基于平稳小波和脉冲耦合神经网络的金属断口多聚焦融合分析方法  

Multi-focus Fusion Analysis Method of Metal Fracture Based on Stationary Wavelet and Pulse Coupled Neural Network

作　　者：洪刚 冯子航 HONG Gang;FENG Zihang(School of Automation, Beijing Institute of Technology, Beijing 100081, China;NCS Testing Technology Co. Ltd., Beijing 100081, China)

机构地区：[1]北京理工大学自动化学院,北京100081 [2]钢研纳克检测技术股份有限公司,北京100081

出　　处：《微型电脑应用》2021年第11期200-203,共4页Microcomputer Applications

摘　　要：为了提高金属断口融合的主观、客观效果,提出一种二维平稳小波与脉冲耦合神经网络(PCNN)结合的方法,对金属断口的多聚焦图像进行融合,用以获得金属材料在断裂过程中的多种信息。利用二维平稳小波具有平移不变性的特点,分解断口图像,低频域进行多尺度、双通道的PCNN融合,高频域采用张量的奇异值分解法融合,最后进行二维平稳小波逆变换,得到融合图像,融合后图像突出了断口形貌中的细节信息,使得断口形貌特征表达更为全面。提出的融合方法,改进了传统方法测量单一、易丢失细节的不足,克服了断口图像分析仪景深受限的缺点,具有一定的实用性。In order to improve the subjective and objective effects of metal fracture fusion,a method combining two dimensional stationary wavelet and pulse coupled neural network(PCNN)is proposed to fuse multi-focus images of metal fracture to obtain various information of metal material in fracture process.Based on the translation invariance of two dimensional stationary wavelet,the fracture image is decomposed.The low-frequency domain is fused by multiscale and dual channel PCNN.The high-frequency domain is fused by tensor singular value decomposition.Finally,the two dimensional stationary wavelet inverse transform is used to get the fused image.The fused image highlights the detail information of fracture morphology,which makes the fracture morphology feature expression more comprehensive.The method overcomes the shortcomings of traditional methods,such as single measurement,easy to lose details,and overcomes the shortcomings of limited depth of field of fracture image analyzer.

关 键 词：脉冲耦合神经网络 多聚焦图像融合 断口图像 平稳小波 

分 类 号：TG115.5[金属学及工艺—物理冶金]