• In terms of spatial resolution higher resolution means smaller pixel size, coarser or lower resolution

means larger pixel size.

• Sensor basically records the radiation reflected by each pixels. Radiation is recorded in different

wavelength. Sensor is also called detector.

• The capability of the detector governs or influences it’s resolution. Depending upon the resolving

capability of detector which is called instantaneous field of view, the size of the pixel on the ground

is determined. In other words the size of a pixel on the ground is determined by the resolving power

of the detector.

• The instantaneous field of view (IFOV) is the solid angle subtended by the pixel at the sensor.

• In case the pixel is rectan gular instead of a square we have to write say 10m x 15m and not 150 sq

meter. In case of square we simply write the pixel size as 1 dimensional. Eg 10m = 10m x 10m

As is clearly visible, smaller the pixel size greater the

spatial resolution.

Spectral Resolution

The electromagnetic radiation emitted by each pixel can fall in a broad range of wavelength, so

depending upon the sensor capability, the measurement of specific range of wavelength is called

spectral resolution.

In other words spectral resolution is number of bands which the

sensor or the detector can detect.

In the example on the left the width of the 1st band 0.3 - 0.7 is 0.4, and in

the second band 0.3-0.5 the width is 0.2.

• Now for spectral resolution, lower is the width, higher is the spectral resolution and vice versa.

 Now, each band represent a range of

wavelength and has specific

application potential i.e. provides us

with some specific information about

the ground object. For example blue

band will be able to map the water

bodies, and helps to differentiate soil or

vegetation from water bodies.

Radiometric Resolution

It is the ability of the sensor to differentiate between very slight difference in energy. The radiometric

resolution is expressed in bits. For example 2bits resolution , 3bits resolution or 4bits resolution.

1bit = 2 ^ 1 = 0, 1

2bits = 2 ^ 2 = 0 ,1, 2, 3

3bits = 2 ^3 = 0, 1, 2, 3, 4, 5, 6, 7

4bits = 2 ^ 4 = 0, 1, 2, …. , 15

8bits = 2 ^ 8 = 0, 1, 2, ……, 255

The pixel having the value 0 means pixel with lowest reflectance, and the pixel with 2^n - 1 is the pixel

with highest reflectance. For example in spectral resolution of 2bits, we have 0, 1, 2, 3 as the

reflectance value. Here 3 represents the pixel with highest reflectance value, and 0 represents the pixel

with highest reflectance value.

• The greater the radiometric resolution of a sensor, the more sensitive it is to detecting small

differences in reflected or emitted energy.

Temporal Resolution

It refers to the time taken by the satellite to revisit the same scene on the Earth’s surface, in other

words, it refers to the time between two successive image of the same location on the Earth’s surface.

Greater (Finer) the temporal resolution, lesser the revisit period (i.e. time to visit the same

location on the surface) and vice versa.

Lower temporal resolution is also called coarser temporal resolution.

Note : Satellite have 3 basic orbit - Polar Orbit,

Equitorial Orbit and Inclined Orbit.

Geo-Stationary satellite focuses on the some

specific location throughout the duration of its

operation. This is because the angular velocity of

Earth’s rotation and satellite’s revolution are same.

The satellite that are geo-stationary about the equator

are also called Equatorial Satellite.

For Polar Satellites

• For polar satellite, Earth revolves from west to east

and these satellite revolves from north to south.

Since Earth is not a perfect

sphere the height of

satellite will vary hencethe

use
of nominal

The orbit of polar satellite is synchronised in such a manner that wherever the satellite is

corresponding to Earth, the local crossing time of the satellite it approximately same through the

duration of its mission. This is what the mean solar time represents in the above picture. The

9.45 am time remains same throughout the duration of the mission of this satellite.

Image Storage and Data Retrieval

There are 3 formats for Data Storage

• BSQ - Band Sequential

• BIL - Bands Interleaved by Line

• BIP - Bands Interleaved by Pixel

Note : Data stored in a format must be retrieved in that particular format.

In BSQ, Col1 is the Pixel1 of Line1, similarly Col2 is the Pixel2 of Line1, and it goes to n. That is each

line is showing the data of each pixel. Here Bands are arranged in sequential manner.

In BIL, here again Col1 is Pixel1, Col2 is Pixel2 and so on. The only difference is that here lines are

arranged sequentially. Here Line1 contains all the bands and so does all the other lines, hence the

name bands interleaved by lines. Here only after Line1 has stored all the bands then Line2 starts

storing bands. Here each Line represents a scene(location) on the ground.

