基于PlanetLab的流媒体QoE测量与分析

发布时间：2018-05-14 03:03

  本文选题：视频质量 + QOE　； 参考：《电子科技大学》2014年硕士论文

【摘要】：随着计算机网络的普及和视频压缩技术的提高,网络多媒体应用在生活中的发展越来越快,用户也也越来越关心流媒体的质量。受到网络的影响,流媒体在传输过程中会发生数据丢失,产生各种视频损伤、卡顿、马赛克现象。这些损伤对用户而言是不希望发生的,对于运营商而言希望能够监测到这些损伤的发生、对这些损伤进行指标化以及做出相应的策略。以往的性能指标主要是从丢包率、带宽大小、延迟、抖动等QoS参数来描述网络服务性能。用户作为流媒体应用的直接使用者,有着最大的话语权,他们的感受是服务质量的最好说明。QoE(Quality of Experience)是从用户角度出发的评价指标,评价系统的整体性能。本文主要研究流媒体的测量与QOE分析,参考如今视频质量模型,提出了一种部分参考的客观视频质量评估方法。主要从图像压缩技术出发,针对图像的空域和时域信息做特征提取操作,并对这些特征信息建立数学模型,最后综合得出最终结果。对于空域的特征信息定义了能量变化值(EVD);时域特征信息使用广义高斯密度函数。对于流媒体应用的模拟主要以PlanetLab众多节点为实验平台,开发并部署可以进行实时流媒体交互、实时流媒体模拟的工具,并基于提出的客观视频质量评估算法实时监控流媒体视频的图像质量。流媒体经过实验平台的传输之后,会形成各种各样的损伤情况。最后通过主观评价得分来验证本方法的正确性。本文的半参考评估方法通过在发送端得到原始视频的特征信息,然后将这些特征信息发送到接收端对比发送端的特征信息会得到一个客观的质量评价,这个评价结果可以用来提供给服务商作为网络服务质量的参考。
[Abstract]:With the popularization of computer network and the improvement of video compression technology, network multimedia applications are developing more and more rapidly in daily life, and users are more and more concerned about the quality of streaming media. Under the influence of the network, streaming media will lose data during the transmission process, resulting in a variety of video damage, Carton, mosaic phenomenon. These injuries are not expected to occur to users, and operators want to be able to monitor the occurrence of these injuries, to target these injuries and to make corresponding strategies. In the past, QoS parameters such as packet loss rate, bandwidth size, delay and jitter were used to describe network service performance. As the direct users of streaming media applications, users have the greatest right to speak. Their feeling is that the best explanation of quality of service .QoS of experience is the evaluation index from the point of view of users and the overall performance of the evaluation system. This paper mainly studies the measurement and QOE analysis of streaming media. Referring to the current video quality model, a partial reference objective video quality evaluation method is proposed. Based on the image compression technology, the feature extraction operation is done for the spatial and temporal information of the image, and the mathematical model of the feature information is established. Finally, the final results are obtained. The energy variation value is defined for the spatial feature information and the generalized Gao Si density function is used for the time domain feature information. The simulation of streaming media application is mainly based on many nodes of PlanetLab. The tools for real-time streaming media interaction and real-time streaming media simulation can be developed and deployed. Based on the proposed objective video quality evaluation algorithm, the video quality of streaming media video is monitored in real time. Streaming media after the transmission of the experimental platform, will form a variety of damage. Finally, the correctness of the method is verified by subjective evaluation scores. The semi-reference evaluation method in this paper obtains the feature information of the original video at the sender, and then sends the feature information to the receiving end to compare the feature information of the sender to get an objective quality evaluation. The evaluation results can be used as a reference for service providers as the quality of service on the network.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN919.8
，

本文编号：1886022