3D Meteorological Scene Simulation Method Based on ClimateNeRF

Climate change is gradually becoming more harmful to human beings and the Earth, and people need to propose timely and targeted solutions to reduce its impact. However, the current visualization of meteorological hazard scenarios is mostly displayed in the form of text and weather maps, which lacks intuition, is not conducive to public understanding and has limited dissemination of meteorological information, and requires a more intuitive way to display the impact of meteorology on the environment. Therefore, this paper proposes a 3D climate scene simulation method based on ClimateNeRF. The method introduces SFM+MVS and Zip-NeRF to improve the scene realism, accuracy and generation efficiency, and realizes more diversified and higher-quality style migration effects by modifying the preset parameters and introducing the attention mechanism. Meanwhile, we adopt C/S architecture to build a parameter regulation panel to improve interactivity and intuitiveness. Experimental results show that our method has good performance in terms of model scene quality, generation speed and intuitiveness. This paper describes in detail the specific implementation process of the method, key technologies and so on. It can enable the public to see more intuitively the impact of climate change on a scenario, improve their attention to meteorological hazards, and make timely plans according to their needs.