重庆MODIS/NDVI时间序列数据优化研究

Research on the Reconstructing of Time-series MODIS/NDVI Data of Chongqing

重庆独特的地理环境与气候特征导致NDVI普遍存在明显的云干扰等导致的噪声。目前最为常用的最大值合成法(MVC)虽然一定程度上减弱云雾及其他因素的影响,但不能保证获取高质量的数据。为了构建重庆地区高质量NDVI时序数据集研究采用时间序列谐波分析法(HANTS)对MODIS/NDVI数据进行时间序列重构;并进一步结合质量控制文件对MODIS/NDVI做进一步订正。选择2001—2010年MODIS/NDVI数据进行序列订正效果分析表明,原始NDVI数据存在较多的噪声,不能直接用于年代际时间特征趋势分析。经过HANTS重构、质量控制文件掩膜合成构建的NDVI数据序列,对因为云的影响导致的原始数据偏低现象给予了有效订正,又较好地保留了原始数据的高质量信息,维持了原始数据应有的整体趋势特征,适合用于进一步的时空特征分析。

Due to effect of cloud and atmospheric conditions, residual noise in the NDVI time-series data will induce erroneous results in our further quantitive analysis in Chongqing. The method of Maximum Value Compositing （MVC） can reduce the effect of cloud, atmospheric and other conditions in some degree, but it also has some shortcomings. In this paper the method of Harmonic Analysis of Time Series （HANTS） and the quality control documents for MODIS / NDVI are used to produce the high quality MODIS/NDVI time-series data. The NDVI time-series data from 2001 to 2010 are selected as thesample data. The results show_that the original NDVI data exists more rmise, cannot be direetly used for trend analysis features of decadal time. The reconstueted MODIS/NDVI time-series data using HANTS method and the quality control documents for MODIS / NDVI can revise the low value of NDVI due to the mask of cloud effectively, and preferably retains the high quality information in original data, and maintain the characteristics of the overall trend of the original MODIS/NDVI time-series data. So the optimized and the reconstructed NDVI data are suitable for further analysis.