This document contains a summary of an advanced image classification workshop presentation. It discusses pixel-based and object-based image classification techniques. Pixel-based classification involves classifying pixels based on their spectral values using supervised or unsupervised classification methods. Supervised classification uses training data to develop algorithms to classify pixels, while unsupervised classification automatically groups pixels into clusters. Object-based classification considers both spectral and spatial characteristics of grouped pixels.
1. Welcome
to the Workshop Presentation on
Advanced Image Classification
BAYES AHMED
PhD Student
University College London (UCL), UK
Workshop at BIP, Bangladesh
13 September 2014
2. IImmaaggee CCllaassssiiffiiccaattiioonn
Image classification refers to the task of extracting
information classes from a multiband raster
image.
The resulting raster from image classification can
be used to create thematic maps.
4. IImmaaggee CCllaassssiiffiiccaattiioonn
Here IMAGE stands for Raster Image
(e.g. Satellite Image)
In general in GIS, we use two types of Image
Classification:
1.Pixel Based
2.Object Based
5. PPiixxeell
Pixel is a physical point (e.g. dot), or the smallest
addressable element (e.g. cell) in a raster image
10. DDiiggiittaall IImmaaggee CCllaassssiiffiiccaattiioonn
Digital image classification uses the
spectral information represented by
the digital numbers in one or more
spectral bands, and attempts to
classify each individual pixel based on
this spectral information.
11. Spectral and Information Classes
Spectral Classes are groups of pixels that are uniform (or
near-similar) with respect to their brightness values in the
different spectral channels of the data.
Information Classes are those categories of interest that the
analyst is actually trying to identify in the imagery.
A broad information class may contain a number of
spectral sub-classes with unique spectral variations. It is
the analyst's job to decide on the utility of the different spectral
classes and their correspondence to useful information classes.
12. SSuuppeerrvviisseedd CCllaassssiiffiiccaattiioonn
There are two general approaches to pixel-based image
classification: supervised and unsupervised.
Supervised Classification: the analyst identifies in the
imagery homogeneous representative samples
(information classes) of interest. These samples are
referred to as training areas.
The selection of appropriate training areas is based on the
analyst's familiarity with the geographical area and their
knowledge of the actual surface cover types present in
the image. Thus, the analyst is "supervising" the
categorization of a set of specific classes.
13. SSuuppeerrvviisseedd CCllaassssiiffiiccaattiioonn
Information classes (i.e., landcover types)
The software system is then
used to develop a statistical
characterization/ algorithm
(mean, variance and
covariance) of the reflectances
for each information class. This
stage is often called signature
development.
14. SSuuppeerrvviisseedd CCllaassssiiffiiccaattiioonn
The image is then classified by examining the reflectance for
each pixel and making a decision about which of the
signatures it resembles most. There are several techniques
for making these decisions, called classifiers.
Classifiers: Minimum distance to means (MINDIST),
maximum likelihood (MAXLIKE), linear discriminant analysis
(FISHER), Bayesian (BAYCLASS), multi-layer perceptron
(MLP) neural network, self-organizing map (SOM) neural
network; Mahalanobis typicalities (MAHALCLASS), Dempster-
Shafer belief (BELCLASS), linear spectral unmixing (UNMIX),
fuzzy (FUZCLASS), spectral angle mapper (HYPERSAM),
minimum distance to means (HYPERMIN), linear spectral
unmixing (HYPERUNMIX), orthogonal subspace projection
(HYPEROSP), and absorption area analysis
(HYPERABSORB) etc.
15. MMaaxxiimmuumm LLiikkeelliihhoooodd
The maximum likelihood classifier calculates for each class the
probability of the cell belonging to that class given its attribute
values. Each pixel is assigned to the class that has the
highest probability (that is, the maximum likelihood).
17. UUnnssuuppeerrvviisseedd CCllaassssiiffiiccaattiioonn
Unsupervised classification reverses the supervised
classification process.
Spectral classes (or clusters) are grouped first, based
solely on the numerical information in the data, and are then
matched by the analyst to information classes (if possible).
Programs, called clustering algorithms, are used to
determine the natural (statistical) groupings or structures in the
data.
19. Clus Iso Clustteerr UUnnssuuppeerrvviisseedd CCllaassssiiffiiccaattiioonn
Statistically, clusters are naturally occurring groupings in
the data.
The Iso prefix of the isodata clustering algorithm stands for
Iterative Self Organizing (ISO), a method of performing
clustering.
The iso cluster algorithm is an iterative process for computing
the minimum Euclidean distance when assigning each
candidate cell to a cluster.
The specified Number of classes value is the maximum
number of clusters that can result from the clustering process.
20. Clus Iso Clustteerr UUnnssuuppeerrvviisseedd CCllaassssiiffiiccaattiioonn
An empty graph is made with the range of values in the first
band plotted on the x-axis and the range of values in the
second band plotted on the y-axis.
A 45-degree line is drawn and
divided into the number of
classes you specify. The
center point of each of these
line segments is the initial
mean value for the classes.
21. Clus Iso Clustteerr UUnnssuuppeerrvviisseedd CCllaassssiiffiiccaattiioonn
Each sample cell is plotted on the graph, and the distance
from the point to each mean center point on the 45-degree
line is determined. The distance is calculated in attribute space
using the Pythagorean theorem. The sample point is assigned
to the cluster represented by the closest mean center point.
22. Clus Iso Clustteerr UUnnssuuppeerrvviisseedd CCllaassssiiffiiccaattiioonn
The next sample point is plotted, and the above procedure is
repeated for all sample points.
23. Clus Iso Clustteerr UUnnssuuppeerrvviisseedd CCllaassssiiffiiccaattiioonn
The above process will iterate. Before the next iteration, a new
mean center point is calculated for each cluster based on
the values of the cell locations currently assigned to the cluster
in the previous iteration. With the new mean center point for
each cluster, the previous two steps are repeated.
25. Object-bbaasseedd IImmaaggee AAnnaallyyssiiss ((OOBBIIAA))
The pixel-based procedures analyze the spectral
properties of every pixel within the area of interest,
without taking into account the spatial or contextual
information related to the pixel of interest.
OBIA analyzes both the spectral and spatial/contextual
properties of pixels and use a segmentation process
and iterative learning algorithm to achieve a semi-automatic
classification.
It considers – spectral properties (i.e., color), size,
shape, and texture, as well as context from a
neighborhood surrounding the pixels.