The Study Of Annual Maximum Fractional Vegetation Cover Extraction Based On Temporal NDVI Data Of Landsat And MODIS

Vegetation coverage is an important indicator to evaluate ecological change,monitor soil and water loss,and study the degree of desertification.However,due to the limitation of vegetation types and phenology,the vegetation coverage in the same region varies greatly in different temporal remote sensing data.Even for the same crop,limited by topography(such as elevation and slope aspect),the spatial distribution of vegetation coverage in the same temporal remote sensing data is also different.In addition,limited by cloud pollution,it is often impossible to obtain enough remote sensing data in some areas for comparison of multi-phase FVC extraction results.Therefore,the results of vegetation coverage inversion based on multi-temporal remote sensing data will be biased when evaluating the changes of ecological environmental quality such as grassland degradation in a certain region.In this paper,the method of integrating Landsat and MODIS data to extract the maximum vegetation coverage in a year is discussed,and the results are used as one of the basic bases for evaluating the ecological and environmental quality changes such as grassland degradation using multi-temporal remote sensing data,so as to reduce the influence of phenology and topographic conditions on vegetation coverage.A vegetated area near Buha River was selected as the study area.Based on the Landsat and MODIS time series data during the growing season of this area.ESTARFM(Enhanced Spatial and Temporal Adaptive Reference Fusion Model)algorithm was used to complement Landsat NDVI data.Then,The vegetation growth curve of Landsat NDVI was reconstructed based on the Logistic model,and the corresponding dates of annual Maximum NDVI extracted from MODIS MVC(Maximum Value Composite)were introduced to retrieve the Maximum Value of Landsat NDVI pixel by pixel.Finally,the annual maximum FVC values were obtained based on the pixel dichotomy model,and the inter-annual maximum FVC changes in2001,2007,2014 and 2016 in the study area were analyzed.The results are as follows:(1)The Landsat-NDVI data obtained by ESTARFM spatiotemporal fusion algorithm is approximately consistent with the actual Landsat-NDVI values,and the correlation coefficient is0.89.Therefore,Landsat-NDVI data obtained by ESTARFM temporal and spatial fusion algorithm can be added to Landsat-NDVI time series data to make up for the problem of data missing caused by serious cloud pollution in high-altitude areas and low temporal resolution of Landsat images.(2)The maximum vegetation coverage date of the year extracted from MODIS MVC was introduced into the Logistic model to reconstruct vegetation growth curve.It not only reduced the number of initial values assigned to the Logistic model(reduced by 1),but also enhanced the accuracy of Logistic iteration results.(3)The maximum value of the annual NDVI extracted by the five-parameter Logistic model is in good agreement with the Landsat NDVI data of the corresponding date,with the correlation coefficient up to 0.96 and the median error of 0.03.Therefore,the extraction results in this paper better reflect the distribution of annual maximum vegetation coverage in the study area.(4)Compared with MODIS NDVI MVC inversion of annual maximum vegetation coverage,the proposed method not only improves the spatial resolution of inversion results(from 250 m of MODIS to 30 m of Landsat),but also has a clearer response to the details of ground objects.Moreover,the effect of topography and phenology on vegetation coverage was weakened.