I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Digital Image Processing
Basic Concepts
Mrs. Minakshi Kumar
Scientist “SG”
Photogrammetry and Remote Sensing Department
Geospatial Technology and Outreach Programme Group
Indian Institute of Remote Sensing
Indian Space Research Organisation
Department of Space, Government of India
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Presentation Outline
Digital Image
Digital Image Processing
Image Preprocessing
Radiometric Errors & Correction
Line /Column Dropout / Banding
Haze Correction
Sun angle Correction
Geometric Error & Correction
Rectification
Resampling
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
What is an Image ?
An IMAGE is a Pictorial Representation of an object or a scene.
Forms of Images
Analog
Digital
Digital Image Processing Minakshi,PRSD,IIRS 3
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
What is a Digital Image ?
Produced by Electro optical Sensors
Composed of tiny equal areas, or
picture elements abbreviated as pixels
or pels arranged in a rectangular
array
With each pixel is associated a
number known as Digital Number(DN)
or Brightness value (BV) or gray level
which is a record of variation in
radiant energy in discrete form.
An object reflecting more energy • Range of DN Values depend on
records a higher number for itself on Radiometric resolution
the digital image and vice versa. • 0 – Darkest, • 255 – Brightest
(for 8 bit radiometric resolution)
Digital Image Processing Minakshi,PRSD,IIRS 4
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Multi Spectral Remotely Sensed Image
Digital Images of an area captured in different spectral ranges (bands)
A pixel is referred by its column, row, band number.
Digital Image Processing Minakshi,PRSD,IIRS 5
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Digital Image Processing
Digital image processing can be defined as the computer manipulation of digital
values contained in an image for the purposes of image correction, image
enhancement and feature extraction.
A digital image processing system consists of computer Hardware (Personal
Computer) and dedicated Image processing software necessary to analyze
digital image data.
Image Processing Software Functionalities
Data Acquisition/Restoration - Compensates for data errors, i.e Preprocessing
(Radiometric and Geometric)
Image Enhancement - Alters the visual impact of the image on the
interpreter to improve the information content
Information Extraction - Utilizes the decision making capability of computers to recognize
and classify pixels on the basis of their signatures
Digital Image Processing Minakshi,PRSD,IIRS 6
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Major Digital Image Processing Systems
Open Source
Commercial ILWIS (http://www.ilwis.org/index.htm)
ERDAS IMAGINE Opticks
http://opticks.org/confluence/display/opticks/Welcome+To+
ENVI Opticks
IDRISI GRASS (Geographic Resources Analysis Support System
http://grass.osgeo.org/ )
ER Mapper
OSSIM (Open Source Software Image Map
PCI Geomatica www.ossim.org )
eCognition Multispec
https://engineering.purdue.edu/~biehl/MultiSpec/index.html
MATLAB QGIS (A Free and Open Source Geographic Information
Intergraph System http://www.qgis.org/en/site/)
Digital Image Processing Minakshi,PRSD,IIRS 7
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Image Preprocessing
Remote sensing systems may not function perfectly all the
time.
The Earth’s atmosphere, land, and water are complex and
do not lend themselves well to being recorded by remote
sensing devices that have constraints such as spatial,
spectral, temporal, and radiometric resolution.
Consequently, error may creep into the data acquisition
process and can degrade the quality of the remote sensor
data collected.
The two most common types of error encountered in
remotely sensed data are radiometric and geometric.
Radiometric and geometric correction of remotely sensed
data are normally referred to as preprocessing
operations because they are performed prior to information
extraction.
Digital Image Processing Minakshi,PRSD,IIRS 8
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Radiometric errors- Causes
Radiometric errors are caused by detector imbalance and
atmospheric deficiencies.
Radiometric corrections are also called as cosmetic
corrections and are done to improve the visual
appearance of the image.
Common radiometric errors
Periodic line or column drop-outs,
Line or column striping.
Random bad pixels (shot noise),
Partial line or column drop-outs
Atmospheric errors
Digital Image Processing Minakshi,PRSD,IIRS 9
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Line dropout
If a detector fails to function this can result in
an entire line or column of data with no
spectral information.
The bad line or column is commonly called a
line or column drop-out and contains
brightness values equal to zero
Periodic Line Dropout
Partial Line Dropout
Digital Image Processing Minakshi,PRSD,IIRS 10
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Correction for Missing Lines /columns
It is first necessary to locate each bad line in the dataset.
A simple thresholding algorithm makes a pass through the dataset and flags
any scan line having a mean brightness value at or near zero.
