A method for on-orbit radiometric calibration of the thermal infrared channel of a meteorological satellite

Abstract

This invention discloses an on-orbit radiometric calibration method for the thermal infrared channel of a meteorological satellite, comprising the following steps: selecting a suitable working time for the field calibration test based on satellite transit information and weather conditions at the calibration site; using an unmanned surface vessel (USV) carrying an infrared radiometer CE312, releasing the USV via a dock approximately two hours before satellite transit, and navigating on a pre-set route on the lake surface. This invention significantly saves manpower, material resources, and financial investment by utilizing the USV to automatically observe and acquire water surface radiation brightness temperature; by combining this with the acquisition of encrypted atmospheric profile data from radiosondes, the calibration frequency can be greatly increased, and calibration accuracy can be further improved; by comparing the entrance pupil brightness temperature information obtained from the thermal infrared channels of various satellites, the spatial distribution of brightness temperature during the synchronous test period at the calibration site and the degree of influence of cloud and weather conditions on the lake surface can be analyzed, providing significant reference for the selection of the synchronous testing area.

Description

Technical Field

[0001] This invention relates to the field of absolute radiometric calibration technology for infrared remote sensing sensors, and more specifically, to an on-orbit radiometric calibration method for the thermal infrared channel of a meteorological satellite. Background Technology

[0002] After half a century of effort, my country has successfully launched 20 meteorological satellites, with 9 currently in orbit. This has established a comprehensive Earth observation capability that combines imaging and detection, covering visible, infrared, and microwave spectral bands, making China one of the few countries in the world with both polar-orbiting and geostationary meteorological satellite series. The new generation of polar-orbiting and geostationary meteorological satellites—Fengyun-3 and Fengyun-4—has upgraded my country's meteorological satellite observation system, with some of its comprehensive observation capabilities reaching international leading levels. Both series of meteorological satellites have infrared channel Earth observation capabilities.

[0003] Absolute radiometric calibration of remote sensing sensors serves as a bridge between instrument counts and the actual surface parameters they reflect. It is the starting point for the quantification of remote sensing information and the foundation for quantitative inversion of surface biophysical parameters and the establishment of remote sensing models. After satellite launch, changes in the operating environment and status, as well as the aging of components over long-term operation, can alter the pre-launch calibration coefficients. Therefore, on-orbit field calibration is necessary to ensure the reliability and accuracy of remote sensing data applications. Regarding field radiometric calibration, the Dunhuang land calibration test site and the Qinghai Lake water surface calibration test site have been established, and field radiometric calibration has been conducted for my country's Fengyun series meteorological satellites, oceanographic satellites, resource satellites, environmental disaster reduction satellites, Gaofen series satellites, and military satellites, respectively. In terms of cross-calibration, currently, foreign satellite sensors are mainly used to cross-calibrate my country's meteorological and resource satellite sensors, achieving good results.

[0004] Real-time on-orbit absolute radiometric calibration of infrared remote sensing sensors has not yet been fully resolved both domestically and internationally, and it directly affects the quantitative application level of satellite infrared remote sensing data. The development of on-board calibration systems for my country's remote sensing satellites is relatively lagging, especially in the development of on-board calibration devices for infrared channels. Due to limitations in satellite platform and onboard instrument design, most currently operational infrared payloads still use post-insertion optical paths for blackbody calibration, failing to meet the requirements of absolute radiometric calibration. Furthermore, national standards and accuracy testing for on-board blackbodies are incomplete, resulting in a vague overall accuracy evaluation of infrared calibration. In addition, site radiometric calibration is limited by the number of field visits and weather conditions, confined to the radiometric calibration of only a portion of remote sensors. Moreover, the calibration methods used are limited, and the test site climate conditions are not ideal, failing to meet the needs of practical applications. Furthermore, insufficient investment in experimental remote sensing resources, incomplete types of experimental sites, weak basic research, and inadequate interdisciplinary collaboration have hindered the theoretical and technological innovation and the development of fully quantitative systems. In cross-calibration, the method itself has strict requirements regarding channel settings, channel spectral response functions, spatial resolution, transit time, revisit period, and geometric registration accuracy between the reference and calibrated sensors, resulting in a high implementation threshold. Furthermore, calibration accuracy heavily depends on the absolute radiometric calibration accuracy of the reference sensor itself. These factors directly impact satellite data preprocessing and standard product production; the lack of verification of the measured physical parameters reduces the reliability and application of data products, failing to meet all the requirements for providing models or algorithms for applications. Moreover, on-orbit and field radiometric absolute calibration requires significant investment of manpower, resources, and funding. Therefore, it is necessary to overcome problems such as imperfect on-board calibration, high workload and low frequency of field calibration (only 1-2 field tests per year), large workload of synchronous field observation tests, numerous observation parameters, complex data processing procedures, difficulty in controlling accuracy, and high implementation thresholds for cross-calibration.

