Analysis And Research On Broadcast Television Transmission Fault Prediction Based On Convolutional Neural Network

With the development of radio and television industry,radio and television wireless transmission field has experienced from the era of electronic tube to the era of solid-state and then to the digital era.Radio and television wireless transmission stations have gradually adapted to the development of the times,introduced the technology of each period and applied in the work.At present,the monitoring system of each transmission station has some monitoring equipment through acquisition.The key voltage and current values,which are identified as the capability of equipment failure when they exceed the threshold value set by human,cannot distinguish some soft failures from real failures.For this reason,in the actual equipment repair and maintenance,it is still necessary for technical personnel to accurately determine the location of failure points by combining the monitoring chart and their own experience.Therefore,it is necessary to study and adopt an artificial intelligence method to check the monitoring chart,identify the fault situation and eliminate the false alarm by imitating the human visual system.In this paper,according to the needs of the project,the problem of transmission fault judgment and classification of radio and television transmitting station is studied.Based on the recognition and processing of thetypical incident power monitoring diagram of the transmitter in the radio and television wireless transmitter monitoring system,this paper introduces the deep learning technology to judge and classify the transmission fault of the radio and television transmitter.Firstly,the data of voltage value,current value,state value,audio and video signal of the collected equipment are uploaded to the upper computer through the lower computer(various data acquisition controllers),and then the data are analyzed by the model designed based on vgg-16 model architecture,and then the hidden fault points are found out.In this paper,the typical transmitter incident power monitoring chart is processed and used as the input of CNN model.Through multiple iterative training,the corresponding super parameters of CNN model are optimized.Experiments show that the optimized CNN model can easily distinguish the soft fault and the real fault in the identification of the incident power monitoring diagram of radio and television wireless transmitter,and the recognition accuracy has reached about98%,which better meets the actual needs.