OLI 1 Coastal 0.43 – 0.45 30 Coastal and aerosol studies

Bathymetric mapping, distinguishing soil from vegetation, and deciduous from
OLI 2 Blue 0.45 – 0.51 30
coniferous vegetation
OLI 3 Green 0.53 – 0.59 30 Emphasizes peak vegetation, which is useful for assessing plant vigor

OLI 4 Red 0.63 – 0.67 30 Discriminates vegetation slopes

OLI 5 NIR 0.85 – 0.88 30 Emphasizes biomass content and shorelines

OLI 6 SWIR 1 1.57 – 1.65 30 Discriminates moisture content of soil and vegetation; penetrates thin clouds
Improved ability to track moisture content of soil and vegetation and thin cloud
OLI 7 SWIR 2 2.11 – 2.29 30
penetration
OLI 8 Pan 0.50 – 0.68 15 15 meter resolution, sharper image definition

OLI 9 Cirrus 1.36 – 1.38 30 Improved detection of cirrus cloud contamination

TIRS 10 TIRS 1 10.60 – 11.19 30 (100) 100 meter resolution, thermal mapping and estimated soil moisture

TIRS 11 TIRS 2 11.50 – 12.51 30 (100) 100 meter resolution, thermal mapping and estimated soil moisture

31 10 2021 | page. 16
 Spatial Resolution
 Every raster layer in a GIS has pixels (cells) of a fixed size that
determine its spatial resolution. This becomes apparent
when you look at an image at a small scale and then zoom in
to a large scale
 Images with a pixel size covering a small area are called
‘high resolution’ images because it collects huge details in
the image.
 Images with a pixel size covering a large area are called ‘low
resolution’ images it collects small details in the image. This satellite image looks good when using a small scale

 Several factors determine the spatial resolution of an image.

1- the capabilities of the sensor used to take an image.
 For example SPOT5 satellites can take images where
each pixel is 10 m x 10 m.
 MODIS satellite takes images only at 500 m x 500 m
per pixel.
 In aerial photography, pixel sizes of 50 cm x 50 cm are
not uncommon.

at a large scale you can see the individual pixels

31 10 2021 | page. 17
  Computed rasters using spatial analysis (rainfall), the spatial density of the weather stations information determine the
spatial resolution.

 Storage requirements is important as well. Think of a raster that is 3 x 3 pixels requires to store 9 numbers in the
computer’s memory. However, a raster layer for South Africa with pixels of 1 km x 1 km (South Africa is around
1,219,090 km 2). Which means your computer would need to store over a million numbers on its hard disk in order to
hold all of the information. Making the pixel size smaller would greatly increase the amount of storage needed.

 Sometimes using a low spatial resolution is useful when you want to work with a large area and are not interested in
looking at any one area in a lot of detail. The cloud maps you see on the weather report, are an example of this –– it’s
useful to see the clouds across the whole country.

 On the other hand, using low resolution raster data can be problematic if you are interested in a small region because
you probably won’t be able to make out any individual features from the image.

31 10 2021 | page. 18
 Spectral resolution
 camera or camera on a cellphone use electronic sensors to detect red, green and blue light (RGB). When the
picture is displayed on a screen or printed out, RGB information is combined to show you an image that your
eyes can interpret. While the information is still in digital format though, this RGB information is stored in
separate colour bands.

 Our eyes can only see RGB wavelengths, however, the electronic sensors in cameras are able to detect
wavelengths that our eyes cannot. Of course in a hand held camera it probably doesn’t make sense to record
information from the non-visible parts of the spectrum since most people just want to look at pictures.

 non-visible parts of the light spectrum are often referred to as multi-spectral images can be also detected in
advanced cameras (remote sensing science, in satellites and airplanes). the non-visible parts of the spectrum
can be very useful. For example, measuring infra-red light can be useful in identifying water bodies.

 Because having images containing multiple bands of light is so useful in GIS, raster data are often provided as
multi-band images. Each band in the image is like a separate layer. The number of bands in a raster image is
referred to as its spectral resolution.
 If an image consists of only one band, it is often called a grayscale image. With grayscale images, you can
apply false colouring to make the differences in values in the pixels more obvious. Images with false colouring
applied are often referred to as pseudocolour images.

31 10 2021 | page. 19
 Representation accuracy of map features (curved and fractured lines) depends on the
earth surface roughness and its distribution of pixels

31 10 2021 | page. 20
 Raster to vector conversion
 In our discussion of vector data, we explained that often raster data are used as a backdrop layer, which is
then used as a base from which vector features can be digitised.

 Another approach is to use advanced computer programs to automatically extract vector features from
images.

 Some features such as roads show in an image as a sudden change of colour from neighbouring pixels. The
computer program looks for such colour changes and creates vector features as a result. This kind of
functionality is normally only available in very specialised (and often expensive) GIS software.

Vector to raster conversion

 One side effect of this is that attribute data (that is attributes associated with the original vector data) will be lost
when the conversion takes place.

 Having vectors converted to raster format can be useful though when you want to give GIS data to non GIS
users. With the simpler raster formats, the person you give the raster image to can simply view it as an image
on their computer without needing any special GIS software.

31 10 2021 | page. 21
The basic steps to produce raster maps from vector maps
manually.

31 10 2021 | page. 22
The basic steps to produce raster maps from vector maps
manually.
Here, it is necessary to study the data contained in the map and solve the mixed pixel
problem. Ordinary letters or numbers are used to symbolize the prevailing type of land
use.

