GEOG272-Lab6 (2)
docx
keyboard_arrow_up
School
University of Maryland *
*We aren’t endorsed by this school
Course
272
Subject
Geography
Date
Feb 20, 2024
Type
docx
Pages
3
Uploaded by Josephmass
Lab 6 Unsupervised Classification
Lab Data:
GEOG 272 Lab 2 Data.zip
. Make sure to download, unzip and save all the files
into your own folder.
Overview
We will be using the usual Landsat image to create an unsupervised image classification for
this exercise. In particularly, this lab will:
1)
generate clusters using an unsupervised classification approach
(ISODATA), and 2)
evaluate the clusters in scatterplots and
class statistics.
Please complete the tasks and answer the questions written in RED.
Part One: Unsupervised Classification (Clustering)
ISODATA is a widely used clustering algorithm which makes a large number of passes
through the remote sensing dataset. The ISODATA utility repeats the clustering of the
image until either a maximum number of iterations have been performed, or a maximum
percentage of unchanged pixels have been reached between two iterations. Performing an
unsupervised classification is simpler than a supervised classification, because the
signatures are automatically generated by the ISODATA algorithm. However, as stated
before, the analyst must have ground reference information and knowledge of the terrain,
or useful ancillary data if this approach is to be successful.
To begin the unsupervised classification, click on the Classification icon and then select
Unsupervised.
Fill in the input and output information in the
Unsupervised
Classification
dialog box. Set the Number of Classes from minimum of 1 to maximum 5
(you may enter a higher maximum number of classes to examine differences in the results)
and the Maximum Iterations to 5. Maximum Iterations is the number of times that the
ISODATA utility will re-cluster the data. It prevents the utility from running too long, or from
getting stuck in a cycle without reaching the convergence threshold. The convergence
threshold is the maximum percentage of pixels whose cluster assignments can go
-1-
unchanged between iterations. This prevents the ISODATA utility from running indefinitely.
Leave everything else in its default state. When you have entered all relevant information
click OK to begin the process. After you have run the unsupervised classification with the above parameters, try rerunning
another classification but change the Maximum Iterations to 20. Compare your output files
and answer the following: [1
]
What is the difference in your results between the unsupervised classification with 5
iterations and the unsupervised classification with 20 iterations? Include a screenshot.
Part Two: Cluster Identification
To aid in evaluation we will need to view the results of the clustering so that we may see
how the clusters are arranged in feature space and thereby make informed decisions about
how to correctly label the cluster. Open your unsupervised classification results (using the
results from your run with 20 iterations) and visually compare them to your Landsat TM
image. You will look at signature information to help you interpret and label the results of
the unsupervised classification based on spectral patterns. Highlight the base image layer, select Tools > 2D
Scatter Plots in the Toolbar on the top
of the window. Assign the Red band to your x axis and the NIR band to the y-axis. Click on
your image window and move around the display to visually display the unsupervised
classes on your scatter plot. Use your knowledge of the spectral properties of each land
cover type in the scatter plot to help you identify and label the unsupervised classes. You
will also want to view the statistic information for each of your classes.
In the ENVI toolbox, select Classification
, Post-classification
, Class Statistics
.
Classification Input File is the classified image; Statistics Input File is the raw TM image that
you ran the classification on. Select all classes and check histograms. Now you can look at
the statistics for each class individually to help you correctly identify the ground cover
(
Stats for ). Use this information to complete the following:
[2] Make a screenshot of the 2D scatterplot. How did it help you identify and label the
unsupervised classification results?
-2-
[3] Make a screenshot of the mean and standard deviation of each band for each of your
five classes. Label each class with its corresponding land cover type on your screenshot.
[4] Consider the advantages and disadvantages of using either the supervised or
unsupervised classification approach to answer the following: 1)
Which classification approach did a better job in your practice?
2)
When would one approach be more appropriate than the other?
3) Is there anything missing in your final classification outputs that was visually
apparent in your Landsat scene? (considering both supervised and unsupervised
classification results)
-3-
Your preview ends here
Eager to read complete document? Join bartleby learn and gain access to the full version
- Access to all documents
- Unlimited textbook solutions
- 24/7 expert homework help