Landuse Dynamic Monitoring And Prediction Under CA-Markov Model

With the development of space technology, sensor technology, digital image processing technology and other related technologies technology, modern remote sensing technology has entered a real-time, dynamic, fast, accurately new stage which conducts earth observation and remote sensing technology is also increasingly used in land use dynamic monitoring.In this paper, by the images of a Wanbei county’s 2004,2009 and the same period of 2014 TM (2014 to OLI), used maximum likelihood method to employ by combining ENVI software with remote sensing technology and obtain three land use classification map. Compared the second national land data with classification results, the results suggest that classification is feasible. Through the analysis the sizes of each class vary of different stages in each reasonable allocation and infer weights formula, additional, calculate weights of each reasonable allocation by utilizing such formula while class changes. Moreover, use MCE model of IDRISI software to generate a type of land use suitability image and superimpose to generate a set of images. Based on transition probability matrix of the land use types in different period and predict land use situation to study area of 2014 and 2019. Compared the classification with the prediction results in weights and equal weights of 2014 and defined prediction accuracy. By comparison, defined the superiority of weights and according to the land use types evolution trend to put forward series of recommendations to land planning of study area.