In this lab you will learn how to mosaic images and conduct unsupervised classification in ERDAS Imagine.
There are two parts in this lab. For Part I, you
will mosaic several sets of images, including air photos and LandSat images. You will extract thematic land-cover information
from a TM image (germtm.img) using an unsupervised
classification method in Part II. The image files for this lab can be found in I:\Students\Instructors\Geoffrey_Duh\GEOG4581\Lab6.
Copy these files to your working folder before you begin. These files are:
2-2.img
2-3.img
air-photo-1.img
air-photo-2.img
germtm.img
wasia1_mss.img
wasia2_mss.img
wasia3_tm.img
Part
Complete the tutorial exercise in ERDAS Tour Guide (pages 311 to 335). Produce pictures showing the original individual images and the mosaicked images for air-photo-1 and 2 and for wasia1_mss, 2_mss, and 3_tm. Make sure you clearly label individual pictures.
Part II. Unsupervised Classification
Read pages 243-259 (up to RGB Clustering) of the ERDAS
Field Guide, paying particular attention to pages 254-259 (Unsupervised Training
section). Open the Tour Guide and skim pages 135 to 142 on Classification.
Complete the tutorial exercise on pages 135 to 142 and make sure you define each of the 10 spectral classes using the classification scheme shown in the table below. Some spectral classes will have the same land-cover class name. Produce a map composition of the classified map that includes a legend with the land-cover class names. Put your name on it and submit it with your report. Write one paragraph describing how the map was created including a brief description of the ISODATA process and what the different menu choices mean within the ERDAS ISODATA menu.
Land-cover classification |
Water |
|
Crop/Pasture |
Impervious/Bare soil |
Residential |
Others |