一种基于小波谱的分形函数奇异性确定方法  

Method for Determining Singularity of Fractal Function Based on Wavelet Spectrum

作　　者：王兆瑞[1] 吕善伟[1] 中村武恒[2] 

机构地区：[1]北京航空航天大学电子信息工程学院,北京100083 [2]京都大学电子工程系,日本615-8510

出　　处：《电子学报》2008年第3期482-485,共4页Acta Electronica Sinica

摘　　要：基于最大重叠离散小波谱的定义,本文提出了一种确定分形信号局部奇异性指数的算法,构造了一种类似于奇异性谱的直方图,并用之来描述信号奇异性的全局统计分布.算法的有效性通过数字试验及在真实心率数据中的应用得到了验证.Based on the definition of wavelet spectrum of the maximal overlap discrete wavelet transform, we propose a novel algorithm for determining the local singularity exponents, which provides the local scaling information of interest,and construct a singularity-spectrum-like histogram to describe the global statistical distn＇bution of the local singularity strengths. The validity and efficiency are verified by numerical experiment and application to real heart rate data.

关 键 词：奇异性谱 局部HSlder指数 小波变换 

分 类 号：TN911[电子电信—通信与信息系统]