CASIA OpenIR  > 毕业生  > 硕士学位论文
Alternative TitleResearch & Application of image mosaic and virtual mouse on mobile device
Thesis Advisor卢汉清
Degree Grantor中国科学院研究生院
Place of Conferral中国科学院自动化研究所
Degree Discipline模式识别与智能系统
Keyword虚拟鼠标 图像拼接 尺度空间 运动估计 Virtual Mouse Image Mosaic Scale-space Sift K-d Tree Motion Estimation
Abstract摘要 随着移动终端软件和硬件的发展,计算机视觉技术在移动设备上的应用逐渐变成可能。视觉技术与移动终端的结合是一个重要的技术体现。目前已经有很多这方面的系统,比如手势识别,人脸识别等都有在手机上应用的例子。计算机视觉在移动设备上应用的局限性在于对图像的计算量很大,并且需要摄像机支持,对cpu的计算能力和手机电池的容量都有很高的要求,目前只有少数高端的智能手机能满足这方面的需求。我们的工作也是基于智能手机平台展开的。 本文主要贡献体现在两个方面: 本文的第一个工作是一个手机图像拼接系统,这个系统极大的扩展了手机摄像机的功能,它可能把摄像机拍摄的图像序列拼接成一幅全景图。尤其对于拼接文本更加实用。当用户看到一篇有用的文档时有时需要把它保存下来,可是摄像机很难清晰的把这样的一幅图像完整的拍摄下来,使用图像拼接功能我们只需要像扫描机一样使摄像机连续的在文档上移动,采集图像序列,拼接后的结果将是一幅表示整个文档的全景图。在这个工作中,我们使用了非常稳定的图像特征点算法,和快速的基于高维空间搜索数结构的特征点匹配算法。提高的拼接算法的精确性和速度。 第二个是利用对摄像机进行运动估计的方法在手机上实现了虚拟的鼠标。该方法利用了手机摄像机采集的视频数据实时的算出摄像机的位移量,从而可以驱动虚拟鼠标在手机屏幕菜单上移动,实现了鼠标功能。在这个工作里,我们采用了一些方法提高了计算速度,使得系统能基本上达到实时。本工作的贡献在于第一次提出了虚拟鼠标的概念,并提出了快速算法使得该系统能在计算能力有限的手机上做到实时。 关键字:虚拟鼠标,图像拼接,尺度空间,SIFT, K-dtree, 运动估计
Other AbstractAbstract Thanks to the rapid development of the software and hardware of computer, the application of computer vision technology on mobile phone becomes imaginable. One of the cross point between the vision technology and mobile phone is to use this technology to enhance the human machine interaction ability of the mobile phone. At present, there are many such systems, such as gesture and face recognition on mobile device. The basic drawbacks of these application is its high compute complicity and sometimes it require an on board camera. So the compute ability of the cpu must meet the task and the capability of the phone’s battery must be high. Our contribution is based on two works: The first is a document image mosaic system based on mobile platform,this System can mosaic pieces of small document images into large panorama image. It extends the camera function on the phone. When a user has seen a piece of text he wants to save it and read in the future. But the view angle of the on board camera is narrow, so it is very difficult to get a clear image in one shot. Use our system the user only need to capture part of the document. He can scan the document use the on board camera to get large set of the document image. We help the user to mosaic these images into one. In the second work, we present a novel virtual mouse system for mobile phones. The system is based on computer vision techniques. It first captures an image using the camera of the mobile phone, then extracts feature points of the scene and tracks them. By using this algorithm we can estimate the camera motion and give the translation vector to our virtual mouse interface. We use NOKIA 6630 camera phone as the developing platform which has a Mega Pixel camera on the back of it. The experiments show the promising results. The on board camera is used to capture the scene frame by frame, and then pixel correlation method is used to get the displacement vector. When users move the phone, the scene’s motion is opposite to that of the phone. keywords:virtual mouse, image mosaic, scale-space, SIFT, K-d tree, motion estimation ;
Other Identifier200328014604123
Document Type学位论文
Recommended Citation
GB/T 7714
盖永波. 移动设备上图像拼接研究与虚拟鼠标实现[D]. 中国科学院自动化研究所. 中国科学院研究生院,2006.
Files in This Item:
File Name/Size DocType Version Access License
CASIA_20032801460412(1476KB) 暂不开放CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[盖永波]'s Articles
Baidu academic
Similar articles in Baidu academic
[盖永波]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[盖永波]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.