摘要
为了精确地提取焊接缺陷,进一步提高缺陷检测的准确性,提出了一种基于改进ChanVese(CV)模型和脉冲耦合神经网络(pulse coupled neural network,PCNN)的非下采样Shearlet变换(non-subsampled Shearlet transform,NSST)域焊接缺陷提取方法。首先,对焊接缺陷图像进行NSST分解,对得到的低频分量采用PCNN提取出缺陷的主要区域;然后,利用背景抑制后的低频分量和高频分量构造出高频特征图像,并对其进行粗分割,再利用改进的CV模型寻找最优轮廓,提取出缺陷精细轮廓;最后,融合缺陷的主要区域和精细轮廓信息得到最终的结果。实验结果表明,与其他缺陷提取法相比,所用方法提取的缺陷结构更为完整,缺陷轮廓更为精细。
In order to extract welding defect more accurately and further improve the accuracy of defect detection,a welding defect extraction method based on improved Chan-Vese(CV)model and pulse coupled neural network(PCNN)in the non-subsampled Shearlet transform(NSST)domain is proposed.Firstly,a welding defect image is decomposed by NSST.The main region of defect is obtained through processing low-frequency component by using PCNN.Then,high-frequency feature image is constructed through low-frequency after background suppression and high-frequency,and improved CV model is used to search optimalcontour of defect after coarse segmentation.Finally,the final defect is extracted by fusing main region and fine contour of welding defect.Compared with recently proposed defect extraction methods,the extracted welding defect using the proposed method has more complete structure and optimal contour.
出处
《光学仪器》
2015年第1期57-64,共8页
Optical Instruments
基金
国家自然科学基金(61071163
61271327)
南京航空航天大学博士学位论文创新与创优基金(BCXJ14-08)
江苏省研究生培养创新工程(KYLX_0277)
中央高校基本科研业务费专项资金资助
江苏高校优势学科建设工程(PADA)