论文部分内容阅读
提出了一种定位全自动LED引线键合机焊点位置的新方法。该方法先用图形处理器(graphic processing unit,GPU)对获取的LED微芯片图像进行基于均值查找的自适应窗口大小的中值滤波处理,再通过自适应阈值的图像分割算法确定微芯片的潜在区域,然后将所有潜在区域以位置关系分组存储,利用灰度基本对称特征筛选出每组中的最佳潜在区域,再对选定区域进行加权处理,最后计算质心以精确定位LED微芯片焊点位置。实验结果表明,该检测方法速度快、定位准,而且对LED微芯片形状的一致性无要求,更适合自动化操作,从而提高生产效率。
A new method of locating the solder joints of automatic LED wire bonding machine is proposed. In the method, a mean-search-based adaptive window size median filtering process is applied to a captured LED microchip image using a graphic processing unit (GPU), and the potential of the microchip is determined through an adaptive threshold image segmentation algorithm Then, all the potential areas are stored in groups according to the position relationship. The best potential areas in each group are screened by using the basic symmetry of gray scale, and the selected areas are weighted. Finally, the center of mass is calculated to accurately locate the LED chip joints position. Experimental results show that the detection method is fast, accurate positioning, and the consistency of the shape of the LED microchip no requirements, more suitable for automation, thereby increasing production efficiency.