Research On Queue Scheduling Algorithm Of Video Surveillance Intelligent Detection System

With the rapid development of the Internet technology, the traditional net service can’t meet people’s need any more. The high-advanced multimedia service has become the mainstream on the net currently and it is also an integrate service with the ability to dispose characters, prints and videos. While how to provide high-quality service of distinguishing service has become a bottleneck in Internet development.The video surveillance intelligent detection system is such a system which can detect the picture, presetting bit, video, equipment and PTZ so that it can detect the breakdown of each passageway. This system has a higher demand for the quality of the service. It should be capable to distinguish real-time performances from non-real ones. It needs to differentiate the quality of priority service polices. At the same time, it also needs to guarantee the equality and efficiency.There are two main QoS service models, IntServ and DiffServ. Queens scheduling algorithm based on DiffServ model is the key technique to guarantee the service of QoS. Thus the thesis adopted queue scheduling algorithm to improve the service quality of the video surveillance intelligent detection system. The few main scheduling algorithms all have their drawbacks. WFQ algorithm performs best in terms of fairness and it could provide the scheduling with different weights. But when faced with complex service means, WFQ performs poorly. Its real-time application must line which costs a large amount of time and may lead to internet congestion.My thesis raised an improvement project for GPS-based WFQ on the foundation of DiffServ model. I combined token barrels technology with PQ and adopt the weight dynamic regulation strategy to offset WFQ’s disadvantage which can’t distinguish the real-time application. What’s more, it could provide a stable service rate which could guarantee the scheduling fairness.1) The thesis raised the strategy to combine WFQ and PQ in view of WFQ’s drawback in distinguishing the real-time application. An absolute priority queen was added into the algorithm to make it capable to dispose the service with high real-time need.2) The time delay increases sharply when data bursts. The algorithm regulate strategy using queen weight. In order to prevent major fluctuations and decrease the congestion rate. the algorithm distribute weight to queen again.3) When burst data appears, the queen adopts token barrels technology to stabilize algorithm’s service rate. It also decreases the delay jitter in the algorithm.The result of the experiment on video surveillance intelligent detection system suggests that the optimized WFQ algorithm still have the ability to provide different weight. It can also distinguish real-time application and provide stable regulation rate. Therefore the influence of burst data towards service delay and bandwidth allocation has been decreased.