Height Fitting by Radial Neural Network for the Construction Quality Control of Face Rockfill Dam

According to the characteristics of face rockfill dam surface and the points collected by the surveying robot are much and their height difference is small,to achieve optimal fitting effect compared with real height,a method was proposed for fitting height by radial basis function(RBF) neural network based on multiquadric(MQ) function and optimal smooth factor by means of mathematical statistics and progressive approach method.To verify its feasibility,we carried out the surface height fitting of face rockfill dam with the data from surveying robot.The results show that the method can get better fitting effect and has higher precision than many common fitting methods.So the method maybe more suitable for surface height fitting of face rockfill dam and can provide the reliable data base to monitor the rolled thickness and construction quality for face rockfill dam.