A Fault Detection Method of Terminal Display Equipment Fusing Semantic Understanding
A fault detection method that integrates semantic understanding is proposed to address the fault images of display equipment in terminal buildings.Firstly,a rolling text stitching technique is designed to extract interface text information.Then,based on the airport business background,semantic rules are integrated to train a fault classification model,which realizes intelli-gent detection of abnormal display interfaces and display information ambiguities in display devices.Finally,the neural network model compression technology is used to lightweight the model and deploy it on SOM-RK3399 embedded devices.The experiment shows that the classification accuracy of the detection method integrating semantic understanding module reaches 88.74%.This method can effectively solve the shortcomings of traditional fault detection techniques,improve fault detection efficiency,and re-duce manual inspection situations.
semantic understandingrolling text stitchingfailure detectionmodel lightweight
国家科技期刊平台
NSTL
万方数据
