程鑫, 凌鹤, 李硕. 汽车焊缝密封胶条的光条中心提取算法研究[J]. 应用光学, 2019, 40(5): 823-828. DOI: 10.5768/JAO201940.0502006
引用本文: 程鑫, 凌鹤, 李硕. 汽车焊缝密封胶条的光条中心提取算法研究[J]. 应用光学, 2019, 40(5): 823-828. DOI: 10.5768/JAO201940.0502006
CHENG Xin, LING He, LI Shuo. Light bar center extraction algorithm for automotive weld sealant strips[J]. Journal of Applied Optics, 2019, 40(5): 823-828. DOI: 10.5768/JAO201940.0502006
Citation: CHENG Xin, LING He, LI Shuo. Light bar center extraction algorithm for automotive weld sealant strips[J]. Journal of Applied Optics, 2019, 40(5): 823-828. DOI: 10.5768/JAO201940.0502006

汽车焊缝密封胶条的光条中心提取算法研究

Light bar center extraction algorithm for automotive weld sealant strips

  • 摘要: 汽车焊缝涂胶能够提高车身密封水平, 需要对胶条进行检测并且获得三维信息。激光三角法是应用广泛的检测方法, 其关键技术之一是准确提取光条中心。但针对黑色被检胶体, 光条中心提取鲁棒性差, 因此提出了一种基于光条分割的Hessian矩阵定位和高斯曲线拟合的算法。首先, 将整体图像对比度提高, 运用去噪算法消除干扰的噪声; 然后用极值法初步定位光条, 结合Hessian矩阵对激光条的中心进行定位; 最后利用高斯曲线准确拟合出激光条的中心。实验研究表明, 在对光条图像添加噪声幅度为100的白噪声之后进行处理, 同样能够将精度控制在亚像素级别, 能够提高黑色被检测物体的光条的稳定性。

     

    Abstract: Automotive weld gluing can improve the body seal level, and it is necessary to detect the strip and obtain three-dimensional information. Laser triangulation is a widely used detection method. One of its key technologies is to accurately extract the center of the strip. However, for the black detection colloid, the strip center extraction is poorly robust, so an algorithm based on strip segmentation for Hessian matrix positioning and Gaussian curve fitting is proposed. Firstly, the overall image contrast is improved, and the noise is removed by the denoising algorithm. Then the extreme value method is used to locate the light strip, and the Hessian matrix is used to locate the center of the laser strip. Finally, the center of the laser strip is accurately fitted by the Gaussian curve.Experimental studies show that after adding white noise with a noise amplitude of 100 to the light bar image, the precision can be also controlled at the sub-pixel level, and the stability of the light bar of the black detected object can be improved.

     

/

返回文章
返回