[0005] There is currently no effective solution to the above problems. Summary of the Invention

[0006] In view of the above-mentioned technical problems in related technologies, the present invention proposes an on-orbit radiometric calibration method for the thermal infrared channel of a meteorological satellite, which can overcome the above-mentioned shortcomings of the prior art.

[0007] To achieve the above-mentioned technical objectives, the technical solution of the present invention is implemented as follows:

[0008] A method for on-orbit radiometric calibration of the thermal infrared channel of a meteorological satellite includes the following steps:

[0009] S1 selects the appropriate working time for the field calibration test based on satellite overpass information and weather conditions at the calibration site;

[0010] S2 uses an unmanned surface vessel (USV) equipped with an infrared radiometer CE312. About two hours before the satellite passes overhead, the USV is released from the dock and sails on the lake according to a pre-set route. It observes the lake surface radiance along the prescribed route to obtain continuous observation water surface brightness and temperature information, and ensures synchronization with as many satellites as possible during the synchronization period. After completion, it returns to the base according to the predetermined route.

[0011] S3 calibration data processing and analysis:

[0012] S31 Extract Satellite Entrance Brightness Temperature Information: Select the overpass synchronous image data of each satellite's thermal infrared payload or thermal infrared channel, and calculate the corresponding entrance brightness temperature information according to the radiometric calibration coefficient of each satellite. Compare the entrance brightness temperature information results obtained from the thermal infrared channels of each satellite, analyze the spatial distribution of brightness temperature during the synchronous test period at the calibration site and the degree of influence of clouds and weather conditions on the lake surface, and select a suitable synchronization area according to the actual situation.

[0013] S32 Extracts Surface Radiance Information: The spectral response functions of different loads are processed and matched. At the same time, the infrared radiometer CE312 is calibrated as a blackbody before and after the experiment. The measurement parameters are converted to obtain the spectral radiance at a specific wavelength.

[0014] S33 processes ground-based atmospheric observation data: it uses a solar photometer to invert and calculate aerosol optical thickness, which is obtained using the Langley method under a specific distribution to obtain the relationship between aerosol optical thickness and wavelength; it uses a radiosonde carried by an atmospheric sounding balloon to track the balloon's changing position and corresponding observation parameters to obtain the real-time profile of these parameters.

[0015] S34 Atmospheric Radiation Simulation: Atmospheric radiation simulation is performed using an atmospheric radiative transfer model to obtain the quantitative impact of the atmosphere on the upward radiation over the lake surface; based on the spectral response functions of the thermal infrared channels of different satellite infrared payloads, the transmittance information is convolved with the channel spectral response functions to obtain the transmittance and path radiation information of different channels, and the impact of the atmosphere on the thermal infrared radiation of the Earth's surface is quantified based on the transmittance and path radiation information.

[0016] S35 performs spectral matching between satellite sensors and the ground-based infrared radiometer CE312.

[0017] Furthermore, the spectral matching factor is calculated to reduce the difference in spectral response between ground-based observation instruments and satellite thermal infrared payloads.

