Thermal Remote Sensing:

In thermal remote sensing: we measure the radiations 'emitted' from the surface of the
target, as opposed to optical remote sensing where we measure the radiations 'reflected' by
the target under consideration.

Thermal properties of materials:

1. Emissivity of materials:
Emissivity: Emissivity, e, is the ratio between the radiant flux exiting a real-world
selective radiating body (Fr) and a blackbody at the same temperature (Fb):
𝐹𝑟
e = 𝐹𝑏
 All selectively radiating bodies have emissivities ranging from 0 to <1 that
fluctuate depending upon the wavelengths of energy being considered.
 Some materials like distilled water have emissivities close to one (0.99) over the
wavelength interval from 8 - 14 mm. Others such as polished aluminum (0.08)
and stainless steel (0.16) have very low emissivities.
 Two rocks lying next to one another on the ground could have the same true
kinetic temperature but have different apparent temperatures when sensed by a
thermal radiometer simply because their emissivities are different. The
emissivity of an object may be influenced by a number factors, including: colour,
moisture content, surface roughness, compaction, viewing angle.
2. Thermal capacity (c) is the ability of a material to store heat. It is measured as the
number of calories required to raise a gram of material (e.g. water) 1 ˚C (cal g-1 ˚C-1).
Water has the highest thermal capacity (1.00). It stores heat very well relative to all the
other materials.
𝑄
C =𝑚∗𝑇

where Q = amount of energy m = mass of substance T = Difference in final & initial

temperature

3. Thermal conductivity (K): The rate at which heat will pass through a specific thickness
of a substance, measured as the calories delivered in 1 second across a 1 centimeter
square area through a thickness of 1 cm at a temperature gradient of 1 degree
Centigrade (units: calories per centimeter per second per degree Centigrade). The
conductivity of a material is variable due to soil moisture and particle size. Many rocks
and soils are extremely poor conductors of heat.
4. Thermal inertia (P) is a measurement of the thermal response of a material to
temperature changes and is measured in calories per square centimeter per second
 square root per degree Celsius (cal cm-2 sec -1/2 ˚C-1). Thermal inertia is computed using
the equation:
P = (K x p x c)1/2

where K is thermal conductivity, p is density (g cm-3), and c is thermal capacity.

Density is the most important property in this equation because thermal inertia generally
increases linearly with increasing material density.

The resistance of a material to temperature change, indicated by the time dependent

variations in temperature during a full heating/cooling cycle (a 24-hour day for the Earth);

Applications of Thermal infrared remote sensing

i) Geology: Denser rocks such as basalt and sandstone have higher thermal inertia and on
night thermal IR images they show warm signature compared to less dense rocks such as
siltstone, cinders etc. Hence differentiation of rock types is possible.

ii) Military: Unusual concentration of troops, weaponry, military vehicles, jungle trails etc.
can be identified by the tonal changes in the thermal imagery.

iii) Hydrology: Hot industrial effluents that pollute water may be detected and the
diffusion pattern may be mapped.

iv) Agriculture: Thermal images have been utilized for the identification of crop species and
soil types, for detecting crop diseases

v) Botany: Such knowledge is useful for assessing plant health, age, relative water supply or
degree of irrigation.

vi) Forestry: Forest fires beneath a forest cover story can be located with thermal IR. During
extensive fires, damage caused can be assessed even under excessive smoke conditions.

Thermal data sets: LANDSAT and ASTER

LANDSAT

Dynamic surface water extent, Burned area, and Fractional snow cover.

ASTER (Advanced Space-borne Thermal Emission and Reflection Radiometer)

Global Digital Elevation Model (ASTER GDEM), Global Water Bodies Database, Land surface
temperature
 MICROWAVE REMOTE SENSING

Analyzing the information collected by the sensors that operate in the microwave portion of
the electromagnetic spectrum is called as Microwave Remote Sensing. Microwave portion
of the spectrum includes wavelengths within the approximate range of 1mm to 1m. These
longer waves have the capability of penetrating through the clouds thus overcoming the
atmospheric effects that we often encounter in the visible and infrared regions of EMR.

There are two distinctive features that characterize microwave energy from a remote
sensing standpoint:

1. Microwaves are capable of penetrating the atmosphere under virtually all conditions.
Depending on the wavelengths involved, microwave energy can “see through” haze, light
rain and snow, clouds, and smoke.

