Tensor-based land utilization simulation method, system and device and storage medium

Abstract

The invention discloses a tensor-based land utilization simulation method, system and device and a storage medium. The method comprises the steps of obtaining land utilization data, driving factor data, a conversion cost table and an inertia intensity table; training a convolutional neural network to obtain a fitting model, and further obtaining a conversion probability tensor; executing neighborhood influence calculation to obtain a neighborhood influence tensor; executing conversion cost calculation to obtain a conversion cost tensor; executing inertia coefficient calculation to obtain an inertia coefficient tensor; obtaining a combined probability tensor according to the conversion probability tensor, the neighborhood influence tensor, the conversion cost tensor and the inertia coefficient tensor; and executing tensor wheel disc selection on the combined probability tensor to obtain a simulation result. According to the method, each part in the land utilization simulation process based on the cellular automaton is formalized into the tensor, and the tensor framework is applied to accelerate the operation, so that the efficient land utilization simulation process is realized while the complex parallel algorithm design and CUDA programming are avoided.

Description

[0001] technology neighborhood [0002] The invention relates to the field of geographical information science, in particular to a tensor-based land use simulation method, system, device and storage medium. Background technique [0003] With the further development of society, the concept of land planning also changes with the change of industrial structure; therefore, the study of the spatial and temporal change trend of land in a specific area has important reference significance for future planning. [0004] Cellular automata, also known as CA, as a space-time discrete dynamic model with strong spatial computing capabilities, plays an important role in complexity science; the multi-class land use simulation method based on cellular automata, according to Neighborhood state and a series of conversion rules to simulate the complex temporal and spatial changes of land use have become an important tool for studying land use change and its socio-economic and natural environmenta...

AI Technical Summary

Problems solved by technology

These methods can indeed improve the performance of the CA method, but the realization of these advanced algorithms requires more computer knowledge and experience, which is usually lacked by geographers

Images