MPEG4视频编码算法的研究与实现
|
摘要
多媒体技术正在迅速发展，应用领域不断拓广。与之相适应，各种多媒体数据压缩编码标准也在不断地发展和完善。MPEG4是现在最重要最有影响的多媒体数据压缩编码国际标准之一。基于对象的编码思想使其具有高压缩比、可扩展性、可交互性等许多优点。MPEG4代表着未来多媒体数据压缩编码的发展趋势。
本文的实际应用背景是实现一个基于MPEG4标准的远程视频监控系统。远程视频监控要求对视频图像进行实时的压缩编码以及根据网络状况自适应地对压缩视频流进行流化传输，因此视频编码器必须能够进行快速的编码和比特率控制。基于上面的背景，本文第一部分针对MPEG4标准的核心部分——视频的压缩编码进行了两方面的重点研究。首先在对传统的运动估计算法进行研究和改进的基础上，提出了四步搜索块匹配的运动估计算法，减少了运动估计的时间，提高了编码效率；其次在分析MPEG2比特率控制的基础上，提出了一种新的比特率控制算法——平均响应比特率控制算法，该算法能够快速有效的实现自适应编码。
本文的第二部分以上面的研究成果为基础，结合ProjectMayo组织提供的视频压缩基本算法的实现代码，完成MPEG4视频编码器和解码器原型的软件实现，为远程视频监控系统的实现奠定基础。同时，为了将视频编码器和解码器原型更好的用于远程视频监控系统，本文提出了对原型进一步改进的方法，即在对视频图像压缩编码之前加入运动检测的预处理过程，从而进一步提高编码效率。
另外，本文对MPEG4标准中静态Sprite对象的生成和编码方法做了比较深入的研究，为将其用于背景图像变化少的一类问题的应用奠定了一定的理论基础。
Responding to multimedia technology rapidly developing and application domain promptly enlarging,many multimedia compression and coding standards progress and perfect continuously. Among them ,MPEG4 plays an important role in the multimedia field. Due to it's object-based idea,it has the virtue of high compression ratio,scalability,inter-activity with users and so on. For this reason,MPEG4 indicates the tendency of the multimedia compression and coding technique.
Based the research background of the remote video surveillance which makes real-time and adaptive compression and streaming video sequence necessary ,this paper first does the deep research on video compression and coding which is the hardcore part of MPEG4 mainly from two facets. The first facet is to put forward four-step block matching algorithm which can deduce the time of motion estimation and improve the coding efficiency,based on the traditional motion estimation algorithms. The second facet is to propose a new rate control algorithm,that is average-reaction rate control algorithm,based on the rate control of MPEG2. The new rate control algorithm can achieve rapid and efficient adaptive coding.
Based on the above research and the sourcecodes provided by the basic algorithms of video coding provided by the organization of ProjeictMayo and MPEG4,we implement the prototype of video encoder and decoder. At the same time ,this paper does some improvement on the prototype so as to make it be used in the remote video surveillance more properly. The improvement method which can achieve higher encoding efficiency is to add the pretreatment of motion detection before the encoding of the video pictures.
Finally,this paper does deep research on the static Sprite object in MPEG4. It will be the theory basis of the special applications in which the background of video pictures does no rapid motions.
本文的实际应用背景是实现一个基于MPEG4标准的远程视频监控系统。远程视频监控要求对视频图像进行实时的压缩编码以及根据网络状况自适应地对压缩视频流进行流化传输，因此视频编码器必须能够进行快速的编码和比特率控制。基于上面的背景，本文第一部分针对MPEG4标准的核心部分——视频的压缩编码进行了两方面的重点研究。首先在对传统的运动估计算法进行研究和改进的基础上，提出了四步搜索块匹配的运动估计算法，减少了运动估计的时间，提高了编码效率；其次在分析MPEG2比特率控制的基础上，提出了一种新的比特率控制算法——平均响应比特率控制算法，该算法能够快速有效的实现自适应编码。
本文的第二部分以上面的研究成果为基础，结合ProjectMayo组织提供的视频压缩基本算法的实现代码，完成MPEG4视频编码器和解码器原型的软件实现，为远程视频监控系统的实现奠定基础。同时，为了将视频编码器和解码器原型更好的用于远程视频监控系统，本文提出了对原型进一步改进的方法，即在对视频图像压缩编码之前加入运动检测的预处理过程，从而进一步提高编码效率。
另外，本文对MPEG4标准中静态Sprite对象的生成和编码方法做了比较深入的研究，为将其用于背景图像变化少的一类问题的应用奠定了一定的理论基础。
Responding to multimedia technology rapidly developing and application domain promptly enlarging,many multimedia compression and coding standards progress and perfect continuously. Among them ,MPEG4 plays an important role in the multimedia field. Due to it's object-based idea,it has the virtue of high compression ratio,scalability,inter-activity with users and so on. For this reason,MPEG4 indicates the tendency of the multimedia compression and coding technique.