[0018] Furthermore, during synchronization, a satellite observation zenith angle of less than 15 degrees was selected.

[0019] Furthermore, the standard deviation between ground observation pixels during synchronous observation was less than 10%.

[0020] Furthermore, in step S31, the spectral radiance observed by the satellite can be converted into a brightness temperature value according to Planck's formula:

[0021]

[0022] Where, the formula C1 = 1.191 will -12 W / (cm 2 · / (cm1 male) -1 ) 4 C2 = 1.439 K·cm is the first radiation constant; C2 = 1.439 K·cm is the second radiation constant; σ is the wavenumber (cm⁻¹), T is the temperature (K), and L(σ) is the spectral radiance (W / (cm²)). 2 ·sr·cm -1 )).

[0023] Furthermore, in step S31, the lake surface for calibration experiment is selected based on the cloud-affected area and the synchronization area. The selection criteria are to avoid the influence of clouds and cloud edges and to ensure that the lake surface area has uniform temperature changes. The synchronization area must be within the observation track range. If the track range is affected by clouds, a suitable nearby area is selected.

[0024] Further, in step S33, the aerosol optical thickness is calculated as follows: Using the measured aerosol optical thickness at two wavelengths, the following is calculated:

[0025]

[0026] Then, the optical thickness of the aerosol at the required wavelength is inverted using the measured α and β values.

[0027] Furthermore, in step S33, the corresponding observation parameters include temperature, humidity, air pressure, wind direction, and wind speed.

[0028] Furthermore, it also includes: S4 inputting observation data from unmanned surface vessels, synchronous image data from meteorological satellites passing over the area, and radiosonde profile data measured during the synchronization period into the MODTRAN model to simulate and obtain the apparent radiance of the top-of-atmosphere channel.

[0029] Furthermore, it also includes: S5 calculates the absolute radiometric calibration coefficient of the channel to be calibrated by linear regression based on the apparent radiance at the top of the atmosphere of the channel to be calibrated at the matching point obtained by simulation and the average value of the extracted meteorological satellite DN value.

[0030] The beneficial effects of this invention are as follows: By utilizing unmanned surface vessels to automatically observe and obtain water surface radiation brightness temperature, it replaces the traditional field measurement work of personnel collecting water surface radiation information by boat on the lake, which can greatly save manpower, material resources, and financial resources; by combining it with the acquisition of encrypted atmospheric profile data, the calibration frequency can be greatly increased, and the calibration accuracy can be further improved; by comparing the entrance pupil brightness temperature information obtained from the thermal infrared channels of various satellites, the spatial distribution of brightness temperature during the synchronous test period at the calibration site and the degree of influence of cloud and weather conditions on the lake surface can be analyzed, which has great reference value for the selection of the synchronous area. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 This is a flowchart of the radiometric calibration data processing and calibration calculation process for the on-orbit radiometric calibration method of the thermal infrared channel of a meteorological satellite according to an embodiment of the present invention;

[0033] Figure 2 This is a front view of the infrared radiometer unmanned surface vessel mounted on the meteorological satellite thermal infrared channel on-orbit radiometric calibration method according to an embodiment of the present invention.

[0034] Figure 3 This is a top view of the infrared radiometer unmanned surface vessel mounted on the meteorological satellite thermal infrared channel on-orbit radiometric calibration method according to an embodiment of the present invention.

[0035] Figure 4 The location of the Qinghai Lake calibration field is described in the on-orbit radiometric calibration method for the thermal infrared channel of a meteorological satellite according to an embodiment of the present invention.

[0036] Figure 5 This is the technical process for site radiometric calibration of the thermal infrared payload in the on-orbit radiometric calibration method for the thermal infrared channel of a meteorological satellite, as described in an embodiment of the present invention.

[0037] Figure 6 This is a schematic diagram of the synchronous test observation area and flight path planning design for the on-orbit radiometric calibration method of the thermal infrared channel of a meteorological satellite according to an embodiment of the present invention;

[0038] Figure 7 This is an image of the TERRA / AQUA transit observation field over Qinghai Lake, based on the on-orbit radiometric calibration method for the thermal infrared channel of a meteorological satellite according to an embodiment of the present invention.

