【作者】 杨明；

【导师】 肖德贵； 周志军；

【作者基本信息】 湖南大学 ， 计算机技术（专业学位）， 2017， 硕士

【摘要】 图像处理是计算机通过一系列算法对图像进行分析,以达到研究人员所需图像结果的技术。那么通过实验来验证研究人员算法可行性和处理效果是必不可少的过程。又由于图像处理算法种类繁多,常用的算法使用率高,那么实验人员就会面临着重复编写这些高使用率算法的难题,大大降低了自己的实验效率。因此一个操作便捷、可扩展性强、实验效率高的图像处理实验平台对于研究人员是非常必要的。构件技术是软件复用领域的一个重难点,它能够通过它的复用特性,大大的提高软件开发效率,缩短开发周期,减少工作量。本文设计开发了基于构件的图像处理实验平台。该平台有着构件化的算法组合功能,同时支持用户添加新的算法构件,使用户可以将新添加的算法构件和平台现有的算法构件进行组合生成算法通道,帮助实验人员解决对一些常用算法进行重复编码的难题,提高实验效率。而且对实验结果有着可视化的输入输出,使实验人员可以实时直观的看到实验效果。该平台采用拖拽组合算法构件的操作形式,有良好的交互能力,操作便捷。在设计开发过程中也遇到了许多问题,作者通过提出装饰组装机制和构件化组态模型予以解决。主要研究内容如下:1.提出一种基于装饰的组装机制。该组装机制通过创建包装构件,动态的扩展构件的功能来完成组装。由于图像处理实验平台是针对流式的算法通道来对图片进行处理,而每个构件是一种处理算法。因此,针对该平台的特殊性以及专业性,该组装机制充分切合了实验平台的算法组合需求以及拖拽组合的操作模式。2.提出了一种基于构件的组态模型。由于该实验平台有着其独特的拖拽组合算法构件来形成算法通道的操作形式,以及图像处理算法多样化流程规约的领域特性。借鉴了工控领域中组态的概念和特性,提出了基于构件的组态模型。该模型基于构件的基础给开发人员以及用户带来了高效率的开发与使用,以及切合图像处理实验平台需求的三层软件体系结构为可扩展性打下基础,同时引入宏观逻辑描述使得实验人员可根据自己的业务需求来对图像处理实验平台算法通道的规则进行限制约束,使用户可以参与到系统的开发中去来保证系统的功能可以和用户的业务需求始终保持一致。3.在构件化组态模型和装饰组装机制的基础上对图像处理实验平台进行设计与实现,同时对部分关键代码进行了展示。最后进行了大量不同角度、不同层面的测试实验,对图像处理实验平台算法构件的组装执行、装饰组装机制、宏观逻辑描述等进行了全面的测试验证。更多还原

【Abstract】 Image processing is a technique by which a computer analyzes an image through a series of algorithms to achieve the desired image results of the researcher.It is essential to experiment with the feasibility and effectiveness of the researcher’s algorithm.There are many different kinds of image processing algorithms with high utilization rate,the experimental staff will face the problem of repeatedly writing these high usage algorithms,greatly reducing their experimental efficiency.Therefore,image processing experiment platform is very necessary for the researchers,which is convenient operation,strong extensibility and high test efficiency.Component technology is a difficult point in the field of software reuse,which can greatly improve the efficiency of software development,shorten the development cycle,and reduce the workload through its reuse characteristics.This paper designs and develops a component-based image processing experiment platform.The platform has a component-based algorithm combination function,and supports the user to add new algorithm components,so that users can combine the newly added algorithm components with the existing algorithm components of the platform to generate algorithm channel,then,it can help the experimental staff to solve the problem of repeated coding for some common algorithms and improve the experimental efficiency.Moreover,the experimental results have a visual input and output,so that the experimental staff can see the experimental results in real time and intuitive.The platform adopts the operating form of drag and drop combination of algorithm components,which has good interactive ability and convenient operation.In the process of design and development,many problems are encountered,and the author solves the problem by proposing the decorative assembly mechanism and the component configuration model.The main contents are as follows:1.Proposed a decorative-based assembly model.The assembly mechanism accomplishes the assembly by creating a wrapper component and dynamically expanding the functionality of the component.Image processing platform is a flow-type algorithm channel to process images,and each component is a processing algorithm.Therefore,In view of the specificity and professionalism of the platform,the assembly mechanism fully complies with the experimental platform of the combination of algorithm platform requirements and the operating mode of drag and drop combination.2.Proposed a component-based configuration model.As the experimental platform has its unique drag and drop combination of algorithm components to form the operation of the algorithm channel,and the domain characteristics of the diversified process specification of the image processing algorithms.Based on the concept and characteristics of the configuration in industrial control field,a component-based configuration model is proposed.The model based on the foundation of the component brings the developers and users efficient development and use,and meets the image processing platform needs of the three-tier software architecture to lay the foundation for scalability,while the introduction of macro logic description makes experimenter restrict the image processing experiment platform algorithm channel rules according to their own business needs,so that users can participate in the development of the system to ensure that the system functions and the business needs of users can always be consistent.3.Based on the component configuration model and decorative assembly mechanism,the image processing experiment platform is designed and implemented,and some key codes are displayed.Finally,a large number of different angles and different levels of test experiments are conducted,and the assembly execution,decorative assembly mechanism,macro logic description and so on of the Image processing experiment platform algorithm component has carried on a comprehensive test verification.更多还原