The Infrared Thermal Image Feature Extraction Of The Solder Joints Of The Printed Circuit Board

Solder joint is an important component of printed circuit board.It not only serves as a channel for electrical connection,but also provides mechanical connection between electronic components and the substrate.Virtual soldering is a common defect in circuit board welding.The existence of virtual soldering will lead to abnormal operation of circuit.The unstable phenomenon of good and bad will bring great hidden dangers to the debugging,use and maintenance of circuit.Therefore,it is particularly important to judge whether virtual soldering exists at each solder joint of circuit board and the degree of virtual soldering.At present,the commonly used non-destructive testing methods,such as X-ray,optical detection,flying needle detection,can not effectively detect such welding defects.Infrared nondestructive testing has many advantages,such as wide application range,non-contact measurement,fast detection speed,high detection accuracy,easy qualitative and quantitative analysis,and easy observation.It is considered as a new method for virtual welding defect detection of solder joints.This paper focuses on the detection and determination of virtual soldering and virtual soldering degree of printed circuit board solder joints by using infrared non-destructive testing technology.In this paper,the principle analysis of infrared nondestructive testing of solder joints,the construction of infrared nondestructive testing system for solder joints,the selection of testing parameters,the processing of infrared image sequence algorithm,the identification and determination of defect features are discussed.The heat transfer process of virtual solder joint structure under laser excitation was analyzed.A one-dimensional analytical model of temperature field distribution was established,and the relationship between surface temperature of solder joint and physical parameters as well as time was determined.On this basis,three infrared image sequence processing algorithms are proposed.In theory,three kinds of algorithm models are introduced.An infrared nondestructive testing system for laser-stimulated virtual welding of solder joints was built.The optimum laser spot size,incident angle,excitation power and excitation time were determined through experimental study.The parameters and software of infrared thermal imager are introduced.Three infrared image sequence algorithms are processed by self-compiling program in MATLAB.Pulse phase method,infrared signal trend analysis and apparent endothermic coefficient method were studied to process and extract features of infrared image sequence.The corresponding relationship between virtual welding degree and characteristic parameters wasestablished as a basis for judging virtual welding degree.