Spatiotemporal Fusion Method of Normalized Difference Vegetation Index Data Based on Different Spatiotemporal Resolutions

Abstract

A spatio-temporal fusion method of normalized vegetation index data based on different spatio-temporal resolutions, the spatio-temporal fusion of normalized vegetation index data with different spatio-temporal resolutions, using noise secondary filtering based on adjacent NDVI observations and high temporal resolution Noise filtering of NDVI median value; combined with linear interpolation and spatial filtering, automatic fusion of NDVI data with different temporal and spatial resolutions to generate high temporal and high spatial resolution NDVI data; and using spatial filtering algorithm to eliminate MODIS pixel boundary effects, and finally get High spatial resolution high temporal resolution NDVI data of the target region. The invention reduces the uncertainty of subsequent fusion processing without considering whether the input data is clear and cloudless; by combining the proximity to the NDVI value of the target pixel and the Euclidean distance to the target pixel, the NDVI value of the fused target pixel is corrected , to improve the continuity of the fused image.

Description

[0001] The present application relates to a method for obtaining normalized vegetation index data, specifically, a method for obtaining normalized vegetation index data with high spatial and temporal resolution after time-space fusion using normalized vegetation index numbers with different temporal and spatial resolutions. Background technique [0002] Vegetation index is composed of data from different bands of satellites, which can simply and effectively reflect the growth status of plants, and has a linear relationship with vegetation coverage. It is an important indicator to identify vegetation growth status and coverage. At present, dozens of vegetation indices have been developed, among which, the normalized difference vegetation index NDVI (Normalized Vegetation Index) can eliminate most of the effects related to instrument calibration, sun angle, terrain, cloud, shadow and atmospheric conditions, and increase the The ability to respond to vegetation is most widely used...

AI Technical Summary

Problems solved by technology

However, the above-mentioned method based on the linear spectral mixture model, the calculated pixel values ​​of medium and high spatial resolutions are only category averages, not real pixel values

The assumption of equal albedo of the same type when calculating the abundance matrix has limitations for surface vegetation with high spatial heterogeneity

[0004] (2) Adaptive fusion method: In order to avoid solving the linear spectral mixture model and consider the spatial variability of pixel reflectance, Gao et al. proposed an adaptive remote sensing image space-time fusion method (Spatial and Temporal Adaptive Reflectance Fusion Model, STARFM ), ignoring the differences in the spectral response functions of different sensors, assuming that the land cover type and system error do not change with time, and fusing the MODIS spectrum into the Landsat image, the simulation effect of the STARFM method on spatially heterogeneous regions needs to be improved

The above data fusion method effectively improves the spatio-temporal resolution of the image, but there are also limitations, and the input requirements for data fusion are relatively high: MODIS-Landsat image pairs with better image quality and closer to the target time are required, but due to the Influenced by bad data or invalid values ​​such as clouds / cloud shadows, these images are difficult to use for fusion, and it is difficult to achieve automatic and efficient fusion for long-term monitoring of surface vegetation in large areas

[0005] In summary, the existing remote sensing data fusion methods are not ideal in areas with complex surface vegetation on the one hand.

On the other hand, the fusion method has stricter requirements on the input data, which are all remote sensing data for clear sky and cloudless, and there is still a bottleneck for the cloud coverage area.

[0006] Therefore, how to provide a spatiotemporal fusion method to obtain NDVI data with high temporal resolution and high spatial resolution, and to solve the problem of difficult fusion of cloud / cloud shadow pollution areas, and reduce the quality of input NDVI data sources The dependence of existing technologies has become a technology that needs to be solved urgently

Images