Inversion Of Grassland Vegetation Greenup Based On High Temporal And Spatial Resolution NDVI Dataset By Remote Sensing

Phenology of grassland vegetation can reflect the rapid response of terrestrial ecosystem to environmental change,and is an important basis for grassland ecological regulation and grazing policy formulation.At present,the spatial resolution of remote sensing images used for grassland vegetation phenology monitoring is mostly low,and the impact of meteorological factors such as clouds,rain and snow on sensor imaging results in a significant reduction in the effective image of a single sensor and a reduction in image temporal resolution.Thus,it is difficult to achieve continuous monitoring in the same region using high temporal resolution images.Spatio-temporal data fusion technology can combine the spatial information of high spatial resolution remote sensing image and the time information of high temporal resolution remote sensing image to generate fusion image with high temporal resolution and high spatial resolution,to some extent make up for the shortcomings of sensor technology and the impact of meteorological environment on the quality of remote sensing image.Although there are many existing spatio-temporal data fusion algorithms,the principles of various algorithms are different,and the advantages and limitations of the algorithms are not the same.In view of this,this paper takes Hulunbuir and Xilinhot in Inner Mongolia as research areas,and studies for resolving the problem of low spatial and temporal resolution of remote sensing data sources in grassland greenup remote sensing monitoring.The main research contents and corresponding research results are as follows:(1)Three kinds of spatio-temporal data fusion algorithms were introduced.Based on CUDA heterogeneous programming technology,the corresponding data fusion tools were developed.The high-temporal resolution NDVI dataset in the study area was formed,which solved the problem of low spatial and temporal resolution of remote sensing image used to inverse the grassland vegetation greenup.(2)Evaluate and analyze the adaptability of different fusion algorithms and strategies to the spatial pattern and spatio-temporal Variance.The results show that each algorithm does not have universality for different geomorphological patterns.When the Temporal to Spatial Variance Ratios is large,the three fusion algorithms can not predict the sudden disturbance information well.The overall fusion accuracy of the STDFA algorithm is the highest,followed by the STARFM and USTARFM algorithms.Compared with reflectance fusion,NDVI fusion has better adaptability to different geomorphological patterns and Spatio-Temporal Variance.(3)Compare and analyze the effects of long-term sequence data constructed by different fusion algorithms in the inversion of vegetation greenup.The results show that the high spatial-temporal resolution NDVI generated by spatial-temporal data fusion can improve the inversion accuracy of grassland vegetation greenup to a certain extent.The greenup of the fusion data of the STARFM algorithm is the closest to the measured value of the ground sampling point.The accuracy of greenup inversion,which are extracted from the fusion daily data and its 8d synthetic data,decreases in varying degrees when the base period Landsat data is missing in the neighbourhood of the greenup,and the STARFM algorithm is least affected by the data missing,followed by the USTARFM algorithm.(4)The accuracy of non-time series data fusion does not determine the accuracy of the algorithm’s fusion in long-term sequence data.Considering the advantages and limitations of each algorithm,the STARFM algorithm based on NDVI fusion strategy is more suitable for the construction of high spatial-temporal resolution NDVI data in grassland.