多源卫星遥感土地覆被产品在南美洲的一致性分析

发布时间：2018-03-29 08:09

本文选题：土地覆被　切入点：空间一致性　出处：《遥感信息》2017年02期

【摘要】：针对不同卫星遥感产品在不同区域缺乏一致性基准的问题,提出类型构成相似性、类别混淆程度、空间一致性及参考程度等4种方法,对比分析不同土地覆被产品间的一致性。鉴于南美洲区域土地覆被空间结构和变化对全球变化研究具有重要意义,利用上述4种方法研究了GLOBCOVER2005、GLOBCOVER2009、GLC2000、MODIS2000、GLOBELAND30-2010等5种全球卫星土地覆被产品在南美洲地区的一致性。结果表明,5种产品对于南美洲土地类型的构成刻画基本一致,且对林地识别的一致性最高;南美洲有近60%的土地具有较高的一致性;5种产品两两比较时,参考精度大致在42.27~87.59%之间,GLOBCOVER2009/GLOBCOVER2005组合的参考精度最高,反映出土地覆被动态变化所引起的误差远小于不同制作机构、不同数据源、不同判读方法所带来的制作误差。
[Abstract]:Aiming at the lack of consistent datum in different satellite remote sensing products in different regions, four methods, such as similarity of types, degree of category confusion, spatial consistency and reference degree, are put forward. Comparative analysis of consistency among different land cover products. Given the importance of land cover spatial structure and change in South America to the study of global change, The consistency of five global satellite land cover products such as GLOBCOVER2009 / GLOBCOVER2009 / GLOBELAND30-2010 and GLOBCOVER2009GLOBELAND30-2010 in South America was studied by using the above four methods. The results showed that the five products were basically consistent in describing the composition of land types in South America and had the highest consistency in forest land identification. Nearly 60% of the land in South America has a high consistency. When comparing five products, the reference accuracy of GLOBCOVER2009 / GLOBCOVER2005 combination is the highest, which shows that the error caused by dynamic change of land cover is much smaller than that caused by different production institutions. The error caused by different data sources and different interpretation methods.
【作者单位】：中国科学院地理科学与资源研究所;资源与环境信息系统国家重点实验室;中国科学院大学;
【基金】：国家重点研发计划项目(2016YFB0501502、2016YFC0503701) 国家科技重大专项(00-Y30B14-9001-14/16)
【分类号】：TP79

【参考文献】