| Fused magnesia is a kind of alkaline magnesia refractory with high purity,high melting point,compact structure and stable chemical properties.It is widely used in metallurgical industry,glass industry,cement industry and other fields.In China,the preparation of fused magnesia is usually carried out by electrofusion,and the magnesite ore is heated and smelted in an fused magnesia furnace to obtain a fused magnesia through a series of physical and chemical changes.At present,the production process of fused magnesium furnace still needs workers to patrol continuously at the production site,so as to take timely measures when abnormal conditions,etc which require manual intervention are found.However,the industrial environment of fused magnesia furnace is bad,on-site inspection is easy to pose a threat to the safety of workers.And the production process is generally carried out at night,the labor intensity of workers on duty at night is large,so it is prone to missed detection and misdetection when judging the working conditions.In view of the above problems,fused magnesium production enterprises urgently need advanced remote monitoring systems to liberate inspection workers from the harsh industrial scene environment.The existing video-based remote video surveillance system is subject to transmission bandwidth load when the video is transmitted remotely,the constraint is difficult to transmit high-definition video.And the video sensor is susceptible to dust,flame glare,water mist and high temperature interference in the scene,which causes the key feature areas such as the furnace mouth flame in the video to be interfered,which may lead to missed detection.To this end,this paper proposes a fused magnesia furnace visualization system,whose main function is to enhance the virtual enhancement of the monitoring video of the fused magnesium furnace production process,to achieve the enhancement of key features and remote monitoring.The key issue in the virtual monitoring of the magnesium furnace production process is how to realize the dynamic reconstruction of the furnace flame.Aiming at this problem,this paper proposes a dynamic texture-based flame visualization technology,in which the dynamic feature analysis and extraction of the key area of the furnace flame is carried out,and the kernel function-based classifier is designed to automatically recognize the visual characteristics of the flame in the current video.The characteristic parameters are then transmitted to the remote monitoring terminal for visual reconstruction of the virtual magnesium furnace.The operator can grasp the realtime production situation through the fused magnesia furnace visualization system.The main work of this paper is as follows:(1)Based on the fact that an experienced operator can directly distinguish the different working conditions by observing the dynamic changes of the brightness,shape and color of the flame,the dynamic characteristics of the furnace flame are analyzed,and a large number of live videos are collected.The dynamic texture data set of the furnace mouth of the magnesium furnace is established.Based on the experience of the field workers,three modes of normal melting,material shortage and feeding are defined for the furnace flame of the magnesium furnace.(2)For the flame mode identification of the fused magnesia furnace,the texture change of the furnace flame is regarded as a dynamic system,and the time series dynamic characteristics and image features of the image sequence are combined to establish the linear dynamic system model based on the flame dynamic texture.The dynamic system model is used to capture the static and dynamic characteristics of the furnace flame.A kernel function based on the main angle of the dynamic texture model subspace is proposed and designed which is used to design the classifier for dynamic texture features.A comparative experiment is performed on the dynamics texture dataset of flame in the furnace mouth of the magnesium furnace to evaluate the performance of the furnace flame modal classifier based on the main angle of the subspace designed in this paper.(3)In view of the lack of monitoring of the video data of the furnace flame which is intuitive visual information for the existing remote monitoring system,this paper has developed a virtual monitoring system for the fused magnesium furnace,which remotely conveys the dynamic texture characteristics of the furnace flame.That method avoids the situation that the direct transmission of video data is limited by the transmission bandwidth and cannot transmit high-definition video.And finally the virtual monitoring system realizes the visual reconstruction of the flame based on dynamic features in the remote terminal.The development of the virtual software of fused magnesium furnace includes two parts:local server and remote client.The functions implemented on the server side include:video feature extraction and remote transmission,database management,industrial live video playback,and group furnace demand trend drawing.The functions implemented on the client side include:remote reception of data,visual reconstruction of furnace dynamic flame,visual reconstruction of static model of fused magnesium furnace related equipment,three-phase current voltage and group furnace demand data display,group furnace demand trend Drawing,working conditions broadcast.In addition,the visualization system also supports multiuser simultaneous monitoring and cross-platform functions,and achieves a combination with virtual reality technology. |
PDF Full Text Request