基于主动形状模型的焊缝定位算法

针对焊接环境对工人造成伤害和焊件材质影响焊接质量的问题,提出了一种应用于机器视觉激光焊缝定位系统的基于主动形状模型的焊缝定位算法.利用支持向量机对焊缝图像局部区域梯度方向信息进行目标图像初定位,运用主动形状模型方法建立特征模型,并计算特征模型和测试样本间马氏距离,从而对形状和姿态参数进行更新得到焊缝坐标.结果表明,基于主动形状模型的焊缝定位算法具有准确性高、抗干扰和误差小的特点,能够满足激光焊缝定位系统对准确性的要求.

Weld positioning algorithm based on active shape model

Aiming at the problem that the welding environment damages the workers’ health and the material quality of weldments affects the welding quality, a weld positioning algorithm based on the active shape model(ASM)was proposed and applied to the laser weld positioning system with the machine vision. The target image initial positioning for the gradient direction information of local region in weld image was performed with the support vector machine(SVM). Meanwhile, the feature model was established with the active shape model(ASM)method, the Markov distance between the feature model and test samples was calculated so as to update the shape and attitude parameters, and the weld coordinate was obtained. The results show that the weld positioning algorithm based on ASM has the characteristics of high accuracy, anti-interference and small errors, and can meet the requirements on the accuracy of laser weld positioning system.