The Detail Restoration And Enhancement Method Of Fluid Animation

In the grid-based fluid animation process, the use of semi-Lagrangian approach to the convective term of Navier-Stokes equations will inevitably introduce numerical dissipation, resulting in gradually lost of fluid details. In order to improve the process, a more accurate interpolation method can be used to simulate the fluid on more accurate grids, but this takes tremendous computation time and resources. From the view of fluid detail restoration and enhancement, this thesis, using fluid detail model and method of detail compensation, solves the numerical dissipation problems. The method of this thesis guarantees the speed of the fluid animation, ensures the physical reality of the simulation, and better meets human visual effects.For general fluid (e.g. water, smoke), the fluid details are displayed in the high frequency part of the fluid velocity field. The fluid vortex area includes abundant high frequency fluid details, has good effect in vision and gets much of concern in fluid visualization domain. Either in fluid detail restoration or in detail enhancement, this thesis concerns mainly about the vortex area of fluid. Firstly the vortex area of fluid is extracted, and then fluid details are compensated to this area locally. The method of this thesis adds details to where they need.This thesis studies the fluid details deeply in the way of digital signal processing, analyzes and extracts fluid details with wavelet technology, combines the turbulence knowledge of fluid dynamics, and proposes a new kind of method of fluid detail restoration and based on frequency of fluid vortex. As in fluid detail enhancement, a local fluid detail enhancement method based on vortex area is proposed. At last, a new modified Purlin noise model is invented by the combination of the knowledge of both vortex model and noise model in this thesis. Basing on this new noise model, this thesis presents an integrated approach of fluid detail restoration and enhancement and compensates the details to fluid velocity field together with the advantages of both detail restoration methods and detail enhancement methods.