Professional Documents
Culture Documents
MGY 002E B4 Removed
MGY 002E B4 Removed
MGY 002E B4 Removed
Structure
12.1 Introduction
Objectives
12.2 Image Enhancement
Contrast Enhancement
Spatial Enhancement
12.3 Image Transformation
Arithmetic Operations
Image Fusion
12.4 Summary
12.5 Unit End Questions
12.6 References
12.7 Further/Suggested Reading
12.8 Answers
12.1 INTRODUCTION
In the previous unit, you have learnt about the different errors affecting the
quality of remotely sensed data and image correction techniques. You can
extract usable information from an image that has been corrected for
radiometric and geometric errors.
Objectives
After studying this unit, you should be able to:
• define image enhancement and image transformation;
• describe the concept of commonly used image enhancement strategies;
• discuss the basics of image transformation and outline the uses of
vegetation index; and
• list uses of image fusion techniques.
23
Processing and
Classification of Remotely 12.2 IMAGE ENHANCEMENT
Sensed Images
Image enhancement simply refers to the techniques used to increase the
human interpretability and ability to derive meaningful information from
remotely sensed images. It improves the quality and clarity of images. The
enhancement operations include removing blurring and noise, increasing
contrast and revealing details. For example, image enhancement of a blurred
and low contrast satellite image of the Himalayan Mountains could be made
clear by increasing contrast range and reducing blurring in the original image.
The original image might have areas of very high and very low intensity,
which mask details. The enhanced image will have improved contrast
The term enhancement is
providing details of the masked image regions.
used to define the changes
carried out in an image to The primary goal of image enhancement is to improve the visual
improve its appearance for interpretability of the images. This is done by increasing the apparent
better visual interpretation. distinction between features in the image. Image enhancement involves use of
There are number of a number of statistical and image manipulation functions available in an image
image enhancement
processing software. Image features are enhanced by the following two
techniques available but
no single technique can be operations:
said to be best because
needs of users varies and • Point Operations: In this operation, value of pixel is enhanced
also a technique which independent of characteristics of neighbourhood pixels within each band.
works for a certain image
for a particular area may • Local (Neighbourhood) Operations: In this operation, value of pixel is
not work for other images enhanced based on neighbouring brightness values.
for the same area.
Image enhancement techniques based on point operations are also called as
Image enhancement contrast enhancement techniques and the enhancement techniques based on
techniques are also known
neighbourhood operations are also known as spatial enhancement techniques.
as radiometric enhance-
ment techniques.
We shall now discuss about the different image enhancement techniques in the
following two sections.
(a) (b)
(a) (b)
Fig. 12.2: An example of gray level thresholding based image enahncement. (a) A LISS
false colour composite generated from green, red and NIR bands and (b)
resultant binary image of gray level thresholding carried out on NIR band 25
Processing and Level Slicing
Classification of Remotely Level or density slicing is an enhancement technique whereby DNs distributed
Sensed Images
along the x axis of an image histogram are divided into a series of analyst
specified intervals or slices and thereby creating a new image showing select
number of slices (classes) in it.
Contrast Stretching
Image display and recording devices often operate over a range of 256 gray
levels (maximum range of numbers represented in 8 bit image data). Image
data in a single image rarely extend over this entire range. Contrast stretching
technique expands the narrow range of pixel values typically present in an
input image to over a wider range of pixel values. Contrast stretching can be
of two types, linear and non-linear.
A linear contrast stretching occurs, when pixel values of an image are
uniformly stretched across a larger display range. By expanding the range of
original input values of the image, the total dynamic range of display device
can be utilised.
Usually, original minimum and maximum gray level values are set to the
lowest and highest pixel values in the display range, respectively. For
example, if maximum pixel value in the original image is 158 and minimum
pixel value is 60 as illustrated in Fig. 12.3a, the ‘stretched’ values would be set
at 255 and 0, respectively (Fig. 12.3b). The intermediate gray level values of
original image would be similarly and proportionately set to new values. These
types of enhancements are best applied to remotely sensed images with normal
or near-normal distribution, meaning, all the brightness values fall within a
narrow range of the image histogram and only one mode is apparent. This
approach of linear contrast stretching is known as minimum-maximum stretch.
