基于线性预测分析的金属材料超声识别方法  

Ultrasonic identification of metal materialsbased on linear prediction analysis

作　　者：刘昱 贺升平 贺西平[1] LIU Yu;HE Shengping;HE Xiping(Shaanxi Key Laboratory of Ultrasonic,College of Physics and Information Technology,Shaanxi Normal University,Xi’an 710119,China;Luzhou High-tech Research Institute,Luzhou 646000,China)

机构地区：[1]陕西师范大学物理学与信息技术学院陕西省超声重点实验室,西安710119 [2]泸州高新技术研究所,四川泸州646000

出　　处：《兵器装备工程学报》2023年第10期307-312,328,共7页Journal of Ordnance Equipment Engineering

基　　金：国家自然科学基金项目(12174241);装备预研共用技术项目(XXXXX040302)。

摘　　要：针对金属材料中超声波的背向散射与其微观结构特征密切相关,即使材料相同,超声信号也会因金属晶粒的尺寸和径向分布的随机性而产生差异的问题。计算了金属样品超声信号的线性预测系数,以此作为其超声指纹,结合拉伊达判别准则确定的阈值,完成了对形状、材料完全一致的金属样品的识别。结果表明:基于线性预测分析的超声金属材料识别方法可以快速、准确地识别出各样品。该方法对贵重金属材料、金属文物的识别保护具有一定的参考应用价值。Microstructure of metal material results in different backscattering of ultrasonic wave,and ultrasonic signals are impacted by the randomness of the size and radial distribution of metal grains.Ultrasonic signals of different individuals are diverse even if their materials are the same.In this paper,the linear prediction coefficients of ultrasonic signals of metal samples are extracted as ultrasonic fingerprints,and the identification threshold is determined based on Pauta criterion.The identification results demonstrate that the method based on linear prediction analysis could identify metal individuals quickly and accurately.This method provides a reference value for the identification and protection of precious metal materials and cultural relics.

关 键 词：超声背向散射 线性预测分析 拉伊达法则 超声指纹 识别阈值 

分 类 号：TB553[理学—物理]