摘要
针对传统的基于漏磁信号识别方法的局限性,以焊缝缺陷的漏磁图像为研究对象,提出了一种新的基于数学形态学的焊缝缺陷的边缘提取方法,对焊缝与缺陷的识别、缺陷定位等方面进行了深入研究,实现了对焊缝缺陷的可视化图像显示。首先利用所设计的新型磁化系统,对实验板上焊缝不同区域分布的矩形槽缺陷进行了连续非接触漏磁扫描,获取了其三维漏磁信号分布图。然后将其转化为二维灰度图形,利用构造的一种数学形态学优化边缘提取算法对灰度图形进行边缘检测。结果表明,根据边缘检测结果的轮廓图,可直观化缺陷形态、位置等信息,定位精度达到96.67%。该方法能较准确地提取漏磁信号图像的焊缝和缺陷边界,实现二者的有效分离,具有良好的适应性和实用性。
Aiming to traditional limitations of the method based on Magnetic Flux Leakage （MFL） signal recognition, takes the MFL imaging of the weld defect as the research object, a new edge extraction method for the weld defect is proposed based on the mathematical morphology, and the identification of the weld defect and the defect location et al are researched, also, the visual im- age of the weld defect is implemented. First, a continuous non-contact scanning method is used in the rectangular slot defect exists in the different regions of the weld by the use of the design of the new magnetization system, and a three-dimensional leakage magnetic field signal distribution is obtained. And then transform into the two-dimensional grayscale graphics, using the configura- tion of a morphology optimized edge extraction algorithm for edge detection on the grayscale. The outcome indicate that according to the diagram shows the contour of the edge detection result can be shown defect shape, location and other information intuitive- ly ,the positioning accuracy reaches 96.67%. This method can extract the MFL signal image of the weld and the defect boundaries accurately, also, the both is separated effectively, and the method has a good adaptability and practicality.
出处
《现代制造工程》
CSCD
北大核心
2014年第2期109-114,122,共7页
Modern Manufacturing Engineering
基金
黑龙江省研究生创新科研项目(YJSCX2012-048HLJ)
东北石油大学青年科学基金项目(2013NQ133)
关键词
焊缝缺陷
漏磁成像
可视化
数学形态学
边缘检测
weld defect
magnetic flux leakage imaging
visualization
mathematical morphology
edge detection
