摘要
结合航空航天行业对铝合金焊接的要求,提出了一种基于KNN算法的铝合金工件阳极化层打磨质量视觉检测系统。系统利用SCARA机器人及固定在其上的图像获取装置,多角度、多方位、多视场地对传送设备上的工件进行拍摄,并通过图像处理技术得出工件的打磨质量检测结论。结合铝合金特点,利用打磨前后工件颜色不同,采用KNN分类算法对工件打磨区域进行分类,得到工件被打磨区域宽度的测量值以及工件打磨的质量,并与采用传统动态阈值分割、边缘提取等图像处理方法进行了对比。最终完成铝合金工件阳极化层打磨质量检测实验,结果表明,系统可以有效地对打磨质量进行检测,并且与其他方法相比有较高的稳定性。
In order to meet the requirements of aerospace industry for welding of aluminum alloys,this paper proposes a visual inspection system for the quality of anodized layer of aluminum alloy workpieces based on KNN algorithm.Firstly,the system used the SCARA robot and the image acquisition device fixed on it to photograph the workpieces on the conveyor equipment from multi-angle,multi-azimuth,and multi-view,and then,it applied the image processing technology to draw conclusions on the quality of the grinding.Next,in combination with the characteristics of aluminum alloys and different colors of workpieces before and after grinding,the KNN classification algorithm was used to classify the grinding areas so as to obtain the measured value of the width of the grinding area as well as the quality of grinding,which were compared with those results acquired by using conventional image processing methods.Finally,the quality test of the anodizing layer of aluminum alloy workpieces was completed.The results show that the system can effectively detect the quality of grinding,and has higher stability compared with other methods.
作者
李明洋
唐国宝
李乾
刘尔彬
高永卓
LI Mingyang;TANG Guobao;LI Qian;LIU Erbin;GAO Yongzhuo(State Key Laboratory of Robotics and System,Harbin Institute of Technology,Harbin 150080,China;Guangzhou Risong Technology Co.,Ltd.,Guangzhou 510760,China)
出处
《机械与电子》
2018年第6期45-49,共5页
Machinery & Electronics
基金
广东省重大科技专项项目(2015B010918001)
关键词
KNN分类算法
机器人
阳极化层打磨
质量检测
KNN classification algorithm
robot
anodizing layer polish
quality inspection