KR20100029529A - Method for calibration of coms using desert and ocean - Google Patents

Abstract

PURPOSE: A method for the calibration of a communications marine meteorological satellite through desert and marine is provided to measure exact radiant luminance by correcting digital count value of the communications marine existing satellite with correction factor. CONSTITUTION: A method for the calibration of a communications marine meteorological satellite through desert and marine comprises following steps. A first radiant luminance is calculated by inputting bidirectional distribution parameter according to wavelength to a radiation transfer model(S160). A second radiant luminance is calculated by inputting an example angular of the object satellite to be corrected, a sun zenith angle, a wavelength region, and marine target data of a standard satellite to the radiation transfer model. Correction factor is calculated by regression-analyzing the first radiant luminance, the second radiant luminance, and a digital count which is the intensity of light.

Description

Method for Calibration of COMS using Desert and Ocean

The present invention relates to a method for correcting the communication ocean phase satellite through the desert and the ocean, and more specifically, the desert and ocean to correct the satellite to be corrected by changing the satellite observation value observed by the digital count value in the satellite body to a value having a physical meaning It relates to a method of correction of communication ocean phase satellite through.

The light that reaches the satellite sensor is observed as an integer digital count (DC) value, and to take advantage of the satellite observation value, a radiation calibration process that converts the radiation into a physically meaningful radiance is performed. need.

The ideal way to make accurate radiation corrections is to have a constant source of radiation correction in the satellite vehicle.

However, since communication, ocean and meteorological satellites (COMS) have no satellite internal calibration resources for the visible channel, the calibration of the visible channel must be performed with values outside the satellite for the correction of the visible channel during satellite operation. do.

Inter-satellite correction is used for the correction of this visible channel.

Inter-satellite correction is a method of calculating the correction coefficient by comparing the radiance of other satellites that coincide with the COMS in time and space.

However, if the wavelength range and width of the two satellite sensors are different, it should be corrected, and this relationship has a problem of being affected by the cloud, the surface and the atmospheric condition. In addition, there is a problem that the radiation luminance between two satellites may not clearly show a linear relationship when there is a navigation error.

An object of the present invention for solving the above problems is because the surface reflectivity is large, the impact of the aerosol is small, the desert target does not need to provide accurate aerosol values from other satellites, even if the surface reflectivity is homogeneous, even without the bidirectional distribution function data Calculate the radiance using the radiative transfer model through the ocean target, calculate the correction coefficient by regression analysis with the digital count value observed from the communication ocean phase satellite, and then calculate the digital count of the communication ocean satellite satellite observed in real time. It is to provide a method of calibrating the communication ocean topography through the desert and the ocean so that the corrected radiance can be measured by correcting the value with a correction coefficient.

In order to achieve the above object, a method of correcting a communication ocean phase satellite through the desert and the ocean of the present invention includes a viewing angle, a solar zenith angle and a wavelength range of a satellite to be corrected, and vertical temperature, pressure, humidity, Water vapor volume, ozone volume, aerosol optical thickness, and bidirectional distribution function parameters for wavelengths of 2.5 nanometers are input to the radiative transfer model to calculate the first radiance, the viewing angle of the satellite to be corrected, the solar zenith angle, and the wavelength range and the reference satellite The first radiation calculated by calculating the radiative luminance and calculating the second radiance by inputting the target data, the total precipitation amount, the ozone amount, the aerosol optical thickness, the wind speed, the wind direction, and the dye concentration of the seawater into the radiative transfer model. Regression of the digital count which is the intensity of light obtained by observing the luminance and the second radiated luminance and the satellite and ocean targets to be calibrated Seats to provide a correction method of the COMS through the desert and ocean to a correction coefficient calculating step of obtaining a correction coefficient.

를 모디스 밴드 반응함수를 이용하여 모디스의 7개의 밴드와 41개의 지표면 형태에 따라 다른 값을 가지는 알베도

로 변환하고, 모디스에서 제공하는 7개의 밴드에서의 양방향성 분포함수 파라미터

를 이용하여 밴드, 시간, 표적 위치에 따라 다른 값을 갖는 태양의 고도가 10°일 때의 블랙 스카이 알베도

를 산출하는 제1단계, 제1단계에서의

와

를

와 같이 회귀분석하여 41개의 지표면 형태 중에서 상관관계가 가장 높을 때의 지표면 형태를 선정하고, 이때의 기울기 m과 절편 b를 이용하여 임의의 파장 λ2에서

