一种蓝藻水华提取空间尺度效应分析  

Scale Analysis of Cyanobacteria Bloom from MODIS and HJ-1A/1B Observations

作　　者：李晓俊[1,2] 马泽忠[1,2] 周志跃[1,2] 

机构地区：[1]重庆市国土资源和房屋勘测规划院,重庆400000 [2]重庆市土地利用与遥感监测工程技术研究中心,重庆400000

出　　处：《遥感信息》2016年第5期108-113,共6页Remote Sensing Information

基　　金：国家"十二五"科技支撑项目(2012BAJ23B05)

摘　　要：针对不同尺度遥感数据监测蓝藻水华必然会带来尺度效应的问题,以太湖为研究区,基于准同步的MODIS L1B产品数据以及HJ-1A/1B数据,分析两者获取水华信息的空间尺度差异,应用地统计学中的半变异函数对蓝藻水华分布的空间异质性进行研究,并进一步分析2种尺度下水华提取的空间尺度效应与空间异质性的关系。发现由NDVI计算方式非线性引起的差异可以忽略不计;在30m尺度上随机因素对空间异质性影响很低,样本区的空间异质性主要是由空间自相关性的降低引起的;2种尺度上水华提取差异与空间异质性的关系可以用清晰的线性关系表达。该研究对于下一步修正MODIS 250m尺度上的水华提取精度具有重要意义。There is scale difference when retrieving cyanobacteria bloom information from images with different spatial resolution.In this paper,we selected Taihu lake as the research area and statistical samples which have the same extent but different algae coverage proportion.Then we used semi-variogram to describe the spatial heterogeneity in statistical samples.Several conclusions are drawn：error caused by the NDVI calculation method is negligible;random factors of spatial heterogeneity can be ignored within 30 mscale.The main factor causing spatial heterogeneity is spatial auto-correlation.We also found that spatial heterogeneity has good regression relation with the scale difference between two sensors.It will be important to correct the errors of extraction of cyanobacteria bloom using images at large scale.

关 键 词：蓝藻水华 尺度效应 空间异质性 半变异函数 太湖 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]