Research On Surface Defect Detection Technology Of Diode Glass Bulb

As an important material of diode package,glass bulb has been widely used because of its good heat resistance,mechanical strength,moisture resistance and reliability compared with the diode performance of epoxy resin package.Due to the limitations of the production process in the actual production process will have gas lines,cracks,edges and corners and other defects,undermined the uniformity of glass products,light transmission and thermal shock resistance.Seriously affecting the electrical insulation performance of the diode.Surface defect detection methods based on image processing and pattern recognition technologies are becoming more sophisticated and become an important method for automatically identifying surface defects.In this paper,image processing techniques and pattern recognition techniques are applied to the defect detection of diode bulb to realize the automatic detection of diode surface defects.This paper introduces the latest research progress of glass flaw detection,expounds the causes of glass flaw defects and the traditional detection methods.A complete diode flaw detection scheme is proposed,in which the hardware platform is used for diode glass image acquisition.The software algorithm design mainly includes the pretreatment,feature extraction and defect identification of the diode glass bulb,and the detailed analysis of the method Key technical and difficult issues.In the processing stage,adaptive median filtering is used to perform the blanching and morphological filtering of the collected glass bulb images.In the feature extraction stage,texture features,moment invariant features,and gray feature extraction algorithms are used,and based on the improved neural Network algorithm to extract the features of many kinds of glass images with different defects.In the end of this paper,a large number of diode glass images collected in the field are extracted and the surface defects of diode glass are identified by neural network.Through the experimental verification and testing,the detection method proposed in this paper can effectively detect the existence of cracks on the surface of the glass bulb,breakage,debris on the inner wall of the glass bulb has a good recognition effect.The overall recognition rate as high as 91%,experimental results show that the method has some engineering value.