Research On The Surface Defects Detection Of Flexible Integrated Circuit Substrate

Flexible Integrated Circuit Substrate is a printed circuit board,which is widely used in aerospace and electronic information industry.With the miniaturization of related products,the accuracy of the surface circuit portion of the flexible integrated circuit substrate has reached the micron level and continue to develop towards the nano level,so the accuracy requirements for the appearance inspection of the flexible integrated circuit substrate are become stricter.This paper design a surface defect detection system based on dual vision imaging system.Aiming at the surface defects of the copper-clad substrate,the high-precision detection algorithm is studied.The main research work is as follows:(1)Based on the on-site requirements for the appearance inspection of high-density flexible integrated circuit substrates,the calculation methods of key component selection parameters are studied,and a suitable dual computer vision imaging system is designed,combined with a high-precision three-axis motion control platform,adsorption mechanism and mechanical platform to build a complete surface defect detection system of flexible integrated circuit substrate.(2)An image correction algorithm combining Hessian matrix eigenvalues and gradient fields is proposed to obtain the image direction field and realize the correction of the flexible substrate image.(3)According to the color characteristics of the surface of the flexible substrate,an image segmentation algorithm based on color space is proposed.The multi-channel local histogram adaptive threshold method is used to segment the flexible substrate image copper foil area,text area and background area.(4)This paper proposes a multi-resolution image contrast algorithm based on similarity.First,a multi-resolution image is constructed by Gaussian pyramid,and then the image is compared with the principle of structural similarity and discrete cosine transform,so as to realize the abnormal detection of the surface of the flexible integrated circuit substrate.(5)Aiming at several common defects on the appearance surface of high-density flexible integrated circuit substrate,a defect classification and positioning algorithm based on deep learning is proposed to realize the type identification and positioning of defect targets in abnormal area images.This article provides a high-precision surface defect detection system for the production process of high-density flexible integrated circuit substrates,which solves the problem of automatic detection and has practical engineering value.