摘要
在Computer Vision(CV)领域应用模型的开发过程中,领域专家和技术人员之间的沟通存在很大问题,同时,种类繁多的CV算子间的组合调用较为困难,难以集成.为解决以上问题,本文基于Xtext框架,采用模型驱动开发的方式,设计实现了面向计算机视觉的领域特定语言--Computer Vision Domain Specific Language(CVDSL). CVDSL整合了常用CV算子,为CV领域的应用模型开发提供了统一的平台.本文首先为CV领域构建了领域元模型;其次在领域元模型的基础上,本文定义了语法规则体系,并针对领域相关语法规则库,定制了语法检查器;最后本文设计实现了可将CVDSL转换为C++的代码自动生成器.本文通过具体的CV领域应用模型的开发和DSL定性评估框架(FQAD),对CVDSL进行了评估,验证了CVDSL的优点.
There are many problems in the communication between domain experts and technicians during the development of Computer Vision( CV) application model and it is difficult to integrate and invoke various CV operators together. Therefore,to solve these problems,this paper proposes a model-driven development method,designs and implements Computer Vision Domain Specific Language(CVDSL) based on the Xtextframework. CVDSL integrates some common CV operators,and provides a unified platform for developing application models in the CV field. In this paper,the domain meta model is built for the CV field at first. Secondly,the grammar rule system is defined based on the domain meta model,and then the grammar checker is customized for the domain related grammar rule base. And this paper designs an automatic code generator which converts CVDSL to C + +. In this paper,the CVDSL is evaluated by the Framework for Qualitative Assessment of DSLs( FQAD) and specific CV domain application model developing is completed by the CVDSL,which both verify the advantages of CVDSL.
作者
姜靖
郑启龙
JIANG Jing;ZHENG Qi-long(School of Computer Science,University of Science and Technology of China,Hefei 230026,China;National High Performance Computing Center,Hefei 230026,China)
出处
《小型微型计算机系统》
CSCD
北大核心
2020年第3期617-624,共8页
Journal of Chinese Computer Systems
基金
国家核高基重大专项项目(2012ZX01034-001-001)资助.
