Abstract: As a method of time-frequency analysis，spectrogram will be affected by cross-terms while analysing signals，especially for signals who are aneared，and also frequency resolutions will be influenced.This paper proposes combining fractional Fourier transform with spectrogram towards multi-component signals analysis.First，it obtains the optimal factional orders by using the second order fractional Fourier transform moments，and makes a fractional Fourier transform for multi-component signals according to the gained order，then gives spectrogram analysis based on the FrFT.Simulative examples indicate that the proposed method can effectively distinguish multi-component chirp signals，and the cross-terms can be suppressed availably.

摘要： 作为时频分析方法的一种，谱图对多分量信号分析时受交叉项影响，特别是当信号相隔很近时尤为严重，而且频率分辨率会受影响。给出了结合分数阶Fourier变换（FrFT）对多分量信号进行谱图分析的方法。首先利用分数阶二阶矩极值点而找到相应的最优旋转阶数，对所给多分量信号按此阶数做分数阶Fourier变换，再在此基础上做谱图分析。仿真实例表明，该方法对初始频率、调频率很接近的多分量的chirp信号能有效识别，交叉项可得到较好的抑制。