(a)
(b)
(c)
Fig. 12.3: Illustration of principle of contrast stretch. Note: with a typical application of a
linear stretch, areas that were dark-toned will appear darker and areas that were
bright toned will appear even brighter. Linear contrast enhancement also makes
26 subtle variations within data more obvious (source: redrawn after Jensen, 1996)
One major disadvantage of a linear stretch is that it assigns an equal number of Image Enhancement and
Transformation
display levels for infrequently occurring values as it does for frequently
occurring values. This can result in certain features being enhanced, while
other features, which may be of greater importance, will not be sufficiently
enhanced. In these circumstances, we use non-linear contrast enhancement
techniques as discussed below.
Contrary to the linear stretch, a non-linear stretch expands one portion of gray
scale while compressing other portions. This sort of enhancement is typically
applied with purpose of highlighting one part of the brightness range. While
spatial information is preserved, quantitative radiometric information can be
lost. Two of the most common of non-linear contract stretching methods are
discussed below:
a) Histogram Equalisation
Histogram equalisation is one of the most useful forms of non-linear contrast
enhancement. When an image’s histogram is equalised, all pixel values of the
image are re-distributed so there are approximately an equal number of pixels
to each of the user-specified output gray scale classes (e.g., 32, 64, and 256). The
contrast is increased at the most populated range of brightness values of
histogram (peaks). In Fig. 12.3c, you can see how histogram equalisation at
different thresholds increases contrast of the image. It automatically reduces
contrast in very light or dark parts of the image associated with the tails of a
normally distributed histogram. Histogram equalisation can also separate pixels
into distinct groups, if there are very few output values over a wide range.
Histogram equalisation is effective only when original image has poor contrast
to start with, otherwise histogram equalisation may degrade the image quality.
b) Logarithmic Stretch
A logarithmic or square root stretch compresses the higher brightness values
within an image and disproportionately expands the darker values. It is applied
by taking the log (square root) of original brightness values. Fig. 12.4 shows
the difference between a linear and a logarithmic stretch. Thus, the original
darker values are given a greater relative contrast than original values at the
higher end of the brightness scale. This would be done in order to allow better
visual discrimination among the darker features of the image, while
maintaining some limited recognition of brighter features. The logarithmic
stretch is a more emphatic version of the square root stretch. Each of these can
be inverted (log to inverse log; square root to squared), so as to favour the
bright end of scale instead.
Logarithmic approach
stretches DN values in
darker part of histogram
whereas inverse
logarithmic approach
stretches DN values in
brighter part.
Fig. 12.4: Logic of linear, logarithmic and inverse log contrast stretch. BV indicates
brightness value (source: redrawn after Jensen, 1996)
27
Processing and The type of non-linear enhancement technique that is chosen to be performed
Classification of Remotely
Sensed Images
on an image depends on the nature of original data and what features the
analyst is interested in. For example, water features, which are typically darker
in tone, would benefit from log or square root stretches, while the land
features, which are brighter, would still be somewhat recognisable.
(a) (b)
Fig. 12.5: (a) Example of low frequency image and (b) example of high frequency image
The most commonly used filters for digital image filtering rely on the concept
of convolution and operate in the image domain for reasons of simplicity and
processing efficiency. For example, low-pass filters are mainly used to
smoothen image features and to remove noise but often at the cost of
degrading image spatial resolution (blurring). In order to remove random noise
with the minimum degradation of resolution, various edge-preserved filters
have been developed such as the adaptive median filter.
Digital image filtering is useful for enhancing lineaments that may represent
significant geological structures such as faults, veins or dykes. It can also
enhance image texture for discrimination of lithology and drainage patterns.
Digital image filtering is used for the land use studies, highlighting textures of
urban features, road systems and agricultural areas.
• image fusion.
29
Processing and 12.3.1 Arithmetic Operations
Classification of Remotely
Sensed Images One of the techniques of image transformation is to apply simple arithmetic/
logical operations to the image data. Arithmetic operations of addition,
subtraction, multiplication and division can be performed on two or more
images of the same geographical area. The prerequisite for the operation is that
all input images should be co-registered. These input images may either be
individual bands from images acquired at different dates or separate image
bands from a single multispectral image.
Fig. 12.6: Example of image subtraction operation showing the resultant image and its
new values obtained from subtraction of image B from image A
Image division or band ratio is probably the widely used arithmetic operation.
