Research On Image Processing Algorithm For Defect Detection Of Fused Silica Crucible

With the rapid development of the world and China’s solar photovoltaic industry,the demand for solar polysilicon key supporting materials,namely quartz ceramic crucibles for polycrystalline silicon ingots and smelting and purification,has increased rapidly.The flaws in the production process lead to various defects in the production process,which not only affects the quality of the crucible,but also causes serious economic losses to the subsequent production enterprises.Therefore,surface inspection has become a good yield and the core technology of enterprise market competitiveness.At present,the detection method of defects is mainly based on manual visual inspection.Traditional manual inspection cannot meet the requirements of real-time online detection of automated production lines.Machine vision as an emerging non-destructive testing method has been used more and more in the field of detection.The image processing technology based on machine vision-based inspection technology will become the future development trend.According to the quality inspection standards and testing requirements of crucible products,this thesis designs and implements the image acquisition hardware system of the detection system,selects appropriate cameras,lenses and light sources according to the defect detection standards and requirements,and determines the Light source illumination according to the characteristics of crucible defects,set up an image acquisition device for the surface of the crucible.The software system mainly includes two aspects,one is the realization of the whole mechanism and function module of the software,and the other is the research and implementation of the image processing algorithm for defect detection.The overall structure of the software is based on the MFC framework of Visual Studio 2010 software,loading the camera acquisition function module,and using the feedback type exposure acquisition algorithm to achieve image acquisition.In the image processing algorithm part,the multithreading technology is used to separately open the thread for image detection,to speed up the operation,perform pre-processing operations such as filtering and de-noising,region segmentation on the acquired image,and then perform image according to the standard deviation based background difference method.The segmentation is performed,and then the feature parameters of the region of interest are extracted,and the defect regions meeting the defect feature parameters are identified.The defect features that meet the conditions are displayed on the display interface of the software for easy to observe the detection effect,and the processing results are also transmitted to the PLC for data statistics.After testing in the actual environment of the site,the test results show that the image processing algorithm for the detection of fused silica defects can accurately,quickly and stably detect the image of crucible,and the detection efficiency and accuracy are better than manual detection,which can be applied well.In practice,it is a good image processing algorithm that can be applied to detect flaws in defects.