[0039] Figure 8 This is the imaging status of the FY-3D transit observation field of Qinghai Lake according to the on-orbit radiometric calibration method of the meteorological satellite thermal infrared channel described in the embodiments of the present invention;

[0040] Figure 9 This is a spatial distribution map of the brightness temperature of channel 24 of FY-3D according to the on-orbit radiometric calibration method of the thermal infrared channel of the meteorological satellite according to an embodiment of the present invention (from left to right: August 18 and August 20).

[0041] Figure 10 The image shows the pixel brightness temperature distribution in the track synchronization area of ​​the on-orbit radiometric calibration method for the thermal infrared channel of a meteorological satellite according to an embodiment of the present invention (from left to right: August 18, August 19, and August 20).

[0042] Figure 11 The brightness temperature spatial distribution of the 4th channel of FY3C in the on-orbit radiometric calibration method of the thermal infrared channel of the meteorological satellite according to the embodiment of the present invention (distributed from left to right on August 18 and August 20);

[0043] Figure 12 The image shows the spatial distribution of brightness temperature of channel 12 of FY4A in the on-orbit radiometric calibration method for the thermal infrared channel of a meteorological satellite according to an embodiment of the present invention (from left to right: August 18 and August 20).

[0044] Figure 13 The spatial distribution of brightness temperature of TERRA / AQUA MODIS channel 31 is shown in the on-orbit radiometric calibration method for the thermal infrared channel of a meteorological satellite according to an embodiment of the present invention (from left to right: August 18 and August 20).

[0045] Figure 14 The lake surface brightness temperature observed by CE312 during the field observation of the on-orbit radiometric calibration method of the meteorological satellite thermal infrared channel according to the embodiment of the present invention;

[0046] Figure 15 This refers to the synchronous observation of the lake surface during satellite transit according to the on-orbit radiometric calibration method of the meteorological satellite thermal infrared channel described in the embodiments of the present invention.

[0047] Figure 16 This is an aerosol optical thickness map at 550 nm during synchronous observation of the on-orbit radiometric calibration method for the thermal infrared channel of a meteorological satellite according to an embodiment of the present invention.

[0048] Figure 17 These are atmospheric temperature and humidity profiles obtained from a weather balloon using the on-orbit radiometric calibration method for the thermal infrared channel of a meteorological satellite according to an embodiment of the present invention.

[0049] Figure 18This is an image showing the atmospheric transmittance of FY3D and MODIS on August 18, based on the on-orbit radiometric calibration method for the thermal infrared channel of a meteorological satellite according to an embodiment of the present invention.

[0050] Figure 19 This is an atmospheric path radiation diagram of FY3D and MODIS on August 18, based on the on-orbit radiometric calibration method for the thermal infrared channel of a meteorological satellite according to an embodiment of the present invention.

[0051] Figure 20 This is a comparison of different spectral response functions between the ground and satellite thermal infrared channels in the on-orbit radiometric calibration method for the thermal infrared channel of a meteorological satellite according to an embodiment of the present invention;

[0052] Figure 21 This invention relates to the ground-based observation of CE312 and FY3C satellite infrared channel spectral matching using the on-orbit radiometric calibration method for the thermal infrared channel of meteorological satellites according to embodiments of the present invention.

[0053] Figure 22 This invention relates to the ground-based observation of CE312 and FY3D satellite infrared channel spectral matching using the on-orbit radiometric calibration method for the thermal infrared channel of meteorological satellites according to an embodiment of the present invention.

[0054] Figure 23 This invention relates to the ground-based observation of CE312 and FY4A satellite infrared channel spectral matching using the on-orbit radiometric calibration method for the thermal infrared channel of meteorological satellites according to embodiments of the present invention.

