Land utilization simulation and prediction method

Abstract

The invention discloses a land utilization simulation and prediction method. The method comprises the following steps of collecting land utilization data, performing preprocessing to obtain an initial data set, and dividing the initial data set into a training set and a test set based on a sliding window technology; constructing a ConvLSTM-CA model on the basis of a ConvLSTM algorithm and a city expansion CA model, and training the ConvLSTM-CA model by taking the training set as input to obtain a target ConvLSTM-CA model; and calculating the conversion probability of the land utilization type according to the target ConvLSTM-CA model and the land utilization data, and based on a maximum probability conversion principle, obtaining a land utilization simulation prediction result. According to the prediction method, the utilization rate and the application value of the prediction network are greatly improved, and meanwhile, the ConvLSTM-CA model is adopted, so that the simulation effect is ensured, and meanwhile, the simulation precision is greatly improved.

Description

technical field [0001] The invention belongs to the field of land simulation prediction, in particular to a land use simulation and prediction method. Background technique [0002] With the rapid growth of my country's economy, the number of urban population continues to increase, and the scale of urban expansion is gradually increasing. Today's new data environment has large volumes of data, multiple data categories, wider coverage of data scales, faster updates, stronger timeliness, higher accuracy, and more humanistic. The evolution of city scale is affected by many aspects such as social economy, infrastructure, population, etc., so the process of urban evolution is very complicated. Predicting changes in future land use, so as to pre-arrange land development, utilization, improvement and protection in time and space, has important practical significance for the allocation and sustainable use of regional land resources. With the popularity and maturity of technologies ...

AI Technical Summary

Problems solved by technology

Although a variety of linear and nonlinear methods have been applied in the construction of CA model conversion rules, these methods are based on the information mining of a single cell to construct conversion rules, even if there are neighborhood constraints in the rules. Since the change of the cell is not only affected by the state of the adjacent cell, but also affected by the spatial variable information of the adjacent cell, it is difficult to achieve a more accurate expression without considering the variable information characteristics of the adjacent cell. evolution simulation

Images