Research On Progressive Mesh Simplification Algorithm For Mobile Devices

Limited by processing capacity,screen size,network bandwidth,mobile devices can notrender high-density meshes,but need low-density meshes maintaining the key features of the models.For applications such as 3D games and 3D maps,the resolution of the model needs to be adjusted according to the change of viewpoint,which requires the multi-resolution display of the mesh models.Progressive mesh represents the original triangular mesh as a simple basic mesh and a set of progressive records.It is suitable for multi-resolution display and progressive transmission,because the basic mesh can be restored to the original mesh without accuracy loss by processing the progressive records;and during the restoration process,meshes of different resolution can be generated.Among the edge-collapsing mesh simplification algorithms,QEM algorithm has fast speed,can generate uniform grids,but loses key features of the models after highly simplified.The algorithms taking the curvature-like features as edge-collapsing cost keep the sharp parts of meshes well,but have high risk of face-flipping which may breaking the structure of meshes.The integrated methods,taking more than two properties into the measurement of the edge-collapsing cost,are time consuming and hard to balance the properties.According to the requirements of mobile applications,aiming to solving the problem that a lot of simplification algorithms can not balance the simplification speed and quality well,a mesh simplification algorithm combining quadratic error matrices and curvature-like features is proposed to construct progressive meshes.In the definition of edge-collapsing cost,parameter w is set to dynamically control the relative influence of curvature-like feature and quadratic error on the cost according to the degree of model simplification,so as to achieve better simplification effect.Normal vector linear iteration method is used to simplify the calculation of edge-collapsing cost and improve the speed of simplification.This thesis evaluates the mesh simplification algorithm from two aspects: simplification speed and simplification quality.The evaluation index of simplification speed is the initialization time and simplification time of the model.The evaluation index of simplification quality is visual effect and the global error of the simplified mesh measured by bidirectional Hausdorff distance.In the experiment,QEM algorithm,Melax's algorithm and this algorithm are compared and evaluated with the above four indicators,and the effectiveness,superiority and universality of this algorithm are analyzed.The experimental results show that the proposed algorithm can maintain the key features of the model,significantly reduce the mesh structure damage caused by face-flipping,and simplify the mesh quickly.It is insensitive to the uniformity of the original mesh,and has the ability to simplify large-scale meshes,and performs better on the mesh with sharp features.