Question 1

The number of pixels used in training does not affect the quality of image classification.

Accepted Answer

A) True 
 B)False  False

Question 2

Georeferencing an image to geographical coordinates (latitude and longitude) is an operation of

Accepted Answer

A)  filtering 
B)  linear stretching 
C)  radiometric correction 
D)  geometric correction 
A)  filtering 
B)  linear stretching 
C)  radiometric correction 
D)  geometric correction D

Question 3

Image classification is based on

Accepted Answer

A)  tone or colour 
B)  digital numbers 
C)  site or association 
D)  shape and texture 
A)  tone or colour 
B)  digital numbers 
C)  site or association 
D)  shape and texture B

Question 4

With the classification error matrix below, what is the overall accuracy of the land use classification?
$\begin{array}{|l|l|l|l|l|}
\hline & \text { Urban } & \text { Forest } & \text { Cropland } & \begin{array}{l}
\text { Row } \
\text { Total }
\end{array} \
\hline \text { Urban } & 7 & 2 & 1 & 10 \
\hline \text { Forest } & 2 & 25 & 8 & 35 \
\hline \text { Cropland } & 3 & 1 & 51 & 55 \
\hline \begin{array}{l}
\text { Column } \
\text { Total }
\end{array} & 12 & 28 & 60 & 100 \
\hline
\end{array}$

Accepted Answer

A)  100% 
B)  27.7% 
C)  29% 
D)  83% 
A)  100% 
B)  27.7% 
C)  29% 
D)  83%

Question 5

Cubic convolution is a geometric correction method, which calculates the DN value of a rectified pixel

Accepted Answer

A)  as the value of its nearest pixel in the input image 
B)  as the average value of the four nearest pixels in the input image 
C)  as the average value of the sixteen nearest pixels in the input image 
D)  using a cubic polynomial function 
A)  as the value of its nearest pixel in the input image 
B)  as the average value of the four nearest pixels in the input image 
C)  as the average value of the sixteen nearest pixels in the input image 
D)  using a cubic polynomial function

Question 6

Which of the following is not a potential problem with unsupervised classification?

Accepted Answer

A)  some spectral classes may be meaningless 
B)  a training area may contain a mix of different thematic classes 
C)  reference data used for assigning classes by the analyst may contain errors 
D)  a single thematic class may be split among two spectral classes 
A)  some spectral classes may be meaningless 
B)  a training area may contain a mix of different thematic classes 
C)  reference data used for assigning classes by the analyst may contain errors 
D)  a single thematic class may be split among two spectral classes

Question 7

On a colour infrared image (or photo),

Accepted Answer

A)  targets with high near-infrared reflectance appear red 
B)  targets with high red reflectance appear red 
C)  targets with high green reflectance appear red 
D)  targets with high blue reflectance appear red 
A)  targets with high near-infrared reflectance appear red 
B)  targets with high red reflectance appear red 
C)  targets with high green reflectance appear red 
D)  targets with high blue reflectance appear red

Question 8

Microwaves are the longest wavelengths used for remote sensing

Accepted Answer

A) True 
 B)False

Question 9

With linear stretching,

Accepted Answer

A)  the image would appear darker 
B)  the image would appear lighter 
C)  light tonal areas would appear darker and dark areas would appear lighter 
D)  light tonal areas would appear lighter and dark areas would appear darker 
A)  the image would appear darker 
B)  the image would appear lighter 
C)  light tonal areas would appear darker and dark areas would appear lighter 
D)  light tonal areas would appear lighter and dark areas would appear darker

Question 10

In a near IR image, forest appears

Accepted Answer

A)  darker than water body 
B)  brighter than water body 
C)  as bright as water body 
A)  darker than water body 
B)  brighter than water body 
C)  as bright as water body

Question 11

In supervised classification, training areas are

Accepted Answer

A)  representative samples of different thematic classes selected by the image analyst 
B)  representative samples of different thematic classes selected by the computer 
C)  spectral clusters identified by the image analyst 
D)  spectral clusters identified by the computer 
A)  representative samples of different thematic classes selected by the image analyst 
B)  representative samples of different thematic classes selected by the computer 
C)  spectral clusters identified by the image analyst 
D)  spectral clusters identified by the computer

Question 12

Training areas must

Accepted Answer

A)  be drawn from a single area 
B)  capture the spectral variability of a class 
C)  capture the spectral boundary between classes 
D)  be evenly distributed in the image 
A)  be drawn from a single area 
B)  capture the spectral variability of a class 
C)  capture the spectral boundary between classes 
D)  be evenly distributed in the image

Question 13

Which band of the spectrum is the best for separating water and land?

Accepted Answer

A)  red 
B)  blue 
C)  green 
D)  near IR 
A)  red 
B)  blue 
C)  green 
D)  near IR

Question 14

The spectral reflectance of a particular object varies with wavelength

Accepted Answer

A) True 
 B)False

Question 15

Atmospheric windows are

Accepted Answer

A)  those areas of the spectrum that are not severely influenced by atmospheric absorption 
B)  those areas of the spectrum that are blocked by atmospheric absorption 
C)  those areas of the spectrum in which little atmospheric scattering occurs 
D)  those areas of the spectrum that are not useful for remote sensing 
A)  those areas of the spectrum that are not severely influenced by atmospheric absorption 
B)  those areas of the spectrum that are blocked by atmospheric absorption 
C)  those areas of the spectrum in which little atmospheric scattering occurs 
D)  those areas of the spectrum that are not useful for remote sensing