Once identified, it is then possible to evaluate the output for a pixel in the
preceding line (BVi –1,j,k) and succeeding line (BVi+1,j,k) and assign the output
pixel (BVi,j,k) in the drop-out line
Replacement by either the preceding or the succeeding line
BVI,J=BVI,J-1 OR BVI,J=BVI,J+1
BVI,J =missing pixel value of pixel I scan line J
Averaging of the neighbouring pixel values
BVI,J=(BVI,J-1 + BVI,J+1)/2
Replacing the line with other highly correlated band.
Digital Image Processing Minakshi,PRSD,IIRS 11
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Line Striping (Banding)
The response of some of the detectors may shift towards lower or higher end
causing the presence of a systematic horizontal / vertical banding pattern
Banding is an cosmetic defect and it interferes with the visual appreciation of
the patterns and features on the image
Variation in gain and offset of each sensor (linear sensor characteristic) as the
sensor deteriorates in time
Digital Image Processing Minakshi,PRSD,IIRS 12
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Line striping correction
A sensor is called ideal when there is a linear relationship
between input and the output.
Correction uses a linear expression to model the relationship
between input & output values.
Assumes that mean and standard deviation of data from each
detector should be same.
Linear sensor model : n sensors
y = a.x + b D (t ) D(t ) n 2 1
a = gain Measured
DN
b = offset
D (t m 2
)
x = input
y = output Input
Digital Image Processing Minakshi,PRSD,IIRS 13
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Atmosphere induced errors
HAZE
Scattered light reaching the
sensor from the atmosphere
Additive effect, reducing
CONTRAST
SKYLIGHT
Scattered light reaching the
sensor after being reflected from
the Earth’s surface
Multiplicative effect
SUNANGLE
Time/Seasonal effect changing
the atmospheric path
Multiplicative effect
Digital Image Processing Minakshi,PRSD,IIRS 14
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Atmospheric Haze Effect
with haze
without haze with haze
DN values of objects in a single band
Object1: DN = 20 DN = 20 + 20
Object2: DN = 40 DN = 40 + 20
Contrast: 40/20 = 2X 60/40 = 1.5 X without haze
Digital Image Processing Minakshi,PRSD,IIRS 15
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Haze Correction- Dark Object Subtraction
Histogram Minimum Method
Assumption: infrared bands are not affected by Haze
Identify black bodies: clear water and shadow zones with zero
reflectance in the infrared bands
Identify DN values at shorter wavelength bands of the same pixel
positions. These DN are entirely due to haze
Subtract the minimum of the DN values related to black bodies of a
particular band from all the pixel values of that band
Digital Image Processing Minakshi,PRSD,IIRS 16
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Sun Angle Correction
The position of sun relative to earth changes depending on
time of day and day of year.
Solar elevation angle: Time- and location dependent
Sun elevation correction accounts for the seasonal position
of the sun relative to the earth
Image data acquired under different solar illumination
angles need to be normalized to a constant solar position
In the northern hemisphere the solar elevation angle is
smaller in winter than in summer
The solar zenith angle is equal to 90 degree minus the
solar elevation angle
Irradiance varies with the seasonal changes in solar
elevation angle and the changing distance between the
earth and sun
Correction necessary for mosaicking and change detection
Digital Image Processing Minakshi,PRSD,IIRS 17
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Sun Angle Correction
Image data acquired under different solar illumination angles are normalized
by calculating pixel brightness values assuming the sun was at the zenith on
each date of sensing.
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.
DN
DN '
SIN ( )
Two Images with different Sun-angles Corrected Mosaic
Digital Image Processing Minakshi,PRSD,IIRS 18
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Quiz Time
1. A Digital Image Composed of tiny equal areas called as ___________arranged in a rectangular
array
i. Picture Element
ii. Pixel
iii. Pel
iv. All of above
2. The size of the tiny area in the digital Image depends on __________ and the value depends on
_________ of Sensor.
i. Spatial , Radiometric Resolution
ii. Radiometric, Spectral Resolution
iii. Spectral Resolution, Swath of Sensor
iv. None of the Above
3. A digital Image affected with Haze is corrected by Subtracting the minimum of the DN values
related to black bodies of a particular band from all the pixel values of that band
1. True
2. False
4. Scattered light reaching the sensor after being reflected from the Earth’s surface is called
1. Haze
2. Skylight
3. Irradiance
4. None of the above.
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Geometric Errors and Corrections
The transformation of remotely sensed images so that it has
a scale and projections of a map is called geometric
correction.
