Cumuliform Cloud Animation Control Based on Natural Images

Cloud is an important element in synthetic outdoor scenes. Many methods have been developed for simulating realistic clouds. The image-based cloud modeling methods can estimate cloud shape similar to the input image. However, simulating cloud evolution among multiple specified shapes remains a challenging task. This paper presents a method to simulate the evolution of cloud between two given natural images, which are treated as the initial and target shapes, respectively. Atmospheric hydrodynamics combined with particle system is used to generate clouds physical movement. Then, taking the target shape as constraint, the driving forces are calculated and the control over the cloud movement is realized. Finally, we apply the PBF approach to ensure the cloud's incompressibility. The experiment results show that the proposed method can effectively simulate cloud evolution between discrepant initial and target shapes.