2. Microwave reflections or emissions from earth materials bear no direct relationship to

their counterparts in the visible or thermal portions of the spectrum. For example, surfaces
that appear “rough” in the visible portion of the spectrum may be “smooth” as seen by
microwaves. In general, microwave responses afford us a markedly different “view” of the
environment—one far removed from the views experienced by sensing light or heat.

There are two types of microwave remote sensing; active and passive. The active type
receives the backscattering which is reflected from the transmitted microwave which is
incident on the ground surface. Synthetic aperture radar (SAR), microwave scatterometers,
radar altimeters etc. are active microwave sensors. The passive type receives the microwave
radiation emitted from objects on the ground. The microwave radiometer is one of the
passive microwave sensors.
 DIGITAL IMAGE PROCESSING

Based upon the mechanism used in detecting and recording, the remotely sensed data
products may be broadly classified into two types: Photographic Images, Digital Images

Analog:Photographs are acquired in the optical regions of electromagnetic spectrum, i.e. 0.3
– 0.9 µm. Four different types of light sensitive film emulsion bases are used to obtain
photographs. These are black and white, colour, black and white infrared and colour
infrared. However, in aerial photography black and white film is normally used. Photographs
may be enlarged to any extent without loosing information contents or the contrast.

Digital Images: A digital image consists of discrete picture elements called pixels. Each one
of the pixels in an image has an intensity value and an address in two-dimensional image
space. A digital number (DN) represents the average intensity value of a pixel. It is
dependent upon the electromagnetic energy received by the sensor and the intensity levels
used to describe its range.

Types of image interpretation:

Elements of Visual Interpretation:

1. Tone: - refers to the relative brightness or colour of objects in an image. Generally,

tone is the fundamental element for distinguishing between different targets or
features. Variations in tone also allows the elements of shape, texture, and pattern
of objects to be distinguished.
2. Size: scale of an image is another important characteristic of individual objects. It
helps in distinctively identifying the industrial complexes with residential dwellings.
3. Shape: an outline of an individual object provides important clues in the
interpretation of remote sensing images. The shape of some of the objects is so
distinctive that make them easy to identify. For example, the shape of the Sansad
Bhawan is typically distinct from many other built-up features. Similarly, a railway
 line can be readily distinguished from a road due to its long continuous linearity in
shape
4. Texture: The texture refers to the minor variations in tones of grey or hues of colour.
For example, dense residential areas in a large city form fine texture due to the
concentration of the houses in a smaller area and the low-density residential areas
produce a coarse texture. Similarly, in high resolution images the sugarcane plants
produce coarse texture in comparison to the fine texture of rice or wheat plants.

5. Shadow: is also helpful in interpretation as it may provide an idea of the profile and
relative height of a target or targets which may make identification easier. For
example, the Qutub Minar located in Delhi, minarets of mosques, overhead water
tanks.

6. Pattern: Spatial arrangements of visibly discernible [Link] arrangements can

easily be identified from the images through the utilisation of the pattern they form.
For example, planned residential areas with the same size and layout plan of the
dwelling units in an urban area can easily be identified if their pattern is followed.
7. Association: The association refers to the relationship between the objects and their
surroundings along with their geographical location. For example, an educational
institution always finds its association with its location in or near a residential area as
well as the location of a playground within the same premises.

Digital Interpretation: Digital interpretation facilitates quantitative analysis of digital data

with the help of computers to extract information about the earth surface. Digital
interpretation is popularly known as ‘Image Processing’. Image processing deals with image
correction, image enhancement and information extraction.

Categories of Digital image processing

1. Image rectification and restoration (pre-processing): These operations aim to

correct distorted or degraded image data to create a more faithful representation of
the original scene and to improve an image’s utility for further manipulation.
2. Image enhancement: methods applied for effective display of image to facilitate
visual interpretation. Image enhancement involves techniques for heightening the
visual distinctions among features in a scene, ultimately increasing the amount of
information that can be interpreted from the data.
3. Image classification: Automatic categorization of pixels into various land-cover
classes. Though digital image interpretation has the capability to classify earth
surface features with the help of a computer, it must be supplemented with ground
truthing, that is, verification of the interpreted regions with actual information
gathered from the ground by field survey.
 4. Data merging and GIS integration: These procedures are used to combine image
data for a given geographic area with other geographically referenced data sets for
the same area.
5. Biophysical modelling: remotely sensed data are often used along with GIS
techniques to facilitate environmental modelling. Eg to estimate varied parameters
as crop yield, water depth, pollution concentration.
6. Image transmission and compression: it refers to compression of increasingly high
volume of data available from remote sensing system for their distribution and
transmission over the net.

