**4. Discussions and Conclusions**

We generated simulated seismic data based on the finite difference method and used the modified U-Net to successfully predict the underground velocity field and the location of anomalies from seismic waveforms, and then used CNN to post-process the generated images. Experimental results show that the effect of multi-shot-point prediction is not necessarily better than that of single shot point. After numerical verification, the predicted velocity field and abnormal point position have very little error with the ground truth. It should be noted that traditional VMB normally needs weeks or months to reconstruct the velocity model; however, our algorithm only needs days, which is shown as Table 1. Next, we will use real seismic data to verify and refine our model.

**Table 1.** Time of calculation in one anomaly without fault experiment.


There are still some problems with our model. First of all, as shown in Figures 10, 12 and 14, as the depth increases, the error of the velocity prediction will gradually increase. In addition, the data used in each of our experiments correspond to 1000 sets of artificially generated information. It is necessary to increase the amount of information and use real seismic data to improve the applicability of the model to the actual situation. Finally, although U-Net can extract image features well and achieve the required training effects, this processing is based on images and not directly obtained from seismic data training. The image resolution will seriously affect the training effect. We will continue to find ways to overcome these problems.

**Author Contributions:** Methodology, J.J. (Jiwei Jia) and J.J. (Jian Jiao); Resources, P.Y.; Software, Z.L. (Ziqian Li); Validation, Z.L. (Zheng Lu); Writing—Original Draft, Z.L. (Ziqian Li). All authors have read and agreed to the published version of the manuscript.

**Funding:** The authors are supported by the Natural Science Foundation of Jilin Province (Grant No. 20210101481JC), the Shanghai Municipal Science and Technology Major Project (Grant No. 2021SHZDZX0103), and the Fundamental Research Funds for the Central Universities (Grant No. 93K172020K27).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data that support the findings of this study are available from the corresponding author upon reasonable request.

**Conflicts of Interest:** The authors declare no conflict of interest.
