Performance Analysis And Optimization Research On Printed Circuits Board Automatic Optical Inspection

Automatic Optical Inspection system (AOI) is one of the key equipment of Surface Mount Technology (SMT). As many researches have been carried out, AOI theory and technology have been used widely. However, there are still some problems such as low adaptability, low veracity and low efficiency, which need to be solved.First, based on the principle of AOI, a reconstructive model, which includes component layer, defect layer and feature layer, is presented. After classifying the defect layer and modeling the feature layer, reconstruction of the inspection algorithm is implemented. And then object oriented technology is adopted to implement the reconstruction of software. At last, the reconstructive hardware structure is designed and the reconstruction of hardware is implemented. Via the reconstruction of the inspection algorithm, software and hardware, the adaptability of AOI is improved.First, based on the feature of the AOI testing, a criterion for feature evaluation and selection is proposed. And then based on the risk minimization Bayesian, the classification of AOI inspection is designed. At last, an algorithm based on statistic learning for PCB assembling defects inspection is produced to improve the veracity of AOI.Iterative self-organizing data clustering is adopted to plan the FOV of AOI. And then SAA based on regional research is adopted to optimize the inspection path of AOI. At last, a fast inspection algorithm is proposed to improve the efficiency of AOI.Based on the research of this paper, AOI is designed and produced. And experiments are carried out. Experimental results show that the AOI has high adaptability, veracity and efficiency. The researches of this paper are valid.