基于多描述编码的视频组播网络带宽优化分配算法

为了最大限度的满足各种类型视频应用的带宽要求,提高不同类型视频用户的满意度,解决不同类型应用的视频流对网络带宽的需求差异很大的问题,该文提出了一个基于多描述编码的带宽分配模拟退火算法;实验表明,该算法比其他带宽分配算法更能充分利用带宽资源且能有效提高网络视频用户的整体满意度。     

交通监控视频压缩与网络传输系统设计

针对现有交通监控视频中采用的H.264算法复杂度高、运算量大、误码跨帧传播等问题,提出了基于专用编解码芯片ADV212硬件实现JPEG2000标准的视频压缩传输系统的设计方法。设计以ADV7180对PAL制式模拟视频进行采集,以ADV212完成对采集后视频数据的压缩处理,最后通过DM9000A实现对压缩数据的以太网传输。FPGA作为系统主控,主要实现对各芯片的初始化及逻辑控制。实验结果表明,在保证压缩图像具有较高峰值信噪比的情况下,系统对速率为166Mbps的视频图像进行压缩处理后输出速率为8Mbps,有效减轻了网络传输带宽压力,且便于网络终端进行视频数据存储。同时系统还具有集成度高、体积小、性价比高、压缩比可调等优点,支持多种视频监控场合使用。     

基于缓冲占用比无人机视频自适应传输算法

针对目前基于发送速率的自适应视频传输算法不能有效利用和适应网络传输,提出了利用数据缓冲占用比的无人机载视频自适应无线传输方法;通过计算发送缓存内的数据占空比的变化判断网络发送能力,进而调节下次采集的视频帧率,或在调节下次采集视频帧率之前调整压缩编码率,由此自适应的设置最大化传输速率;相对于传统基于线性增加发送速率的方法,文章算法根据数据占空比自动调节发送速率和编码率,使平均发送速率提高了0.25 Mbit/s,带宽利用率提高了8%,实现了在不同网络状况下发送速率和编码率的自适应最大化,通过有效利用网络传输,提高视频数据传输的实时性、可靠性和完整性。     

基于帧的多描述视频编码及其错误隐藏

为了解决视频流在不可靠网络上的错误传播问题,使用基于帧的多描述视频编码,提出采用预测的预处理和后处理过程方案,实现了描述间的冗余插入;实现了几种不同复杂度,不同性能的错误掩盖算法以适应多样化的网络传输环境.仿真实验结果表明,这种编码系统能有效控制视频流的错误传播,并且编码后的数据流更能适应各种网络传输状况.     

基于帧的多描述视频编码及其错误隐藏

为了解决视频流在不可靠网络上的错误传播问题,使用基于帧的多描述视频编码,提出采用预测的预处理和后处理过程方案,实现了描述间的冗余插入|实现了几种不同复杂度,不同性能的错误掩盖算法以适应多样化的网络传输环境.仿真实验结果表明,这种编码系统能有效控制视频流的错误传播,并且编码后的数据流更能适应各种网络传输状况.     

WDM波长路由光网络的波长分配算法

由于WDM网络中的传输和交换是以波长为粒度进行的。因此波长分配算法就成为光网络的研究热点。有效的算法可大大提高波长的利用率，并可充分挖掘WDM的带宽资源。介绍了WDM全光网络中静态和动态波长分配算法，尤其是结合大型网孔网络详述了几种动态波长分配算法。针对波长变换器在全光网络中的应用。提出了一种可以解决任意类型光网络的分层图算法，它可以一次性解决路由选择和波长分配问题，大大提高了该算法的效率。是未来光网络中非常有前景的一种波长分配算法。     

