PRE-PROCESSING

Radiometric Correction
 LESSON OUTCOME 2
DESCRIBE THE PROCESS OF PRE-PROCESSING INCLUDING
RADIOMETRIC CORRECTION AND GEOMETRIC CORRECTION
RADIOMETRIC CORRECTION
▪ Introduction
▪ Internal and External Errors
▪ Types of Radiometric Corrections
▪ Noise Removal
▪ Stripping/Banding
▪ Line Drop
▪ Random Bit Error
▪ Atmospheric Correction
▪ Haze Compensation/Reduction
INTRODUCTION TO RADIOMETRIC CORRECTION
▪ Radiometric correction is carried out when an image data is
recorded by the sensors contain errors in the measured brightness
values of the pixels
▪ Aim: to reduce the influence of errors or inconsistencies in image
brightness value that may limit one’s ability to interpret or
quantitatively process and analyze digital remotely sensed images.
▪ It also include correcting the data for sensors irregularities and
unwanted sensors, or atmospheric noise and converting the data so
they can accurately represent the reflected or emitted radiation
measured by sensor.
INTERNAL & EXTERNAL ERRORS
(SOURCES & CORRECTIONS)
SOURCES OF ERRORS CORRECTIONS
Internal
introduced by the remote sensing system. Corrected based on prelaunchor in-flight
Uncalibrated or miscalibrated sensor may calibration measurements.
cause errors. Sun elevation correction and earth-sun
e.g. Noise distance correction
Detect sensor error (sensor irregulaties or
unwanted signal)
Noise removal
External
introduced by phenomena that vary in nature Atmospheric effects
through space and time. Background
External variables that cause errors are Absolute correction
atmosphere, terrain elevation, slope, and
Relative correction
aspect.
Topographic effects
Topographic normalization
TYPES OF RADIOMETRIC CORRECTIONS
FROM INTERNAL ERROR
▪ Noise Removal
▪ electronic noise - both random and periodic
FROM EXTERNAL ERROR
▪ Sun elevation correction and earth-sun distance correction
▪ Haze compensation
▪ Sun-angle correction
▪ for comparison and mosaic images acquired from different time of the
year
▪ Correction for atmospheric scattering
▪ subtract the haze DN values from different bands DN
 INTERNAL ERRORS
NOISE
▪ Radiometric error also referred to as “noise”, which
could ne considered any undesirable spatial or
temporal variations in image brightness (not image
surface)
▪ Image noise is any unwanted disturbance in image
data that is due to limitations in the sensing, signal
digitization, or data recording process.
▪ Potential source: electronic interference between
sensor components
 NOISE REMOVAL
▪ Noise can either degrade or totally mask the true
radiometric information content of a digital image
▪ Common form of noises and errors (System errors)
include:-
▪ Stripping/Banding
▪ Line Drop
▪ Bit Error
 STRIPPING OR BANDING
 Occurs due to non-identical detector response –
variation and drift.
▪ was commonly in early Landsat MSS data of the six (6) MSS’s
detectors. The drift was different for each of the six detectors, cause
the same brightness was detected differently by the different
detectors.
▪ With time the response of some detectors may drift to higher or lower
levels, resulting in relatively higher or lower values along every sixth
line in the image data
▪ Line striping is corrected by
▪ (i) destripping (linear method) or using
▪ (ii) histograms matching per detector
 LINE DROP
▪ Occurs due to recording problems when one of the detectors of the
sensor in question either gives wrong data or stops functioning
(systems malfunction permanently or temporarily).
▪ The results is a zero brightness value in each pixel of the particular line
and will appear black in the image.
▪ There is no way to restore these values but it is possible to improve the
visual interpretation of the data based on spatial autocorrelation.
▪ This relies on the fact that points close in geographical space tend to have
similar values of the variable of interest. This is the basis for the estimation
of missing brightness values in scan lines.
▪Dropped line are normally corrected by
▪ (i) replacing the line by pixel values in the line above
or below
▪ (ii) or with average of the two
 RANDOM BAD PIXEL
▪ Occurs when an individual detector does not record spectral data
for an individual pixel.
▪ When this occurs randomly, it is called a bad pixel.
▪ It also can be characterized by non systematic variations in grey
level from pixel to pixel.
▪ Often referred as being “spikey” and caused “salt and pepper” appearance.
▪ Numerous random bad pixels found within the scene is called shot noise
because it appears that the image was shot by a shotgun.
▪ Normally these bad pixels contain values of 0 or 255 (in 8- bit data) in one or
more of the bands
 BVa = (15+17+16+16+14+15+16+13) = 15
8
BVb = (14+16+19+14+15+16+17+17) = 16
8
BVa BVb
▪ The noisy pixel can be replaced by averaging its neighboring values.
▪ The mean of the eight (n) surrounding pixels is computed using the equation
and the value substituted for BVi,j,k in the corrected image:
 EXTERNAL ERRORS
ATMOSPHERIC EFFECT
▪ A value recorded at a given pixel location on a remotely-
sensed image is not a record of the true ground-leaving
radiance at that point.
▪ for the magnitude of the ground-leaving signal is attenuated due
to atmospheric absorption and
▪ its directional properties are altered by scattering.
▪ A full radiometric correction would include
▪ conversion of pixel digital value to apparent (at-sensor)
radiance
▪ subtraction of the atmospheric contribution
▪ topographic normalization and
▪ sensor calibration.
▪ The end product is an estimate of the true ground-leaving
radiance at a pixel.
Atmospheric Effect on Brightness Values
 ATMOSPHERIC EFFECT DUE TO SCATTERING
RAYLEIGH SCATTERING

Rayleigh Scattering
•Caused by particles much smaller than a wavelength
•Declines with the fourth power of the wavelength
•Responsible for blue skies, red sunsets
•Key element in radiometric correction of images
 HAZE COMPENSATION/REDUCTION
▪ Aerial and satellite images often contain haze. Presence of haze
reduces image contrast and makes visual examination of images
difficult.
▪ It is due to Rayleigh scattering - Particle size responsible for effect
smaller than the radiation’s wavelength (e.g. oxygen and nitrogen)
▪ Scattering is wavelength dependent and is more pronounced in shorter
wavelengths and negligible in the NIR
▪ Haze has an additive effect resulting in higher DN values.
▪ Purpose of haze compensation in multispectral data : to
observe the radiance recorded over target areas of zero
reflectance
▪ For example, the reflectance of deep clear water is zero
in NIR region of the spectrum
▪ Therefore, any signal observed over such an area represents
the path radiance
▪ This value can be subtracted from all the pixels in that band
HAZE-EXAMPLE INDONESIA

(a) Before haze removal (b) After haze removal