Water dominates Winner takes all Edges separate

W W G W G G W E G
W W G W W G W E G
W W G W G G E E G
31 10 2021 | page. 23
 Raster analysis
 There are many analytical tools that can be run on raster data which cannot be used with vector data. For
example, rasters can be used to model water flow over the land surface. This information can be used to
calculate where watersheds and stream networks exist, based on the terrain.

 Raster data are also often used in agriculture and forestry to manage crop production. For example with a
satellite image of a farmer’s lands, you can identify areas where the plants are growing poorly and then use
that information to apply more fertilizer on the affected areas only. Foresters use raster data to estimate how
much timber can be harvested from an area.

 Raster data is also very important for disaster management. Analysis of Digital Elevation Models (a kind of
raster where each pixel contains the height above sea level) can then be used to identify areas that are likely
to be flooded. This can then be used to target rescue and relief efforts to areas where it is needed the most.

Common problems / things to be aware of

As we have already mentioned, high resolution raster data can require large
amounts of computer storage.

31 10 2021 | page. 24
Storing Raster Data.
1- Run Length Encoding
An ASCII file is created using a word processor software, where data is
entered on the basis of the Run Length Encoding, which is data entry in the
form of a pair of numbers, the first number expresses the number of repeated
pixels, and the other expresses the pixel value (digital number – DN)
ASCI files after execution

31 10 2021 | page. 25
 1

Run-Length Codes allows the data to be

reduced to 50% of the storage space

12

Run-Length Codes
1 7:F, 1:S, 4:G
7:F, 1:S, 4:G
6:F, 1:M, 1:S, 4:G
5:F, 2:M, 1:S, 4:G
2:F, 4:M, 2:S, 4:G
1:F, 4:M, 2:S, 5:G
4:M, 2:S, 1:M, 5:G
3:M, 1:S, 3:M, 5:G
2:M, 2:S, 3:M, 5:G
2:M, 1:S, 3:M, 6:G
1:M, 2:S, 3:M, 6:G
12 2:S, 4:M, 6:G

31 10 2021 | page. 26
Storing Raster Data.
2- QuadTree
 Divide the area
recursively and in order
into four equal sections.

 Inspection for the

homogeneous quadrant that
has the same
characteristic of the
phenomenon studied.

 Heterogeneous quadrants
apply the same division
to obtain smaller
quadrants.
31 10 2021 | page. 27
 QuadTree is based on spatial autocorrelation, which is
based on the probability of similarity of adjacent
places for a characteristic. This similarity
decreases as the places are further away from each
other.

Continue in the division until reaching the quadrants

that achieve the required spatial accuracy, and thus
we will get squares of different dimensions but have
the same attribute and special symbols are placed to
assign the attribute,
Finally, we have obtained a tree structure for the data. This method is
used in many GIS systems and provides a significant reduction in the size
of represented matrices.

31 10 2021 | page. 28
 Advantages and disadvantages of raster data
Disadvantages of Raster Data Advantages of Raster Data
1- Storing information in this way is less accurate and less 1- Easy to store and organize. Because the number of pixels
detailed than vector data, because the pixel may not cover and their dimensions are known, and because the pixels are
all the phenomena. Generalization in the geographical coded to distinguish them from each other (the number for
location - pixel, so the extent of clarity lies in increasing and the row and column). The pixel values ​can be converted to
decreasing the number of pixels. colours or shades, making them easier to understand
2- It takes up a large amount of memory, because all the (Choropleth or thematic maps)
cells are recorded, whether they are covering the 2- Ease of data manipulation and change (pixels are known
phenomenon or empty. and defined). Processing is facilitated by increasing the cells,
3- The shape of the phenomenon is not beautiful and but the time taken increases with the increase in the
inaccurate, unlike the phenomena with a precise vector number of cells.
system. The pixels are less beautiful and accurate as the size 3- Compatible with remotely sensed information (aerial
of the pixels increases and their number decreases, and vice photos are also satellite images).
versa. 4- Its programs are less accurate, less expensive, and less
complex than vector GIS
5- Layers are easy to put on top of each other (have to be
similar pixel resolution)
6- It facilitates the use of modelling, because modelling is
based on the generalization to determine their behaviour.
And cells are a generalization process
31 10 2021 | page. 29
 Comparison Vector GIS and Raster GIS

1. Raster Maps 1. Vector Maps

1. Requires a lot of storage capacity 1. Requires small storage capacity

2. There are no topological relationships at 2. Topological relationships can be
the element level established
3. It is handled at the pixel level 3. It is handled at the feature level
4. Editing and editing needs smart 4. Editing, modifying, and querying for
software features is easy
5. Economical, less expensive and quickly 5. More expensive and time consuming
completed 6. more accurate in general
6. The accuracy is low and depends on the
pixel dimension

31 10 2021 | page. 30
 What have we learned?
 Raster data are a grid of regularly sized pixels.

 Raster data are good for showing continually varying information.

 The size of pixels in a raster determines its spatial resolution.

 Raster images can contain one or more bands, each covering the same spatial area,
but containing different information.

 When raster data contains bands from different parts of the electromagnetic spectrum,
they are called multi-spectral images.

 Three of the bands of a multi-spectral image can be shown in the colours Red, Green
and Blue so that we can see them.

 Images with a single band are called grayscale images.

 Single band, grayscale images can be shown in pseudocolour by the GIS.

 Raster images can consume a large amount of storage space.

31 10 2021 | page. 31