遥感应用的本质是投入专业知识从对地观测影像中提炼专题信息,并以之服务于各类分析与决策的过程。其中,信息计算是整个遥感应用服务技术链的基底。本文在传承地学信息图谱理论的基础上,提出了遥感"图-谱"信息耦合的空间认知理论,构建了"像元—基元—目标—格局"为一体的遥感信息图谱计算的理论方法体系,将其分为"像元级"和"对象级"两个层次,并阐述了高性能计算环境支持下,遥感信息图谱计算平台的设计开发思路及目前研发进展,总结了遥感信息图谱计算的发展趋势和重点研究问题。
The essential issue of remote sensing application is to extract thematic information and domain knowledge from earth observation images,and thereby these information to be used in various processes of analysis and decision-making.However,information computation is the basement of the entire remote sensing application service technology system.Based on the theory of geo-informatic Tupu,we proposed the spatial cognition theory coupled with remote sensing information Tupu,constructed the theoretical and methodological system of remote sensing information Tupu in the way of "pixel-object-target-pattern",which can be divided into two levels of "pixel-level" and "object-level".This paper also discussed the idea of the designing and developing the remote sensing Tupu computation platform based on the high performance computation environment,and summarized the developing trend and important research issues of remote sensing information Tupu computation.