云环境下基于多虚拟机的在线视频监控系统设计

针对常规视频监控系统在线实时性不强,海量视频数据传输迟滞,任务管理单一等问题,提出了构建在云计算环下基于多虚拟机技术在线视频监控系统 ,利用云计算平台中的物理资源与服务资源提升在线视频监控系统数据处理能力,虚拟机可同时处理大量的视频监控数据,并将视频数据以云存储的方式存储于云端服务器,降低了设备建设成本,可根据不同用户需求定制相关服务。本系统基于云计算平台设计,应用数十台乃至数百台虚拟机对在线视频监控数据进行处理,设计实现了云平台下在线视频监控系统的结构设计、以太网通信接口设计、服务器硬件配置和虚拟机控制。在软件设计方面通过对各虚拟机资源利用率的计算而动态分配资源,从而可以有效减少网络传输系统状态信息的带宽开销。通过系统功能与性能测试表明,在常规公共网络10M带宽的情况下,本系统在线视频监控数据的传输延迟时间相比于传统视频监控减少了85%以上,监控视频数据量减少了75%以上。     

嵌入式网络数字带宽测量监控系统设计与实现

网络带宽表征网络传输信息的能力,是衡量网络服务质量的重要指标。为了实现对网络数字带宽测量的低成本和网络化的需求,提出了一种基于嵌入式系统的网络数字带宽测量监控系统设计与实现方法,依据以太网为基础,由带宽测量模块、采用MIC29302BU芯片为核心的电源模块,基于EEPROM芯片的网络接入模块、嵌入式网络数字监控系统自检模块以及模数转换模块组成,分析了该系统进行网络数字带宽测量的流程以及具体的实现程序代码。实验结果说明,该监控系统实现了对海量数字带宽的准确测量,并且具有较高的监控效率和精度。     

基于FPGA视频合成系统设计与实现

为实现将输入的六路BT656视频无失真地合并成一路BT1120视频输出,采用将FPGA技术和视频合成系统相结合的设计方法,用verilog语言设计完成有效视频数据的抽取、SRAM乒乓操作以及FPGA对于视频的拼接处理方法。该系统由视频输入解码模块、存储模块、输出解码模块、i2c模块以及时钟管理模块组成。经算法仿真和逻辑综合,该设计可以实现视频合成的基本功能,满足视频监控系统的实时性要求.综合结果表明该设计占用FPGA片上逻辑资源少,系统运行频率高。     

基于运动剧烈程度的无参考视频质量评价模型

网络视频质量评估具有无参考性、实时性、网络传输状态依赖性和主观视觉性等需求.本文以网络丢包引发的视频失真为研究重点,针对不同解码类型帧上的网络丢包引发的视频失真持续效应不同和网络丢包引发的人眼视觉感受与丢包所在帧视频内容的运动剧烈程度有着显著的关联等现象,提出了一种基于运动剧烈程度的无参考视频质量评价模型.该模型无需原始参考视频序列亦无需视频解码,对客户端接收到的码流分析其由丢包引起的视频损伤,标记受损宏块,建立受损宏块的失真持续效应和运动剧烈程度与视频质量间的关联,并完成视频质量评估.实验结果表明,该方法计算量小,实时性高,与主观评价结果一致性程度较高.     

Improving Perceived Quality of Live Adaptative Video Streaming

Quality of service (QoS) requirements for live streaming are most required for video-on-demand (VoD), where they are more sensitive to variations in delay, jitter, and packet loss. Dynamic Adaptive Streaming over HTTP (DASH) is the most popular technology for live streaming and VoD, where it has been massively deployed on the Internet. DASH is an over-the-top application using unmanaged networks to distribute content with the best possible quality. Widely, it uses large reception buffers in order to keep a seamless playback for VoD applications. However, the use of large buffers in live streaming services is not allowed because of the induced delay. Hence, network congestion caused by insufficient queues could decrease the user-perceived video quality. Active Queue Management (AQM) arises as an alternative to control the congestion in a router’s queue, pressing the TCP traffic sources to reduce their transmission rate when it detects incipient congestion. As a consequence, the DASH client tends to decrease the quality of the streamed video. In this article, we evaluate the performance of recent AQM strategies for real-time adaptive video streaming and propose a new AQM algorithm using Long Short-Term Memory (LSTM) neural networks to improve the user-perceived video quality. The LSTM forecast the trend of queue delay to allow earlier packet discard in order to avoid the network congestion. The results show that the proposed method outperforms the competing AQM algorithms, mainly in scenarios where there are congested networks.     

