Design and Implementation of CNN-Based DTTB Image Intelligent Fault Diagnosis Platform

Conventional methodologies for monitoring digital terrestrial television broadcasting (DTTB) predominantly concentrate on assessing transmitted signal indices and transport stream (TS) quality. Alarms are activated when the monitored parameters exceed predefined thresholds. Nonetheless, these monitoring techniques fall short in their capacity to comprehensively gauge the end-users’ viewing experience. In a concerted effort to rectify this issue, an DTTB image intelligent fault diagnosis platform has been meticulously devised to oversee and diagnose broadcast quality through visual representation. The platform is meticulously engineered upon a Browser/Server (B/S) three-tier architecture and leverages convolutional neural network (CNN) as the core technology to establish a virtual digital artificial intelligence (AI) inspector. It harnesses streaming media services to facilitate swift and concurrent inspections of stations spanning the entire province. Experimental and deployment findings underscore the profound enhancement in monitoring precision and efficiency conferred by this platform, markedly augmenting the support capabilities of the DTTB operation and maintenance system.