A FrameworkforImplicit Surfaces Reconstruction formLargeClouds ofPoints
Ashraf Saad Hussein · 2007
Thispaperpresents anintegrated framework for surfaces fromlarge scale clouds ofpoints considering any surface reconstruction capable ofhandling large scale clouds of distributed memoryhigh-performance computer. On the points. Thisframework isbasedontwoproposed methods for other hand, theproposed AMR-based polygonization method implicit surface fitting andpolygonization toconvert a cloudof unorganized points into anoptimized surface. Theproposedfitting optimizes thepolygonization ofthereconstructed surface by method employs thePartition ofUnity (POU)method associated adapting domainsampling tothesurface geometric details. withtheRadial Basis Functions (RBF) overadistributed computing environment tofacilitate andspeedup fitting oflarge scale clouds Theoutline ofthepaper isasfollows. First, wepresent the without anydatareduction topreserve allthesurface details. related worktotheimplicit surface fitting andpolygonization Moreover, aninnovative Adaptive MeshRefinement (AMR)based insection 2.Thearchitecture oftheproposed framework is method isproposedfor implicit surface polygonization. This method described indetails in section 3.The accuracyand steers adaptive volumesampling viaa series ofoptimization criteria toprovide accurate andoptimized surfaces withminimum performance oftheproposed framework areintensively number ofpolygons. Theexperimental results fortheconsidered studied insection 4.Finally, section 5gives theconclusion test models showed anaverage reduction of60% infitting time andanoverview ofthefuture work. using 16processing nodesand90%inpolygonization timeonthe master nodeonlyagainst other traditional methods withbetter II.RELATED WORK performance. . AMR-based~~~~~~~~ poyoiain mto.Cneunl, it last strategy depends onthepartition ofunity method((10) faiitts enane an spesu'eosruto fipii and(9)) wheretheradial basis function isemployed tosolve