基于RBF神经网络的B样条曲面重构

B-spline Surface Reconstruction Based on RBF Neural Network

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,并在B样条曲面上做了仿真试验.实验结果表明:该模型不仅能够有效地逼近不完善的、带有噪声的曲面,而且拟合精度高、网络的训练速度快.说明了径向基函数神经网络应用于曲面重构问题的可行性.

This paper use RBFNN model to reconstruct free surface. Surface Reconstruction established for the RBF network model, and has done a simulation test on the B-spline surface. The results show that the model can not only effectively approach imperfect, with noise surface, and have a high-precision network training speed. Described the feasibility of RBFNN on surface reconstruction.