를

에 대입하여 임의의 파장 λ2에서의 지표면 알베도

를 구하는 제2단계 및 제2단계에서 산출된

를

(여기서, α_{λ 1}와 α_λ3은 λ2에 가까운 두 개의 모디스 밴드의 각 중심파장인 λ1 및 λ 3에서의 각 알베도, α_λ2는 임의의 파장 λ2에서의 지표면 알베도, f_λ1과 f_λ3은 모디스에서 제공하는 λ1 및 λ3에서의 양방향성 분포함수 파라미터, f_λ2는 임의의 파장 λ2에서의 양방향성 분포함수 파라미터)에 대입하여 임의의 파장 λ2에서의 양방향성 분포함수 파라미터가 산출되는 제3단계를 포함할 수 있다.The bidirectional distribution function parameter for each wavelength at 2.5 nanometer intervals is provided by the auxiliary spectral library, a black sky albedo with a 10 ° altitude of the sun with different values depending on the wavelength and 41 surface geometries.

Albedo with different values according to 7 bands and 41 surface types of Modis using Modis band response function

Bidirectional distribution function parameter in 7 bands

Black Sky Albedo at 10 ° altitude with different values depending on band, time and target location

In the first step, the first step of calculating

Wow

To

Regression analysis is used to select the surface morphology at the highest correlation among 41 surface morphologies, and use the slope m and the intercept b at any wavelength λ2.

To

Surface albedo at any wavelength λ 2 by substituting for

Calculated in the second and second steps

To

(Wherein, α _{λ 1} and α _λ3 are each albedo at each center wavelength of λ1 and λ 3 of the two modiseu band close to λ2, α _λ2 is the ground surface at any wavelength λ2 albedo, f _λ1 and f _λ3 is modiseu The bidirectional distribution function parameter at λ1 and λ3 provided by, and f _λ2 may be substituted into the bidirectional distribution function parameter at any wavelength λ2) to include a third step of calculating the bidirectional distribution function parameter at any wavelength λ2. have.

In the calculating of the correction coefficient, DC (t), which is the digital count value of observing the desert target and the marine target according to time, and L _c , which is the first radiation brightness and the second radiation brightness, are obtained.

(여기서, DC₀는 보정대상위성이 우주공간을 바라보았을 때의 복사 휘도)에 의해 회귀분석하여 보정계수 c_g를 산출할 수 있다.

(Where DC ₀ is the radiant luminance when the satellite to be viewed in space) can be calculated by regression to calculate the correction factor c _g .

As described above, according to the present invention, since the surface reflectivity is large, the desert target does not need to be provided from other satellites due to the low aerosol influence, and the marine target does not need the bidirectional distribution function because the surface reflectivity is homogeneous. Calculate the radiance using the radiative propagation model, calculate the correction coefficient by regression analysis with the digital count value observed from the communication ocean phase satellite, and then correct the digital count value of the communication ocean satellite satellite observed in real time. It is possible to provide a method for correcting the communication ocean topography through the desert and the ocean so that the correct radiance can be measured by correcting the

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily carry out the present invention.

The desert target in the present invention has the advantage that the surface reflectivity is large, so the influence of the aerosol is small, it is not necessary to provide accurate aerosol values from other satellites, marine targets do not need the data of the bidirectional distribution function because the surface reflectivity is homogeneous There is this. Through these desert targets and marine targets, radiative luminance is calculated in the radiative transfer model, and the digital count values and regression analysis observed from the communication ocean phase satellite can be used to calculate the correction coefficient required for actual satellite operation.

In order to calibrate the communication ocean topography through the desert and the ocean according to an embodiment of the present invention, the viewing angle, solar zenith angle, wavelength range, vertical temperature, pressure, humidity, water vapor amount, Input the ozone amount, aerosol optical thickness, and bidirectional distribution function parameter for each wavelength of 2.5 nanometers into the radiative transfer model to calculate the first radiance, and the marine target of view angle, solar zenith angle, wavelength range, and reference satellite of the calibration target satellite. The second radiation intensity is calculated by inputting the total precipitation amount, ozone amount, aerosol optical thickness, wind speed, wind direction, and dye concentration of seawater as data.

Then, the correction coefficient is obtained by regression analysis of the digital count, which is the intensity of light, obtained by observing the first target luminance, the second radiation luminance, the desert target, and the marine target by the calibration target satellite.

For reference, the correction coefficient can be obtained in three ways, the first correction coefficient using only the desert target, the second correction coefficient using only the marine target, and the third correction coefficient using both the desert target and the marine target. will be.

Thus, the third correction coefficient is used to correct the communication ocean phase satellite, which is the correction target satellite.