Question 16

Remote sensing is only interested in measuring the radiation

Accepted Answer

A)  absorbed by targets 
B)  transmitted through targets 
C)  reflected from targets 
A)  absorbed by targets 
B)  transmitted through targets 
C)  reflected from targets

Question 17

Aerial photography is better than satellite remote sensing because

Accepted Answer

A)  aerial photography allows repetitive imaging of the Earth's surface on a continuing basis 
B)  aerial photography has a revisit period of several days 
C)  aerial photography can produce digital images 
D)  aerial photography has higher spectral resolutions 
E)  aerial photography provides the highest spatial resolution 
A)  aerial photography allows repetitive imaging of the Earth's surface on a continuing basis 
B)  aerial photography has a revisit period of several days 
C)  aerial photography can produce digital images 
D)  aerial photography has higher spectral resolutions 
E)  aerial photography provides the highest spatial resolution

Question 18

According to the classification error matrix below, how many training pixels are omission errors and how many are commission errors with the urban class?
$\begin{array}{|l|l|l|l|l|}
\hline & \text { Urban } & \text { Forest } & \text { Cropland } & \begin{array}{l}
\text { Row } \
\text { Total }
\end{array} \
\hline \text { Urban } & 7 & 2 & 1 & 10 \
\hline \text { Forest } & 2 & 25 & 8 & 35 \
\hline \text { Cropland } & 3 & 1 & 51 & 55 \
\hline \begin{array}{l}
\text { Column } \
\text { Total }
\end{array} & 12 & 28 & 60 & 100 \
\hline
\end{array}$

Accepted Answer

A)  3, 5 (omission, commission) 
B)  5, 3 
C)  7, 12 
D)  7, 10 
A)  3, 5 (omission, commission) 
B)  5, 3 
C)  7, 12 
D)  7, 10

Question 19

Converting DN values to radiance values is an operation of

Accepted Answer

A)  filtering 
B)  linear stretching 
C)  radiometric correction 
D)  geometric correction 
A)  filtering 
B)  linear stretching 
C)  radiometric correction 
D)  geometric correction

Question 20

With histogram equalisation, DN values are assigned

Accepted Answer

A)  as their frequencies of occurrence 
B)  based on their frequencies of occurrence 
C)  as their original DN values divided by their frequencies of occurrence 
D)  as their original DN values multiplied by their frequencies of occurrence 
A)  as their frequencies of occurrence 
B)  based on their frequencies of occurrence 
C)  as their original DN values divided by their frequencies of occurrence 
D)  as their original DN values multiplied by their frequencies of occurrence

The number of pixels used in training does not affect the quality of image classification.

Georeferencing an image to geographical coordinates (latitude and longitude) is an operation of

Image classification is based on

With the classification error matrix below, what is the overall accuracy of the land use classification? Urban Forest Cropland Row Total Urban 7 2 1 10 Forest 2 25 8 35 Cropland 3 1 51 55 Column Total 12 28 60 100

Cubic convolution is a geometric correction method, which calculates the DN value of a rectified pixel

Which of the following is not a potential problem with unsupervised classification?

On a colour infrared image (or photo),

Microwaves are the longest wavelengths used for remote sensing

With linear stretching,

In a near IR image, forest appears

In supervised classification, training areas are

Training areas must

Which band of the spectrum is the best for separating water and land?

The spectral reflectance of a particular object varies with wavelength

Atmospheric windows are

Remote sensing is only interested in measuring the radiation

Aerial photography is better than satellite remote sensing because

According to the classification error matrix below, how many training pixels are omission errors and how many are commission errors with the urban class? Urban Forest Cropland Row Total Urban 7 2 1 10 Forest 2 25 8 35 Cropland 3 1 51 55 Column Total 12 28 60 100

Converting DN values to radiance values is an operation of

With histogram equalisation, DN values are assigned

GIS Overview

Spatial Data and Data Management

Spatial Data Input and Manipulation

Spatial Analysis

Spatial Data Exploration With Statistics

Terrain Analysis

Spatial Visualisation

Spatial Modeling and Decision Analysis

Environmental Applications of GIS

Filters

Exam 6: Remote Sensing Data Analysis

The number of pixels used in training does not affect the quality of image classification.

Georeferencing an image to geographical coordinates (latitude and longitude) is an operation of

Image classification is based on

With the classification error matrix below, what is the overall accuracy of the land use classification? Urban Forest Cropland Row Total Urban 7 2 1 10 Forest 2 25 8 35 Cropland 3 1 51 55 Column Total 12 28 60 100

Cubic convolution is a geometric correction method, which calculates the DN value of a rectified pixel

Which of the following is not a potential problem with unsupervised classification?

On a colour infrared image (or photo),

Microwaves are the longest wavelengths used for remote sensing

With linear stretching,

In a near IR image, forest appears

In supervised classification, training areas are

Training areas must

Which band of the spectrum is the best for separating water and land?

The spectral reflectance of a particular object varies with wavelength

Atmospheric windows are

Remote sensing is only interested in measuring the radiation

Aerial photography is better than satellite remote sensing because

According to the classification error matrix below, how many training pixels are omission errors and how many are commission errors with the urban class? Urban Forest Cropland Row Total Urban 7 2 1 10 Forest 2 25 8 35 Cropland 3 1 51 55 Column Total 12 28 60 100

Converting DN values to radiance values is an operation of

With histogram equalisation, DN values are assigned

GIS Overview

Spatial Data and Data Management

Spatial Data Input and Manipulation

Spatial Analysis

Spatial Data Exploration With Statistics

Terrain Analysis

Spatial Visualisation

Spatial Modeling and Decision Analysis

Environmental Applications of GIS

Filters