基于边缘计算的智能视频分析算法研究

Intelligent Video Analysis Algorithm based on Mobile Edge Computing

为适应摄像头在智慧城市、智能交通、自动驾驶等新兴领域应用部署愈加广泛的需求,视频分析需更高精度、更低延时地响应分析结果。然而,这种高精度的分析同时也带来了巨大的计算资源需求,计算资源受限的摄像头无法胜任分析任务。边缘计算不仅可以解决本地摄像头计算资源问题,还可以显著降低向云端传输视频流数据的时间。本文探讨了利用深度强化学习方法,在边缘节点辅助摄像头集群视频分析任务场景下,根据当前网络系统条件动态决策,卸载部分指定摄像头上的分析任务,以在满足任务响应延时的约束前提下,最大化一段时间内任务分析的精度。仿真结果表明,本文提出的方法在任务的响应延时和准确度方面获得了良好效果。

With the increasing deployment of cameras in emerging fields such as smart cities,intelligent transportation,and autonomous driving,there is a growing demand for video analysis with higher accuracy and lower latency.However,achieving high-precision analysis poses significant challenges due to the substantial computational resources required,which can overwhelm resource-constrained cameras.Edge computing offers a solution by addressing the computational resource limitations of local cameras and reducing the time required to transmit video stream data to the cloud.This paper discusses the use of deep reinforcement learning methods to dynamically allocate video analysis tasks in a camera cluster assisted by edge nodes.The goal is to maximize the accuracy of task analysis within a given time while meeting the constraints of task response latency.Simulation results demonstrate that the proposed method achieves favorable performance in terms of task response latency and accuracy.