In order to calculate the first correction factor, for the area that was identified as 100% Confidence Clear in the scene analysis of the COMS Meteorological Data Processing System (CMDPS) through the desert marine target 6S (Second Simulation of the Satellite Signal in the Solar Spectrum) (Vermote et al, 1997) shows the angle of view, solar zenith angle, wavelength range, and atmospheric and surface information of the satellite to be calibrated. Enter to calculate the first radiance. The first correction coefficient may be obtained by regression analysis of the digital count value obtained by observing the desert target in the first radiation luminance and the communication ocean phase satellite (COMS).

In order to calculate the second correction factor, the view angle, solar zenith angle, wavelength range, total precipitable water, ozone amount, aerosol, and aerosol velocity of the satellite to be calibrated with respect to the marine target are calculated. Wind Speed, Wind Direction, and Pigment Concentration information of the seawater are input to the radiation transfer model 6S to calculate the second radiation intensity.

In addition, the second correction coefficient may be obtained by regression analysis of the digital count value obtained by observing the marine target in the second radiation luminance and the communication ocean phase satellite.

Here, the wavelength range refers to the wavelength range of the sensor installed in the COMS.

1 is a flowchart illustrating a process of calculating a correction factor using a desert target according to an embodiment of the present invention, FIG. 2 is an exemplary diagram showing a first condition for selecting a desert target, and FIG. 3 is a desert target selected. 4 is an exemplary view showing a third condition for selecting a desert target, and FIG. 5 is an exemplary view showing a state in which a desert target is selected.

In order to calculate the first correction coefficient using the desert target, first, a desert target must be selected (step S110), and three conditions must be satisfied to select a desert target.

As a first condition, the surface reflectivity must be large. Since the error due to the surface reflectivity is due to the relative change in the surface reflectivity, there is less error in the bright surface having the large surface reflectivity.

를 사용하였다.

는 양방향성 분포함수(Bidirectional Reflectance Distribution Function; BRDF)를 위성 청전각, 태양 천정각 및 상대 방위각에 대해서 적분한 값으로 직달 및 산란광이 지표면에 도달하여 반사되는 알베도(Albedo)를 의미하며 수학식 1과 같이 결정된다.Bright sky selection includes white sky albedo

Was used.

Is an integrated value of the Bidirectional Reflectance Distribution Function (BRDF) for the satellite hearing angle, solar zenith angle, and relative azimuth angle, which means Albedo reflected by direct and scattered light reaching the surface of the earth. Is determined.

Here, f is a coefficient calculated according to the wavelength in the modis, iso (Isotropic) is isotropic, vol (Volumetric) is the dimensional effect related to the volume of the plant, geo (Geometric) is the plant group This is the case when the steric effect related to shadow is affected.

를 평균하여 도 2에 나타내었다.Thus, two years of 2005 and 2006 in three bands of Band 1, Band 2 and Band 4

2 is shown as an average.

As a second condition, the change in the surface reflectivity with time should be small. Because the surface BRDF coefficient is input every 8 days, if the surface reflectivity changes with time, it causes errors.

의 시간에 대한 표준 편차를 시간 평균 나눈 값을 사용하며, 이를 도 3에 나타내었다.For target selection with less change in reflectivity over time,

The standard deviation over time is divided by the time average, which is shown in FIG. 3.

As a third condition, the surface reflectivity must be spatially uniform. If the surface reflectivity is spatially non-uniform, the error of the positional information of the satellite pixel data degrades the accuracy of the correction.

Target selection with less spatial variation is shown in FIG. 4 using the method of Mitchell's method, which is obtained by dividing the spatial standard deviation by the spatial mean in a 20 km × 20 km space.

By setting the criteria for these three conditions, as shown in Fig. 5, a correction target can be selected. Here, the yellow target (Y) gave the strictest criteria, the green target (G) applied the loose criteria, and the blue target (B) applied the looser criteria.

Therefore, eleven targets (A, B, C, D, E, F, G, H, I, J, K) of the regions that meet the most stringent criteria were applied to calculate the first correction factor.

Therefore, as shown in FIG. 5, 11 targets are selected as desert targets for calculating the first correction coefficient, and blue pixels for the desert targets must be selected (step S120).

Chuncheon pixels utilize only 100% clear area data selected from CMDPS scene analysis. That is, the digital count value and the albedo value are averaged by only collecting 100% clear areas among pixels within 0.05 ° with respect to a given correction target.

Subsequently, the surface information calculates the surface surface BRDF for each wavelength of 2.5 nm based on the BRDF provided by the Moderate Resolution Imaging Spectroradiometer (MODIS), which is a weather sensor mounted on a terra satellite, which is the reference satellite in the present invention. (Step S130).

