Research On Defect Detecting Technology Of Printed Circuit Borad

Along with the higher complexity and output of Printed Circuit Board(PCB) manufacture recent years, the traditional methods of PCB defect inspection such as got-up inspection can't meet demands of inspection gradually because of restricted contact, high cost or low efficiency. The Automatic Optic Inspection (AOI) based on image-processing had showed its advantages in inspection methods. In face of costly inspection device overseas and intense requirements of PCB's manufacture, research on a kind of rapid and useful technology of Automatic Optic Inspection have important academic and economic value. The author researched on PCB's defect detecting technology based on image processing in order to realize the high-precision and fast detecting of PCB defects.The author established a framework of defect detecting system based on image processing, constructed the structure of software, researched on PCB's defect detecting algorithm in pivotal function modules of inspection software and application of algorithm. The inspection system used the manner of un-contact, real-time, lossless inspection, which was composed of several parts, including movement control, illumination, image collecting and computer image processing system. Inspection software of system included modules that system self-checking and initialization,standard image reversion and saving it into gallery, defects checking and defects analyze. In these modules, standard image, defects checking and defects analyzing modules were pivotal function modules.Standard image reversion and save it into gallery module was accomplished before checking in batches, which realized reversion and preservation of standard image, by some measures such as drawing in EMS memory and save pictures by pieces, it solved the problem of getting,transforming,showing and preserving a lot of data caused by high precise in good time.Defect detecting was to find position of defects, contained image pretreatment,matching,contrast,eliminating noise and showing defects, the problem that noise weaken the speciality of PCB circuitry was resolved by segmenting with self-adapted threshold value and eliminating noise by fast open operation, and the conflict of multiprocessing taches,a lot of data in high precise image and fast checking were resolved by segmenting image into two gray level to compress data,building galleries for transferring and processing image fast,limiting the range of processing in every piece of picture when matching and noise-eliminating,integrating the procedures of checking into the same scan circle to process in synchronization.Defect analyze was to accumulate the characteristics,recognise defects and get the elementary evaluation of quality of PCB, analyze contained geometric characteristics'getting, defect recognition by choosing in many kinds of characteristics and elementary evaluation of quality of the whole PCB based on analysis. The Conflict between many characteristics and fast recognition was resolved and the time of analyze was decreased by limiting the range of sign, dendriform delaminating and judging gradually the defects which is recognized by many characteristics choosed.Experiment proved that defect detecting algorithms of PCB in this article can inspect fast and practicably.