[0055] In the picture: 1. Weather station, 2. Probe, 3. Telescopic pole, 4. Control box, 5. Industrial computer. Detailed Implementation

[0056] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the scope of protection of the present invention.

[0057] like Figure 1-3 As shown in Figure 5, an on-orbit radiometric calibration method for the thermal infrared channel of a meteorological satellite according to an embodiment of the present invention comprises two parts:

[0058] The first part is the acquisition of space-to-ground synchronized measurement data by the unmanned surface vessel (USV) surface observation system. The USV uses solar energy for propulsion, giving it long endurance. Equipped with meteorological and underwater observation instruments, it can acquire data on surface meteorological elements, water temperature, and salinity. The navigation mode supports automatic and manual modes, and can be switched between them at any time. In buoy mode, it automatically reaches the detection location without releasing the vessel or setting up anchors. It can continuously measure surface air temperature, air pressure, humidity, wind speed, wind direction, and sea temperature in sea state 5, and can survive in sea states 6 and above. The land-based subsystem monitors, receives, processes, and distributes data in real time.

[0059] The infrared radiometer is mounted on an unmanned surface vessel (USV) platform, with its probe extending approximately 2 meters out of the hull and observing vertically downwards at a height of about 1.5 meters above the lake surface. Based on the infrared radiometer's 25° field of view and the 1.5-meter height above the lake surface, the water surface radiance information within a 0.3474 m² area measured by the infrared radiometer during the observation can be calculated (e.g., Figure 2 and Figure 3 (As shown). The infrared radiometer measures water meters in multiple bands, with a temperature range of -80 to 50℃, a temperature resolution of 0.01℃, a response time of 0.1s, a field of view of 25°, an operating temperature of -20 to 50℃, and a data repeatability better than 99.65%.

[0060] Based on the meteorological satellite transit conditions, a satellite-to-ground synchronous observation experiment was designed. The unmanned surface vessel (USV) was debugged and radiometer equipment was installed at the shore-based test site. About two hours before the satellite transit, the USV was released from the dock and sailed on the lake surface in the satellite-to-ground synchronous observation area along a pre-set route, and the infrared radiance of the lake surface was observed. During the satellite-to-ground synchronous observation, as much lake surface observation data as possible was collected as possible under the condition that the radiometer performance allowed. About two hours after the satellite transit, the USV returned to the dock along the predetermined route and was recovered.

[0061] The second part is the data processing and calibration calculation of on-orbit site radiometric calibration data from the thermal infrared channel of a meteorological satellite based on synchronous observations from an unmanned surface vessel. Figure 1 This is a flowchart of radiation calibration data processing and calibration calculation. (For example...) Figure 1 As shown, the absolute radiometric calibration method proposed in this invention includes the following steps:

[0062] Step 1: Unmanned Surface Vessel Observation Data Reading and Processing

[0063] Data such as latitude and longitude, infrared radiation brightness temperature, water temperature, air temperature, air pressure, wind direction, wind speed, and observation time are read from the observation data of the unmanned surface vessel.

[0064] Step Two: Meteorological Satellite Data Reading and Processing

[0065] The infrared observation image data of the channel to be calibrated, the latitude and longitude of each pixel, the satellite observation zenith angle, azimuth angle, spectral response function of the channel to be calibrated, and the satellite observation time are read from meteorological satellite data.

[0066] Step 3: Matching unmanned surface vessel (USV) ground-to-space synchronous observation data with satellite data

[0067] Using the latitude, longitude, and observation time of the unmanned surface vessel (USV) satellite-ground synchronous observation data as a reference, spatial and temporal matching processing is performed on the meteorological satellite observation data. Matching points in cloudless areas are selected, and USV satellite-ground synchronous observation data and meteorological satellite observation data at these matching points are obtained. The average of the DN values ​​of the pixels in the surrounding 3×3 window is taken from the meteorological satellite data at the obtained matching points.

[0068] Step 4: Sounding data reading and processing

[0069] Read geopotential height, temperature profile, humidity profile, and pressure profile data, as well as latitude, longitude, and time, from the radiosonde data.