Band ratio is an image transformation technique that is applied to enhance the
contrast between features by dividing DN for pixels in one image band by DN
of pixels in the other image band. Band ratioing technique is generally applied
on a multispectral image. Ratio is an effective technique for selectively
enhancing spectral features as shown in Fig. 12.7. Ratio images derived from
different band pairs are often used to generate ratio colour composites in a
Red, Blue and Green (RGB) display. Many indices, such as the normalised
difference vegetation index (NDVI) have been developed based on both
differencing and ratio operations.
(a) (b)
30
Image Enhancement and
Transformation
(c) (d)
Fig. 12.7: True colour and band ratio images. (a) Landsat TM true colour composite
(RGB); (b) Thematic Mapper band (TM) 3/TM4: This ratio has defined
barren lands and urban area uniquely but it could not define water body,
forests and croplands and (c) TM2/TM3: this ratio has distinguished
croplands, barren lands sharply but it has not separated croplands, forests
and water body. Both forests and water body have appeared as lighter tone
and barren land appeared as dark tone and it did not enhance urban area.
Since chlorophyll has strong reflectance in the band 2 (0.52 -0.60 um) region
and strong absorption in the band 3 (0.63 -0.69 um) region, vegetation has
appeared as brighter tone and (d) TM3/TM5: This ratio enhanced barren
lands, highways, street patterns within the urban areas and urban built-up or
cemented areas. It could not enhance clear water but it enhanced turbid water.
This ratio is useful for observing differences in water turbidity. Barren lands,
highways, urban and built-up areas have appeared as lighter tone and forests,
water body and croplands appeared as dark tone
Band ratio is an effective technique for suppressing topographic shadows. For
a given incident angle of solar radiation, the radiation energy received by a
land surface depends on the angle between the land surface and incident
radiation (Fig. 12.8). Therefore, solar illumination on a land surface varies
with terrain slope and aspect, which results in topographic shadows. In a
remotely sensed image, the spectral information is often occluded by sharp
variations of topographic shadowing. DNs in different spectral bands of a
multispectral image are proportional to the solar radiation received by the land
surface and its spectral reflectance.
Fig. 12.8: Ratio of pixel values in NIR region to the corresponding pixel value in the
visible red region of the spectrum. The ratios for the illuminated and shaded
slopes are very similar, although pixel value differs by a factor of more than
two. Hence, an image made up of NIR:R ratio values at pixel positions will
exhibit a much reduced shadow or topographic effect (source: redrawn after
Liu and Mason, 2009) 31
Processing and Using combination of band ratio and subtraction techniques several indices
Classification of Remotely
Sensed Images
have been developed which are widely used for geological, ecological and
other applications. Of the several indices, the vegetation indices such as
normalised differential vegetation index (NDVI) are more popular. Vegetation
index (VI) is a number that is generated by some combination of remote
sensing bands and may have some relationship to the amount of vegetation in
a given image pixel. The index is computed using several spectral bands that
are sensitive to chlorophyll concentration and photosynthetic activity.
Of the many vegetation indices we will discuss in detail about NDVI, which is
most widely used. NDVI is the acronym for Normalised Difference Vegetation
Index. It is a numerical indicator that uses visible and NIR bands of the
electromagnetic (EM) spectrum. It is mainly adopted to analyse remote
sensing measurements and assess the ‘greenness’ of the target. NDVI is a very
important tool to study vegetation from remote sensing satellite as most of the
multispectral satellite sensors have visible and infrared channels, which can be
utilised to calculate NDVI. For example, NDVI can be calculated using the red
band and infrared band of LISS III data, by applying the formula given below:
NDVI works in the following ways. The pigment in plant leaves, chlorophyll
strongly absorbs visible light (from 0.4 to 0.7 μm) for use in photosynthesis.
The cell structure of leaves, on the other hand, strongly reflects NIR light
(from 0.7 to 1.1 μm). The more leaves a plant has, more these wavelengths of
light are affected, respectively. We can measure intensity of light coming off
the Earth in visible and NIR wavelengths and quantify photosynthetic capacity
of vegetation in a given pixel of land surface. In general, if there is much more
reflected radiation in NIR wavelengths than in visible wavelengths then
vegetation in that pixel is likely to be dense and may contain some types of
forest. If there is very little difference in the intensity of visible and NIR
32
wavelengths reflected, then vegetation is probably sparse and may consist of Image Enhancement and
Transformation
grassland, tundra, or desert. Since we know the behaviour of plants across EM
spectrum, we can derive NDVI information by focusing on the satellite bands
that are most sensitive to vegetation information (NIR and red). Therefore,
bigger the difference between NIR and red reflectance, more vegetation there
has to be.