Preprocessing: (Image rectification and restoration)

Pre-processing functions involve those operations that are normally required prior to the
main data analysis and extraction of information, and are generally grouped as radiometric
or geometric corrections. These methods correct the data for distortions or degradations
that stem from the image acquisition process and for the faithful representation of the
scene recorded in that image.

Geometric corrections: Raw digital images usually contain geometric distortions so

significant that they cannot be used directly as a map base without subsequent processing.
The sources of these distortions range from variations in the altitude and velocity of the
sensor platform to factors such as panoramic distortion, earth curvature,atmospheric
refraction, relief displacement, and nonlinearities in the sweep of a sensor’s IFOV.
Geometric correctionsis carried out by geo-referencing or geometric registration.

Radiometric corrections: If remotely sensed images taken at different times of the year are
being compared, it is usually necessary to apply a sun elevation correction and an earth–
sun distance correction.

The sun elevation correction accounts for the seasonal position of the sun relative to the
earth. The correction is usually applied by dividing each pixel value in a scene by the sine of
the solar elevation angle for the particular time and location of imaging.

The earth–sun distance correction is applied to normalize for the seasonal changes in the
distance between the earth and the sun. The earth–sun distance is usually expressed in
astronomical units. The irradiance from the sun decreases as the square of the earth–sun
distance.

Ignoring atmospheric effects, the combined influence of solar zenith angle and earth–sun
distance on the irradiance incident on the earth’s surface can be expressed as

E= E0cosθ0/d2

Where,
E = normalizedsolarirradiance

E0= solarirradianceat meanearth–sundistance

θ0= sun’sanglefromthezenith

d =earth–sundistance, inastronomicalunits

Noise Removal:

Image noise is any unwanted disturbance in image data that is due to limitations in the
sensing, signal digitization, or data recording process. The potential sources of noise range
from periodic malfunction of a detector. Noise can degrade or totally mask the true
radiometric information content of a digital image.

For example, multispectral sensors that sweep multiple scan lines simultaneously often
produce data containing systematic striping or line drop. In this situation, a number of
adjacent pixels along a line (or an entire line) may contain spurious DNs, often values of 0 or
“no data.” This problem is normally addressed by replacing the defective DNs with the
average of the values for the pixels occurring in the lines just above and below.

Image enhancement technique:

The primary goal of image enhancement is to improve the visual interpretability of an image
by increasing the apparent distinction between the features in the scene. Enhancement
operations are normally applied to image data after the appropriate pre-processing steps
have been performed. Noise removal, in particular, is an important precursor to most
enhancements. The most commonly applied digital enhancement techniques.

1. Contrast manipulation. Gray-level thresholding, level slicing, and contrast stretching.

2. Spatial feature manipulation. Spatial filtering, edge enhancement, and Fourier analysis.

3. Multi-image manipulation. Multispectral band ratioing and differencing, vegetation and

other indices, principal components, intensity–hue–saturation (IHS), and decorrelation
stretching.

Contrast manipulation

Gray-Level Thresholding

Gray- level thresholding is used to segment an input image into two classes one for those
pixels having DN values below an analyst defined value and another for those above this
value. Thus, thresholding on an image creates a binary image contains pixels of only two DN
values, 0 and 255. In following example (Fig. 11.6.), let threshold value 185 is applied. In the
input image, all the pixels having DN values less than 185 will appear pure black (DN
value=0) and all the pixels having DN values greater than 185 will appear white (DN
value=255) in the output image.

