CASIA OpenIR  > 毕业生  > 硕士学位论文
Alternative TitleThe Design and Implementation of Networked Image Analysis Services Platform
Thesis Advisor李子青
Degree Grantor中国科学院大学
Place of Conferral中国科学院自动化研究所
Degree Discipline计算机技术
Keyword监控影像 Hadoop 网络服务 并行处理 分布式存储 Surveillance Video Hadoop Web-services Parallel Processing Distributed Storage
Abstract影像数据作为司法鉴定三大类之一,对刑事侦查、司法鉴定工作有重要意义,但是,在实际场景中,由于现场环境、设备条件局限,于犯罪现场获取的监控录像的影像质量往往很差,这就导致了许多影像资料不能作为诉讼证据。面对这些日益增长的“低质量”影像数据,如何提供便捷、有效的海量影像分析和管理服务,为不同系统提供灵活的影像质量改善功能,对办案人员从中发现出更多的线索和证据,更加有效地打击各类犯罪行为有重要意义。    针对公安刑侦破案中的监控视频和图像,本文提出了海量影像数据并行处理与分布式存储方案,并实现了面向多系统的网络化影像分析服务平台。该平台可以保证不同系统按需调取相应服务,对海量的受损图像进行一些实时或非实时的复原和主观优化处理提高画质来协助工作,能够改善提取到的犯罪现场影像的质量,并有效存储这些线索和证据,从而更加有效地打击各类犯罪行为。    本文主要研究解决如下三个影像分析处理中的实际问题:    1)面向监控视频的实用影像分析处理算法的实现、封装与集成    2)实现监控影像的并行处理与分布式存储共享    3)整合分散的影像处理功能成为可操作的标准服务    针对上诉问题,本文从分析网络化影像分析服务平台的需求出发,研究基于Hadoop的影像处理与存储和基于Web Services的监控影像处理算法服务封装与发布管理,以使其能够满足监控影像分析服务平台的数据量大、访问量多与算法复杂等特点。
Other AbstractBeing one of the three methods for judicial expertise, image data is vital to the criminal detection and judicial expertise. However, limited by the various environments or equipments, quality of the surveillance videos obtained from the crime scene is usually under average, making amounts of the videos unqualified for the effective evidence in litigation process. Facing the increasing number of "poor quality" image data dilemma, to provide a convenient and effective access to image analysis and management, as well as a flexible and versatile access to the quality improvement in different systems, is of vast importance to help the investigators to discover more clues and evidence to crack down on the crime.    Aiming at the surveillance videos and images of public security forensic investigation, this thesis proposes the schemes of mass image data parallel processing and distributed storage, and realizes the networking image analysis service platform oriented to the multi-system as well. This platform ensures that different systems can retrieve relevant services according to the demands, together with improving the mass-damaged-image’s quality by means of real-time or non-real-time recover as well as subjective processing to assist the investigation. Besides, it can also improve the retrieved image’s quality of crime scenes and effectively store the evidences, thus strongly fighting against crimes.    This thesis mainly studies and solves three practical problems:    1)Realizing and integrating the surveillance video oriented processing algorithms    2)Realizing parallel processing and distributed storage of massive surveillance video    3)Integrating disperse image processing functions to form operable standard services    With regard to the problems above, this topic is intended to investigate the demands of networking image analysis service platform. It researches image analysis and storage based on Hadoop as well as the surveillance image processing arithmetic encapsulation and release management based on Web Services for the sake of meeting the requirements such as great capacity of data, large visitor volume and complex arithmetic.
Other Identifier2010E8009070043
Document Type学位论文
Recommended Citation
GB/T 7714
黄永卿. 网络化影像分析服务平台设计与实现[D]. 中国科学院自动化研究所. 中国科学院大学,2013.
Files in This Item:
File Name/Size DocType Version Access License
CASIA_2010E800907004(2812KB) 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
[黄永卿]'s Articles
Baidu academic
Similar articles in Baidu academic
[黄永卿]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[黄永卿]'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.