摘要
受最大连接数、数据带宽等因素的影响,图像传输质量波动性较大,导致现有视频监控系统传输效率较低。基于EPOLL机制完成实时监控传输系统的设计与实现,建立EPOLL机制模型,分析摄像头与Linux内核架构的关系,定义图像传输协议,实现客户端图像帧请求及数据解析,最后于GTK+平台实时显示。实测数据验证,相对于传统的SELECT和POLL模型,EPOLL模型的适应性更强、传输效率更高、稳定性更好,为视频采集、视频监控等实时应用提供了一定的参考价值和技术支撑。
Affected by factors such as maximum number of connections and data bandwidth,image transmission quality fluctuates greatly,and it leads to that existing video surveillance systems have lower transmission efficiency.Based on the EPOLL mechanism,this paper completes the design and implementation of real-time monitoring transmission system.Firstly,we establish the EPOLL mechanism model,analyze the relationship between the camera and the Linux kernel architecture.Secondly,we define image transfer protocol for client image frame request and data parsing.Finally,we display it timely on the GTK+platform.The measured data prove that the EPOLL model is more adaptable,more efficient,and stabler than the traditional SELECT and POLL models.It provides certain reference value and technical support for real-time applications such as video capture and video surveillance.
作者
张绪红
黄睿
ZHANG Xuhong;HUANG Rui(School of Automation,Guangdong Polytechnic Normal University,Guangzhou 510665,China;Department of Computer Science,Guangdong University of Education,Guangzhou 510303,China)
出处
《实验室研究与探索》
CAS
北大核心
2020年第1期134-139,共6页
Research and Exploration In Laboratory
基金
广东省自然科学基金项目(2017A030313364)。