D2D-assisted user-centric adaptive video transmission in next generation cellular networks

Next-generation cellular networks need to provide seamless connectivity with higher data rates, increased capacity, and enhanced network coverage. As multimedia service demands in various heterogeneous devices grow rapidly compared to the underlying network’s capacity and bandwidth, the adaptation in multimedia streaming services is essential for providing satisfactory Quality of Experience (QoE). This paper develops a Device-to-Device (D2D)-assisted Utility-based Adaptive Multimedia (video) Streaming scheme (UAMS) using D2D communication in a 5th Generation (5G) cellular network where low-battery users may extend their streaming duration by spending lower reception energy with the help of D2D-assisted communication. The adaptation algorithm considers four utility functions: quality, power consumption, packet error ratio, and remaining battery of the user devices to adapt the bitrate dynamically and augment viewers’ experience. We formulate an optimization problem to maximize the joint utility function to provide the best adaptive multimedia content selected for transmission to the end-users either directly or via D2D Relay Nodes (DRNs) in every scheduling interval. We use a graph theoretic approach for choosing the best DRNs. Extensive simulations show the efficacy of the proposed scheme in terms of saved battery energy, churn rate, and QoE metrics compared to a few well-known existing schemes in the literature that do not use D2D communication.     

基于网络传输的实时图像融合系统设计

 为了实现图像融合系统远距离实时传输,提出一种基于TI公司TMS320DM642和TMS320DM365处理器平台的高速网络传输图像融合系统的设计方案。设计基于TMS320DM642融合模块和基于TMS320DM365网络传输模块,解决了图像融合模块和网络传输模块间的数据通信问题;在融合模块中实现图像配准和图像实时融合,融合图像数据在网络传输模块中进行H-264算法压缩并根据RTP(real-time transport protocol,实时传输协议)协议实现RTP封包处理,传输码率约为460 kb/s;利用线程间管道通信技术,实现了一种新的基于RTP/RTSP(real time streaming protocol,实时流传输协议)嵌入式流媒体服务器的搭建方法。实验结果表明：系统总体图像延迟为300 ms左右。     

基于卷积神经网络的焊缝缺陷图像分类研究

为有效地对焊缝缺陷进行分类，从而判断焊接质量的等级，对传统卷积神经网络进行改进，提出一种多尺度压缩激励网络模型（SINet）。将4组两两串联的3×3卷积模块与Inception模块、压缩激励模块（SE block）相结合。通过多尺度压缩激励模块（SI module）将卷积层中的特征进行多尺度融合和特征重标定以提高分类准确率，并用全局平均池化层代替全连接层减少模型参数。此外考虑到焊接缺陷数量不平衡对准确率的影响，采用深度卷积对抗生成网络（DCGAN）进行数据集的平衡处理，并在该数据集上验证模型的有效性。与传统卷积神经网络相比，该模型具有良好的性能，在测试集上准确率达到96.77%，同时模型的参数个数也明显减少。结果表明该方法对焊缝缺陷图像能进行有效地分类。     

基于云计算的统一通信警务协同研究与应用

针对统一通信的资金高投入,为了提高统一通信的普及性,提出了托管型统一通信方案;整体方案基于云计算架构,由业务层、呼叫控制层、终端接入层和业务与网络管理层组成,同时分析了统一通信业务,包括基本语音业务、多媒体会议、协同类业务、统一消息、终端业务等;研究了统一通信语音网关,设计中继网关提高组网能力,设计接入网关扩展终端业务;探索了云架构的实现技术,包括P2PSIP技术、虚拟化技术、视频编解码技术等;P2P架构能够使系统动态适应节点失效的情况,提高整个系统的可靠性与鲁棒性;针对线路可提供带宽的不确定性及不稳定性,为保证视频质量,视频会议采用可伸缩视频编码技术;研究成果应用于公安智能警务协同系统,提高了语音系统的用户感知;同时分析了行业现状,展望了未来前景。     

