基于RANSAC算法的2 D Metrology在BGA质量检测中的应用研究

摘要

For BGA chip positioning problem in quality detection,this paper proposes a 2D Metrology chip positioning method based on RANSAC algorithm. First carries on the 2D Metrology modeling,the second object is added to the 2D Metrology model,and then set 2D Metrology model parameters,using 2D Metrology,finally get the 2D Metrology model object positioning results. Based on RANSAC algorithm, this method can accurately identify the model of object boundary characteristics, simple, high precision, good robustness and strong anti-jamming capability. The experiment show that the proposed method is more effective than the chip positioning method based on least square method.%针对BGA质量检测中芯片图像的识别定位问题,提出了一种基于RANSAC算法的2 D Metrology 芯片定位方法。首先进行2 D Metrology 建模,其次添加对象到2 D Metrology模型,然后设置2D Metrology模型参数,应用2D Metrology,最后得到2D Metrology模型的对象定位边结果。该方法基于RANSAC算法,能精确识别模型对象中的边缘轮廓特征,简单易行、定位精度高、鲁棒性好、抗干扰能力强。文中通过实验,与常用的基于最小二乘法的芯片定位方法进行对比,验证了该方法的有效性。