While media audiovisual programs have brought great convenience to people's study, work and life, there are also a lot of bad information flooding them. This research mainly discusses the design of media audio-visual program monitoring system based on network topology. The system mainly monitors audiovisual programs through the network topology scanning module. For security reasons, the system network structure is divided into two parts: internal network and external network. The supervisory system is located in the internal network and is physically separated from the external network. In the initial scanning process, first obtain the starting device information in the configuration file, which is the routing device by default, read the relevant information in the MIB through the SNMP agent, and analyze the neighboring routing device and its downstream switching devices and hosts. Information and other topological information, and transmit the topological information to the data organization and distribution module. In order to avoid a large load on the network operation of the target autonomous domain, it is recommended to scan a larger network once an hour, or you can make your own decision based on the actual situation. In order to be able to quickly and accurately discover the changes in scanned data, the system uses MD5 verification to generate a "fingerprint" for each record to identify the data. Through this algorithm, the amount of information in the database is effectively reduced, and the efficiency of retrieval is improved, and the system is optimized. In order to reduce the misjudgment of legal shadow marks in shadow marks, after verifying 10,000 sets of sample shadow mark images and adjusting the contrast recognition threshold, the missed judgement rate can be less than 2%. The system designed in this research meets the needs of program supervision.

媒体视听节目为人们的学习，工作和生活带来了极大的便利，但同时也充斥着许多不良信息。本研究主要讨论了基于网络拓扑的媒体视听节目监控系统的设计。系统主要通过网络拓扑扫描模块监视视听节目。出于安全原因，系统网络结构分为两部分：内部网络和外部网络。监控系统位于内部网络中，并与外部网络物理隔离。在初始扫描过程中，首先获取配置文件中的启动设备信息，默认为路由设备，通过SNMP Agent读取MIB中的相关信息，并分析邻居路由设备及其下游交换设备和主机。信息和其他拓扑信息，并将拓扑信息传输到数据组织和分发模块。为了避免给目标自治域的网络操作带来很大的负担，建议每小时扫描一次更大的网络，或者您可以根据实际情况自行决定。为了能够快速而准确地发现扫描数据的变化，系统使用MD5验证为每个记录生成“指纹”以识别数据。通过该算法，有效减少了数据库中的信息量，提高了检索效率，并对系统进行了优化。为了减少对阴影标记中合法阴影标记的误判，在验证了10,000套样本阴影标记图像并调整了对比度识别阈值之后，漏判率可以小于2％。本研究设计的系统满足程序监督的需求。