An improved ANN method for GNSS elevation surface fitting
In the study of elevation datum transformation model in the construction of refined regional elevation system,machine learning algorithm is mainly used to replace the traditional surface fitting method,that is,nonlinear machine learning model is used to realize the fitting of GNSS elevation system.In this paper,the ant colony optimi-zation algorithm,genetic algorithm and particle swarm optimization algorithm are used to optimize the ANN model.The measured GNSS and leveling data of a mining observation station are used to verify the effect of the optimization algorithm.The experimental results show that,in the case of large observation area and irregular height,the opti-mized ANN model has achieved good results.The ANN model optimized by the particle swarm optimization algo-rithm is more suitable for the GNSS elevation surface fitting in a small area,which effectively improves the accuracy of elevation fitting.
NETL
NSTL
万方数据
