Short-time difference spectrum analysis method

According to the short-time difference spectrum analysis method, transient components in signals are represented and extracted through time-frequency analysis, firstly, window signals are intercepted through a sliding window, then discrete Fourier transform is carried out on the window signals to obtain a short-time Fourier transform matrix, then a short-time difference spectrum matrix is calculated, and a time-frequency graph of short-time difference spectrum analysis is drawn; and finally, performing short-time Fourier inversion on the short-time difference spectrum matrix, reconstructing a time domain signal of a transient component in the signal, and realizing targeted extraction of the transient component by time-frequency domain filtering. The method can effectively overcome the influence of steady-state harmonic components in engineering signals, depicts transient components caused by defects, improves the detection capability of early anomalies, and has important engineering application value. 短时差谱分析方