6 is a flowchart showing a process of calculating indicator information.

The BRDF provided by Modis is the value in seven bands. However, in order to calculate the radiative luminance of the satellite altitude in the radiation transfer model, the wavelength-specific BRDF values at 2.5 nm intervals are required within seven bands.

For this purpose, we used wavelength-specific albedo for 41 surface geometries from the ASTER Spectral Library.

For reference, the ester spectrum library used in the present invention used data provided on the Internet website 'http://speclib.jpl.nasa.gov/'. Thus, the Internet website 'http://speclib.jpl.nasa.gov/scripts/lib/asp/SoilResp.asp' has wavelength-specific data on various surface types. Of the forms, 41 surface geometries with albedo wavelengths of 0.4 to 14.0112 micrometers were used.

First, the wavelength-specific albedo provided by the ester spectral library and the BRDF provided by Modis must be converted into the same physical quantity.

는 태양의 고도가 10°일 때의 블랙 스카이 알베도(Black Sky Albedo)이며, 각각의 파장(λ)과 지표면 형태(type)에 따라 다른 값을 가진다. 그리고, 수학식 2에서와 같이, 모디스 밴드 반응함수(Φ)를 이용하여 밴드(band)와 지표면 형태(type)에 따라 다른 값을 가지는 알베도

로 변환한다(단계 S131).Wavelength-specific albedo provided by the ester spectral library

Is the Black Sky Albedo at an altitude of 10 ° and has different values for each wavelength (λ) and surface type. And, as in Equation 2, using the modis band response function Φ Albedo having a different value depending on the band (band) and the surface type (type)

To step S131.

That is, Equation 2 is calculated for each of the seven bands of the modis, and the denominator is the sum of the modis band response functions from the first wavelength to the end wavelength in the wavelength range of any one of the seven bands. In this case, the molecule is obtained by adding the albedo provided by the ester spectral library of the wavelength value to the modis band response function of the wavelength value from the first wavelength to the end wavelength in the wavelength range of the band.

는 파장과 41개의 지표면 형태에 따라 다른 값을 가지는

를, 모디스의 7개 밴드 중 그 파장이 속하는 밴드와 41개의 지표면 형태에 따라 값이 달라지는 알베도로 변환시켜 평균을 산출한 것이다.therefore,

Has different values depending on the wavelength and 41 surface types.

The average of the seven bands of Modis is converted to an albedo whose value varies according to the band to which the wavelength belongs and the shape of the 41 surface surfaces.

를 태양 고도가 10°일 때의 블랙 스카이 알베도인

로 산출하여,

와

를 동일한 물리량이 되도록 한다. 모디스에서 제공하는 7개의 밴드에서의 BRDF를 이용하여 블랙 스카이 알베도로 산출하는 것은 당업자에게 자명하다(단계 S132).Then, BRDF in the seven bands provided by Modis

Black Sky Albedoin at an altitude of 10 °

Calculated as

Wow

Let be the same physical quantity. It is obvious to those skilled in the art to calculate the black sky albedo using BRDF in seven bands provided by Modis (step S132).

Here, BRDF, that is, f is a coefficient calculated and provided according to a band in a modis, and is provided in a 0.05 degree grid at intervals of 16 days for seven bands provided in a modis, and a band, time, and target are provided. The value varies depending on the target.

또한 밴드(Band), 시간(Time), 표적 위치(Target)에 따라 다른 값을 가지게 된다.therefore,

In addition, it has different values depending on the band, time, and target position.

In addition, the target position may be represented by latitude and longitude values as a position on earth with respect to the desert target and the marine target selected for performing the calibration.

와

를, 수학식 3에서와 같이, 회귀분석하여 41개의 지표면 형태 중에서 상관관계가 가장 높을 때의 지표면 형태(type)를 선정한다. 즉, 41개의 지표면 각각에 대하여 회귀분석을 실시하여 상관관계가 가장 높을 때의 지표면 형태를 선정하는 것이다(단계 S133).next,

Wow

As shown in Equation 3, regression analysis selects the surface type at the highest correlation among 41 surface types. In other words, a regression analysis is performed on each of the 41 surface surfaces to select the surface form at the highest correlation (step S133).

에 사용하여

와 일치하는 파장별 지표면 알베도

를 산출하게 된다(단계 S134).And, as in Equation 4, the slope m and the intercept value b at this time

Using to

Surface-specific albedo

Is calculated (step S134).