Density or Level Slicing

Level slicing is similar to grey level thresholding. In grey level thresholding, the input image
pixels are segmented in two groups lower and higher of a defined DN value or level, in case
of level slicing a number of levels (m) are used, which segmented the image in a number of
groups (m+1). Thus, in density slicing or level slicing, the image histogram is divided into an
analyst specified intervals or slices. All of the DN values in the input image falling within a
specified are then displayed at a single DN in the output image. Consequently, if four slices
are established, the output image will contain four different grey levels.

Application of density slice reduces the details of an image, but it also reduces the noise in
the image. This is useful in highlighting the features in the imagery.

Contrast Stretching

One of the most useful method of contrast enhancement is contrast stretching. 8 bit image
data contains 256 (=28) levels in grey scale of DN values. Often remote sensing sensors fail
to collect reflectance/radiation covering the entire 256 levels, normally confined to a
narrow range of DN values, which produce a low contrast image. Let an input image have
brightness values between 60 and 158. Contrast stretching operation expands the narrow
range (60 to 158) of DN values of the input image to the full range (0 to 255) of DN values in
the output image.

Contrast stretch is categorized as a) linear and b) non-linear.

Spatial Filtering

Image filtering:

Spatial image enhancement is used to increase the image quality through filtering. Spatial
frequency is the rate of change of image values or DN values in the scene in a unit distance.
Filtering operation emphasizes or de-emphasizes this spatial frequency for better image
interpretation. Thus filters highlight or suppress specific features based on their spatial
frequency hence it depends on the image value of neighbour’s. Different kinds of filters are
explained below.

Convolution filter: In convolution filtering a window also known as operators or kernels

having odd numbers of pixels (e.g., 3x3, 5x5, 7x7) used to move throughout the input image,
the DN value of the output image is obtained by a mathematical manipulation. A
convolution filter is classified as low-pass filter and high-pass filter.

Low-pass filter: A low-pass filter emphasizes low frequency or homogeneous areas and
reduces the smaller details in the image. It is one of the methods of smoothing operation. It
reduces the noise in the imagery. Low-pass filter normally sub-divided as average (mean),
mode and median filter.

Average filter: In case of average or mean filter the output DN value is obtained by
multiplying each coefficient of the kernel with the corresponding DN in the input image and
adding all them all and dividing by the sum of kernels value. The coefficients of the kernel
range from equal value to different weighted value.

Mode filter: A kernel is fitted in the input image, and computes the mode value within this
kernel. The resultant value is the final value of the center pixel of the kernel.

Median filter: Median filter is similar to mode filter. In this case the median value of the
kernel is computed which replaces the center pixel DN value. Median filter is very useful for
removing random noise, salt-pepper pattern noise, speckle noise in RADAR imagery.

High-pass filter: High-pass filter is the opposite of low-pass filter. In contrast to low-pass
filter, high-pass filter enhance the variations in the imagery. It increases the spatial
frequency, thus sharpens the details in the imagery. High frequency filter works in the
similar way of average filter, uses a kernel may be normal or weighted. Another way of
getting the output of high-pass filtering is subtraction of the low-pass filtered image from
the original image. The following 5x5 image explains the operation of low pass and high pass
filters.
Multispectral image classification:

The overall objective of image classification procedures is to automatically categorize all

pixels in an image into land cover classes or themes. Often this is done using spectral
patterns; that is, pixels that share similar combinations of spectral reflectance or emissivity
are grouped together in classes that are assumed to represent particular categories of
surface features. There are two types of classification i.e., supervised and unsupervised
classification.
Supervised:

Supervised classification is based on the idea that a user can select sample pixels in an
image that are representative of specific classes (land cover/land use) and then direct the
image processing software to use these training sites as references for the classification of
all other pixels in the image. Training sites are selected based on the knowledge of the user.
Training fields are areas of known identity delineated on the digital image.

Unsupervised
Unsupervised classification (commonly referred to as clustering) is a method of extracting
land-cover information.
Compared to supervised classification, unsupervised classification normally requires only a
minimal amount of initial input from the analyst. This is because clustering does not
normally require training data. The process where numerical operations are performed that
search for natural groupings of the spectral properties of pixels. The clustering process
results in a classification map consisting of ‘m’ spectral classes. The analyst then assigns or
transform the spectral classes into thematic information classes of interest (e.g., forest,
agriculture). In unsupervised classification, the software does most of the processing on its
own generally resulting in more categories than the user is interested in.