[0070] Step 5: Sounding data matching and processing

[0071] Based on the matching results of unmanned surface vessel space-ground synchronous observation data and meteorological satellite data, the latitude and longitude information of the matching point is obtained, and the radiosonde profile data corresponding to the time and space of the matching point is extracted from the radiosonde dataset according to the latitude and longitude.

[0072] Step Six: Radiative Transfer Simulation

[0073] The data obtained from the unmanned surface vessel's ground-to-space synchronous observations at each matching point, the satellite's zenith and azimuth angle data, the atmospheric temperature, humidity and pressure profile data, and the spectral response function and time information of the uncalibrated channel of the meteorological satellite remote sensing sensor are input into the radiative transfer calculation model to simulate and obtain the apparent radiance of the top-of-atmosphere channel at each matching point.

[0074] Step 7: Absolute Radiation Calibration Calculation

[0075] Based on the apparent radiance at the top of the atmosphere of the channel to be calibrated at the matching point obtained by simulation, the average DN value of the extracted 3×3 window pixels of meteorological satellite, and the DN value of cold air observation by meteorological satellite, the absolute radiometric calibration coefficient of the channel to be calibrated is obtained by linear regression calculation.

[0076] To facilitate understanding of the above technical solutions of the present invention, the following detailed description of the above technical solutions of the present invention will be provided through specific usage methods.

[0077] In practical application, the on-orbit radiometric calibration method for the thermal infrared channel of a meteorological satellite according to the present invention is illustrated with a specific example:

[0078] 1. Experimental observations

[0079] To assess the on-orbit operation and attenuation of the FY series meteorological satellites and to evaluate the rationality of their radiometric calibration coefficients, a synchronous ground-based observation operational test of the thermal infrared payload was conducted at the Qinghai Lake calibration site from August 14 to August 22, 2019. The synchronous test was organized and led by the National Satellite Meteorological Center of the China Meteorological Administration, with the participation of multiple institutions including the Chinese Academy of Sciences and the National Institute of Metrology. The objectives of this test were: 1) to complete the site calibration test of the domestically produced operational satellites FY-3C, FY-3D, and FY-4A and to evaluate their radiometric calibration status; 2) to evaluate the feasibility of using unmanned surface vessels (USVs) in field calibration tests by cross-referencing their field calibration with that of the Terra / Aqua-MODIS satellite thermal infrared payloads.

[0080] The radiometric calibration site for the thermal infrared payload was chosen to be the Qinghai Lake standard calibration site, one of the most commonly used calibration sites in China. Located in northeastern Qinghai Province, this site is a crucial part of China's remote sensing satellite radiometric correction field. Qinghai Lake is my country's largest inland saltwater lake, with an area of ​​approximately 4635 km², a length of about 106 km east to west, a width of about 63 km north to south, and an average depth of 19 m. With an average altitude of 3196 m, surrounded by mountains, Qinghai Lake receives 2981.2 hours of sunshine annually, has a relative humidity of 69%, and boasts a clean and dry atmosphere with high stability. The annual aerosol content is 0.1 K, indicating a low aerosol particle content. Spectrometer measurements revealed that Qinghai Lake exhibits high, uniform, and stable water cleanliness, with a uniform surface temperature distribution. The observed brightness temperature deviation between similar water surfaces was less than 0.1 K within 30-60 minutes, making it an ideal infrared radiation target source suitable for conducting site radiometric calibration assessments and alternative calibration work for domestic and international thermal infrared sensors. Figure 4 It is the Qinghai Lake thermal infrared calibration field.

[0081] 2. Field Calibration Work Plan

[0082] The basic workflow for site calibration is as follows: Figure 5 Based primarily on satellite transit information and weather conditions at the calibration site, the appropriate working time for the field calibration experiment is selected, and corresponding surface and atmospheric parameter observation equipment is deployed. Within half an hour before and after the satellite transit calibration field, surface radiance, atmospheric parameters, and other environmental parameters are simultaneously acquired. The equivalent values ​​at the satellite remote sensor's entrance pupil are calculated using an atmospheric radiative transfer model, and combined with the count values ​​simultaneously observed by the imager in the radiation calibration field area, the payload calibration coefficients are calculated and evaluated.

