摘要
最近十多年来 ,具有专家系统外壳的图象理解软件的开发一直是一个重要的研究课题。然而 ,多数基于知识的图象理解系统或者局限于诸如分割之类的特定图象操作上 ,或者局限于像理解建筑物的航空图象这样特定的应用。由于这些系统的目标和知识结构的局限性 ,因而不能将它们推广到其它领域。此研究就是试图解决此问题 :首先将Agent构造环境扩展为一个MultiAgent的图象理解系统 ,而此系统是知识入口的开发工具 ,它为用户提供一个类似于专家系统外壳的界面。这样一方面能加速图象理解系统的开发 ,另一方面方便那些缺乏图象理解知识的人开发图象理解应用程序。
The development of software that would be to image understanding systems what expert system shells are to expert systems has been the subject of considerable enquiry over the last ten years. However, most knowledge-based image understanding systems were limited either to particular imaging operation, such as segmentation, or to particular applications, like understanding aerial images of buildings. These systems, by virtue of the restricted nature of their goals and knowledge structures, could not be generalized to other domains. The purpose of our research is for solving this problem: firstly, we extend ABE (Agent Building Environment) to a image understanding system of multi agent, and it is the development tool for knowledge entry, it provides users with an interface that is similar to expert system shells. Thus, image understanding based on multi agent would both speed up system development and allow personnel less knowledgeable about image understanding to development applications.
出处
《计算机应用研究》
CSCD
北大核心
2001年第10期54-57,共4页
Application Research of Computers
基金
国家973项目 (G1 9980 30 60 0 )
福建省自然科学基金资助
