多邊形物體的多重解析度模型

Abstract

在虛擬實境的應用中 , 穩定且快速的顯示速率是獲得平滑且真實的 視覺感覺的關鍵因素. 在這篇論文中 , 我們提出了一個自動地產生多層 次的精密度多邊形物體的演算法.這個方法包括了五個步驟 , 第一個步驟 是找出多邊形模型的邊界盒 , 第二個步驟是調整式切割這個邊界盒 , 第 三個步驟是計算多邊形面與子空間的邊界的交點 , 第四個步驟是收集原 始座標點以組成一些聚集並且將交點設定為代表點 , 第五個步驟則是連 結這些代表點以獲得簡化模型. 靠著限制經由調整式切割原始模型的邊界 盒所產生的子空間個數 , 本演算法便能夠讓使用者只事先指定簡化比率. 而後程式便能有效的化簡多邊形模型到指定的簡化比率(在有限的誤差範 圍內). In virtual reality applications and constant frame rate is critical for achieving smooth and realistic visual perception. In this thesis, we rpopose a method that reads as an input the polygonal mesh anda user-specified reduction rate, and then produces a reduced polygonal mesh satisfying the redduction rate.This method consists five steps. Step one is to adaptively subdivide the bounding box.Step threeis to compute the intersection points between the polygonal meshes and voxel's edges.Step four is to associate each polygonal vertex with an intersection point to form a cluster in which each intersectionpoint is representative point. Step five is to connect the represevtative points to avtain a simplifiedmodel. By limiting only the number of candidate voxels resulting from the subdivision ,the program isable to reduce the polygonal mesh to a level that is specified by the user.