Institutional Repository
| 基于人脸检测的自动摄像跟踪系统开发研究 | |
| Alternative Title | Development and Research on Automatic Camera Shooting and Tracing System based on Face Detection |
| 任雍 | |
| Major | 计算机应用技术 |
| Supervisor | 方贵明 |
| 2008-06-04 | |
| Degree Grantor | 中国科学院软件研究所 |
| Degree Level | 硕士 |
| Place of Degree Grantor | 软件研究所 |
| Keyword | 自动跟踪摄像 人脸检测 肤色检测 类哈尔特征 Adaboost 级联分类器 |
| Classification | 暂无 |
| English Abstract | 自动摄像跟踪系统，是一种通过传感器检测或者数字图像处理的方法，控制摄像机自动的对运动中的人物或特定物体实施跟踪拍摄的系统，融合了计算机网络通信、计算机视觉、传感器网络等多个领域的技术，包含一系列软件和硬件设备，可以广泛运用在安全保卫视频监控，会议摄像，课堂和讲座摄像，表演摄像等场景，是近年来需求增长较大的领域。 人脸检测是数字图像处理和计算机视觉的一个分支问题，指的是在输入的图像中判断是否有人脸存在，并确定出所有人脸的位置、大小甚至姿势朝向的过程，作为进一步控制和处理的依据。这是计算机视觉方向最热门的技术专题之一，拥有非常广阔的应用空间。 目前还没有成熟的自动跟踪摄像系统应用于安全保卫或一般民用摄像领域，已有的一些类似系统也不能完全达到自动跟踪人物摄像的要求，性能距离全自动化无人干涉的跟踪摄像还有相当差距。 本文旨在运用计算机网络作为图像数据和控制信息媒体，运用数字图像处理领域计算机视觉中人脸检测和人脸跟踪的技术作为自动人物识别的依据，设计人物检测、人物跟踪、摄像机自动跟踪拍摄、人物位置运动估算、简单运动估计等算法策略，研究和开发一套完整的自动摄像跟踪系统软件，支持网络远程访问和控制，在无需人为操作和调整的情况下，实现对人物脸部自动跟踪拍摄的功能。这个系统适用于一般需求的安保视频监控和课堂自动摄像的应用。 本系统相对于已有的性能较好的类似系统——红外视频监控系统有许多明显优势，在达到很好的智能化、实时性、响应速度和准确性的同时，实现了设备的简化，大大降低了成本；同时在现场部署时不需要重新配置，后期维护简便；并且能够实现连续位置跟踪；系统稳定性优良等等。 |
| Call Number | 暂无 |
| Abstract | Automatic Camera Shooting and Tracing System, which is the combination of computer networks, computer vision and wireless sensor networks, including a series of software and hardware devices, detects object by sensor or digital image process. It can be used in the application of security monitoring, meeting and conference camera shooting and so on. Face detection, which is a part of the area of digital image process and computer vision, detects face from the inputted picture, calculating the position, size and pose in order to do more process. It is one of the most popular problems in the research of computer vision. Nowadays the similar systems can not completely match the needs of security monitoring and automatic camera shooting. This paper introduces the automatic camera shooting and tracing system based on face detection technology. This system which contains face detection, face tracing, automatic camera control algorithm, person position and motion calculation, basic motion estimation and so on, can automatically shoot the moving face without artificial manipulation. This system achieves desirable instantaneous property, response resolution and accuracy. Furthermore, the implementation complexity is effectively simplified, thus, the cost is significantly reduced. In addition, configuration renew is not necessary when the deployment and maintenance is subsequently simplified. Last but not least, continuously tracking is attained based on this system, robustness can be satisfied and so on. |
| Department | 多媒体通信和网络工程研究中心 |
| Content Type | 学位论文 |
| URI | http://ir.iscas.ac.cn/handle/311060/5766 |
| Collection | 多媒体通信和网络工程研究中心 |
| Recommended Citation GB/T 7714 | 任雍. 基于人脸检测的自动摄像跟踪系统开发研究[D]. 软件研究所. 中国科学院软件研究所,2008. |
| Files in This Item: | ||||||
| File Name/Size | DocType | Version | Access | License | ||
| 10001_20052801502906（3596KB） | 开放获取 | -- | Application Full Text | |||
Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.
Edit Comment