The goal of image enhancement is
to improve the visual
interpretability of an image.
Enhancements are used to make it
easier for visual interpretation and
understanding of imagery.
Image enhancement techniques
are used to highlight features of
interest.
Example: Through contrast
variation or Contrast manipulation
There is a need to visualize image data at most stages of the
remote sensing.
An understanding of how we perceive color is required at two
main stages in the remote sensing process.
It is required to produce ideal ‘pictures’ from:
multispectral image data on the computer screen (additive
principle) or
as a printed hard copy (Subtractive principle)
The theory of color perception plays an important role in the
subsequent interpretation of images.
8.2.1. Perception of Colors
Color is the perceptual result of light in the visible region of
the spectrum incident upon the retina (human eyes).
Color perception theory is applied whenever colors are
reproduced, for example in color photography, TV, printing
and computer animation.
Tri-stimuli model
The eye’s general sensitivity is to the wave lengths
between 0.4 to 0.7 μm.
Various three dimension spaces are used to describe and
define color.
A. RGB (Red, Green and Blue)
B. IHS (Intensity, Hue and Saturation)
The IHS system, which refers to intensity, hue and
saturation, more naturally reflected our sensation of color.
Intensity describes whether a color is light or dark or
simply describes total brightness of a color.
Hue refers to the names that we give to colors: red, green,
yellow, orange, purple etc. or the dominant (average)
wave length of light contributing to a color.
Saturation describes purity of color relative to gray
C. YMC (Yellow, Magenta and Cyan )
As RGB is used in computer and TV display, YMC color
description is used in printing hard copy
The principles are subtractive with the colored filters ;
yellow, magenta and cyan.
8.3. Image display and Image composites
Natural Color Composites
True color composite is made, where the RGB channels
relate to the red, green and blue wavelength bands of a
camera or multispectral scanner
Natural color composite of OLI bands of Landsat 8 image
of path 170 and row 52.
Color composite (RGB):
Red = band 4 (0.64 to 0.67 μm)
Green = band 3 (0.53 to 0.59 μm)
Blue = band 2 (0.45 to 0.51 μm )
Image Composites
Each band of a multispectral image can be displayed one band at a time as a grayscale
image, or as a combination of three bands at a time as a color composite image.
The three primary colors of light are red, green, and blue (RGB).
Computer screens can display an image composed of three different bands, by using a
different primary color for each band.
When we combine these three images, the result is a color image with each pixel’s color
determined by combination of RGB of different brightness.
Natural or True Color Composites
A natural or true color composite is an image displaying a
combination of the visible red, green and blue bands to the
corresponding red, green and blue channels on the computer display.
The resulting composite resembles what would be observed naturally
by the human eye:
vegetation appears green,
water dark is blue to black and
bare ground and impervious surfaces appear light gray and
brown.
Many people prefer true color composites, as colors appear natural
to our eyes, but often subtle differences in features are difficult to
recognize.
Natural color images can be low in contrast and somewhat hazy due
the scattering of blue light by the atmosphere.
False Color Composites
False color images are a representation of a multispectral image
produced using any bands other than visible red, green and blue as
the red, green and blue components of the display.
False color composites allow us to visualize wavelengths that the
human eye can not see (i.e. near-infrared and beyond).
Using bands such as near infrared highlights the spectral differences
and often increases the interpretability of the data.
There are many different false colored composites that can be used
to highlight different features. See the heading below for more
information about common band combinations for false color
composites.
Band Combinations for Landsat 8
Band combinations are selected for a number of reasons and it is
helpful to understand the spectral reflectance profiles of features
you are interested in studying. For example in the NIR false color
composite shown above healthy vegetation appears bright red as
they reflect more near infrared than green. Another common
combination uses the shortwave infrared (shown as red), the near
infrared (green), and the green visible band (shown as blue). In this
false color composite vegetation appear bright green, bare ground
appears reddish and snow appears bright blue. Though there are
many possible combinations of wavelength bands, the table to the
left is a list of some that are commonly used. The band
combinations are listed by band number in order of red, green, blue
(RGB):
False Color Composites
False color composite of OLI bands of Landsat 8 image of path
170 and row 52.
Color composite (RGB):
Red = band 6 (1.57 to 1.65 μm)
Green = band 5 ( 0.85 to 0.88 μm )
Blue = band 4 (0.64 to 0.67 μm )
False color composite of OLI bands of Landsat 8 image of path 170
and row 52.
Color composite (RGB):
Red = band 7 (2.11 to 2.29 μm)
Green = band 6 (1.57 to 1.65 μm )
Blue = band 4 (0.64 to 0.67 μm )
Common Band Combinations in RGB comparisons for Landsat 7 or
Landsat 5, and Landsat 8
Classification is a process by which a set of items is
grouped into classes based on common characteristics.
Classification of satellite image data is based on placing
pixels with similar values into groups and identifying the
common characteristics of the items represented by these
pixels.
The process of image classification typically involves five
steps:
1. Selection and preparation of the image data.
cover types to be classified
the most appropriate sensor
the most appropriate date(s) of acquisition and
the most appropriate wavelength bands
2. Definition of the clusters in the feature space.
supervised classification and
unsupervised classification
3. Selection of classification algorithm.
Once defined the spectral classes in the feature space,
the operator needs to decide on how the pixels (based
on their DN values) are assigned to the classes.
4. Running the actual classification.
based on its DN-values, each individual pixel in the
image is assigned to one of the defined classes.
5. Validation of the result
Once the classified image has been produced its quality
is assessed by comparing it to reference data (ground
truth).
8.4.1. Types of image classification: spectral classification
Classification is the process of assigning pixels to categories
(organize data into categories) based on their numerical properties
(DN, radiance, reflectance, etc…)
There are three types of image classification
A. Supervised
B. Unsupervised
C. Hybrid
1. Supervised Classification:
An operator defines the spectral characteristics of the classes
by identifying sample areas (training areas).
Operator should be familiar with the area of interest
The clusters, as selected by the operator:
Should forma representative data set for a given class
Should not or limitedly overlap with the other clusters,
otherwise, a reliable separation is not possible.
Use ERDAS
Steps in supervised classification
Decide number of classes
Training stage:
collect numerical training areas on spectral response
characterizing the feature pattern from the training
data.
Classification stage:
assigning pixels to corresponding classes.
delineating class boundaries using statistical methods.
compare each unknown pixel to spectral of most
similar category
Output stage:
present results, maps, tables, and GIS data files
There are different types of classifier algorithms in
supervised classification for assigning pixels to
informational classes
Maximum likelihood
Minimum distance method
Parallelepiped
2. Unsupervised Classification:
Assignment of pixels to spectral classes without prior
knowledge of the existence or names of these classes
Clustering algorithms are used to partition the feature
space into a number of clusters.
Unsupervised classification’s main purpose is to produce
spectral groupings based on certain similarities.
In one of the most common approaches, the user has to
define the maximum number of clusters in a data set.
Image classification workflow