NDVI algorithm subtracts red reflectance values from NIR and divides it by
the sum of NIR and red bands as mentioned in the equation above.
This normalised index formulation allows us to cope with the fact that two
identical patches of vegetation could have different values if one were, for
example in bright sunshine and another under a cloudy sky. The bright pixels
would all have larger values and therefore, a larger absolute difference
between the bands. This is avoided by dividing by the sum of reflectance
values.
NDVI has found a wide application in vegetative studies as it has been used to
estimate crop yields, pasture performance and rangeland carrying capacities
among others. It provides a crude estimate of vegetation health and a means of
monitoring changes in vegetation over time (Fig. 12.9). It can be used to
detect seasonal changes in green biomass but can also be used to detect
changes to human activities (logging) or natural disturbances such as wild fire.
NDVI has been found to be useful for continental or global scale vegetation
monitoring because it can compensate for changing illumination conditions,
surface slope and viewing aspect.
Fig. 12.9: Healthy vegetation absorbs most of the visible light that hits it and reflects a
large portion of NIR light. Unhealthy or sparse vegetation reflects more visible
light and less NIR light
33
Processing and It is often directly related to other ground parameters such as percent of
Classification of Remotely
Sensed Images
ground cover, photosynthetic activity of the plant, surface water, leaf area
index and the amount of biomass.
The most popular and used satellite instrument for collecting NDVI is US
National Oceanic and Atmospheric Administration’s Advanced Very High
Resolution Radiometer (AVHRR) satellite. It is sensitive to wavelengths from
0.55 - 0.7 μm and 0.73 - 1.0 μm, both of which are idealised in NDVI
calculation. AVHRR’s detectors measure intensity of light being reflected from
the different bands. Landsat TM is also used to calculate NDVI but because its
band wavelengths differ (uses bands 3 and 4), it is most often used to create
images with greater detail covering less area. Moderate Resolution Imaging
Spectroradiometer (MODIS) sensor of NASA also has an NDVI standard
product.
Despite its several uses, there are some limitations of NDVI as listed below:
Temporal Resolution: It is important to get the right balance between
temporal resolution of data and time scale of variation in the quantity
measured e.g., maximum value composites of NDVI data are rarely less than
decadal. Consequently, NDVI should not be used to investigate short-term
events like forest fires. It is more appropriate to use NDVI data to monitor
longer-term events like the growth of vegetation through a season, or annual
rates of deforestation.
Land Cover Types: With the exception of very large irrigation schemes and
commercial agriculture, an NDVI pixel very rarely covers a single
homogeneous agricultural region. Instead it may cover roads, buildings, bare
soil, small water bodies, natural vegetation and agriculture, all within one
pixel. An NDVI pixel is the sum of the radiation reflected from all land cover
types within the area covered by the pixel. NDVI is an indicator of the
condition of overall vegetation in an area, including natural vegetation and
agriculture. In rain-fed agriculture, natural vegetation may follow similar
patterns to the agriculture. More often however, agriculture is more
susceptible to adverse conditions and follows different growth cycles. When
looking at NDVI, always remember that you are looking at general conditions
and not necessarily the condition of a specific crop.
Sparse Vegetation and Soil Type: Light reflected from the soil can have a
significant effect on NDVI values (changing the values by up to 20%).
Generally, greater the radiance reflected from the soil, lower NDVI values.
The soil type, therefore, is an important factor. Given two soil types, one with
a greater reflectivity but with similar vegetative conditions, soil with greater
reflectivity will on average produce lower NDVI values.
Off-Nadir Effects: As a satellite sensor scans across the Earth, there is only
one point, in the centre of the scan that is directly underneath the sensor
(called sub-point or nadir). The distance from the sensor to the ground
34
increases away from the sub-point. This results in increased atmospheric Image Enhancement and
Transformation
interference as the light must pass through more atmosphere before reaching
the sensor, and therefore reduced NDVI values.
......................................................................................................................
......................................................................................................................
......................................................................................................................
......................................................................................................................
......................................................................................................................
......................................................................................................................
......................................................................................................................
