Research On Quality Inspection System For PCB Bare Board Based On Machine Vision

Printed circuit boards(PCB)are the foundation of the modern information industry and their quality directly affects the quality and cost of many electronic products and devices.This work focuses on the defects of missing and offset holes in PCB bare board,short circuit of circuit diagram,circuit breaks,burrs,defects,and voids,and a small amount of solder and a large amount of defects in solder joint quality.A high-efficiency,high-precision,low-cost machine vision-based PCB bare board quality inspection system for small and medium-sized enterprise production lines was designed.Hence,on-line quality inspection of PCB bare boards was achieved to early recognise the defects,improve product quality and reduce costs.Firstly,the visual platform is used to obtain high-quality images.The PCB bare board images are preprocessed on the basis of obtaining high-quality images.Removing the noise in the PCB bare board image through median filtering,then using the gray-scale conversion method to enhance important information of circuit diagrams and solder joints..Finally,the images of circuit diagrams and solder joints can be well separated from the background board to obtain circuit diagrams and solder joint images via the global threshold of image segmentation.Secondly,the detection and recognition algorithms for three kinds of quality defects on PCB bare boards were studied separately.1)For the positioning holes on the PCB bare board used for the positioning the circuit board in the welding process,the quality of the positioning holes directly affects the PCB board for accurate automatic patching or automatic plug-in.This work proposes an improved random Hough transform circle detection algorithm for real-time and efficient detection of the positioning holes on PCB bare board.It can also extract the number of positioning holes center,radius and other relevant parameters of these holes to determine,if there are missing holes.2)For the circuit diagram on the PCB bare board used to fix the components,connect the circuit,and facilitate the component installation,the quality defect is an indispensable part of the PCB bare board quality defect.With regard to the five defects of the circuit diagram,this work uses a multi-feature joint identification method that combines the number of regions,Euler number,and area area to classify and recognize the defect images detected by the difference shadow method.3)The quality of PCB solder joints directly affects the quality of the applied electronic products.This chapter analyzes the characteristics of solder joints,selects three representative characteristic variables as the input vector of the fuzzy neural network structure,and its output represents solder joints.Therefore,the quality of the joint can be detected.This article carries on out the image processing experiment on the OpenCV program platform,then the examination results can be displayed through Labview.This improves developing time and leads to excellent results.