| The ceramic dielectric resonator filter is an important part of the 5G communication base station.In the production of the filter,defects such as PCB board bending,silver plating damage,and ceramic body cracking may occur.These defects will affect the performance of the filter.In order to ensure the quality of the produced filter requires testing of the 5G ceramic filter.The research content of this paper is to design the defect detection algorithm of the filter ceramic surface and the filter point cloud processing algorithm.This paper designs a point cloud processing algorithm for the point cloud file of the filter obtained by the structured light scanning system,successfully measures the flatness and coplanarity of the filter PCB,and writes the core part of the point cloud processing algorithm into a callable function to realize the quick import of point cloud processing algorithms for different new products.After analyzing the two-dimensional image on the side of the filter,based on the computer vision image processing algorithm,a defect detection algorithm for the surface of the filter ceramic body was written,and the algorithm was packaged into an independent function.The detection algorithm can identify two types of defects: bottom exposed and cracks on the ceramic surface of the filter.After experimental verification,the algorithm's correct recognition rate of exposed defects is 94.87%,and the probability of judging the correct sample as a defective sample is 0.55%,which meets the needs of the inspection project.The experimental results of the crack defect detection algorithm are: the detection rates of the two detection surfaces are 92.11% and 90%,and the false detection rates are2.97% and 2.19%,respectively.This paper also designed an optimization algorithm for cracks.After two optimizations,the false detection rate of cracks on the non-character side dropped by 0.89% and 0.94%,respectively,and the final false detection rate on the non-character surface dropped to 1.14%. |
PDF Full Text Request