는 밴드, 지표면 형태(type)에 따라 값이 달라지는 것이고,

는 밴드, 표적 위치(target), 시간에 따라 값이 달라지므로, m과 b는 지표면 형태, 표적 위치, 시간에 따라 각각 다른 값이 구해질 수 있다.here,

Is the value varies depending on the band and surface type,

Since values vary depending on the band, target location, and time, m and b may be different values depending on the surface type, target location, and time.

However, since the surface type at the highest correlation is selected, m and b have different values depending on the target position and time.

를 가져와, 수학식 4에 대입하여 임의의 파장 λ2에서의 지표면 알베도

를 산출할 수 있는데, 지표면 알베도의 파장별 변화양상과 양방향성 분포함수 파라미터의 파장별 변화양상이 동일하므로,

(여기서, α_{λ 1}와 α_λ3은 λ2에 가까운 두 개의 모디스 밴드의 각 중심파장인 λ1 및 λ3에서의 각 알베도, α_λ2는 임의의 파장 λ2에서의 지표면 알베도, f_λ1과 f_λ3은 모디스에서 제공하는 λ1 및 λ3에서의 양방향성 분포함수 파라미터, f_λ2는 임의의 파장 λ2에서의 양방향성 분포함수 파라미터)와 같이 나타낼 수 있다.Thus, albedo at any wavelength λ 2 in the ester spectral library

, And substitute in Equation 4 to provide the surface albedo at an arbitrary wavelength λ2.

Since the change pattern for each wavelength of the surface albedo and the change pattern for each wavelength of the bidirectional distribution function parameter are the same,

(Where α _{λ 1} and α _{λ 3} are the respective albedo at _{λ 1} and _λ 3, the central wavelengths of the two modis bands close to _{λ 2} , α _{λ 2} is the surface albedo at any wavelength λ 2, and f λ ₁ and f λ ₃ are the modis The bidirectional distribution function parameter at λ1 and λ3 provided, f _λ2 can be expressed as a bidirectional distribution function parameter at any wavelength λ2).

를 아래의 수학식 5에 대입하면, α_λ1와 α_λ3, f_λ1과 f_λ3 는 모디스에서 제공받는 자료로서 정해진 값이므로, 임의의 파장 λ2에서의 BRDF

를 산출할 수 있게 된다(단계 S135).Calculated as above

Is replaced by Equation 5 below, α λ ₁ and α λ _3, and f λ ₁ and f λ ₃ are values determined as data provided by Modis, and thus BRDF at arbitrary wavelength λ 2.

Can be calculated (step S135).

When the indicator information is calculated in this way, the atmospheric information should be calculated (step S140).

Atmospheric information is based on the vertical temperature, pressure, and humidity of the US62 Atmospheric Model, the total water vapor at the National Centers for Environmental Prediction (NCEP), and the total ozone volume at the Organization Meteorologique Internationale (OMI). Use

For reference, the atmospheric model refers to vertical temperature, pressure, and humidity data, and the atmospheric model includes mid-latitude winter, summer, tropical, extreme summer, and winter, and the US62 atmospheric model refers to the US standard atmosphere in 1962 (McClatchey). et al. 1972: McClatchey, RA, RW Fenn, JEA Selby, FE Volz, and JS Garing, 1972: Optical properties of the atmosphere.3rd ed.AFCRL Environ.Res.Papers No. 411, 108 pp.).

In the case of aerosols, continental aerosols in the radiation transfer model are used, and the aerosol optical thickness (AOT) is fixed at 0.1.

The fixed value of aerosol information is that when the surface reflectivity is large, the effect of the aerosol is small, and when the observation results of the Aerosol Robotic Network (AERONET) near the calibration target are monitored for a long time, This is because the value is about 0.1 except when the value suddenly increases.

In addition, large AOT values will not turn out to be 100% clear in CMDPS scene analysis.

In this way, when the indicator information and the atmospheric information about the desert target is obtained, the viewing angle, the solar zenith angle, the wavelength range, the indicator information and the atmospheric information of the satellite to be corrected are input to the radiation transfer model (step S150), and the radiation from the desert target is input. The first copy luminance, which is the luminance, is calculated (step S160).

For reference, it is apparent to those skilled in the art that when the above-described input data is inputted to the 6S program, which is a copy transfer model, the radiated luminance is calculated as a result.

Then, the digital count value and the first copy luminance observed in the COMS at regular time intervals are regressively analyzed (step S170) to calculate a first correction coefficient (step S180).

The first correction coefficient may be calculated by Equation 7 to be described later, where L _c is the first radiation luminance, DC (t) is a digital count obtained by observing a desert target, and DC ₀ is the difference between the correction bands. The radiative luminance when looking at outer space, c _g, is the first correction factor.

