摘要
为克服离散小波变换平移不变性差的不足,通过构造的Hilbert变换对滤波器组,实现了双树复小波变换。建立了基于复小波变换系数的Power-Law检测器。推导了检测统计量的渐近性能和偏移系数,由输入信噪比损失给出了检测器参数的选择方法。利用仿真与实测数据进行了试验,并与基于离散小波变换的Power-Law检测器进行了比较。本文提出的小波方法能很好实现平移不变性,建立的检测器具有更大的偏移系数,在相同的信噪比和虚警概率下能够得到更高的检测概率。
By the Hilbert transform pair filters dual-tree complex wavelet transform is achieved to overcome the translation-invarianee deficiency of the discrete wavelet transform. A novel Power-Law detector based on complex wavelet transform coefficients is proposed. The asymp- totic performance and the deflection coefficients of the detector are deduced. Using input signal-to-noise ratio, the method for choosing the parameter of the proposed detector is presented. The simulation and real data are used to test the performance of the detector. A comparison between the proposed detector and the detector based on discrete wavelet transform is given. Experimental results show that the proposed wavelet method has preferable shift-invariance and the novel detector has a much larger deflection coefficient. With the same signal-to- noise ratio and false alarm probability, the novel detector can achieve higher detection probability compared with the detector based on discrete wavelet transform.
出处
《数据采集与处理》
CSCD
北大核心
2009年第1期23-27,共5页
Journal of Data Acquisition and Processing
基金
国防预研基金(2010301010101)资助项目
