A Hyperspectral Thermal Infrared Emissivity Inversion Method Based on Correlation Wavelet Reconstruction

Abstract

The invention discloses a hyperspectral thermal infrared emissivity inversion method based on correlation wavelet reconstruction, which uses the emissivity normalization method to calculate the initial temperature T 0 ; with T 0 As the center, establish a temperature gradient look-up table, and calculate the initial emissivity curve ε for each temperature 0 , take ε 0 Filtered emissivity curve ε 1 and ε 0 The ε corresponding to the maximum linear correlation coefficient of 0 As a preliminary estimate of the emissivity curve ε 2 ; for ε 2 Perform multi-scale wavelet decomposition and reconstruction, calculate the linear correlation coefficient between each high-frequency reconstruction component and atmospheric down-going radiation, and add each high-frequency reconstruction component with different scale coefficients on the basis of the low-frequency reconstruction component to obtain the emissivity curve Inversion result ε. Based on the correlation of multi-scale wavelet reconstruction components, the invention reduces the estimation deviation of atmospheric radiation down-radiation and the inversion error caused by filtering in the separation process of hyperspectral thermal infrared temperature emissivity, and realizes the calculation of hyperspectral thermal infrared emissivity. Accurate inversion.

Description

(1) Technical field [0001] The invention relates to a hyperspectral thermal infrared emissivity inversion method based on correlation wavelet reconstruction, belongs to the technical field of hyperspectral data processing methods and applications, and is suitable for hyperspectral thermal infrared remote sensing inversion research. (2) Background technology [0002] The radiation energy of the thermal radiation of the ground object is determined by the emissivity of the ground object itself and the temperature of the ground object. How to invert the surface temperature and surface emissivity from the radiance information received by the sensor is the difficulty of thermal infrared remote sensing. [0003] The development of hyperspectral has created conditions for solving thermal infrared remote sensing inversion. There are many hyperspectral bands, which can describe fine spectral curves. The existing hyperspectral thermal infrared temperature and emissivity high-precisio...

AI Technical Summary

Problems solved by technology

The traditional temperature emissivity separation method based on wavelet transform believes that in the wavelet domain, high-frequency wavelet coefficients usually represent noise, while low-frequency wavelet coefficients contain most of the information in the original signal, so it is proposed to extract the surface emissivity from low-frequency wavelet coefficients, The emissivity spectral features contained in the high-frequency wavelet coefficients are lost

Images