基于蚁群算法的铝合金MIG焊熔池边缘检测

Welding pool's edge detection of aluminum MIG based on ant colony algorithm

针对铝合金MIG焊熔池边缘提取问题,建立了视觉传感和采集的实验系统并提出和设计了基于蚁群算法的熔池图像处理程序。介绍了系统与算法的结构,同时与基于Sobel算子、Canny算子的熔池边缘提取结果进行了对比。结果表明蚁群算法在熔池图像的分割和特征提取性能上具有一定的优异性。通过该算法提取的铝合金MIG焊接熔池清晰边缘,可以有效地克服熔池图像信号中的噪声,阴极雾化区的影响,为实现铝合金MIG焊过程实时检测与控制奠定了理论基础。

A digital vision sensing and data acquisition system is designed and established basing on the ant-colony-algorithm （ACA）welding pool image processing technique to improve the edge detection of MIG welding pool on Aluminum Alloy.The system structure and algorithm procedure are discussed.The comparison between sobel detector,canny detector and ant-colony-algorithm detector is performed which indicates that ant-colony-algorithm detector has obvious advantages in welding pool image processing.Its ability to require clear edge information from MIG welding pool image on Aluminum Alloy and effectively avoid the influences of noise from negative pole atomization has provided us a new possible solution of MIG process welding pool image processing.