[0083] Site radiometric calibration requires precise satellite-ground synchronous observation experiments. To ensure the validity of calibration observation information during satellite transit, long-term continuous monitoring of the observed ground targets is necessary to analyze their long-wave radiation changes under various influencing factors and identify the essential conditions for stable observation for synchronous experiments. Qinghai Lake's water cleanliness and stability meet the experimental requirements. Further analysis of atmospheric effects allows for the assessment of satellite entrance pupil radiance and evaluation of onboard payloads. Site calibration synchronous observations require at least three valid experimental observations.

[0084] To ensure the accuracy of the site radiometric calibration of the thermal infrared load, the following points should be noted:

[0085] (1) Perform precise CE312 blackbody calibration before and after site work;

[0086] (2) Calculate the spectral matching factor to reduce the difference in spectral response between ground observation instruments and satellite thermal infrared payloads;

[0087] (3) High-precision atmospheric profile information is used for atmospheric radiative transfer mode simulation;

[0088] (4) During synchronization, select a smaller satellite observation zenith angle, less than 15 degrees;

[0089] (5) The standard deviation between ground observation pixels during synchronous observation is less than 10%.

[0090] 3. Application of unmanned surface vessels in field calibration tests

[0091] 3.1 Automatic Observation Conditions of Unmanned Surface Vessels

[0092] The unmanned surface vessel (USV) uses solar energy for propulsion, enabling long-duration operation. Equipped with meteorological and underwater observation instruments, it can acquire data on surface meteorological elements and water temperature and salinity. The navigation mode supports automatic and manual modes, which can be switched at any time. When operating as a buoy, it automatically reaches the detection location without needing to release the vessel or set up anchors. It can continuously measure sea surface temperature, air pressure, humidity, wind speed, wind direction, and sea temperature in sea state 5, and can survive in sea states 6 and above. The land-based subsystem monitors, receives, processes, and distributes data in real time.

[0093] Specific technical specifications:

[0094] (1) Hull dimensions: Length * Width * Height 4200 * 1900 * 1400 mm

[0095] (2) Activity radius: 2500km

[0096] (3) Wind resistance: Continuous operation in sea state 5 and survival in sea state 6 and above.

[0097] (4) Power supply method: a combination of solar and battery power supply

[0098] (5) Navigation: Satellite navigation

[0099] (6) Communications: Wi-Fi, cellular networks and satellite communications

[0100] The unmanned surface vessel (USV) observations were conducted using a manually monitored approach during the experiment. A reasonable flight path was set based on the satellite's transit time, the USV's range, and the lake surface conditions. The flight path had to take into account weather conditions and the actual operating conditions of the USV, with a slightly longer range set under clear skies while ensuring sufficient power and speed. Simultaneously, considering the observation characteristics of CE312, two flights were scheduled, one in the morning and one in the afternoon. The observation range of both flights met the spatial requirements for synchronous observation with the corresponding satellite.

[0101] 3.2 Design of Infrared Radiometer Observation Based on Unmanned Surface Vessel

[0102] To save manpower and material costs and conduct continuous and stable observations, this experiment used unmanned surface vessels (USVs) instead of large ships for automatic lake surface observations, and utilized the BeiDou satellite system to transmit observation data in real time. This field calibration experiment was the first to use an USV equipped with an infrared radiometer for synchronous satellite-to-ground observations, a method pioneered in China. The experiment required overcoming various technical challenges in unmanned control, including experimental positioning, trajectory planning, data transmission, and coordination between the USV and its onboard instruments. Therefore, in addition to meeting the requirements of satellite calibration, the experiment also conducted various types of tests and analyses on the observations, attempting to provide a new working method for future calibration experiments.

