摘要
根据径向基函数(RBF)神经网络可以用任意精度逼近任何非线性函数,以及强大的抗噪、修复能力等优点,该文采用RBF神经网络模型进行自由曲面重构,建立了适合曲面重构的径向基函数网络模型。进行了理论分析,并在非均匀有理B样条(NURBS)曲面上做了仿真试验。结果表明:该模型不仅能够有效地逼近不完善的、带有噪声的曲面,而且学习速度很快,提高了对破损、不完全曲面重建的效率和精度,得到的曲面光顺性好。
Based on RBFNN' s advantages of approaching any no -linear function by arbitrary precision,powerful antinoise and the capability of repair and so on,this paper adopts RBFNN model to reconstruct free-form surface,and constitutes RBFNN model fitting to surface reconstruction.The paper makes theoretical analysis and makes experiment on the NURBS surface.The result indicates that this model not only can efficiently approach the surface which is not perfect and has noise,but also the speed of learning is very quick.This can improve efficiency and precision of dirty,imperfect surface reconstruction,and the smooth of surface which is got is good.
出处
《计算机工程与应用》
CSCD
北大核心
2005年第11期66-69,共4页
Computer Engineering and Applications
基金
北京市教育委员会项目(编号:KM200410028013)
