The Research Of Bleeding Simulation In Virtual Brain Surgery Training System

Virtual reality technology is widely used in many fields.The virtual surgery training system is an important application of this technology in the field of medical education.It combines virtual reality technology and medical surgery theory to create an interactive environment which is close to actual surgery.The medical brain tumor operation is complicated,and each step has strict specifications.Therefore,the research of the virtual brain surgery training system has practical significance.This system has many difficulties such as its large number of component modules,high accuracy requirements for each module and complex data coupling.At present,there is no system that can meet all the requirements of virtual brain surgery training system.The bleeding module is one of the keys to embody the fidelity,integrity and immersion of the virtual brain surgery scene,which can also provide operational training for blood aspiration treatment.However,the current research on the fluid and blood characteristics of bleeding simulation is insufficient,and the effect of real-time simulation is lacking in authenticity.In this thesis,the acceleration of the bleeding simulation dynamics model,the optimization of the model combined with physical characteristics,and the rendering of blood and scene are studied.An improved bleeding simulation model is proposed to improve the calculation speed,accuracy and visualization of the simulation.At the same time,it can simulate the bleeding in several kinds of brain surgery,and integrate the established bleeding module into the virtual brain surgery training system.The main contents are as follows:1.Build an SPH(Smoothed Particle Hydrodynamics,SPH)blood physics model based on an accelerated search algorithm and adaptive step size.In this thesis,the Navier-Stokes(N-S)equation is used as the basic equation for controlling blood movement.The N-S equation is solved by the SPH method that maintains its own mass conservation,and a blood physics model with fluid properties is established.Aiming at the problem of occupying a large amount of computation space caused by the frequent calling of the neighbor particle search algorithm in SPH method,an improved algorithm for reducing the repeated search is proposed.The computation time is reduced by 13.4%compared with the space grid method when using this algorithm to search 8000 particles,and the efficiency of the SPH method is improved.In addition,this thesis proposes an adaptive time step method,which replaces the method of updating all particles in the fixed step size in the traditional SPH method.The next time step is determined by factors such as particle velocity,force and surface particles,which further enhances the efficiency in bleeding simulation.2.Optimization of the bleeding model combined with physical characteristics.The bleeding model is optimized from four aspects:maintaining blood density,adding surface tension,improving force instability,and introducing blood coagulation characteristics.Aiming at the problem of exeessive fluctuation of particle density caused by the governing equation in SPH method,a method based on WCSPH algorithm for controlling density fluctuation is proposed.In order to achieve surface tension,this thesis adds tension to the surface particles of the bleeding model based on the fluid surface particles tracking method.In order to deal with the problem of stress instability caused by insufficient number of adjacent particles,a model of artificial repulsive force based on potential energy function is proposed to apply repulsion to particles with too close distance.Effectively improve particle aggregation.In addition,this thesis models the blood coagulation characteristics and proposes a viscous force method based on dynamic viscosity coefficient.The viscosity coefficient is related to temperatUre.The temperature decrease in the simulation will make the relative motion between blood particles more viscous.3.Virtual brain surgery bleeding simulation rendering is based on scene reconstruction of MC(Marching Cubes)algorithm.In order to render the real-time performance,the simple simulation model is used in the general blood flow simulation research,which loses the authenticity of the blood flow simulation effect.This thesis uses the MC algorithm to draw the blood surface.At the same time,a pre-processing hollow element MC algorithm is proposed to improve the computational efficiency.The virtual tissue is surface-drawn according to the three-dimensional data of the real tissue.And through the subsequent rendering of the OpenGL programmable pipeline,a good bleeding surface effect is obtained and a more realistic simulation result is achieved.