基于时频分析的战场声信号主成分特征提取技术

Principal component feature extraction technology based on time-frequency analysis for battlefield acoustics signal

信号的时频分布描述了信号从时域到频域的变换,较为全面地表征了信号的特征。主成分分析是统计学中分析数据的一种有效方法。本文将时频分析的方法应用于声目标的特征提取及分类;在保证信息的相对完整性的基础上,利用基于时频分析的主成分特征提取技术对4类战场目标的声信号进行了特征提取。经仿真实验验证,信号的时频分布较好地体现了各类声目标在时-频域的分布规律,主成分分析方法有效地压缩了数据量。结果表明,各目标的类间可分性测度值较大,具有良好的可分性。

The time-frequency distribution of signal can depict the exchange from time to frequency domain for signal, and it can represent the characteristic of signal entirely. Principal component analysis （PCA） is a kind of effective method in statistics. This paper presented analysis on the feature extraction of the battlefield acoustics using principal component analysis （PCA）, based on time-frequency distribution. The acoustic data was field data taken from four kinds of battlefield targets, and the signals in battlefield acoustics were extracted in the method mentioned above. Simulation experiment indicated that the time-frequency distribution of the signal can exhibit the distribution principle of all kinds of acoustics targets. The experiment results indicate that the PCA can reduce the data volume and the four kinds of targets can be well distinguished.