基于熵的Gabor变换窗函数宽度自适应选择算法

Adaptive Window Width Selection Algorithm for Gabor Transform Based on Entropy

该文针对Gabor变换中窗函数宽度选择的问题,提出了以提高Gabor表示的聚集性和时频分辨率为目的的窗函数宽度自适应选择算法。提出对香农熵的取值范围进行改进,使其更适合度量时频分布的聚集性,进而根据熵度量实现了与信号非平稳性相适应的最优窗函数宽度选择。仿真结果表明该算法对单分量及多分量信号都能有效地选择最优窗函数宽度,能够获得聚集性好、时频分辨率高的Gabor表示,并具有很好的抗噪性能。

To resolve the issue of window width selection, an adaptive algorithm for Gabor transform is proposed, which improves the concentration and time-frequency resolution of Gabor representation. The value range of Shannon entropy is mended to make it more adequate for measuring concentration of time-frequency distribution. Moreover, basing on entropy, an optimal window width can be searched adaptively. Simulation results show that the proposed algorithm chooses the optimal window width for mono-component signal or multi-component signal, giving the best Gabor representation with high concentration and time-frequency resolution. Additionally, the algorithm behaves well under low signal noise ratio.