摘要
根据药柱X射线图像的特点,通过引进模糊理论,利用了广义多层次模糊算子增强算法,使图像呈现出不同灰度层次的边缘信息。采用广义模糊增强算子,克服了Pal和King复杂的变换和逆变换。针对Pal和King算法中对隶属度阈值的设置存在不足,提出了迭代法确定最佳隶属阈值的方法,使选取阈值具有自适应性。从实验的结果来看,针对气孔缺陷能得到清晰的缺陷边缘,为以后的识别打下良好的基础。
Based on characteristic of X-ray detection images of solid grain, it use generalized multi-level operator fuzzy enhancement algorithm, to make different images on the brink of gray-level informetion. Use of generalized fuzzy operator simplifies complex and transformation calculation in Pal King algorithm. In order to overcome disadvantages of Pal King algorithm in membership degree threshold value setting, a membership degree threshold optimal value determination based on iteration. The experimental results indicate that the method can accomplish defect extraction accurately, which will lay a good foundation for hole flaw feature parameter extraction and recognition.
出处
《弹箭与制导学报》
CSCD
北大核心
2009年第4期238-240,共3页
Journal of Projectiles,Rockets,Missiles and Guidance
基金
山西省科技攻关项目(2006031160)资助
关键词
药柱
无损检测
模糊理论
隶属度
迭代法
grain
nondestructive test
fuzzy theory
membership degree
iteration
