Change Pattern Of Land Cover And Its Driving Force From 1990 To 2015 In China Indochina Peninsula Economic Corridor

Since China put forward The Belt and Road(B&R)initiative,the areas along the B&R have increasingly become the focus of domestic and foreign scholars.China Indochina Peninsula Economic Corridor(CICPEC),as a Bridge for economic and trade exchanges between China and ASEAN in the six commercial corridors of B&R,has a significant strategic position.As the basic data of CICPEC planning and construction,land cover plays an essential role in corridor construction and ecological environment assessment.The study of land cover change,as a hot topic in current academic circles,can effectively reflect the close relationship between human activities and the natural environment.The extraction and analysis of land cover and its change information are also of great significance for the study of earth system science.With the rapid development of remote sensing technology,more and more scholars use satellite remote sensing images to study land cover.With the development of computer technology and artificial intelligence,more and more artificial intelligence algorithms are introduced into the field of land cover mapping.At the same time,applying remote sensing data to land cover change detection research can effectively overcome the problems of high cost and low efficiency of traditional change detection methods.Therefore,the introduction of new technologies and methods into remote sensing mapping of land cover in the areas along the CICPEC route can greatly improve the accuracy of land cover classification.On this basis,the analysis of the spatial and temporal pattern changes and driving forces of the classification results can better serve the green B&R initiative.In this paper,the 100km buffer zone along the CICPEC is taken as the research object,and the Landsat TM / OLI remote sensing images in 1990,2000,2010 and 2015 are used as the data source.The multi-classifier iterative classification model and multiindex coupled change detection method are developed to generate the CICPEC 1990-2015 long time series land cover data set.Based on this,the spatial and temporal pattern changes of land cover in the study area from 1990 to 2015 were analyzed using land cover change amplitude,dynamic attitude,and transfer matrix.Besides,the driving forces of land cover change in CICPEC were analyzed from three aspects: nature,population,and social economy.The results are as follows:(1)Proposed an Object-oriented Iterative Classification method based on the Multiple Classifiers Ensemble(OIC-MCE)method.The spectral features,spectral index,terrain feature,texture feature,and shape feature of the image in the study area are fully used to perform object-oriented multi-resolution segmentation of the remote sensing images.Based on this,five common classifiers of Random Forest,Support Vector Machine,K-Nearest Neighbor,Decision Tree,and Naive Bayes are integrated to construct an Object-oriented Iterative Classification method based on the Multiple Classifiers Ensemble(OIC-MCE).The land cover data sets of 1990,2000,2010,and 2015 were obtained through the model.(2)Developed a Multi-Index Integrated Change Analysis(MIICA)method.This method was constructed by ?NDVI,?NDBI,?NDWI,?NBR,VS,and CVA,and obtaining the three-phase change data from 1990 to 2015.At the same time,the Post Classification Comparison(PCC)was used to compare the land cover results obtained through OIC-MCE,and the change information was extracted.The two sets of change detection results are integrated to get the final land cover change data.(3)Produced the CICPEC land cover data set from 1990 to 2015.After verification,the classification accuracy is above 85%,and the Kappa coefficient is above 0.8.(4)Analyzed the dynamic change pattern of land cover in CICPEC in the past 25 years(1990-2015).We concluded that,in the past 25 years,the proportion of the land cover area of the first-level category in the study area.From the top to the low portion are as follows: Forest lands> Cultivated lands> Water bodies> Shrublands> Grasslands> Impervious surface > Wetlands> Barren lands.In addition,the spatial and temporal pattern changes of land cover in the study area were analyzed by using the change amplitude,dynamic attitude,and transfer matrix of land cover.It is concluded that the types of land cover that have changed the most in the past 25 years are Forest lands,Impervious surface,and Cultivated lands,which is reflected by the trend of Forest lands decrease but Impervious surface and Cultivated lands increase.(5)A preliminary discussion on the driving force system of CICPEC land cover change.Explore the driving factors of land cover change in the study area through three aspects of nature,population,and socio-economic.The most important factors are population and socio-economic factors.Through the analysis of population,GDP,and timber export volume of countries along the CICPEC,it is concluded that with the growth of population and economy,it will bring a certain degree of land cover changes.