Brought to you by EarthWeb
IT Library Logo

Click Here!
Click Here!

Search the site:
 
EXPERT SEARCH -----
Programming Languages
Databases
Security
Web Services
Network Services
Middleware
Components
Operating Systems
User Interfaces
Groupware & Collaboration
Content Management
Productivity Applications
Hardware
Fun & Games

EarthWeb Direct EarthWeb Direct Fatbrain Auctions Support Source Answers

EarthWeb sites
Crossnodes
Datamation
Developer.com
DICE
EarthWeb.com
EarthWeb Direct
ERP Hub
Gamelan
GoCertify.com
HTMLGoodies
Intranet Journal
IT Knowledge
IT Library
JavaGoodies
JARS
JavaScripts.com
open source IT
RoadCoders
Y2K Info

Previous Table of Contents Next


2.11. MACERAL ANALYSIS

Catalina et al.
AITEMIN/INCAR
Spain, 1993

This expert system built in Spain is particularly interesting in this discourse for two reasons. First of all, it focuses on a different mineral than most other systems within this category. Next, it combines image analysis with traditional expert systems techniques.

In contrast to most people's beliefs, coal is not a homogeneous substance. It contains several constituents analogous to the minerals of inorganic rocks. These are called macerals, and are the fossilized remains of plant tissues. According to Catalina et al. (Catalina 94), macerals are recognized by their reflectance as well as by their morphology, which can be only recognized using a microscope.

Several components and features can be recognized when coal is analyzed in this way. In addition to more than 20 maceral types, other minerals such as pyrite and quartz may be identified. Furthermore, defects in coal preparation and empty spaces in the coal can be present.

Understanding the maceral composition is very important in determining the best use of each type of coal found. Coal that is going to be traded or applied must also be classified according to the United Nations standard "International Codification System for Medium and High Rank Coals."

Reflectance of light is an important factor when analyzing coal samples. The proportion of direct incident light reflected from a plane-polished surface is the dominating parameter in determining the constituency, since reflectance bands for some components overlap other parameters, which must also be considered. Assessment of morphology can support the analysis in order to distinguish between macerals. Maceral analysis requires both experience and time.

The knowledge-based system created supports the maceral analysis in three ways. It provides a component for image processing and a rule base built up around the actual maceral analysis. In addition, it contains a third component that manages the whole analytic process, thus alleviating the human from a lengthy and tedious task of both observing, analyzing, and concluding.

The activation of the image processing system and the rule-based part is controlled entirely by the management component. It controls the dialog between the user and the system, and it manages the microscope controller. The image processing system is fairly independent, but provides the basis for the expert system rules.

Originally, only the rule base was created. It was patterned according to most of the traditional backward-chaining systems. The goal was to imitate the reasoning of a coal petrography expert when undertaking a maceral analysis without being personally able to see the sample. Queries were also produced to guide the observation that the user makes when using the microscope.

In the latest version, a direct link between the rule-based part and the image processing system was established. In this way, a complete system for automatic maceral analysis was put to test.

According to Catalina et al., the coal maceral analysis system is able to successfully work through the following steps:

  • Selection of the area of the sample to be analyzed
  • Coal sample scanning, field by field, following a square grid
  • Microscope focusing, average image acquisition, and lighting nonuniformity correction
  • The ES determines the component occupying the center of each image, posing questions to the specialized image processing system
  • Presentation of results

The original system was developed under UNIX using an in-house built shell that was programmed in C++ for the work on the knowledge base. The rules have a LISP-type format and there is a set of specialized functions to support the project-specific requirements. The graphical interface was written for X Window and Motif.


Previous Table of Contents Next

footer nav
Use of this site is subject certain Terms & Conditions.
Copyright (c) 1996-1999 EarthWeb, Inc.. All rights reserved. Reproduction in whole or in part in any form or medium without express written permission of EarthWeb is prohibited. Please read our privacy policy for details.