Research On PCB Solder Joint Visual Inspection And Chip Recognition For Digital Micro-spraying Process Packaging

With the development of science and technology,electronic devices tend to be smaller and more complicated in function.Printed circuit boards(PCBs)are also gradually becoming smaller and highly integrated with complex circuits.In order to ensure better performance,longer life and better adaptability of electronic devices,it is essential to package PCB to protect the electronic devices from damp,collision,corrosion and other external influences as much as possible.PCB packaging refers to spraying material onto the circuit board components and pin parts evenly according to the requirements of uniform in order to achieve isolation with the external environment to protect the chip from the external environment damage.In order to improve the precision and automation of PCB packaging,save the investment in advance molding of PCB packaging and reduce the waste of packaging materials,after fully investigating the research results of PCB-based image processing at home and abroad,PCB solder joint positioning and electronic chip character recognition algorithms are researched to meet the technical requirements for PCB packaging using digital micro-injection processes.The specific research contents of this paper are as follows:Firstly,the PCB packaging system is designed,which includes the design and selection of machine vision system.Furthermore,PCB image preprocessing technology is studied,including image filtering,image enhancement and color model conversion,which provides a basis for better and more accurate image processing and improvement of processing efficiency and quality.Secondly,the multi-shaped solder joint positioning algorithm for different kinds of PCBs is researched.Several proposed PCB solder joint positioning algorithms are studied,including the template matching based localization algorithm,the gray feature based localization algorithm and the color threshold based localization algorithm.A new solder joint localization method based on color features and neighborhood information is proposed in this paper,which is able to complete the whole positioning and segmentation of multi-shaped solder joints effectively and accurately.Thirdly,the positioning,character segmentation and recognition of electronic chips on PCB are researched respectively.First,the color distribution characteristics of the electronic chip in the PCB are analyzed and a positioning algorithm based on analysis of the color features and connected components is proposed.Then,the characteristics of electronic chip character distribution are studied and character localization algorithm based on edge projection and projection differentiation is proposed,which can effectively achieve the progressive division of multi-line characters one by one.At last,the BP neural network and the convolutional neural network are studied separately.The network systems are designed to train the chip character samples.The trained networks can realize the chip character recognition.Finally,the electronic chip package block is modeled and the digital micro-jet curing 3D printing system is used to perform the solder joint packaging and the electronic chip packaging experiments,respectively.