In BIP, under each pixel P1 B1, P1 B2, P1 B3,… fills first. That is we move in this way

For Line1 - > For Pixel1 -> B1, B2, … . Then for Pixel2 -> B1, B2… and so on.

Once this is completed then we start to fill Line2.

For example, If we have 20 bands and we know that we can retrieve the data from 3 Bands, then BSQ

is the format we will use, If we want to know data about a particular scene(location) then we will use

BIL.

Geometric Error Correction

The term geometric means relating to the size and the shape of the pixel.

• Remotely sensed imagery exhibits internal and external geometric errors.

• These errors can be systematic(predictable) or nonsystematic(random).

• Systematic geometric errors are easier to identify and correct than random geometric error.
Here Internal errors means error related to malfunctioning of the sensor itself. External errors

means error not relating to the sensor, i.e. error generated due to external causes like the

atmosphere, pressure etc.

When we say that there is a geometric errors essentially, it means the the shape and size of the pixel

has changed from what was desired. For example if the actual value is 25m x 25m the error may be

some deviation like 26m x 24m.

These are the distortions that follow from errors along a particular axis.

Following are the shape of the pixels that results from the given distortions.
The GCP in the above picture means Ground Control Point. It represents the correct geographic co-

ordinates. The purpose is now to implement a transformation process so that the correct coordinates

on the ground are imposed on the distorted images.

Spatial Rectification

The logic is then to fill a rectified

output matrix with values from an

unrectified input image matrix using

output-to-input (inverse) mapping

logic. Output-to-input inverse

mapping logic is the preferred

methodology because it results in a

rectified output matrix with values at

every pixel location.

SPATIAL INTERPOLATION LOGIC

In order to actually geometrically correct the original distorted image, a procedure

called resampling is used to determine the digital values to place in the new pixel

locations of the corrected output image.

The resampling process calculates the new pixel values from the original digital

pixel values in the uncorrected image. There are three common methods for

resampling:

• nearest neighbour

• bilinear interpolation

• cubic convolution
Radiometric Errors and Correction

• Detector Error (Internal Systematic Error)

• Atmospheric Errors

• Error generated due to conversion from Dan to radiance

• Error generated due to conversion from radiance to reflectance

• BRDF Error.
It’s Correction

1. Correction for detector errors

• Line Dropping

• Destriping

2. Atmospheric Corrections

• Histogram Adjustment

• Atomospheric Radiative transfer models

3. Conversion from DN to radiance correction

4. Conversion from radiance to reflectance corrections

5. BRDF correction.

Line Drop

When is sensor operates, it has a set of detectors arranged in form of panel of detectors. So a panel

may detect 10, 16 or more Lines from the Earth’s surface, Then the next lines depending upon the

sensor resolution.

When a rocket is launched, it may so happen that the array of detectors starts to malfunction and

does not record any information at all. As shown below in the image.
 For such Lines there will be no emmitance or

reflectance value recorded by the sensor. THese

Lines will have 0 values. To remove these zero values

We take the avg of the preceding and successive

Line Values. As shown below by taking averages. This method is called

Line Dropout. Note : The data here is BIP(Bands interleaved by pixels).

Line Striping

In Line striping similar to Line dropout, the sensor starts to malfunction, but here instead of not

recording information at all it records incorrect information. The partial log of information values

recorded will be either magnified values or compressed values. Error is corrected using the same

method discussed above.

Atmospheric 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 or the cosine of zenith

angle (which is simple 90° minus the

solar elevation angle). The errors

recorded in such a situation are corrected as -

Because of the presence of aerosols

in the atmosphere some of the

incoming radiation are scattered. Not

only that but also the radiation

reflected from the pixels on the

ground are also scattered, so the

radiation that is detected by the

satellite deviates from what should

have been received. So whatever is

received does not correctly represent

the characteristics of the actual pixel on the ground.

In addition to this it may also happen that a part of the reflected energy by some object(tree in the

illustration) due to its geometric properties, is also detected as a part of radiation for some other pixel

on the ground. What is measured by the sensor is called the DN (Digital Number) and depending

upon the radiometric resolution of the sensor a particular DN value is generated. So essentially DN is a

number that is dependent upon the radiometric resolution of the sensor.

Emittance -> Energy having the wavelength of >= 3 micrometer.

The DN value now has to be converted to reflectance or emittance, only then we get the characteristics

of a particular pixel.

Fada Eel Min Radiance a a pixel