Methodological Preliminaries to the Development of an Expert System for Aerial Photo Interpretation

摘要

This report describes an investigation of the psychological aspects that affect the reasoning of expert aerial photo interpreters. The purpose of the study is to help develop an artificial intelligence expert system to aid terrain analysis. Artificial intelligence offers tools for studying and representing expert knowledge and reasoning. This study examines four methods for extracting the expert's knowledge: the Standard Terrain Analysis Method, the Structured Interview Method, the Limited Information Task, and the Method of Tough Cases. Criteria are presented for analyzing the four methods, and criteria that must pertain to aerial photo interpretation in order for expert systems tools to be applicable are described. This report discusses the structure of expert systems in general and the structure of an expert system for photo interpretation in particular.