Multispectral image radiation correction method based on absolute radiometric calibration

Abstract

The invention discloses a multispectral image radiation correction method based on absolute radiometric calibration, comprising the following steps: (1) conducting relative radiation correction; (2) conducting laboratory absolute radiometric calibration: constructing a laboratory radiance-image DN value look-up table of integrating spheres; (3) conducting absolute radiation correction based on the laboratory absolute calibration results: based on the laboratory radiance-image DN value look-up table, carrying out laboratory absolute radiation correction on the relative radiation correction images; (4) conducting ground calibration field on-orbit absolute radiation calibration: constructing a ground calibration field target on-orbit radiance- laboratory radiance look-up table; and (5) conducting absolute radiation correction based on the on-orbit absolute calibration results: based on the on-orbit radiance- laboratory radiance look-up table, carrying out on-orbit absolute radiation calibration on the laboratory absolute radiation correction images. According to the invention, the multispectral images are accurately converted into ground object real radiance from DN values, thus the colors of the multispectral images really reflect the ground object real spectrum characteristics, and the foundation for remote sensing quantification application is laid.

Description

technical field [0001] The invention relates to a multispectral remote sensing image radiation correction method based on absolute radiation calibration, which belongs to the technical field of satellite image processing and is used for optical remote sensing satellite multispectral camera image radiation correction processing. Background technique [0002] Radiation correction of optical remote sensing images is one of the main contents of optical remote sensing satellite camera image processing. Its purpose is to reduce or eliminate various radiation errors of remote sensing images, improve the radiation accuracy of remote sensing images, and enhance the ability of remote sensing images to reflect the real physical characteristics of actual ground objects. , its content mainly includes two aspects: one is to eliminate the inconsistency of the radiation response between the pixels of the image itself, which is called relative radiation correction; the other is to convert the...

AI Technical Summary

Problems solved by technology

[0009] The limitation of the method of directly using the calibration coefficients calculated by laboratory calibration data is that the radiation response characteristics of the camera in the orbit operation stage after satellite launch may be different from that in the pre-launch development stage, resulting in laboratory radiation calibration. The results are often inconsistent with the radiation response state of the camera when the satellite is in orbit. At the same time, the laboratory radiation calibration results do not consider the influence of external environmental factors such as ground object characteristics, atmospheric conditions, and solar illumination conditions, resulting in laboratory radiation calibration results being used for It is difficult to guarantee the absolute radiation correction accuracy during the satellite's in-orbit operation

[0010] The limitation of the method of directly using the calibration coefficient calculated by the on-orbit calibration data of the ground calibration field is that due to the limitation of the characteristics of the camera device, the relationship between the DN value of the image and the radiance is often nonlinear, so a sufficient number of Only the sample points (corresponding to the standard radiance corresponding to different reflectivity targets under certain environmental conditions) can accurately describe this nonlinear relationship; and due to the limitations of the feasibility, practicability, and economy of on-orbit target layout, ground inspection The reflectivity level (corresponding to the standard radiance level) of the targets arranged in the school site is limited, and often cannot meet the requirements of accurately describing the nonlinear relationship between DN value and radiance. Therefore, the DN value of the image is directly converted into the absolute Radiation correction methods often have insufficient correction accuracy, which restricts the improvement of quantitative inversion accuracy

Images