Knowledge Based Approaches
for Remote Sensing Images Interpretation
INTRODUCTION
CLASSICAL APPROACHES
Landsat Image
Expert Samples
Karhunen-Loeve Approach
Bayesian Classification
Classical Method Evaluation
INTRODUCING EXPERT KNOWLEDGE IN A CLASSIFICATION PROCESS
Localization of the Classes
Expert Knowledge and Out-image Data
Expert Rules
Complementary Information
Maps derived from Out-image Information
Image of Slopes
Image of Orientations
Image of Ridges
Distance to Ridges
Image of Valleys
Image of Distances to Valleys
Image of Distances to Villages
Image of Distances to Roads
How to Introduce Expert Knowledge in the Classification Process?
Knowledge Integration: an Updating Approach
Knowledge Integration : a Fusion Approach
Adequation to the Deductive Process
Expert System Approach
General Description of the System
Knowledge Base
Context Computation (Facts Base)
Inference Engine
ICARE: an Expert System for Satellite Images Interpretation
ICARE - Expert System Description
Fact Data Base
Inference Engine
Algorithm
Rules' Conclusions Computation
Practical Example
NEURAL NETWORKS
Theory
Biological Aspect
Artificial Neuron
Neuron Networks
Neuron Network Learning
The
Rule
Generalized
Rule
Unsupervised Learning
The Conterpropagation
The Fuzzy Neurons
Fuzzy Neural Network
Neural Networks for Knowledge Representation
Overview of the System
Knowledge Representation
The Learning Step
The System Operation
Applications
Potentiality Maps for Irrigation
Palni Application
Conclusion
Concluding Remarks
BIBLIOGRAPHY
IRIT-UPS
Rule
Rule 