Histogram clustering-based fusion model of high space-time normalized difference vegetation index NDVI

Abstract

The invention relates to a histogram clustering-based high spatial-temporal normalized difference vegetation index (NDVI) fusion model, which comprises the following steps of: obtaining high spatial resolution NDVI data HSt0 at a moment t0, performing Gaussian filtering processing on the high spatial resolution NDVI data HSt0, and obtaining a classification diagram 1 according to local histogram characteristics; performing up-sampling on the low-spatial-resolution NDVI data obtained at the t0 moment and the tp moment by utilizing the classification diagram 1 and adopting a linear hybrid modelto obtain HSu0 and HSup with the same resolution as HSt0 at different moments; taking HSt0, HSu0 and HSup as input data based on the classification diagram 1, and predicting high-spatial-resolution NDVI data HStp1 at the moment tp for the first time by utilizing a linear hybrid model; generating a classification graph 2 according to the local histogram features of the HSu0 and the HStp1; based onthe classification diagram 2, taking HSt0, HSu0 and HSup as input data, and predicting the high spatial resolution NDVI data HStp2 at the tp moment again by using the linear hybrid model; and based onthe sizes of different new coarse pixels, performing HStp2 prediction for multiple times, and finally generating a high-resolution NDVI at the tp moment.

Description

technical field [0001] The invention relates to the field of remote sensing image data processing or data analysis, in particular to a fusion model of high spatio-temporal normalized difference vegetation index NDVI based on histogram clustering. Background technique [0002] The normalized difference vegetation index (NDVI) enhances the absorption and reflection characteristics of vegetation, and provides a method to estimate the greenness and vitality of vegetation canopy. One of the commonly used indicators. NVDI time series products have received extensive attention and applications from regional to global scales. At present, it is mainly used in various aspects such as monitoring vegetation cover changes, crop growth status, object type identification, crop yield estimation, biomass estimation, surface evapotranspiration, soil moisture monitoring, and climate change. Therefore, long-term series NDVI products are powerful tools for understanding the past of vegetation,...

AI Technical Summary

Problems solved by technology

The above models are limited in the following aspects: (1) auxiliary data or prior knowledge are required, such as the disclosed model of CN 102831310B and the IFSDAF model need to input classification diagrams, and the prediction results will differ with classification diagrams, and increase Increased workload; (2) Two or more pairs of cloud-free benchmark NDVI data are required. For example, although the NDVI-LMGM model can complete NDVI data fusion through a pair of image reuse, it will affect the prediction accuracy; (3) It is difficult to be accurate at the same time Predict phenological changes and land cover changes, such as the NDVI-LMGM model is difficult to predict land cover changes or the prediction effect is poor

Images