Requirements and techniques for an automated change detection system

To enhance the ability of remote sensing system to provide accurate, timely, and complete geospatial information at regional and/or global scale, an automated change detection system has been and will continue to be one of the important yet challenging problems in remote sensing. This research was designed to evaluate the requirements and prototype the algorithms for an automated change detection system at landscape level using various geospatial data sources including multisensor remotely sensed imagery and ancillary data layers. In this paper, the requirements for an operational change information extraction system and its associated techniques are discussed. These techniques are included in three subsystems: automated computer image understanding, multisource data fusion, and database updating and visualization.