摘要
基于爬壁机器人移动平台和单目相机的图像采集系统,设计了一种焊后焊缝图像处理方法,将改进的自适应中值滤波算法与灰度形态学方法结合,实现从信噪比较高的图像中提取特征。采用基于边缘检测和Hough变换的焊缝位置提取算法,经测试识别准确率达70%,且单幅图像平均处理时间为200 ms,能满足管道爬壁机器人行进过程中的实时焊缝跟踪,并提供了一种引导机器人沿焊缝前进的自主定向方案。
The first step to realize automatic welding seam tracking of wall climbing robot is to obtain the position information of weld seam in real time and accurately.Based on the wall climbing robot and monocular camera,an image acquisition system was proposed.In order to extract features from images with high signal noise,an image processing method for after welding seams was designed based on improved adaptive median filtering algorithm and gray scale morphology.Edge detection and Hough transform was adopted to obtain the weld position.Results show that this algorithm can extract weld position with 70%accuracy,and the average processing time is 200 ms,which meet the requirement of real time welding seam tracking.An autonomous orientation scheme is provided based on welding seam tracking to guide the wall climbing robot along the seam.
作者
杨玥旻
闫维新
YANG Yuemin;YAN Weixin(School of Mechanical Engineering,Shanghai Jiao Tong University,Shanghai 200240,China)
出处
《机械与电子》
2021年第3期65-68,74,共5页
Machinery & Electronics
关键词
爬壁机器人
图像处理
特征提取
焊缝跟踪
定向
wall climbing robot
image processing
feature extraction
welding seam tracking
orientation
