免疫算法在焊缝识别中的应用研究  

Research on the application of immune algorithm in the weld seam recognition

作　　者：丁度坤[1] 

机构地区：[1]东莞职业技术学院,广东东莞523808

出　　处：《焊接技术》2016年第5期64-67,7,共4页Welding Technology

基　　金：广东省数控一代项目"六自由度智能化工业机器人关键技术研究"(2013B011301003);东莞市产学研合作项目"六自由工业机器人关键技术研究"(2014509102211);东莞职业技术学院政校行企合作开展科研与服务项目"四自由度工业机器人关键技术研究"(ZXHQ2014d003);东莞职业技术学院院级基金项目"电动汽车充换电系统的研究与开发"(2015a03)经费资助

摘　　要：对在自然光条件下焊缝识别进行了深入的研究,搭建了TIG焊图像采集系统,实时采集焊接图像。在此基础上,对焊接图像的焊缝识别技术进行研究,在传统的图像分割方法中引入免疫算法。首先定义了图像灰度抗体、抗原,构造免疫亲合度函数,并详细描述了人工免疫操作算子,给出了免疫计算算法流程,以实时获取最优的图像分割阈值。最后,进行了相关的试验验证,试验结果表明,与传统的Ostu法相比,基于免疫计算的焊缝识别法能较好地去除焊缝周围由于光线所造成的干扰,较好地提取焊缝信息,为焊缝跟踪提供更为精准的焊缝信息。The recognition of weld seam which was under a natural light condition had been deeply researched. The TIG welding images capture system had been set up to obtain the weld image in real time. On this basis, the technique of weld seam recognition had been deeply researched. The immune algorithm had been introducted to image segmentation. Firstly, the antibody and antigen which were based on the image gray level had been defined. The immune affinity function had been set up at the same time. Then several artificial immune operators had been detailly described. The flow of the whole algorithm were given too. So the optimal image segmentaion value could be determined. In the end, several experiments had been caught out. The experimental results showed that the weld recognition method, which was introduced the immune algorithm, could effectively reduce the disturb producing by the brightness variance and extract the weld seam position, which could supply a more accuarate weld information for seam tracking.

关 键 词：焊缝跟踪 免疫计算 焊接 

分 类 号：TP273.5[自动化与计算机技术—检测技术与自动化装置] TG409[自动化与计算机技术—控制科学与工程]