Research On The Construction Of Tasseled Cap Transform Indices Time Series Data Sets Based On Spatial-Temporal Fusion Algorithm

Remote sensing images with high spatial resolution are usually the primary data source for the study of the estimation of biophysical parameters with a high accuracy. However, satellite sensors cannot provide images with high spatial resolution and high temporal resolution simultaneously, due to the trade-off between spatial, spectral, and temporal resolutions. In a given period, it is difficult to obtain change information about land surface which occurred during a short time within a small area, because of the lack of high quality remote sensing image with high spatial resolution. High temporal resolution images cannot be used to the study of small region due to the limitation of spatial resolution.In recent years, several spatial and temporal fusion algorithms have been developed to combine remote sensing image with various spatial resolutions and temporal densities to meet the demand of time series change detection at higher spatial resolution. In previous studies, spatial and temporal fusion algorithms are based on surface reflectance data, nevertheless, spectral index based on multi spectral reflectance usually has a close correlation to the specific biophysical parameter of land surface and has a more explicit meaning, therefore, time series data sets generate by spatial and temporal fusion algorithms can reflect the land surface change continuously and serve as a new data source for the environmental change research. The tasseled cap transformation (TCT) is a useful tool for compressing spectral data into a few bands associated with physical scene characteristics with minimal information loss, originally constructed for understanding important phenomena of crop development in spectral space. The first three components, brightness, greenness and wetness, are directly associated with the important physical properties of the land surface, such as land cover types, vegetation types and distribution, and soil moisture. Based on the spatial and temporal fusion model, time series tasseled cap transformation indices with time continuity of regional were produced from MODIS and Landsat OLI data, which can be used to indicate vegetation cover state and the characteristics of phenology. The data set can serve as a quantitative indicators for ecological monitoring and environmental change detection.Based on the spatial-temporal fusion algorithm, this paper analyzed the construction method of tasseled cap transform indices time series data sets based on the Landsat OLI and MODIS reflectance data of the Bingcaowan area in GuLang County in Gansu province and the middle west of Zhongwei city in Ningxia Province. The results of previous spatial-temporal fusion algorithm were analyzed firstly, an improved algorithm was developed by improve the selection of similar neighbor pixels and the calculation of weight for similar pixels. Based on the MODIS daily surface reflectance product and Landsat OLI images with high quality of the two study areas, time series data set were produced respectively.The results show that,(1) Accuracy of the fused image based on the existing algorithm was correlated to the land cover types, and for the data of this study, the result of the wetness was more accuracy. The result of the improved algorithm can reflect the distribution characteristic of different land cover types, the fused images have a higher correlation with the reference data.(2) The window size and estimated number of classes determined the selection result of similar neighbor pixels and precision of fused result, but the quality of the fused result is limited effected by different parameters.(3) Characteristic of the TCT indices was consistent with the vegetation phenology correspondingly, and the data sets produced by the spatial and temporal fusion algorithm can serve as a quantitative indicator for ecological monitoring and vegetation phenology.