It is concerned with placing the reflected, emitted, or back-
scattered measurements or derivative products in their proper
planimetric (map) location so they can be associated with
other spatial information in a geographic information system
(GIS)
Include correcting for geometric distortions due to sensor-
Earth geometry variations, and conversion of the data to real
world coordinates (e.g. latitude and longitude) on the Earth's
surface
Digital Image Processing Minakshi,PRSD,IIRS 20
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Earth Rotation Effect - Image Offset (skew)
Sun-synchronous satellites are normally in
fixed orbits that collect a path (or swath) of
imagery as the satellite makes its way from
the north to the south in descending mode.
Meanwhile, the Earth below rotates on its
axis from west to east making one
complete revolution every 24 hours. This
skews the geometry of the imagery
collected
• Dashed line indicate
shape of distorted
image
• Solid line indicates
restored image
Digital Image Processing Minakshi,PRSD,IIRS 21
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Rectification
Is a process of geometrically correcting an image so that it can be
represented on a planar surface , conform to other images or
conform to a map.
That is it is the process by which geometry of an image is made
planimetric.
It is necessary when accurate area , distance and direction
measurements are required to be made from the imagery.
It is achieved by transforming the data from one grid system into
another grid system using a geometric transformation
Grid transformation is achieved by establishing mathematical
relationship between the addresses of pixels in an image with
corresponding coordinates of those pixels on another image or
map or ground.
Digital Image Processing Minakshi,PRSD,IIRS 22
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Image to Map Rectification Procedure
Two basic operations must be performed to geometrically rectify a
remotely sensed image to a map coordinate system:
Geometric Transformation coefficient computation
The geometric relationship between input pixel location (row & column) and
associated map co-ordinates of the same point (x,y) are identified.
Involves selecting Ground Control Points (GCPS) and fitting polynomial
equations using least squares technique.
Intensity Interpolation (Resampling)
A pixel in the rectified image often requires a value from the input pixel grid
that does not fall neatly on a row and column co-ordinate.
For this reason resampling mechanism is used to determine pixel
brightness value.
Digital Image Processing Minakshi,PRSD,IIRS 23
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Ground Control Points (GCPs)
A ground control point (GCP) is a location on the
surface of the Earth (e.g., a road intersection) that can
be identified on the imagery and located accurately on a
map.
There are two distinct sets of coordinates associated
with each GCP:
source or image coordinates specified in i rows and j
columns, and
Reference or map coordinates (e.g., x, y measured in
degrees of latitude and longitude, or meters in a Universal
Transverse Mercator projection).
Digital Image Processing Minakshi,PRSD,IIRS 24
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Ground Control Points (GCPs)
Accurate GCPs are essential for accurate rectification
Well dispersed GCPs result in more reliable rectification
GCPs for Large Scale Imagery
Road intersections, airport runways, towers buildings etc.
for small scale imagery
larger features like Urban area or Geological features can be
used
NOTE : landmarks that can vary (like lakes, other water bodies,
vegetation etc) should not be used.
Sufficiently large number of GCPs should be selected
Requires a minimum number depending on the type of
transformation
Digital Image Processing Minakshi,PRSD,IIRS 25
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Polynomial Coordinate transformation
Polynomial equations are used to convert the source file
coordinates to rectified map coordinates.
Depending upon the distortions in the imagery, the
number of GCPs used, their location relative to one other,
complex polynomial equations are used.
The degree of complexity of the polynomial is expressed as
ORDER of the polynomial.
The order is simply the highest exponent used in the
polynomial
Digital Image Processing Minakshi,PRSD,IIRS 26
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Mathematical Transformations
Linear Transformations/ Affine transformation/ first order
transformation
X = a0 + a1x + a2 y
Y = b0 + b 1 x + b2 y
where
X , Y are the Rectified coordinates (output)
x, y are the source coordinates (input)
A first order transformation can change
Location in x and/or y
Scale in x and/or y
Skew in x and/or y
Rotation
Digital Image Processing Minakshi,PRSD,IIRS 27
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Polynomial transformation
If the coefficients a0 ,a1,a2, b0, b1 and b2 are known then, the
above polynomial can be used to relate and point on map to its
corresponding point on image and vice versa. Hence six
coefficients are required for this transformation (three for X and
three for Y).
So it requires Minimum THREE GCP’s for solving the above
equation.
However the error cannot be estimated with three GCP’s alone.
Hence one additional GCP is taken
Before applying rectification to the entire set of the data, it is
important to determine how well the six coefficients derived from
the least square regression of the initial GCPs account for the
geometric distortion in the input image.
Digital Image Processing Minakshi,PRSD,IIRS 28
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Accuracy of transformation
In this method, we check how good do selected points fit between the map
and the Image?
