一种基于轮廓特征的射频指纹识别方法  

A Method of RF Fingerprint Recognition Based on Contour Feature

作　　者：崔正阳 胡爱群[1] 彭林宁[1] 

机构地区：[1]东南大学信息科学与工程学院,江苏南京210000

出　　处：《信息网络安全》2017年第10期75-80,共6页Netinfo Security

基　　金：国家自然科学基金[61571110];国家重点基础研究发展计划(973计划)[2013CB338003]

摘　　要：无线设备的射频指纹提取与识别是当前的研究热点之一。文章提出了一种基于图像处理的射频指纹星座轨迹图的提取与识别方法。该方法对星座轨迹图运用中值滤波进行降噪规整等预处理,而后进行图像增强与形态学处理,随后提取轮廓和高密度分布集中区等具体特征,运用像素比对、相关矩阵等方法,实现了无线设备的区分与识别。实验结果表明,该方法得到的无线设备识别率可达到90%。The RF fingerprint extraction and identification of wireless devices is one of the hotspots. This paper presents a method of extracting and recognizing CTF based on image processing.This method is applied to the constellation trajectory using median filter for noise reduction and other preconditioning, and then using image enhancement and morphological processing, getting contour and high density distribution of concentrated areas and other specific features. The pixel matching and correlation matrix and other methods are also used to distinguish and identify features to achieve the distinction between wireless devices.The experimental results show that the wireless device recognition rate can reach 90%.

关 键 词：中值滤波 小波分解 数学形态学 相关矩阵 图像识别 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]