基于几何特征分析法的焊缝特征点提取

Extraction of weld feature points based on geometric feature analysis

  • 摘要:
    目的 旨在解决角接、搭接以及V形坡口典型焊缝特征点提取过程复杂的问题。
    方法 提出了一种基于直线拟合的几何特征分析法,利用线结构光的激光条纹在图像里的特征,对图像进行预处理和条纹中心线提取。首先将所有提取的n个条纹中心点按坐标从小到大排序,在这几种焊缝特征上,利用点到直线距离公式判断关键点n0,从而将条纹中心点集进行划分,并对划分后的点集分别进行直线拟合,最后通过两直线求交点得出特征点。
    结果 结果表明,该方法过程简单且能够有效检测典型焊缝上的特征点信息,且特征点最大提取误差为0.273 mm,满足焊缝检测试验要求。
    结论 该方法能快速且准确地划分特征直线所属点集,并能检测出角接、搭接和V形坡口等多种类型焊缝的特征点,具有良好的适用性。

     

    Abstract: Objective The aim is to tackle problem of extraction process of typical weld feature points, such as fillet joint, lap joint and V-groove is complex. Methods A geometric feature analysis method based on line fitting is proposed. Image preprocessing and stripe center line extraction is achieved by using the characteristics of laser stripe of linear structured light in the image. First of all, the extracted stripe center point sets n are sorted according to coordinates from small to large. On these weld features, the key point n0 is determined by formula of point-to-line distance, so that the center point sets are divided, and divided point sets are respectively fitted by lines, finally the feature points are obtained through the intersection point of two lines. Results The results show that the method has simple process and effective in detecting information of feature points on the welds. The max ordinate error is about 0.273 mm, which meets the requirements of weld detection test. Conclusion The method can quickly and exactly divide point sets to which the feature line belongs to, and detect feature points of various welds, such as fillet joint, lap joint and V-groove. The method has good adaptability.

     

/

返回文章
返回