Detection Methods For Surface Mount Chip With Rectangular Leads

In recent decades,with the development of electronic technology and the progress of society and economy,people have put forward higher requirements for the quantity and quality of electronic products,which has imposed great challenges on the electronic manufacturing industry.The core of electronic products is integrated circuit board and the latest manufacturing technology of integrated circuit board is surface mount technology(SMT).In order to meet the challenges of electronic manufacturing industry,machine vision technology is introduced into SMT to improve the accuracy and speed of production.The main application of machine vision technology in SMT is surface mount chip detection,in which the parameter learning,positioning,and defect detection of surface mount chip detection are realized.The surface mount chip with rectangular leads is the most common and widely used chip type.However,up to now,the detection problem of rectangular lead chips has not been well solved and lack of systematic research,which greatly limits the improvement of detection accuracy and efficiency of rectangular lead chip in SMT.Therefore,it is of great theoretical and practical significance to study the detection methods of rectangular lead chip in the actual electronic manufacturing industry.This article takes chip with rectangular leads as the main research object and conducts in-depth research on chip detection systems and detection algorithms in SMT.The main research contents are summarized as follows:Taking the actual necessaries in electronics manufacturing into consideration,this article firstly analyzes the characteristics of rectangular lead chips and several related definitions are defined.Then,based on the characteristics of rectangular lead chip,the actual requirements of the detection application are analyzed and a machine-vision based chip inspection system is designed.In this system,an industrial control computer is used as the main computing platform and the functions of motion control,image acquisition,image processing,and human-computer interaction are realized.Taking the SMT machine as an example,the effectiveness of the machine vision-based detection system in improving the accuracy and speed of electronic manufacturing is analyzed.In the detection algorithm of chip with rectangular leads,extracting the rectangular lead area from a chip image is a common problem.For the problem of interference areas commonly existed in chip images in the electronic manufacturing production field,the image segmentation algorithm based on the active contour method and the image segmentation algorithm based on the mixture model are improved in this article and better robustness of the segmentation is achieved.Based on these two improved segmentation algorithms,this article proposes the image-segmentation based lead region extraction algorithm,which can effectively extract the lead regions from the chip image and eliminate the influence of interference region.For the problems of large parameter errors in traditional parameter learning methods and the influence of subjective factors in manual entry,this article proposes the parameter learning method for chips with rectangular leads based on a layered back-off algorithm.In this method,the parameter learning of chips with rectangular leads is carried out automatically.the high accuracy of chip parameters is obtained,and the influence of error caused by manual entry is eliminated.In this chip parameter learning method,the root and foot of the lead are processed respectively,which can effectively reduce the complexity of the algorithm and saves the computing time greatly.In the surface mount process,the positioning accuracy of the chip determines the surface mount accuracy and also determines the production quality of electronic product.The deformation problems in chip pin size and the relative position of the chip lead in the chip pose a great challenge to the chip positioning algorithm.In order to deal with these deformation problems,an improved chip positioning algorithm based on deformable component models is proposed.In this method,the deformation problems in chip pin size and the relative position of the chip lead in the chip are effectively handled and locations of the chip and its leads are obtained accurately.A very important step before actually mounting of the chip is the defect detection.The main purpose of defect detection is to check whether the chip to be mounted is of the right package and whether there are defects.If the defective chip is mounted on the integrated circuit board,it will make the electronic products unable to achieve the designed function,or even bring serious security risks.In this article,the surface defects of chip with rectangular leads are classified into parameter-related defects and appearancerelated defects.For parameter-related defects,a method by comparing the results of chip positioning with the standard parameters of the chip is proposed;For appearancerelated defects,a texture feature-based defect detection algorithm is proposed.In the texture feature-based defect detection algorithm,Multi-order fractional wavelet packet decomposition is used to extract texture features of images,Variational Bayesian based mixture model algorithm is used to select image features,and support vector machine is used for image classification.The theoretical research and experimental verification of the detection algorithms for chip with rectangular leads are carried out in this article.The proposed detection system and detection algorithms can be applied in actual electronic manufacturing and be extended to the detection of chips with other packages.