Calibration Of Regional MODIS LAI Products Considering The Difference Of Model,Scale And Image Parameter

MODIS LAI is a short-time remote sensing image product for global regional applications and is widely used in various fields.However,when it is applied to a specific area,there will be a significant underestimation or overestimation,which limits its efficiency and scope of use.Therefore,it is necessary to use high spatial resolution data to improve the data precision for regional high-precision applications.However,the low temporal resolution attribute of high spatial resolution data does not satisfy the one-to-one correspondence between the two kinds of data at all times.Therefore,how to use a limited number of high spatial resolution data to refine and quantify the general rule of precision difference between the two data is particularly important.This study is an important scientific issue for the accuracy improvement of remote sensing satellite biological parameter products,and serves the application of low and medium spatial resolution LAI products.The canopy radiative transfer model is used to quantify the uncertainty of the retrival process of MODIS LAI products by using a limited number of Landsat data.This study optimizes the process of enhancing product precision and refines the general rule of the accuracy difference of retrival results of different spatial resolution data.In order to improve the accuracy of the medium spatial resolution LAI products supported by high spatial resolution data and to optimize the validation process and general application of LAI products,the main work and conclusions of this study are as follows:(1)There is a big difference between Landsat8 OLI and MODIS imaging parameters.This study mainly considers the spectral difference and imaging geometric difference between the two images.By means of the point spread function of the sensor,the asymmetric two-dimensional Gaussian function is used to simulate the spatial response.The 30 m resolution Landsat8 OLI reflectivity product is converted to a 480 m reflectance product with the parameters normalized,and the result is in good agreement with the real reflectance data.(2)The 480 m Landsat8 OLI reflectance data is processed by the MODIS LAI retrival algorithm.The algorithm consists of two parts: the main algorithm and the backup algorithm.The main algorithm is mainly based on the three-dimensional radiative transfer theory of canopy radiation transmission and application of the look-up table.If the main algorithm fails,the back-up algorithm(the empirical model of NDVI and LAI)will be started to calculate the result.After verification,the MODIS LAI retrival algorithm has higher precision.(3)In order to solve the problem of scale effect generated by the retrival process,this study uses two-dimensional wavelet transform to decompose the image,and proves the correlation between high frequency information and scale error.This study uses the relationship between the coefficients of high-frequency and fractal dimensions to correct the scale difference of LAI results.The experimental results show that the wavelet-fractal method has a good effect on the correction of LAI.