摘要
以某型号航空发动机叶片为研究对象,介绍并分析了基于分区域自适应中值滤波的X射线数字图像特征提取和基于曲面函数和阈值分割的X射线数字图像特征提取两种方法,提出了一种新型的数学形态学滤波与计算机视觉算法相结合的缺陷自动提取方法。试验结果表明,提出的方法可行,能有效减小缺陷的误判率。
Taking a certain model turbine blade as an example, two kinds of techniques and methods, i.e. defect extraction of digital radiograph based on subarea and self--adaptive median filtering and defect extraction of digital radiograph based on curved--surface function and threshold, were introduced. A new method with a combination of mathematical morphologic opening operation with solid vision algorithm was put forward, with which flaws of tested part can be extracted accurately and automatically. Experimental results indicate that the new method is feasible and false detections can be decreased effectively.
出处
《中国机械工程》
EI
CAS
CSCD
北大核心
2006年第21期2270-2273,共4页
China Mechanical Engineering
基金
国家自然科学基金资助项目(50275008)
关键词
X射线数字图像
涡轮叶片
缺陷提取
无损检测
digital radiograph
turbine blade
defect extraction
nondestructive testing