BP神经网络在焊缝位置识别中的应用

研究了一种焊缝位置识别新方法,在一定工艺条件下,使用视觉传感器采集焊接熔池图像,选取图像中熔池前端部分进行处理,先对其进行中值滤波与灰度变换,在此基础上,获取每一幅熔池图像的质心值、质心位移、质心速度及电弧与焊缝的偏差值作为训练样本数据.以质心值、质心位移和质心速度为输入量,以偏差值为输出量,利用BP神经网络建立其数学模型,最后对该模型进行检验.检验结果表明,该模型能够较准确地描述熔池图像质心与焊缝偏差之间的关系,为进一步实现精确的焊缝跟踪提供了理论和试验依据.

Application of the BP neural network in seam position recognition

A new method to recognize the seam position is researched in the paper.Weld pool images are collected under a certain technical condition by a vision sensor.And the foreside part of the weld pool image is chosen as the processed region.Firstly,this region is processed by the median filter and image gray transform.Then every weld pool image centroid and the centroid displacement,the centroid moving speed and the error between the weld arc and the seam are used as the training sampled data.The centroid,the centroid displacement and the centroid moving speed are made as input data,while the error is the output.And the mathematical model is acquired through BP neural network.Finally,the model is tested and the result shows that the model can correctly describe the relationship between the weld pool image centroid and the seam error.Also,it provides some theoretical and experimental basis for the seam tracking.