GEOG 4/582 Lab 2: Knowledge-Based Classification

Introduction

In this exercise you will use the Erdas Imagine Expert Classifier to combine GIS and remotely sensed data to extract thematic classes.  You will create a decision tree for the classification process. An example of a knowledge-based classification involves using the following classification rules to identify urban vegetation cover:

1. Within city boundaries
2. High NIR and low visible reflectance

Instructions

1. We will do the Imagine Expert Classifier exercise (Chapter 19) in the Erdas Tour Guide. A digital copy of the Tour Guide can be found in the I:\Students\Instructors\Geoffrey_Duh\ERDAS Imagine folder (TourGuide.pdf). You can download the tutorial file (lab02.zip) here. You will need to unzip it. 
2. Use the viewer to view the image files that will be used in this lab. This will give you a better understanding of the area being studied.
3. Complete the tutorial exercise. Produce the documents and maps (screen-captured images) listed below and hand them in by April 27 (Tuesday) before class.

Deliverables

1. Write a short report describing the knowledge you used to classify images in both parts (i.e., create a knowledge base and create a portable knowledge base) of the exercise and include maps of the classification results described in the TourGuide.
2. Modify the rules so that the classification uses TM Band 4 to identify water in the first part of the exercise. You need to modify the highway map variable definition as well. Capture the rule props dialog window of this rule and include a map of the classification.
3. Describe what other rules/ancillary data could be used to improve the classification accuracy of the image used in this exercise.