Abstract: The embedded system has less resources than PC, so the research results for improving video transfer real-time performance are too complex to be suitable for the embedded video surveillance system. This paper proposes a scheme to enhance the real-time performance of video surveillance system and to economize on memory resources. The scheme adopts the linear increment video traffic prediction algorithm and uses the packet paris to detect the available network bandwidth. It proposes a memory allocation algorithm to save memory resources.

Key words: video traffic prediction, network bandwidth detection, embedded video surveillance, real-time performance

摘要： 针对目前嵌入式系统资源有限及提高视频传输实时性方案不适合嵌入式视频监控系统的情况，提出基于视频流量预测和网络带宽检测的提高视频监控系统实时性方案。该方案采用线性规划视频流量预测算法和数据包的方法检测网络带宽，预测视频流量和网络带宽，据此调整视频缓存，达到提高视频实时性且节约系统资源的目标。

关键词: 视频流量预测, 网络带宽检测, 嵌入式视频监控, 实时性

CLC Number: 

• TP391