Abstract
Aiming at the surface reconstruction problem, the surface reconstruction methods based on BP and RBF neural networks are studied respectively, and the two neural network models are compared. The results show that both network models can effectively reconstruct the surface. The BP model has a short running time, but the randomness of the weights and thresholds make the model stable. The RBF model has high reconstruction stability, but the fitting time is longer than the BP model. Both have their own advantages. The specific application should be determined according to the specific research object.