Abstract: With the continuous development of ionospheric detection technology, there are more ionospheric observation data. Using only a single observation datum will cause the problem of low accuracy of ionospheric inversion. In order to improve the inversion accuracy of the ionosphere, the BP neural network technology is used to effectively fuse the vertical total electron content (VTEC) data of ground-based inversion and international reference ionosphere (IRI) model inversion. Ionospheric inversion experiments were performed in temperate regions [35°(N)--45°(N)，60°(E)--80°(E)\]. The results show that the data fusion based on BP neural network technology and ground-based inversion can both achieve high accuracy of ionospheric VTEC. The data fusion method based on BP neural network has higher ionospheric inversion accuracy than ground-based inversion, so data fusion based on BP neural network technology can improve the ionospheric inversion accuracy.

Key words: ionospheric inversion, BP neural network, data fusion, electron content