Defect Detection For Solder Joints Of SMT Based On Machine Vision

In the modern electronic industry,printed circuit board(PCB)as the carrier of electronic components,the mounting quality of on-board components directly affects the performance of the product,so the PCB patch component solder joint quality inspection is an important industrial production line process.With the use of surface mount technology(SMT),the placement of products to more layers and smaller volume,higher density development,but the traditional detection technology in the detection capacity and speed has been difficult to adapt to the needs of the new surface mounting technology.Therefore,it is important to study the PCB solder joint automatic optical patch element detection technology based on machine vision.The surface mounted in the production process,material processing technology and other factors caused by the solder SMD components appear a variety of defects,the defect characteristics vary,defect types can be divided into the lack of bonding,bridging,tin solder is too small,etc..In view of the existing joint patch element quality inspection above problems,based on the wireless transceiver module PCB board as the research object,analyzes the SMD capacitors and IC chip components of solder joints,joint development of SMD components detection system based on machine vision.The research contents include:1)According to the characteristics of the solder joints and the demand of the testing,the whole system is introduced,and the hardware composition and the selection principle of the system are introduced.Finally,select the appropriate light source,lens,capture card,camera and other hardware,build the experimental platform.2)For solder SMD components detection in image matching,the matching method of traditional correlation coefficient matching time is too long and can not rotate,the disadvantages of the circle projection improved fast feature matching method,the rough matching and fine matching after thinking;through comparing the template and sub graph to mean projection vector is the same a gray level matching,selection ofcandidates quasi matching point,and then on the basis of further fine matching,effective to speed up the matching,the algorithm reduces the search points of time.3)The common types of patch elements are two kinds of SMD resistors and multi pin IC chips,and the characteristics of the two types of solder joints are studied.The solder joint chip components through morphological feature extraction and shape feature to identify solder joint defects,proposed image connected domain by detecting the value of the number of IC chip components,compare the parameter statistics for each connected domain area and set the standard image,and calculate the image and standard image correlation coefficient test and discriminant solder defect types.4)Finally,in the Win 7 platform,based on the Mircrosoft Visual Studio 2010 development of the solder joints quality inspection system software,and debugging.Simulation results show that the visual detection system developed in this paper can quickly and accurately detect the existence of several common defects of solder joints,achieve better detection results,and has certain reference value for the follow-up quality testing automation of solder joint.