Study On Land Remediation Partition Based On Self-organizing Feature Map Neural Network And System Clustering

During the period of the 12th Five-Year Plan period, the land remediation work has achieved remarkable results under guidance of land remediation plan, which promoted the comprehensive improvement of the quantity and quality of cultivated land and impelled farmers' continuous increase on income, agricultural efficiency and rural development. It becomes an important livelihood project in supporting the construction of beautiful rural and moderately prosperous society in all respects. Land remediation zoning is the prerequisite and foundation for compiling land remediation plan. At present, every province are compiling the land remediation plan of the 13th Five-Year Plan period. The paper takes Hubei Province as an example to carry out a zoning research and then propose land remediation priorities and measures of different areas, which can provide the basis for the arrangement and planning of land remediation project in Hubei Province.First of all, this paper summarizes the relevant research results at home and abroad,and clarifies the research contents, methods and ideas. Secondly, it expounds the theoretical support of the land remediation division and introduces the connotation,principles and methods of the land remediation division. Thirdly, the factors influencing the land remediation division are analyzed from five aspects: natural resource condition,economic development level, social living standard, ecological environment condition,regional strategic orientation and policy orientation. Then,according to the strategy of land consolidation of Hubei Province and the principle of index construction, eighteen indexes of four factors in natural ecological conditions, socio-economic conditions, land use conditions and land remediation potential were selected to construct the zoning evaluation index system, which are explained in detail.Finally, the clustering method based on self-organizing feature map (SOM) and the system clustering method are used to delineate the land remediation area of Hubei Province.And two different partition schemes are obtained. The partition defined by the SOM-based clustering method contains 11 categories and another one defined by the system clustering method contains 12 categories. On this basis, the partitioning scheme obtained by the two methods is compared and integrated. The regions classified in different categories in two partition are analyzed. Then,the integrated partition is adjusted accordingly. And the optimized land consolidation plan of Hubei Province is finally obtained, including seven land remediation zoning. According to the actual situation of the district and the priorities of the land remediation, the seven types of zoning are summarized as: ?- urban-rural integration remediation areas, ?-agricultural remediation area, ?-comprehensive agricultural remediation area, ?- urban optimized remediation area, ? - precision poverty alleviation remediation area, ? - modem agricultural remediation area, ? - ecological land remediation area.The conclusion of this paper is that the classification scheme of the land remediation in Hubei Province by two kinds of clustering methods is scientific and accurate, which has certain guiding significance for the regional land remediation work in Hubei Province and provide the scientific basis for the compilation of land remediation plan of Hubei province.