Real-time Monitoring System Of High Temperature Melt Based On Image

To address the need of real-time monitoring of high temperature melt production process, basing on the principle of color temerature, radiation thermometry techniques and MATLAB image analysis technology has been studied. Reported the status of image recognition technology og high temperature melr, analyzed the significance of this topic and raised a image segmentation method basing on visible radiation thermometry. A fast, accurate and stable online high temperature melt monitoring system was designed and it is able to complete the identificaiton of high temperature melt, and get the temperature field and geometry information of the melt as well.MATLAB computing software and Siemens WinCC monitoring platform have been selected as a basic platform with color CCD camera selected as a sensor, grey level of three color of visible light is extracted and image collection is accomplished using MATLAB USB camera connection module. After tricolor threshold segmentation has been completed to the colleted image, brightness field of the melt is established through the gray value of three primary color and brightness formula. Set the axis value of the most bright point as the highest temperature point, temperature field of the melt surface is achieved using colorimetric thermometry. Identify the high temperature melt by temperature threshold segmentation and carry out a series of calculation of the melt image, including calculate the centroid of the image, calculate the image deflection, rotate the image, calculate the jet angle, calculate the length, width and area of the image. Transfer the MATLAB calculation results to WinCC monitoring platform via text file, and then complete the online monitoring of high temperature melt image.When selecting the appropriate design scheme, first of all, start from the principle of radiation thermoetry, compare and analyze the calculation result between colorimetric thermometry and three primary color thermometry in detail and the former is selected. Base on this, a temperature threshold segmentaion method is designed basing on visible radiation. This mehold set the temperature threshold using the colorimetric temperature field of melt surface and realized melt image segmentation. Detail compared the image recognition result of the three primary color threshold segmentation, iterative threshold segmentation, K-means clustering segmentation and temperature threshold segmentation in this paper. The analysis shows that temperature threshold segmentaion is superior to the above several other segmentaion scheme. Base on this scheme, developed a melt image detection and analysis tool using MATLAB, and achieved real-time online monitoring of the flame image combining with WinCC. The experiment result shows that the designed high temperature melt image real-time online monitoring and analysis system is good at stable and real-time operation, meet the purpose of high temperature melt image geometry measurement and temperature field monitoring, and create the opportunity and condition to improve product quality.