Once the first correction factor is calculated, the accuracy of the COMS can be verified through Modis and SeaStar satellite-mounted SeaWiFS (Sea-Viewing Wide Field-of-View Sensor) satellites. (Step S190).

7 is an exemplary diagram illustrating a state in which radiation luminance calculated using a desert target and radiation luminance observed from a reference satellite are verified.

As shown in FIG. 7, the 2005 data were used, and the ratios of the radiated luminance calculated using the desert target and the radiated luminance observed from the reference satellite were averaged for 16 days, and most of the results were − It can be said that it is within the 5% to + 5% margin of error.

In order to calculate the second correction coefficient, the viewing angle, solar zenith angle, wavelength range, total precipitation amount, ozone amount, aerosol, wind speed, wind direction, and dye concentration information of the seawater for the marine target are inputted to the radiative transfer model. The second correction coefficient may be obtained by calculating the second radiation luminance and performing a regression analysis on the second radiation luminance and the digital count value obtained by observing the marine target in the COMS.

8 is a flowchart illustrating a process of calculating a correction coefficient using a marine target according to an embodiment of the present invention.

Item Parameter Source Atmospheric Model Ozone OMI Total Precipitable Water NCEP / NCAR Reanalysis Aerosol Model AOT at 550nm MODIS (TERRA) Ocean model Wind speed NCEP / NCAR Reanalysis Wind direction NCEP / NCAR Reanalysis Pigment Concentration SeaWiFS monthly climatology

Table 1 shows the inputs required to calculate the radiative transfer model for marine targets.

As shown in Table 1, input data from COMS and MODIS, except SeaWiFS, National Centers for Environmental Prediction / National Center for Atmospheric Research (NCEP / NCAR) Reanalysis, and OMI, can be used for COMS observations to reduce the effects of navigation errors and cloud pollution. It consists of 0.2 ° × 0.2 ° grid data for the marine region in the range.

In addition, it is important to match the observed pixels of Tera and COMS to obtain the optical thickness (AOT) information of the aerosol and cloud information for the cloud elimination process, which are important factors in the radiation transfer calculation.

Therefore, only the pixels whose observation time difference between the two satellites is -10 minutes to +10 minutes are selected and used. In order to use the pixels observed by COMS for the correction of the visible channel using the ocean, the pixels may cause an error in the calculation process. Therefore, the marine target is selected by selecting pure marine pixels using the Land / Sea Mask (step S210).

In addition, marine targets have an optically clean atmosphere, unaffected by aerosols, and areas with clear sea water that are low in phytoplankton or other suspended matter are suitable.

In addition, it is feasible that the area is less likely to be contaminated by the cloud by being less affected by the cloud, and the area where the input data can be easily secured (step S221). Then, only the pixels satisfying the condition that the wind speed does not exceed 7 m / s and whose aerosol optical thickness does not exceed 0.1 are used (step S230).

Subsequently, the auxiliary data data shown in Table 1 is calculated (step S240).

The ancillary data simulates atmospheric conditions by assuming a US62 atmospheric profile and selecting total precipitation from NCEP / NCAR reanalysis, ozone from OMI, and a marine aerosol model from Shettle and Fenn (1979). In order to calculate the surface reflectivity, the ocean model of Cox and Munk (1954) and the observation geometry (COMeometry) information of COMS are input to the radiation transfer model (step S250), and the second radiation brightness is calculated (step S260).

For reference, it is apparent to those skilled in the art that when the above-described input data is input to the 6S program, which is a copy transfer model, the radiance luminance is calculated as a result value.

Then, a second correction coefficient is calculated (step S280) through a regression analysis between the digital count value observed in the COMS and the second radiation intensity (step S270).

The second correction coefficient may be calculated by Equation 7 to be described later, where L _c is the second radiant luminance, DC (t) is a digital count obtained by observing an oceanic target, and DC ₀ is the universe to be calibrated. The radiance luminance, c _g , when the space is viewed, is the second correction coefficient.

Once the second correction factor is calculated, the accuracy of the COMS can be verified through Modis and SeaStar satellite-mounted SeaWiFS (Sea-Viewing Wide Field-of-View Sensors) mounted on the reference satellites. (Step S290).

FIG. 9 is an exemplary view illustrating a state in which radiation luminances calculated using an ocean target is compared with radiation luminances observed from a reference satellite, and FIG. 10 is a graph of radiation luminances calculated using an ocean target and observed from a reference satellite. FIG. 11 is an exemplary diagram illustrating a state in which radiant luminance calculated using a marine target is compared with radiant luminance observed from a reference satellite at 5-day intervals.

