A KNOWLEDGE-BASED VISION SYSTEM FOR DETECTING LAND CHANGES AT URBAN FRINGES

Remote sensing provides powerful techniques for environmental monitori ng. However, remote sensing image analysis, including detecting land u se/cover changes, needs considerable time and skills from experienced operators. This limits the applications of remote sensing. Automated i mage analysis is required for fuller and more timely use of the valuab le data. Automating digital image analysis is a very challenging task. The major difficulties relate with vision knowledge acquisition and r epresentation, image feature extraction and description, and sensor da ta correction. Building a practical knowledge-based vision system need s a wide range of techniques and skills to integrate the techniques. I n this paper, an experimental vision system for remote sensing land us e/cover change detection is presented. The system was designed to moni tor the processes of urban development. The design strategy, system st ructure and image analysis sequence are addressed. Some experimental r esults are presented. A set of techniques are described in detail whic h were aimed at addressing the key aspects to constructing a knowledge -based vision system: preprocessing, spatial representation and manipu lation, and incorporating ancillary data.