Thin layer prediction method based on waveform indication simulation

The invention provides a thin layer prediction method based on waveform indication simulation, and the method comprises the following steps: firstly, standardizing a logging curve, screening a natural gamma curve which can well recognize sand shale, and employing wavelet transformation to obtain low and medium frequency information which can reflect lithology in a GR curve; well-seismic calibration and synthetic recording are carried out for a drilled well, and dynamic clustering of seismic waveforms is realized by adopting a singular value decomposition method. And analyzing seismic waveform and GR curve common characteristics of the same layer and the same sedimentary microfacies reservoir, and establishing a sedimentary microfacies, seismic waveform and GR curve database at the well point. An original three-dimensional seismic body is simulated channel by channel, a sample group with the highest seismic library correlation is compared, and an initial model is obtained through GR curve correlation weighted