随着勘探程度的深入，油气勘探目标日趋复杂，对速度建模方法提出新的挑战。弹性波全波形反演可以得到纵波速度、横波速度和密度参数，应用前景广阔。但由于存在周期跳跃，全波形反演对初始模型的依赖性较强。如何构建精确的初始模型是一个亟待解决的科学难题。反射波波形反演利用全波形反演中的低频层析项来反演速度模型的长波长分量，是初始模型建立的有效手段。回折波在浅层速度建模方面具有优势。本项目将研究一种新的弹性反射波和回折波波形反演方法，同时获取纵波速度、横波速度和密度低频模型。新方法通过敏感核分解策略提取回折波和反射波低频敏感核，消除高频偏移等时线的影响；采用纵横波分离、旅行时目标函数保证模型参数梯度的光滑性；基于敏感核和串扰分析设计合理的分步反演流程来提高反演精度。本研究可以获得精确的纵波速度、横波速度和密度初始模型，进而大大改善多分量地震资料全波形反演和逆时偏移的精度，具有重要的理论意义和实际应用价值。

With the exploration ongoing, the target of oil and gas exploration is becoming more and more complicated, posing new challenges to the velocity inversion methods. Elastic Full waveform inversion (FWI) can yield accurate P-wave velocity, S-wave velocity and density. Hence, it has a wide application prospect. However, FWI shows strong dependence on the initial model because of cycle skipping. How to build accurate initial model is a significant scientific problem in FWI. Reflection waveform inversion (RWI) retrieves the long-wavelength components of the velocity model by using the tomography term of FWI, and is treated as an effective tool of the initial model building. Diving waves have some advantages in the estimation of shallow velocity model. In this research, we investigate a new elastic reflected and diving wave waveform inversion method. The new method uses the decomposition of sensitivity kernels to extract the low-frequency sensitivity kernels of reflected and diving waves and eliminate the high-frequency migration isochrones. Besides, we employ the P- and S-wave separation and the traveltime-based objective function to guarantee the smoothness of model parameter gradients. Also, we design the proper hierarchical inversion workflow based on the sensitivity kernel and crosstalk analysis to improve the inversion accuracy. This research can provide accurate P-wave velocity, S-wave velocity and density initial models and further enhance the accuracy of FWI and reverse time migration of multi-component seismic data. Consequently, it has important theoretical significance and practical value.