土地覆盖是自然环境与人类活动相互作用的中心,准确而现势性强的土地覆盖数据是科学研究、资源管理和环境监测等应用的基础资料。该研究作为欧盟联合研究中心2000年全球土地覆盖计划(GLC2000)的一部分,利用2000年的1km空间分辨率的SPOT-4VGTS10数据与DEM、积温和降水等通过AHP方法合成的自然因子数据,采用FAO的土地覆盖分类系统(LCCS),通过非监督分类方法制作中国2000年的土地覆盖图。研究结果表明,在HANTS方法去云处理的基础上,结合气候分区,利用一年内每10天最大值合成的NDVI时间序列自然因子数据集可对除干旱区外大部分地区进行很好的分类,对干旱区则采用8月下旬的VGT原始数据取代NDVI数据参加分类,可达到较好的分类结果。

The China land cover database with 1 km spatial resolution，a part of GLC2000，has been developed with the\n10-day composite SPOT VGT data over a period of O1 January 2000 to 31 December 2000 and the multi-spectral data of the\nlast decade of August 2000. The multi-spectral image of the last decade of August 2000 was then used since the NDVI\ndatasets are not sensitivie to the area with poor vegetations in northwest China. Land-cover mapping for large area will face\nthe problem that the study area landscape greatly varies with the climate，terrain and soil，and it is a common phenomenon\nthat the different objects have the same spectral information or the same object has different spectral information. In order\nto resolve this problem，the whole territory of China had been stratified into eight relative homogenous bio-climatic strata\nand one specific Tibet region. The stratification was carried out based on the climate condition(above 10°C accumulated\ntemperature and aridity )，and then all data pre-processing，classification，interpreting and labeling processes had been\ncarried out in each stratum，respectively. This would be good for the accuracy of the result and process of the data.\n In order to remove cloud contamination and interpolate the missing data masked by cloud，the Harmonic Analysis of\nTime Series(HANTS) was applied to pre-process the 36 layer of VGT S10 NDVI dataset，and the ancillary metric derived\nfrom temperature，precipitation and DEM has been developed with the Analytic Hierarchy Process(AHP ) algorithm to im-\nprove the discriminating ability of VGT metrics. With the assistance of Erdas ISODATA algorithm，the classification has\nbeen carried out，and 22 types of land cover had been labeled in the whole China according to the Land Cover Classifica-\ntion System developed by the UN Food and Agriculture Organization(FAO)in the framework of the AFRICOVER project.\nThe accuracy assessment was complemented by comparing with the data derived from the product with the statistic data\nfrom Chinese Statistics Bureau.