To solve linear polynomials we first take four GCP’s to compute the six
coefficients. Its source coordinates in the original input image are say xi and
yi. The position of the same points in reference map in degrees, feet or
meters are say x,y
Now, if we input the map x,y values for the first GCP back into the linear
polynomial equation with all the coefficients in the place, we would get the
computed or retransformed xr and yr values , which are supposed to be
location of this point in input image
Ideally measured and computed values should be equal.
In reality this does not happen.
Digital Image Processing Minakshi,PRSD,IIRS 29
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Root Mean Square (RMS) error
Accuracy is measured by computing Root Mean Square Error (RMS error)
for each of the ground control point
RMS error is the distance between the input (source or measured) location
of a GCP and the retransformed (or computed) location for the same GCP.
RMS error is computed with a Euclidean Distance Equation.
Where
xi and yi are the input
source coordinates and
xr and yr are the
retransformed coordinates
Digital Image Processing Minakshi,PRSD,IIRS 30
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Acceptable RMS error
The amount of RMS error that is tolerated can be thought of as a window
around each source coordinate, inside which a retransformed coordinate is
considered to be correct.
Acceptable RMS error depends upon the
End use of the data
The type of data being used, and
The accuracy of the GCP and the ancillary data.
Normally an RMS error of less than 1 per GCP and a total RMS error of less
than half a pixel (0.5) is acceptable
Digital Image Processing Minakshi,PRSD,IIRS 31
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Intensity Interpolation (Resampling)
Once an image is warped, DNs are to be assigned to the “new”
pixels?
Since the grid of pixels in the source image rarely matches the grid
for the reference image, the pixels are resampled so that new data
file values for the output file can be calculated.
This process involves filling the rectified output grid with brightness
values extracted from a location in the input image and its
reallocation in the appropriate coordinate location in the rectified
output image.
This results in input line and columns numbers as real
numbers ( and not integers)
When this occurs, methods of assigning Brightness values are
Nearest Neighbour
Bilinear
Cubic
Digital Image Processing Minakshi,PRSD,IIRS 32
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Nearest Neighbor
The nearest neighbor approach uses the value of the closest
input pixel for the output pixel value.
The pixel value occupying the closest image file coordinate to
the estimated coordinate will be used for the output pixel value
in the georeferenced image.
Digital Image Processing Minakshi,PRSD,IIRS 33
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Nearest Neighbor
ADVANTAGES:
Output values are the original input values. Other methods of
resampling tend to average surrounding values. This may be
an important consideration when discriminating between
vegetation types or locating boundaries.
Since original data are retained, this method is recommended
before classification.
Easy to compute and therefore fastest to use.
DISADVANTAGES:
Produces a choppy, "stair-stepped" effect. The image has a
rough appearance relative to the original unrectified data.
Data values may be lost, while other values may be
duplicated.
Digital Image Processing Minakshi,PRSD,IIRS 34
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Bilinear Interpolation
The bilinear interpolation approach uses the weighted average of the nearest
four pixels to the output pixel.
4
Zk where Zk are the surrounding four data point values, and
2
k 1 Dk
D 2 are the distances squared from the point in question
k
BVwt 4 (x’, y’) to the these data points.
1
D
k 1
2
ADVANTAGES:
Stair-step effect caused by the nearest
k
neighbor approach is reduced. Image looks
smooth.
DISADVANTAGES:
Alters original data and reduces contrast by
averaging neighboring values together.
Is computationally more extensive than
nearest neighbor.
Digital Image Processing Minakshi,PRSD,IIRS 35
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Cubic Convolution
The cubic convolution approach uses the weighted average of the nearest
sixteen pixels to the output pixel. The output is similar to bilinear
interpolation, but the smoothing effect caused by the averaging of
surrounding input pixel values is more dramatic.
16
Zk
2
k 1 Dk
where Zk are the surrounding four data
point values, and D2k are the distances
BVwt 16
1 squared from the point in question (x’, y’)
2
k 1 Dk
to the these data points.
ADVANTAGES:
Stair-step effect caused by the nearest neighbor
approach is reduced. Image looks smooth.
DISADVANTAGES:
Alters original data and reduces contrast by averaging
neighboring values together.
Is computationally more expensive than nearest
neighbor or bilinear interpolation.
Digital Image Processing Minakshi,PRSD,IIRS 36
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Digital Image Processing Minakshi,PRSD,IIRS 37
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Input Image
Rectified Image
Digital Image Processing Minakshi,PRSD,IIRS 38
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Discussion / Query
I N D I A N I N S T I T U T E O F R E M O T E S E N S I N G, D E H R A D U N
Thank You