Based the research background of the remote video surveillance which makes real-time and adaptive compression and streaming video sequence necessary ,this paper first does the deep research on video compression and coding which is the hardcore part of MPEG4 mainly from two facets. The first facet is to put forward four-step block matching algorithm which can deduce the time of motion estimation and improve the coding efficiency,based on the traditional motion estimation algorithms. The second facet is to propose a new rate control algorithm,that is average-reaction rate control algorithm,based on the rate control of MPEG2. The new rate control algorithm can achieve rapid and efficient adaptive coding.
Based on the above research and the sourcecodes provided by the basic algorithms of video coding provided by the organization of ProjeictMayo and MPEG4,we implement the prototype of video encoder and decoder. At the same time ,this paper does some improvement on the prototype so as to make it be used in the remote video surveillance more properly. The improvement method which can achieve higher encoding efficiency is to add the pretreatment of motion detection before the encoding of the video pictures.
Finally,this paper does deep research on the static Sprite object in MPEG4. It will be the theory basis of the special applications in which the background of video pictures does no rapid motions.
引文
[1] ISO/IEC WG11 MPEG Video Group. MPEG4 Video Verification Model Version 18. 0. The Pisa meeting of ISO/IEC/JTC1/SC29/WG11, No. MPEG2001/N3908, January 2001.
[2] ISO/IEC JTC1/SC29/WG11 Document N2502. Information Technology——Generic Coding of Audio-Visual Objects, Part2: Visual. ISO/IEC 14496-2.Final Draft International Standard.Nov. 1998
[3] Overview of the MPEG4 Standard, ISO/IEC/JTC1/SC29/WG11 N2725,March, 1999.
[4] Sanggyu Park. A New MEPG2 Rate Control Scheme Using Scene Change Detection. ETRI Journal. July 1996.
[5] Dapeng Wu, Y. Thomas Hou, Bo Li, Wenwu Zhu, Ya-Qin Zhang. An End-to-End Approach for Optimal Mode Selection in Internet Streaming Video:Theory and Application.
[6] 钟玉琢，贺玉文，王琪．基于对象的多媒体数据压缩编码国际标准MPEG—4及其校模型．科学出版社，2000，10．
[7] 钟玉琢，蔡莲红，等．多媒体计算机技术基础及应用．北京：高等教育出版社，1999．
[8] (美)吉布森(Gibson，J，D)等著；李煜等译．多媒体数字压缩原理与标准．北京：电子工业出版社，2000，8．
[9] 肖自美．图像信息理论与压缩编码技术．广州：中山大学出版社，2000，01．
[10] 贺小箭．视频压缩技术研究与MPEG标准视频的软件实现．硕士学位论文．1997．
[11] 高文，吴枫．MPEG-4编码的现状和研究．计算机研究与发展．2000．
[12] TEKALP．A．MURAT著；崔之祜等译．数字视频处理．Prentice-Hall．Inc，1996．
[2] ISO/IEC JTC1/SC29/WG11 Document N2502. Information Technology——Generic Coding of Audio-Visual Objects, Part2: Visual. ISO/IEC 14496-2.Final Draft International Standard.Nov. 1998
[3] Overview of the MPEG4 Standard, ISO/IEC/JTC1/SC29/WG11 N2725,March, 1999.
[4] Sanggyu Park. A New MEPG2 Rate Control Scheme Using Scene Change Detection. ETRI Journal. July 1996.
[5] Dapeng Wu, Y. Thomas Hou, Bo Li, Wenwu Zhu, Ya-Qin Zhang. An End-to-End Approach for Optimal Mode Selection in Internet Streaming Video:Theory and Application.
[6] 钟玉琢，贺玉文，王琪．基于对象的多媒体数据压缩编码国际标准MPEG—4及其校模型．科学出版社，2000，10．
[7] 钟玉琢，蔡莲红，等．多媒体计算机技术基础及应用．北京：高等教育出版社，1999．
[8] (美)吉布森(Gibson，J，D)等著；李煜等译．多媒体数字压缩原理与标准．北京：电子工业出版社，2000，8．
[9] 肖自美．图像信息理论与压缩编码技术．广州：中山大学出版社，2000，01．
[10] 贺小箭．视频压缩技术研究与MPEG标准视频的软件实现．硕士学位论文．1997．
[11] 高文，吴枫．MPEG-4编码的现状和研究．计算机研究与发展．2000．
[12] TEKALP．A．MURAT著；崔之祜等译．数字视频处理．Prentice-Hall．Inc，1996．