(a) (b)
(c)
Fig. 12.10: Illustration of image fusion (merging) of IKONOS and Landsat multispectral
image. (a) High spatial resolution IKONOS image, (b) Landsat multispectral
image of the same area and (c) Fused image having spatial information of
IKONOS image and spectral (colour) information of the Landsat image.
Note the improved clarity of features in the fused image
There are many methods which are used for image fusion. These methods can
be broadly divided into two categories - spatial and transform domain fusions.
Some of the popular spatial domain methods include intensity hue saturation
(IHS) transformation, Brovey method, principal component analysis (PCA)
and high pass filtering based methods. Wavelet transform, Laplace pyramid
and curvelet transform based methods are some of the popular transform
domain fusion methods.
In this unit, we will focus on the most widely used IHS transform method.
Intensity is the overall brightness of the scene and varies from black to white
and saturation represents the purity of colour and also varies linearly while hue
represents colour or dominant wavelength of pixel. Intensity (I) component is
similar to the panchromatic image. This peculiarity is used to produce fusion
between panchromatic data having high spatial resolution and multispectral
data characterised by a high spectral resolution and less spatial detail. The
procedure of IHS based image fusion can be summarised in the following
steps (Fig. 12.11):
Fig. 12.11: RGB-IHS encoding and decoding for image fusion (source: redrawn from
Schowengerdt, 1997)
The resultant fused image is a mixture of spectral information from the low-
resolution colour composite and high spatial resolution information from the
panchromatic image, which better shows the image features. Despite being
useful, the major limitation of these image fusion techniques is that they can
distort spectral information of multispectral data while merging. Another
major disadvantage of fusion approaches is that they produce spatial distortion
in the fused image. Spectral distortion becomes a negative factor while we go
for further processing, such as classification.
37
Image Enhancement and
UNIT 13 IMAGE CLASSIFICATION Transformation
Structure
13.1 Introduction
Objectives
13.2 Concept of Image Classification
Approaches to Classification
Stages in Classification
13.3 Unsupervised Classification
K-Means Clustering
ISODATA Clustering
13.4 Supervised Classification
Parallelepiped Classifier
Maximum Likelihood Classifier
13.5 Signature Evaluation
Spectral Signature
Ways of Signature Evaluation
13.6 Overview of Other Classification Methods
13.7 Selection of An Appropriate Classification Method
13.8 Summary
13.9 Unit End Questions
13.10 References
13.11 Further/Suggested Reading
13.12 Answers
13.1 INTRODUCTION
In the previous unit, you have learnt about different image enhancement and
transformation techniques which help us better visualise and interpret
remotely sensed images. As mentioned earlier, all these techniques aid in
providing only better visual information and, therefore, have limited utility.
In this unit, we will move a step further and learn how to make more sense of
the landscape by dividing it into separate classes based on surface
characteristics. This process is known as image classification. It involves
conversion of raster data into finite set of classes that represent surface types
in the imagery. It may be used to identify vegetation types, anthropogenic
structures, mineral resources, etc. or transient changes in any of these features.
Additionally, classified raster image can be converted to vector features (e.g.,
polygons) in order to compare with other data sets or to calculate spatial
attributes (e.g., area, perimeter, etc). Image classification is a very active field
of study broadly related to the field of pattern recognition. In this unit, we will
discuss about different image classification methods, signature evaluation and
the guidelines for selecting a classification method.
41
Processing and Objectives
Classification of Remotely
Sensed Images After studying this unit, you should be able to:
• define image classification;
• describe different image classification approaches and algorithms used in
remote sensing;
• discuss relative advantages and limitations of commonly used
classification algorithms; and
• explain how to evaluate spectral signatures.
43
Processing and There are no specific rules regarding the number of training sites per class but
Classification of Remotely
Sensed Images
it is advisable to take several training sites for each class to be mapped. If you
take very less number of training sites then it may be difficult to obtain a
A general rule of thumb is spectral signature which truly represents that class and if you take large
that training data for a
number of training sites then a significant time may be getting wasted in
class should be 10 x n
where, n is the number of collecting and evaluating signatures with significantly improving the final
bands. You should also signature.
remember that minimum
number of pixels in a Signature Evaluation is the checking of spectral signatures for their
training site for a class representativeness of the class they attempt to describe and also to ensure a
should be n +1 (Jensen,
minimum of spectral overlap between signatures of different classes. We shall
1986).
discuss in detail about signature evaluation in section 13.5.
