基于电磁超声的铁质文物特征提取方法研究

Research on Feature Extraction Method of Iron Cultural Relics Based on Electromagnetic Ultrasound

中文关键词:  电磁超声  铁质文物  堆叠稀疏自编码器  磁致伸缩效应

英文关键词:electromagnetic ultrasound  iron cultural relics  stacked sparse autoencoder  magnetostrictive effect

中文摘要:

      传统文物鉴定通常凭借个人经验，从文物材质、外观等方面着手，对文物真伪进行鉴定，该方法存在一定的局限性。针对铁质文物，根据材料的磁致伸缩特性取决于材料的微观结构的特点，利用具有磁致伸缩效应的电磁超声检测其特征参数，提出了一种铁质文物的特征参数提取方法；利用电磁超声信号幅值随偏置磁场强度变化曲线，通过堆叠稀疏自编码器提取该曲线的特征参数用于文物的特征表达，并使用支持向量机(Support Vector Machine,SVM)分类算法进行辨识。采用该方法对3个不同的样件进行了实验验证，检测准确率达到93.3%，表明该方法对铁质文物的鉴定具有可行性。

英文摘要:

      Traditional cultural relics identification is usually based on personal experience,from the cultural relics material,appearance and other aspects of the authenticity of the identification,this method has some limitations.A feature extraction method of iron cultural relics is proposed relying on electromagnetic ultrasound with magnetostrictive effect to detect the characteristic parameters of iron cultural relics,based on the fact that the magnetostrictive properties of the material depend on the microstructure of the material.Based on the variation curve of the amplitude of electromagnetic and ultrasonic signals with the intensity of the biased magnetic field,the characteristic parameters of the curve are extracted by stacked sparse autoencoder for the feature expression of cultural relics,and support vector machine(SVM) classification algorithm is used for identification.The method is used to verify three different parts,and the detection accuracy reaches 93.3%.Therefore,this method is feasible for the identification of iron cultural relics.