Up till now, many researchers have conducted a study to predict changes in land cover. The algorithms used in prior studies have the advantage of predicting land cover changes on a pixel basis and considering the interaction between local pixels. Howe ...
Up till now, many researchers have conducted a study to predict changes in land cover. The algorithms used in prior studies have the advantage of predicting land cover changes on a pixel basis and considering the interaction between local pixels. However, the focus was mainly on urban areas land cover changes, and there was a lack of theoretical basis for predicting changes in rural areas land cover. There was also a difficulty in building data for analysis of metropolitan areas.
The purpose of this study is to develop algorithms for predicting land cover changes in metropolitan areas while maintaining the advantages of existing algorithms. Based on the probabilistic spatial data integration method, this algorithm can identify the regularity of land cover changes and predict future land cover changes and apply them to urban and rural areas simultaneously. The regularity of land cover changes will use the Relative Facility Function (RFF). The GIS spatial data will be used for this research are multi-period satellite images and various natural, humanities, and social themes. To assess whether the land cover change prediction model is useful for predicting land cover in metropolitan areas, the forecast results and actual changes in rural areas will be compared and evaluated. In addition, a comparison and evaluation of the urban spatial structure and changes in urban areas will be conducted to assess whether significant results have been derived from the cluster characteristics of the changing areas as well as the pixel-unit prediction. The algorithms developed in this study will help assess the usability of human and social scientific research tools. Furthermore, this study set up the research objective to develop and distribute free land cover change prediction programs using open source GIS for social and educational dissemination of research results.
In the first year, land cover mapping and algorithm development using multi-temporal satellite image and GIS spatial subject systems were used to construct basic data for land cover change detection and prediction. In detail, variables and criteria for urban setting were prepared through literature review, and the characteristics of land cover changes according to urbanization/industrialization were analyzed through the collection of statistical data by land cover. Next, land cover mapping applied by the maximum likelihood method and BPLE method were developed using multi-temporal satellite image data, and new classification algorithm was developed to conduct optimal land cover map. In addition, a land cover change matrix was constructed using the post-classification comparison method. Based on this, land cover change was detected and an initial version of the land cover change detection and change area extraction module was developed. Finally, a GIS database for land cover change detection/prediction was constructed and a spatial thematic map was developed.
In the second year, it was aimed to develop a land cover change detection algorithms and programs using multi-temporal land cover and GIS spatial data. In detail, the changes in future land cover by using time series data of land cover were calculated and the future change characteristics of land cover were analyzed. In addition, the actual changes in urban areas and the direction of change in the study area were identified according to the criteria for setting up urban areas. Eventually, we developed an optimal land cover mapping that improved the classification algorithm of land cover in the maximum likelihood method and PBLE method. In addition, we investigated the relation between the multi-temporal land cover and GIS data using frequency distribution function and analyzed the regularity and characteristics of the change. Finally, a land cover change detection program was developed that incorporates GIS spatial data and land cover scheme based on frequency distribution function.
In the third year, land cover change prediction algorithm and program were developed through the integration of remote sensing data and GIS spatial data. For this purpose, the land cover change prediction algorithm was verified using the change prediction result data for each land cover. The characteristics of the area change among the types of urban structures and the need for urban management of adjacent urban areas were identified. In addition, the relevance of various topics was reviewed to land cover change detection and improve accuracy using the empirical frequency distribution function. Finally, based on frequency distribution function, the land cover change prediction program was developed and verified by integrating GIS spatial data and land cover map.