一种基于图像分析的引线曲率测量算法研究

Study on a wire curvature measuring algorithm based on image analysis

在引线成形中,引线的曲率分布对引线成型有着重要影响。提出了一种依据图像分析技术的测量方法,利用机器视觉获得的引线键合的视频图像,对分解的序列图像进行自适应阈值分割、骨架提取、细化处理后,获得引线各数据点坐标,并对数据点进行高斯光顺,进而用点到弦距离累积算法计算引线各点曲率分布。MATLAB仿真结果表明,该方法与现有曲率计算方法相比,计算复杂度小,具有较好鲁棒性;为进一步研究引线成型提供了基础。

The wire curvature distribution has significant influence on wire looping formation.In this paper, a new method for measuring wire curvature distribution which is based on image analysis is introduced.The wire bonding video which is captured by machine vision is used.The video is divided into sequence images, after adaptive threshold segmentation, skeleton extracting and thinning; finally the coordinates of wire data points are obtained.Then the data points were smoothed with a small width Gaussian kernel.The curvature distribution is calculated by means of point-to-chord distance accumulation algorithm.Through simulating with matlab, the method has lower computation complexity and better robustness than the existing method.It provided fundamental for further studying the wire bonding looping formation.