移动终端图像识别应用研究

发布时间：2018-04-24 22:04

本文选题：图像匹配 + SIFT算法　；参考：《北京邮电大学》2016年硕士论文

【摘要】：随着社会的发展,生活中充斥着越来越多的logo,而当前移动市场帮助人们识别这些logo的应用少之又少。本课题旨在实现一个系统帮助人们识别这些陌生元素,为人们生活提供方便。本课题的核心技术是图像匹配,详细介绍了 SIFT和SURF两种特征点提取算法,通过特征点提取、特征点匹配,最终完成图像的识别。在深入研究两种算法的基础上,本课题设计实现了 logo识别系统。该系统分为客户端、服务器端和图像匹配程序共三个部分。客户端的主要作用为logo的采集、展示、查询历史以及用户登录等,并且为了弥补logo数据库的不足,增加用户的参与度,设置了补充logo的功能。这些功能为用户交互提供了一个很好的平台。服务端的主要功能为同客户端进行数据交换,制定交互协议,将客户端采集到的logo与数据库已有的logo进行匹配,将得到的结果反馈给客户端,完成整个系统的流程。图像匹配程序主要完成图像的识别工作,将用户上传的图像与数据库的所有logo进行一一匹配,匹配成功则输出最终结果。目前客户端已在安卓系统上实现,服务器则在Windows服务器上实现,匹配程序使用C/C++语言编写,部署在服务器端。经验证,该系统能够较好的识别生活中的logo,在保证了一定的准确性的前提下,能够保障快速匹配。
[Abstract]:With the development of society, life is full of more and more logos, and the current mobile market helps people identify these logo applications are very few. The purpose of this paper is to implement a system to help people identify these unfamiliar elements and provide convenience for people's life. Image matching is the core technology of this paper. Two feature point extraction algorithms, SIFT and SURF, are introduced in detail. The recognition of image is finished by feature point extraction and feature point matching. Based on the research of two algorithms, this paper designs and implements the logo recognition system. The system is divided into three parts: client, server and image matching program. The main function of the client is the collection, display, query history and user login of logo. In order to make up for the deficiency of logo database and increase the participation of users, the function of supplementing logo is set up. These functions provide a good platform for user interaction. The main function of the server is to exchange data with the client, make the interactive protocol, match the logo collected by the client with the existing logo of the database, feedback the result to the client, and complete the whole system flow. The image matching program mainly completes the recognition of the image. The image uploaded by the user is matched one by one with all the logo of the database, and the final result is output if the match is successful. At present, the client is implemented on Android and the server is implemented on Windows server. The matching program is written in C / C language and deployed on the server side. It has been proved that the system can recognize logoon in life better and can guarantee fast matching on the premise of certain accuracy.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP391.41

【参考文献】