基于最小二乘拟合法的焊点形状检测

自动光学检查是利用图像处理算法进行电路板故障检测与定位的技术。电路板焊点形状不规则故障，涉及到焊点几何形状的测量。本文提出一种新颖的基于最小二乘拟合法的焊点形状检测算法，对焊点形状轮廓图像进行曲线拟合，确定焊点形状近似的理想圆参数，即圆心和半径，以及焊点形状与理想圆之间的偏差，实验证实了该算法的性能。

Solder joint shape inspection using least square fitting

Automatic optical inspection is a technique for fault detection and location of printed circuit board using digital image processing algorithms. Shape irregularity of solder joint is a kind of faults in printed circuit board. Its detection and location involve geometrical shape determination. How to acquire geometrical shape information of solder joint is crucial for the detection and location of shape irregularity of solder joint. In addition, geometrical shape information of solder joint is a measure to evaluate soldering manner, apparatus and technician. A novel algorithm for solder joint shape inspection using least square fitting is proposed in this paper. The profile image of solder joint shape is curve fitted by using least square fitting. Then the approximate ideal circle parameters including center and radius are estimated. The deviation between solder joint shape and the approximate ideal circle is also obtained. Experiments were carried out. Experiment results validate the performance of the proposed algorithm.