针对传统的机器视觉检测系统软件开发周期过长,系统可重构性、可复用性和可扩展性差的缺点,设计了基于构件的层次模块化机器视觉算法库(MVAL)。同时,提出了适合图像处理算法的构件模型和复合方式,并给出了算法构件的运行框架虚拟机的结构以及视觉系统图形化配置实例。算法模块和整体性能测试的结果表明,系统设计满足工业在线检测的实时性和可靠性要求。此设计对相关软件的研发提供了很好的参考。
Aiming at the demerits of traditional machine vision inspection system software,e.g.,long developing period,poor restructibility,reusability and expandability,the hierarchical modular machine vision algorithm library(MVAL) is designed based on structural components.In addition,the model of structural component and composite mode suitable for image processing algorithm are proposed.The operating framework virtual machine of the components and the graphic configuration example of the vision system are given.The test results of the algorithm modules and entire performance indicate that the design of system satisfies the requirements of real-time performance and reliability for industrial online detection.It provides excellent reference for developing similar software.