摘要
针对WVD法中存在的交叉项干扰,以及对正交镜像滤波(QMF)利用不充分的问题,提出了基于QMF分析的信号分选方法。该方法建立在小波变换的多分辨率分析特性和双通道滤波器组设计的基础上,根据正交镜像滤波器组树(QMFBT)原理,对信号进行合理层数的分解,通过提取表达信号调制信息的图像特征,不依赖先验知识,实现有效地分选。仿真实验表明,与几种常见分选算法相比,该算法具有一定的抗噪性能和更高的分选准确性及实时性。不足之处是信号分解层数的确定和图像特征的计算对实时性存在影响。
Aiming at the cross term interference in the WVD method,as well as the problem of insufficient utilization of the quadrature mirror filter（QMF）,a signal sorting method based on QMF analysis was proposed.Based on multi resolution analysis of wavelet transform and basic characteristics of two channel filter bank design on the method based on quadrature mirror filters（QMFBT）tree decomposition principle,reasonable number of signal,image feature information extracted by the expression of signal modulation,without relying on prior knowledge,sorting effectively.Simulation results showed that the proposed algorithm had a certain anti noise performance compared with several common sorting algorithms,and it was more accurate and real-time.The deficiency is the determination of the number of decomposition of the signal and the image feature impact on the real-time calculation.
出处
《探测与控制学报》
CSCD
北大核心
2017年第4期83-89,共7页
Journal of Detection & Control
关键词
信号分选
小波变换
正交镜像滤波器组
图像特征
signal sorting
wavelet transform
quadrature mirror filter banks
image feature
