摘要随着计算机视觉领域和移动设备领域的发展,应用开发人员正致力于将目标识别技术应用到移动手机上。本文构造了一种目标识别系统框架,同时也给出了系统性能评估的方法。本文中,在安卓手机平台上利用快速鲁棒性特征实现了该目标识别框架。以博物馆导航应用作为研究案例,根据提出的系统性能评估方法对系统做出了性能评估,本文给出了相应的测试结果及结果分析。博物馆导航应用方便了游客获得展品信息,为博物馆节省了人力资源。基于所构造的目标识别系统框架,可以开发出其他应用,也可以将目标识别这项工作延伸至其他的手机平台。63962
毕业论文关键词 目标识别 安卓 快速鲁棒性特征 系统性能评估
毕业设计说明书(论文)外文摘要
Title OBJECT RECOGNITION ON ANDROID MOBILE PLATFORM USING SPEEDED UP ROBUST FEATURES
Abstract As the great advances in the field of computer vision and mobile devices, application developers are now taking efforts to put object recognition technology on mobile phones. In this paper, an object recognition framework is constructed and an evaluation methodology to assess the system performance is proposed. Speeded up robust features on the Android mobile platform is used to implement the framework. In this paper, a museum guide application is realized as a case study using the proposed method, the test is done by means of the proposed evaluation methodology, and testing results and their analysis are given. The museum guide application is convenient for visitors to get information about the displays exhibit and help reducing the human resources. Based on the proposed framework, one could extend it to other applications, and the object recognition method proposed in the paper is further extended to other mobile platforms.
Keywords Object recognition Android Speeded up robust features System performance evaluation
1 引言 1
1.1 课题开发的背景 1
1.2 课题开发的意义 1
1.3 论文的结构 2
2 理论及相关技术 3
2.1 目标识别 3
2.1.1 相关背景 3
2.1.2 目标识别框架 3
2.2 SURF算法 4
2.3 匹配算法 7
2.3.1 分类理论 7
2.3.2 匹配算法 7
2.3.3 ROC曲线 8
2.4 JNI技术简介 10
3 系统构架设计 11
3.1 系统构架 11
3.2 手机客户端 11
3.2.1 安卓平台 11
3.2.2 客户端应用程序 12
3.2.3 目标识别算法 13
3.2.4 客户端通信协议 14
3.3 服务器端 14
3.3.1 JavaEE平台 17
3.3.2 数据库 17
3.3.3 匹配算法 17
3.3.4 服务器端通信协议 18
4 案例研究 安卓移动平台上利用快速鲁棒性特征的目标识别:http://www.751com.cn/jisuanji/lunwen_70838.html