Global Land Cover Classification Information Extraction Methods Of Scale

Research on many areas of land cover data, such as global climate change research, landuse classification and coverage changes, and changes of ecological environment andbiodiversity research, surface, and even the touch family property of the census geographic conditions are indispensable data support. Development of high precision, high resolutionglobal land cover data products have been hoping to achieve international. At present, the international developed several sets of global land cover data, the resolution is only 1000 meters and 300 meters, time limitation, data accuracy and spatial resolution and are present,it is difficult to meet the simulation of global change and earth model needs.In China, although there are some related research and data products, but also covers only the domestic area. Because of the urgent need for the study of global change and earth system model in 2010, China officially launched the research of remote sensing mapping and key technologies of global land cover 863 key projects. The use of remote sensing technologyis becoming more and more mature, the design and development of a new generation of high precision global land cover data products, complete sets of production technology system research and practical, fill our country and international blank in this scale, to provide important basic data for the study of global change and sustainable development planning.This thesis relies on the project in 863 American land cover classification information extraction research carried out, by means of remote sensing technology, method of global land cover classification information of single element layer extraction using 30 mTM images,and on this basis to establish a new classification system can better meet the development model of the earth’s land cover data, the formation of scale production technology system.Research has proved that the technology can effectively reduce the classification of the different spectrum and different objects with the same spectrum phenomenon, and greatly improved the efficiency, precision and large range image information extraction.