基于KMEANS的网格简化算法.pdf

大连理工大学学位论文独创性声明作者郑重声明:所呈交的学位论文,,除文中已经注明引用内容和致谢的地方外,本论文不包含其他个人或集体已经发表的研究成果,. 若有不实之处,本人愿意承担相关法律责任. 学位论文题目: 基丑甾幽E趟生鲴1翌整金』立簋i姜作者签名: 算_厶———Et期:—上丛年—』月卫日万方数据大连理工大学硕士学位论文摘要在计算机图形领域,一个三维图形是由大量的多边形网格组成,三维图形越逼真,所需的多边形网格数目就越多,图形的拓扑结构就越复杂,,需要图形的复杂度也不一样,但是现有的网格模型曲面简化算法在较小的时间复杂度的同时,牺牲了聚类后的简化质量,,在采样点的基础上,利用区域增长的方法进行聚类,在聚类的过程中,,,,还能很好地保持图形的边界特征. 关键词:网格聚类;最远测地距;PCA算法;KMeans算法;采样点万方数据基于KMEANS的网格简化算法 Abstract Grid clustering model puter graphics applications fieldsuch puter animation, interactive visualization and virtual reality has broad existing鲥d clustering algorithm may bespend more time plex,but theresultis not thedatapoint clustering urate phenomenon atthe edge of the paper proposes anew clusteringalgorithm based on faceclustering,In order togetting thesampling points,the algorithm utilizesthemost fargeodesic distance adaptive on the sampling points,we utilizesthefunction ofmeasurement toclustering trianglemesh,and then takeadvantage ofPCAalgorithm andK/VIeansalgorithm make pact and theoriginal model data,puting time greatly issimple,fast andgood to keep theboundary set ofillustrationsaregiven toshow theeffectiveness ofthealgorithm. Key Words:mesh clustering Dijkstra distance;PCA algorithm;kmeans sampling point —II 万方数据大连理工大学硕士学位论文目录摘要???????????????????????????????.I Abstract.??.???.?.?.?.?.?.?.?.?.?.?.?.???.?.??..??.??.??..?...?.?...?II 引言???????????????????????????????l 绪论????????????????????????????????????????.5 国内外相关研究进展??????????????????????7 ???????????????????8 2简化收敛误差???????????????????????????1 ?????????????????????.11 ?????????