针对高速高精度贴片机的计算机视觉检测与定位问题,研究了BGA芯片的图像识别对中问题。首先通过分析贴片机生产中BGA的识别任务,指出了实现BGA识别的关键;然后提出了相应的实现算法的总体思想;接着对算法中要用到的点模式匹配算法作了大概的介绍;最后给出了算法的具体步骤。试验结果表明该算法在速度和精度上都能满足实际生产的需要,且相比点模式匹配算法,速度有了较大的提高。
According to the positioning and detecting problems of computer vision in the high speed and high precision SMT placement, the recognition arithmetic of BGA is researched. The paper firstly points out the key of the recognition of BGA by analysing the recognition tasks in the process of SMT placement production. Secondly, the corresponding image recognition arithmetic frame is given. Thirdly, the point pattern matching used in proposed arithmetic is described. Finally, the arithmetic is given out in detail. The result demonstrates that the speed and accuracy of the arithmetic are good enough to meet the practical application requirements. Besides that, the arithmetic performance is better than the point pattern matching arithmetic.