A Knowledge Representation Architecture for Remote Sensing Image Understanding Systems

摘要：

Knowledge plays a very important role in remote sensing image understanding. In this paper, we consider various types of knowledge related to remote - sensing image understanding, and present a knowledge representation(KR) architecture. Knowledge in the KR architecture is classified into six types, including object knowledge, image knowledge, environment knowledge, algorithm knowledge, task knowledge, and integrated knowledge, which combine knowledge from symbolic representations and computational intelligence. We analysis each knowledge type and its representation, especially task knowledge and integrated knowledge. We employ agents for task knowledge representation, which are able to finish special tasks. Meanwhile, task agents bridge the gap between low-level image processing methods and high-level semantic descriptions. The KR architecture provides the basis of knowledge services for remote sensing image understanding systems.

关键词： knowledge representation architecture task agent remote sensing image understanding

作者: Gaopan Huang Yuan Tian Guanqing Chang

作者单位: Integrated Information System Research Center, CASIA Beijing, P, R, China

会议类型: 国际会议

会议名称: 2011 6th Joint International Information Technology and Artificial Intelligence Conference(2011年第六届IEEE联合国际信息技术与人工智能会议 IEEE ITAIC 2011)

会议地点: 重庆

会议语种:英文

页码: 202-205

在线出版日期: 2011-08-20（万方平台首次上网日期，不代表论文的发表时间）