摘要
对基于结构光视觉传感的焊缝视觉信息检测和识别进行了研究。所获取的焊缝图像中存在着大量的飞溅、烟尘和电弧光等噪声干扰。CCD摄像机获取焊缝图像,经图像采集卡A/D转换后送入计算机内存,然后采用各种图像处理方法对图像数据进行处理,焊缝原始图像经过二值化、中值滤波、拉普拉斯锐化、细化一系列图像预处理,准确提取激光条纹信息,提出了几种常见的焊缝形式的检测和识别算法,算法所需时间均不到1ms,提取焊缝特征,实现对焊缝的自动跟踪。
The detection and identification for weld based on structure light vision sensing was studied.The acquisition to the image of weld was influenced by the intense spatters,combusting gas,arc light and so on.The image,taken by CCD camera and transformed by image col- lection card,was sent to a computer and the image data processed by different image processing methods.After the images of weld were pro- cessed through threshold transform,median filtering,LAPLACE sharpening and thinning,the stripes of welding lines could be distilled exact- ly,so the arithmetic was put forward that the arithmetic for several usual familiar seam types will take less than 1 millisecond,which could ac- quire information and realize seam tracking.
出处
《焊接》
北大核心
2007年第1期28-31,共4页
Welding & Joining
基金
国家自然科学资金(50565003)
江西省科技厅重大科技攻关项目(Z03388)
江西省教育厅科技项目(04JKZ02806)
关键词
结构光
图像处理
焊缝检测
焊缝跟踪
structure light
image processing
seam detection
seam tracking