Acoustic streaming in lithotripsy fields: preliminary observation using a particle image velocimetry method

This study considers the acoustic streaming in water produced by a lithotripsy pulse. Particle image velocimetry (PIV) method was employed to visualize the acoustic streaming produced by an electromagnetic shock wave generator using video images of the light scattering particles suspended in water. Visualized streaming features including several local peaks and vortexes around or at the beam focus were easily seen with naked eyes over all settings of the lithotripter from 10 to 18 kV. Magnitudes of the peak streaming velocity measured vary in the range of 10-40 mm s(-1) with charging voltage settings. Since the streaming velocity was estimated on the basis of a series of the video images of particles averaged over 1/60s, the time resolution limited by the video frame rate which is 1-2 orders of magnitude larger than driving acoustic activities, measured velocities are expected to be underestimated and were shown a similar order of magnitude lower than those calculated from a simple theoretical consideration. Despite such an underestimation, it was shown that, as predicted by theory, the magnitude of the streaming velocity measured by the present PIV method was proportional to acoustic intensity. In particular it has almost a linear correlation with peak negative pressures (r=0.98683, p=0.0018).     

Proposal of Rip up & Re allocate Algorithm for Optical Layer 2 Switch Network

We are developing an optical layer-2s witch network that uses both wavelength-division multiplexing and time-division multiplexing technologies for efficient traffic aggregation in metro networks.For efficient traffic aggregation,path bandwidth control is key because it strongly affects bandwidth efficiency.For this paper,we propose a dynamic time-slot allocation method that uses periodic information of difference values of traffic variation.This method can derive near-optimal allocation with lower computational cost,which enlarges the maximum available network size compared with conventional time-slot allocation methods.Numerical results show that the proposed method enables dynamic path control in 1K-node-scale optical layer-2s witch network,which leads to cost-effective metro networks.     

Adaptive dynamic bandwidth allocation algorithm supporting multi-services over Ethernet passive optical networks

In order to ensure fairness of each optical networks unit (ONU) in access network and meet quality of service (QoS) requirements of differentiated service such as voice, video and data, Ethernet passive optical network (EPON) dynamic bandwidth allocation (DBA) algorithm should avoid light load penalty, choose different scheduling scheme according to light loading or heavy loading circumstances, adaptively adjust the order of ONUs at the case heavy loading in order to guarantee QoS. Aiming at this problem, an adaptive dynamic bandwidth allocation algorithm supporting multi-services is presented. The algorithm guarantees QoS through dividing service into voice, video and data putting them into different queues, optical line terminal (OLT) collects all requests of different services from different ONUs, divides the adopted scheduling scheme according to loading, adaptively adjusts the order of ONUs queue based on minimum delay at the case heavy load. Therefore, it can really guarantee QoS. Simulation results of the algorithm show that the algorithm not only can support multi-service and avoid light load penalty, but also can guarantee the fairness of each ONUs and QoS. It has a better performance than other algorithms particularly at the case of heavy loading.     

Multilevel region of interest guided bit allocation for multiview video coding

Multiview video coding (MVC) is an efficient compression scheme, by which the large amount of multiview video data can be effectively processed. However, it ignores the characteristics of the human visual system. In this paper, we propose a multilevel region of interest (ROI) based bit allocation strategy for MVC, which can take advantage of the visual redundancy to improve encoding efficiency. First, the macro block (MB) saliency is derived from depth information of the video sequence, and the multilevel ROI segmentation is conducted based on the MB saliency distribution. Then, the multiview video bit allocation strategy based on multilevel ROI is proposed. We have evaluated system performance with several multiview video sequences in JMVC 8.5 reference software. Experimental results show that the quality of ROIs has a considerable improvement when the bit rate consumed is kept consistent with JMVC. Meanwhile, the proposed MVC method can save bit rate while maintaining the overall image quality.