9 to 11 compare the radiative luminance observed through the modis for the marine target for one year of 2005 with the radiated luminance calculated through the radiative transfer model, and FIG. 9 shows the atmospheric top calculated by the radiative transfer model. The radiative luminance, the radiant luminance observed by the modis, and the difference between these two values are shown.

It is necessary to clearly check whether the calculated value matches the observed value, and the relative difference is expressed in the same manner as in Equation 6 to know how much the calculated value is compared with the observed value.

Here, Observation refers to the radiant luminance observed from the verified reference satellite, such as Modis, and Calculation is the radiant luminance calculated from the radiative transfer model.

Error, which is the difference between the observed radiance and the calculated radiance, shows a difference in values at most within the range of -5% to + 5% over the entire period, as shown in FIG. There were days with a difference of more than 10%, but this is a fairly small period of one year.

FIG. 11 shows the ratio of the calculated radiant luminance to the observed radiant luminance in units of 5 days, and it can be seen that most of them have a value between 0.95 and 1.05 over the entire period, which is the calculated radiant luminance and the observed radiant luminance. It can be seen that the relative difference of is present at a level within -5% to + 5%.

12 is another exemplary view showing a state in which radiation luminance calculated using an ocean target according to an embodiment of the present invention is compared with radiation luminance observed from a reference satellite, and FIG. 13 is radiation luminance calculated using an ocean target. And another example illustrating a state in which radiative luminance observed from and a reference satellite is compared every five days.

Considering that AOT is the most influential variable in the radiance calculation, if you simulate the radiance of the modis again using the AOT provided by Modis, the applicability of the same algorithm to other satellites is not significant. I can't be sure.

Therefore, data processing using SeaWiFS, a well-proven marine sensor other than Modis, is also used only when AOT is less than 0.1 as in Modis data, minimizing the effect of error from aerosol uncertainty and space for the entire target. The error can be reduced by performing an average and a time average over five days.

The ratio of the calculated radiant luminance to the observed radiant luminance shows high agreement by showing values within 0.95 to 1.05 for most periods, as shown in FIGS. 12 and 13.

Subsequently, the digital correction value obtained by observing the desert target and the marine target in the first radiation luminance, the second radiation luminance, and the communication ocean phase satellite obtained as described above is obtained by regression analysis using Equation 7 to obtain a third correction coefficient. Become.

Where L _c is the first radiation brightness and the second radiation brightness, DC (t) is the digital count obtained by observing the desert target and the marine target, DC ₀ is the radiant brightness when the satellite to be calibrated looks at outer space, c _g is the third correction factor.

The third correction coefficient thus calculated is more accurate than the first and second correction coefficients, and thus is used to correct the COMS, the satellite to be corrected.

Therefore, as described above, after the third correction coefficient c _g is calculated, the digital count DC (t) obtained by observing the desert target and the marine target in real time in COMS is substituted into Equation 7 to accurately radiate luminance L _{c in} real time. Can be calculated.

14 is an exemplary view showing a regression analysis of radiant luminance calculated using a desert target and a marine target, and a digital count value of the desert target and the marine target, and FIG. 15 is calculated using the desert target and the marine target. Another exemplary view showing the regression analysis of radiative luminance and digital count values of desert and marine targets.

The correction coefficients of the multi-fuctional transport satelite (MTSAT 1R) were calculated from November 9, 2007 to November 30, 2007, and are shown in FIG. 14.

The correction coefficient calculated using only the desert target according to the present invention is 0.4597, which is about 8% larger than 0.4258, which is actually provided by MTSAT 1R, and shows a correlation of 0.969.

However, when the marine targets analyzed from November 15, 2007 to November 24, 2007 were used with desert targets, the correction factor was 0.4855, which is 14% larger than the actual value provided by MTSAT 1R, and the correlation was 0.993. It can be seen that.

Therefore, it can be seen that the correlation between the correction coefficients calculated by combining the desert target and the marine target shows linearity close to one.

FIG. 15 is a regression analysis using a digital count value observed in SeaWiFS, a desert target, and a marine target together. As shown in FIG. 15, since the correlation is 0.995 to 0.997, the linearity is close to 1. Can be.

Therefore, after analyzing the multipurpose satellite and SeaWiFS, the correlation between the correction coefficient calculated by combining the desert target and the marine target is closest to 1, so the third correction coefficient is higher than the first correction coefficient and the second correction coefficient. You can see that it is correct.

