基于特征识别算法的广播电视节目播放状态智能监控系统研究

Research on Intelligent Monitoring System for Broadcast Status of Radio and Television Programs Based on Feature Recognition Algorithm

在线阅读下载PDF

导出

摘要在智能化和数字化快速发展的背景下,实时监控广播电视节目播放状态对保障节目播出质量和稳定性具有重要作用,但现有监控系统在实时检测和异常识别方面存在精度不足和响应效率低的问题。为此,设计一种基于特征识别算法的广播电视节目播放状态智能监控系统。该系统利用音视频数据特征提取技术、模式匹配算法和异常检测模型,精准识别和分类节目内容,实现对播放状态的实时监测和异常检测。系统在实际测试中表现出较高的识别性能,平均准确率为96.26%,召回率为93.94%,错误率控制在2.8%以下;在并发连接数达到70时的总处理延迟平均为57.3 ms,满足广播级实时性需求。这些技术的结合,显著提升了广播电视节目播放状态监控的智能化水平和可靠性,具有重要的技术价值和应用意义。 Under the background of the rapid development of intelligence and digitalization,real-time monitoring of broadcast status of radio and television programs plays an important role in ensuring the broadcast quality and stability of programs.However,the existing monitoring system has problems of insufficient accuracy and low response efficiency in real-time detection and anomaly recognition.Therefore,an intelligent monitoring system of broadcast state based on feature recognition algorithm is designed.The system uses feature extraction technology of audio and video data,pattern matching algorithm and anomaly detection model to accurately identify and classify program content,and realize real-time monitoring and anomaly detection of playback status.The system demonstrated excellent recognition performance during the actual test,with an average accuracy rate of 96.26%,a recall rate of 93.94%,and an error rate controlled below 2.8%;when the concurrent connection number reached 70,the average total processing delay was 57.3 ms,meeting the real-time requirements of the broadcast level.Through the combination of these technologies,the intelligence level and reliability of the broadcast status monitoring of radio and television programs are significantly improved,which has important technical value and application significance.

作者彭宇涛温忠衡 PENG Yutao;WEN Zhongheng(Guangxi Radio and Television Big Data Technology Co.,Ltd.,Nanning 530023,China;Guangxi Radio and Television Information Network Co.,Ltd.,Nanning Branch,Nanning 530023,China)

机构地区广西广电大数据科技有限公司广西广播电视信息网络股份有限公司南宁分公司

出处《电视技术》 2025年第7期205-209,共5页 Video Engineering

关键词广播电视播放状态监控特征识别算法音视频处理异常检测 radio and television play status monitoring feature recognition algorithm audio and video processing anomaly detection

分类号 TN948.1 [电子电信—信号与信息处理]

作者简介彭宇涛(1969-),男,本科,高级工程师,研究方向为广播电视计算机应用技术。