A study on the dynamic evolution method of spatio-temporalfield data based on tensor decomposition

Abstract

Spatio-temporal field data is one of the primary forms of digital representation of geographic objects, characterized by massive, multidimensional, complex evolution and analysis-oriented, etc. The organization, management, and analysis of the above data have become important bottlenecks of existing information systems. Moreover, the analysis of dynamic evolution behavior of Spatio-temporal field data is less studied. The correlation between attributes and time is not studied enough, making it difficult to carry out smooth and accurate dynamic simulation and expression in the scene. Therefore, this paper proposes a dynamic organization and description scheme of Spatio-temporal field data based on tensor Tucker decomposition. Firstly, the Spatio-temporal field data is organized based on tensor Tucker decomposition, and the original Spatio-temporal field data is chunked according to time. The Spatio-temporal field data sub-packaging mechanism and dynamic evolution description scheme are designed based on the description document and the field data structure itself; Design time-space field data attributes and the expression relationship of time attributes. Finally, the Spatio-temporal deformation field data model in ground subsidence monitoring is used for applied research. The experimental results show that the dynamic evolution method of Spatio-temporal field data based on tensor decomposition proposed in this paper can achieve accurate dynamic simulation representation of Spatio-temporal field data in geographic scenes.