Decision Making is the process of assigning all the image pixels into thematic
classes using evaluated signatures. It is achieved using algorithms, which are
known as decision rules. The decision rules set certain criteria. When
signature of a candidate pixel passes the criteria set for a particular class, it is
assigned to that class. Pixels failing to satisfy any criteria remain unclassified.
We shall discuss in detail about the decision rules in the next two sections.
The term classifier is
widely used as a synonym
of the term decision rule. Training Signature Evaluation Decision making
Algorithm Find inherent Pixel classification Spectral class Analyst labels clusters
clusters data classes based on clusters map (may involve grouping
of clusters)
Informational
class map
The data point at the center of a cluster is known as a centroid. In most of the
image processing software, each centroid is an existing data point in the given
input data set, picked at random, such that all centroids are unique. Initially, a
randomised set of clusters are produced. Each centroid is thereafter set to the
arithmetic mean of cluster it defines. The process of classification
and centroid adjustment is repeated until the values of centroids stabilise. The
final centroids are used to produce final classification or clustering of input
data, effectively turning set of initially anonymous data points into a set of
data points, each with a class identity.
Advantage
• the main advantage of this algorithm is its simplicity and speed which
allows it to run on large datasets.
Disadvantages
• it does not yield the same result with each run, since the resulting clusters
depend on the initial random assignments
• it is sensitive to outliers, so, for such datasets k-medians clustering is used Outliers in remote sensing
images represent observed
and pixel values that are
• one of the main disadvantages to k-means is the fact that one must specify significantly different from
their neighbourhood pixel
the number of clusters as an input to algorithm.
values.
13.3.2 ISODATA Clustering
ISODATA (Iterative Self-Organising Data Analysis Technique) clustering
method is an extension of k-means clustering method (ERDAS, 1999). It
represents an iterative classification algorithm and is useful when one is not
sure of the number of clusters present in an image. It is iterative because it
makes a large number of passes through the remote sensing dataset until
specified results are obtained. Good results are obtained if all bands in remote
sensing image have similar data ranges. It includes automated merging of
similar clusters and splitting of heterogeneous clusters.
The clustering method requires us to input maximum number of clusters that
you want, a convergence threshold and maximum number of iterations to be
performed. ISODATA clustering takes place in the following steps:
45
Processing and • k arbitrary cluster means are established
Classification of Remotely
Sensed Images • all pixels are relocated into the closest clusters by computing distance
between pixel and cluster
• centroids of all clusters are recalculated and above step is repeated until
the threshold convergence and
• if the number of clusters are within the specified number and distances
between the clusters meet a prescribed threshold, then only clustering is
considered complete.
Advantages
• it is good at finding “true” clusters within the data
• it is not biased to the top pixels in the image
• it does not require image data to be normally distributed and
• cluster signatures can be saved, which can be easily incorporated and
manipulated along with supervised spectral signatures.
Disadvantages
• it is time consuming and
• it requires maximum number of clusters, convergence threshold and
maximum number of iteration as an input to algorithm.
Fig. 13.3: Using supervised classification, pixels are classified into different categories
46
This is based on the statistics of training areas representing different ground Image Classification
objects (Fig. 13.4) selected subjectively by users on the basis of their own
knowledge or experience. Classification is controlled by users’ knowledge but,
on the other hand, is constrained and may even be biased by their subjective
view. Classification can, therefore, be misguided by inappropriate or
inaccurate training, area information and/or incomplete user knowledge. Steps
involved in supervised classification are given in Fig. 13.5.
47
Processing and
Classification of Remotely
Sensed Images
Fig. 13.6: Using the parallelepiped approach, pixel 1 is classified as forest and pixel 2 is
classified as urban
Advantages
• it is a simple and computationally inexpensive method and
• it does not assume a class statistical distribution and includes class
variance.
Disadvantages
• it is least accurate method
• it does not adapt well to elongated (high-covariance) clusters
• it often produces overlapping classes, requiring a second classification
step
• it also becomes more cumbersome with increasing number of channels
and
• pixels falling outside the defined parallelepiped remain unclassified.
