Design Of Defect Detection System For Printed Circuit Board

As the basis for various underlying hardware circuits,the quality of the printed circuit board is related to the stable operation of the entire hardware system.With the development of manufacturing technology of PCB,PCBs have different colors and complex backgrounds,making defect detection even more difficult.However,traditional PCB defect detection techniques are challenged by low detection efficiencies,high false alarm rates,high cost,and dependence on hardware sensing.In order to effectively solve the above problems,this paper combined the actual needs of project engineering and designed a PCB defect detection system to provide an effective technical guarantee for PCB defect detection.In this paper,the defect detection refers to PCB surface defect detection.The main content of the research presented in this paper is as follows:In this paper,a set of acquisition system of digital imaging for PCB board was developed in light of the real demand for PCB defect detection.First,the system’s general design was finished and the primary components were decided.Second,the embedded system’s software and hardware were designed in accordance with the functional specifications,primarily focusing on the circuit and software layout of crucial modules,just like the motor,light source,communication and key data gathering.There are many problems in existing detecting PCB’s defect,including less defect datasets,complex background colors,low efficiency and high false detection rate.Therefore,a PCB defect data set was constructed in combination with a digital imaging device and the corresponding data set was expanded to solve the problem of insufficient defect datasets in this paper.The YOLOv5 defect detection model was choosen as the baseline.To make model focuse on target area and reduce the inflence of background,CBAM attention module was added in defect detection model.With the modified of loss funtion,the regression capability of bounding box was improving.For improving the detection of small object,the detection head was improved.Aiming at how to obtain high-definition PCB images,process the results of PCB defect detection and follow-up management,this paper designed the printed circuit board defect detection client software.Firstly,it combined the information collected by the lower computer with the camera to complete the digital acquisition of the PCB.Secondly,the image preprocessing algorithm was used to enhance the detailed information of the image to prepare for the subsequent defect detection.PCB defect detection was realized by invoking the YOLOv5 defect detection model.Finally,defect information was stored using the database and the corresponding quality inspection report generated.Through the various tests for embedded hardware and software and client software,it is proved that the defect detection system for PCB,designed in this paper,has perfect functions and stable performance.At the same time,it can effectively complete the PCB inspection defect.It effectively solves the problems of low accuracy and high false detection rate of traditional PCB defect detection.Simultaneously,combined with the database function,it can provide effective feedback to PCB manufacturers on relevant issues,further optimising PCB production and manufacturing processes.It has good practical value and development prospect.