中文摘要：

自然界中有大量的物理系统,其中的物理量随时间演化的规律可以用有限多个正弦函数叠加来描述,即物理量是带限分离谱连续信号.因此,可用频谱识别方法来寻找物理系统的演化规律.研究发现,时域矩形窗截出的物理信号,它的连续频谱的频域采样正好是物理量采样数据的DFT给出的频谱;在物理量被时域矩形窗截出的信号的连续频谱幅频曲线上,所有主极大值及其位置正好是物理量各频率成分的振幅和频率.给出了基于采样数据DFT的二次逼近算法,可以快速地从物理量采样数据DFT频谱中提取原物理量（信号）的频谱信息,进而给出物理量随时间演化的规律.

英文摘要：

There are a lot of physical systems in the nature,the physical quantities in the nature can be descirbed by using band-limited separation spectral continuous signal.Therefore,the spectrum recognition is an important way to find physical quantity evolution.This study found that truncation signal by time domain rectangular window,its frequency domain sampling is the DFT of time domain sampling data.Its all main maximum and the location of the amplitude-frequency curve is the amplitude and frequency of a physical quantities,respectively.Based on DFT spectrum,aquadratic approximation algorithm of spectrum identification is presented,it can be used to extract high precision frequency spectrum information of physical quantities,and to construct the time evolution of physical quantities.