The increasing integration of renewable energy sources has caused spatially and temporally uneven inertia distribution in power systems. This paper proposes a regional inertia estimation and virtual inertia configuration method using dynamic mode decomposition (DMD). A regional equivalent inertia model is established based on inertia estimation theory for renewable-penetrated power systems. The power system is partitioned via spectral clustering, and frequency measurement nodes are selected according to Pearson correlation coefficients of regional nodes. Regional inertia is then estimated through the DMD method. Critical regional inertia requirements are calculated based on frequency security constraints to configure virtual inertia for renewable energy and energy storage systems. An improved IEEE 10-machine 39-node model with wind-storage systems is implemented in PSCAD for validation. Simulation results demonstrate that the proposed method achieves regional inertia estimation with errors below 5%. The online virtual inertia configuration strategy for wind turbines and energy storage systems effectively prevents frequency violations and enhances frequency security and stability.