最大熵谱估计在昆虫鸣声特征提取中的应用

Application of Maximum Entropy Spectrum Estimation in Getting Characteristic Parameters of Insect’s Song

鸣虫类的鸣声特征广泛应用于昆虫的种类鉴别、地理分布和区系研究中。采用最大熵谱估计法处理昆虫鸣声信号获取昆虫鸣声特征参数,并采用Burg算法实现信号处理。论文以蟋蟀鸣声样本验证该算法,试验证明,对于昆虫鸣声,经典谱估计法的处理结果几乎没有意义,而用Burg法实现的最大熵谱估计法具有高分辨率、谱线平滑、适合短序列信号等特点,能够有效提取鸣声信号的特征参数。

The song’s character of insect is abroad applied such as the study of identifying species, geography. Maximum entropy spectrum estimation is applied in processing of insect’s song, and Burg algorithm is used to realize signal processing. At last, cricket’s is processed by two methods, one is FFT, and another is Burg algorithm. It has been found, traditional spectrum estimation way is useless in insect’s song, but the method of maximum entropy spectrum estimation based on Burg has the advantage of high distinguish ability, smooth spectral line, and suiting short data and so on., so in this method , the song’s characteristic parameters are get effectively.