In the above description, the present invention has been described with reference to preferred embodiments, but the present invention is not necessarily limited thereto, and a person having ordinary skill in the art to which the present invention pertains does not depart from the technical spirit of the present invention. It will be readily appreciated that various substitutions, modifications and variations can be made.

1 is a flowchart illustrating a process of calculating a correction coefficient using a desert target according to an embodiment of the present invention;

2 is an exemplary view showing a first condition for selecting a desert target;

3 is an exemplary view showing a second condition for selecting a desert target;

4 is an exemplary view showing a third condition for selecting a desert target;

5 is an exemplary view showing a state in which a desert target is selected;

6 is a flowchart illustrating a process of calculating indicator information;

7 is an exemplary diagram illustrating a state in which radiation luminance calculated using a desert target is compared with radiation luminance observed from a reference satellite;

8 is a flowchart illustrating a process of calculating a correction coefficient using a marine target according to an embodiment of the present invention;

9 is an exemplary diagram illustrating a state in which radiation luminance calculated using a marine target is compared with radiation luminance observed from a reference satellite;

10 is an exemplary diagram illustrating a difference between radiative luminance calculated using a marine target and radiative luminance observed from a reference satellite.

 11 is an exemplary diagram illustrating a state in which radiant luminance calculated using a marine target is compared with radiant luminance observed from a reference satellite at 5-day intervals.

 12 is another exemplary view illustrating a state in which radiation luminance calculated using a marine target according to an embodiment of the present invention is compared with radiation luminance observed from a reference satellite;

FIG. 13 is another exemplified diagram illustrating a state in which radiant luminance calculated using an ocean target and a radiant luminance observed from a reference satellite are compared at 5-day intervals; FIG.

14 is an exemplary view showing a state in which regression analysis is performed on a radiant luminance calculated using a desert target and a marine target and a digital count value of the desert target and the marine target;

15 is another exemplary diagram showing a state in which a regression analysis of a radiant luminance calculated using a desert target and a marine target and a digital count value of the desert target and the marine target are observed;

Claims (3)

Radiation transfer model for viewing angle, solar zenith angle and wavelength range of calibration target satellite, vertical temperature, pressure, humidity, water vapor, ozone amount, aerosol optical thickness and wavelength at 2.5 nanometers Calculates the first radiance and transmits the total angle of precipitation, ozone amount, aerosol optical thickness, wind speed, wind direction, and pigment concentration of seawater, which are the target angles of viewing satellites, solar zenith angle, and wavelength range and reference satellites A radiant luminance calculating step of inputting a model to calculate a second radiated luminance; And Compensation coefficients are calculated by regression analysis of the first and second radiance luminances calculated through the radiative luminance calculation step, and the digital counts, which are light intensities obtained by observation of the target and ocean targets of the desert target and the marine target, by regression analysis. step Method of correction of communication maritime satellite through the desert and the ocean comprising a. The method of claim 1, wherein the bidirectional distribution function parameter for each wavelength at 2.5 nanometer intervals, Black Sky Albedo provided by the Ester Spectrum Library, a supplementary source, with an altitude of 10 ° for the Sun, with different values for wavelength and 41 surface geometries.

Albedo with different values according to 7 bands and 41 surface types of Modis using Modis band response function

Bidirectional distribution function parameter in 7 bands

Black Sky Albedo at 10 ° altitude with different values depending on band, time and target location

Calculating a first step; In the first stage

Wow

To

Regression analysis is used to select the surface morphology at the highest correlation among 41 surface morphologies, and use the slope m and the intercept b at any wavelength λ2.

To

Surface albedo at any wavelength λ 2 by substituting for

Obtaining a second step; And Calculated in the second step

To

(Wherein, α _λ1 and α _λ3 are each albedo at the λ1 and λ3 each center wavelength of the two modiseu band close to λ2, α _λ2 is the ground surface at any wavelength λ2 albedo, f _λ1 and f _λ3 are provided modiseu A third step of calculating a bidirectional distribution function parameter at an arbitrary wavelength λ2 by substituting the bidirectional distribution function parameter at λ1 and λ3, f _λ2 is a bidirectional distribution function parameter at an arbitrary wavelength λ2). Method of correction of communication ocean phase satellite through the desert and the ocean, characterized in that it comprises a. The method of claim 1, wherein the calculating of the correction coefficients, DC (t), the digital count value of observing the desert target and the marine target over time, and L _c , the first and second radiant luminance,

(Where DC ₀ is a radiative luminance when the satellite to be viewed in space) calculates a correction factor c _g by regression analysis.

Images