Research On Online Quality Inspection Of Digital Surface Mount Products

The level of modernization in today's society is rapidly increasing.The SMT industry is no exception.The types,shapes,and sizes of PCBs have changed a lot.The requirements for inspection technology are also increasing.Traditional testing methods cannot meet the needs of modern production..With the comprehensive promotion of computer vision technology in recent years,machine vision technology has been applied to many fields,and the machine vision defect detection technology has completely replaced manual visual inspection.This technology has also become the development direction of PCB surface defect detection.This paper proposes a dual-frequency FTP and iterative-based phase unwrapping algorithm based on the related technical theory and key technologies of machine vision detection algorithms,and compares it with single-frequency FTP,two-step phase shift algorithm and four-step phase shift algorithm.Experiments show that the two algorithms have better detection accuracy and detection speed when measuring the step surface,and have wider application range.They can be applied to the measurement of simple smooth surfaces and complex step surfaces at the same time.An efficient detection algorithm for actual measurements.This paper classifies PCB surface defects and designs detection schemes for PCB board IG lead appearance inspection,component inspection,optical character inspection,solder joint and solder resist inspection and shell appearance inspection.Different features include: spot detection algorithm,OCR,quantity statistics,etc.At the same time,the PCB board was positioned and OK/NG determined.The experiment verified the feasibility of each program.In order to improve the automation and intelligence of the SMT production line,this paper designs several real-time data acquisition schemes for the production line,which can collect and store various production data of the production line through software acquisition,sensor hardware acquisition and database or file collection.After data collection,it is classified and processed.Then five key data of patch pressure,solder paste temperature,solder paste thickness,cooling rate and heating time are used for the construction of quality optimization model.Finally,a five-element linear regression model is obtained.The analysis draws the weight of various influencing factors and provides technical improvement programs for the PCB production process.