Advantages
• it is one of the most accurate methods
• it overcomes unclassified pixel problem (subject to threshold values)
• it provides a consistent way to separate pixels in overlap zones between
classes and
• assignment of pixels to classes can be weighted by prior knowledge of the
likelihood that a class is correct.
Disadvantages
• cluster distributions are assumed to be Gaussian in each class and band.
Algorithm requires enough pixels in each training area to describe a
normal population and assumes class covariance matrices are similar
• classes not assigned to training sets tend to be misclassified – a particular
problem for mixtures
• it is reliant on the accuracy of training data. Changes in training set of any
one class can affect the decision boundaries with other classes
• it is relatively computationally expensive and
• it is also not practical with imaging spectrometer data.
49
Processing and Objectives
Classification of Remotely
Sensed Images After studying this unit, you should be able to:
• define accuracy assessment;
• discuss need for accuracy assessment;
• generate a error matrix for interpreted outputs;
• explain the role of sampling size in accuracy assessment; and
• list measures for accuracy assessment.
14.2.1 Definition
Accuracy defines Accuracy is referred to in many different contexts. In the context of image
correctness and it interpretation, accuracy assessment determines the quality of information
measures the degree of
derived from remotely sensed data. Assessment can be either qualitative or
agreement between a
standard that assumed to quantitative. In qualitative assessment, you determine if a map ‘looks right’ by
be correct and a map comparing what you see in the map or image with what you see on the ground.
created from an image. A However, quantitative assessments attempt to identify and measure remote
visually interpreted map or sensing based map errors. In such assessments, you compare map data with
classified image is only
ground truth data, which is assumed to be 100% correct.
said to be highly accurate,
when it corresponds Accuracy of image classification is most often reported as a percentage correct
closely with the assumed and is represented in terms of consumer’s accuracy and producer’s accuracy.
standard.
The consumer’s accuracy (CA) is computed using the number of correctly
classified pixels to the total number of pixels assigned to a particular category.
60 It takes errors of commission into account by telling the consumer that, for all
areas identified as category X, a certain percentage are actually correct. The Accuracy Assessment
producer’s accuracy (PA) informs the image analyst of the number of pixels
correctly classified in a particular category as a percentage of the total number The term consumer’s
of pixels actually belonging to that category in the image. Producer’s accuracy accuracy is used when a
measures errors of omission. classified image is
examined from the user’s
14.2.2 Need for Accuracy Assessment point of view. Producer’s
accuracy is used when
The need for assessing accuracy of a map generated from any remotely sensed same is viewed from
product has become a universal requirement and an integral part of any analyst’s perspective.
classification project. The user community needs to know accuracy of the
classified image data being used. Moreover, different projects have different
accuracy requirement and only those classified images which are above a
certain level of accuracy can be used. Furthermore, accuracy becomes a
critical issue while working in a Geographical Information System (GIS)
framework where you use several layers of remotely sensed data. In such
cases, it would be very important to know the overall accuracy which is
dependent upon knowing the accuracy of each of data layers.
There are a number of reasons why assessment of accuracy is so important.
Some of them are given below:
• accuracy assessment allows self-evaluation and to learn from mistakes in
the classification process
• it provides quantitative comparison of various methods, algorithms and
analysts and
• it also ensures greater reliability of the resulting maps/spatial information
to use in decision-making process.
The need for accuracy assessment is emphasised in literature as well as in
anecdotal evidence. For example, maps of wetlands from various states of
India (e.g., Jammu and Kashmir, Rajasthan, Tamil Nadu, West Bengal) have
been made by several central, state and local agencies using techniques that
included satellite images, aerial photographs and field data. Simply comparing
the various wetland maps would yield little agreement about location, size and
extent of these. In the absence of a valid accuracy assessment you may never
know which of these maps to use.
A map using remotely sensed or other spatial data cannot be regarded as the
final product without taking necessary steps towards assessing accuracy or
validity of that map.
A number of methods exist to investigate accuracy/error in spatial data
including visual inspection, non-site-specific analysis, generating difference
images, error budget analysis and quantitative accuracy assessment.
Fig. 14.1: (a) Non-site-specific accuracy in which two images are compared based on
their total areas. Note that the area of image 1(i.e. A+B+C) is equal to the area
of image 2 (i.e. A+B+C) and (b) site-specific accuracy in which two images are
compared on a site-by-site (i.e. cell-by-cell or pixel by pixel) (source: modified
62 from Campbell, 1996)