RBF Neural Network Arithmetic and Applications in Surface Interpolation Reconstruction

摘要

Aiming at problems such as: surface interpolation reconstruction of points cloud data, surface hole filling and two simple surface connection, a neural network arithmetic was employed. Based on radial basis function neural network, simulated annealing was employed to adjust the network weights. The new arithmetic can approach any nonlinear function by arbitrary precision, and also keep the network from getting into local minimum for global optimization feature of simulated annealing. MATLAB program was compiled, experiments on points cloud data have been done employing this arithmetic, the result shows that this arithmetic can efficiently approach the surface with 10~(-4) mm error precision, and also the learning speed is quick and reconstruction surface is smooth.