摘要
研究图像边缘检测技术,针对存在的孔洞缺陷和噪声影响边缘定位精度,通过ICT图像缺陷边缘检测,具有成像直观、分辨率高等优点。传统采用的边缘检测方法一般都定位在单一的传统检测算法,由于传统的边缘检测算法存在边缘检测模糊或对噪声敏感的缺点。为了解决上述问题,提出了一种检测结构件内部孔洞型缺陷边缘的算法,利用区间值模糊集中的模糊归熵将孔洞缺陷的模糊边缘提取出来,提出一种区域标记算法,得到最终的精确缺陷边缘。试验结果表明,所提出的边缘检测算法不仅很好地解决了噪声问题,而且更具有很好的边缘提取能力,在边缘连续的同时又保证了边缘定位的准确性,明显提高了图像的质量。
Edge detection of ICT is an important method in non-destructive testing,having the advantages of visual imaging and high resolution.At present,the edge detection algorithm has been all confined in single traditional detection method.Due to the disadvantage of noise-sensitive in traditional detection method,in this paper,we introduce a new method to detect the edge of the hole detection.First,extracting the edge of the hole based on the principle of entropy on interval-valued fuzzy sets,then making use of the the label processing,in order to make the final edge accurately.Experimental results show that the proposed edge detection algorithm not only can solve the problem of edge noise,but also has the good capability in edge extraction,and ensures the accuracy of edge position while the edge is continuous.
出处
《计算机仿真》
CSCD
北大核心
2011年第7期265-267,271,共4页
Computer Simulation
基金
河北省自然科学基金(E2007000386)
关键词
边缘检测
区间值模糊集
归熵
区域标记
Edge detection
Interval-valued fuzzy sets
Normalized entrop
Label processing