[0103] The CE312 is mounted on an unmanned surface vessel (USV) platform, with its probe extending approximately 2 meters out of the hull, observing vertically downwards at a height of about 1.5 meters above the lake surface. During the observation, the CE312 measured the lake surface at a depth of 0.3474 meters. 2 The CE312 provides surface radiance information within its range. It operates in four bands, with a temperature range of -80 to 50°C, a temperature resolution of 0.01°C, a response time of 0.1 s, a field of view of 25°, an operating temperature range of -20 to 50°C, and a data repeatability better than 99.65%.

[0104] The test was designed based on the satellite transit conditions. The unmanned surface vessel was debugged and the equipment was installed at the test site. About two hours before the satellite transit, the unmanned surface vessel was released from the dock and sailed on the lake according to the pre-set route. The lake surface radiance was observed according to the prescribed route, and as many satellites as possible were synchronized during the synchronization period. After completion, it returned to the base according to the predetermined route.

[0105] The experiment involved route planning based on the unmanned surface vessel's (USV) control range and endurance. Considering the limitations of the CE312 instrument's synchronous observation data storage capacity, a single flight was set at 3-4 hours, divided into morning and afternoon observation periods. The first synchronous test of the USV was conducted on August 16, 2019, at the Qinghai Lake fishery wharf. The test showed that the USV could continuously navigate under solar power; however, due to speed limitations, the return time had to be precisely set for the next release, so long-range and rapid positioning observations could not yet be guaranteed. Its advantages, however, include stable speed, minimal water surface impact, and the ability to acquire good continuous observation of water surface brightness and temperature information. For the synchronous test trajectory on August 18, 2019, the Qinghai Lake fishery administration wharf was selected as the USV release area, with a 30km range from the wharf as the flight control zone. Simultaneously, other ground-based observation instruments were deployed based on the wharf's observation site to ensure that synchronous observation parameters were located in the same area.

[0106] 4. Field conditions during synchronization

[0107] 4.1 Weather conditions during the observation period

[0108] The synchronous observation experiment period was from August 14 to August 22, 2019. On August 14 and 15, personnel and equipment participating in the synchronous observation experiment arrived at the observation site successively and carried out corresponding instrument debugging and preparatory work for the synchronous experiment. From August 16 to August 22, the synchronous experiment was conducted, including the deployment of ground equipment and observation equipment on the lake surface vessel, the installation of atmospheric sounding equipment, and real-time monitoring of satellite transit information and corresponding cloud image changes. Experiments were carried out according to each transit time.

[0109] The overall atmospheric conditions were good during the satellite's transit. The weather conditions during the test are listed in Table 1. From the 16th to the 18th, clear skies were guaranteed for most of the time. Cloud cover increased slightly on the 19th and 20th, and cloudy skies after the 21st were not suitable for conducting the test. Two actual conditions were fully considered when conducting the test at the observation field: (1) the atmospheric conditions in the Qinghai Lake area change rapidly, and the cloud cover over the lake changes significantly in a short period of time; (2) the unmanned surface vessel synchronous test does not require high manpower costs and can conduct continuous long-term automatic observations. Therefore, the test can be conducted in all weather conditions, and synchronous observation tests can be carried out according to the satellite transit time to ensure that as much synchronous observation data as possible is available at the ideal transit time to meet the needs of the calibration test. The actual work has also confirmed that this test arrangement is reasonable.

[0110] Table 1. Weather conditions during the synchronous test in the Qinghai Lake area.

[0111] date highest temperature minimum temperature weather Wind direction and wind force Air Quality Index Wednesday, August 14, 2019 25℃ 8℃ clear Northeast wind, level 1 32 Excellent Thursday, August 15, 2019 26℃ 9℃ clear Southeast wind, level 2 29 Excellent Friday, August 16, 2019 27℃ 10℃ clear Southeast wind, level 2 35 Excellent Saturday, August 17, 2019 26℃ 11℃ clear Southeast wind, level 2 34 Excellent Sunday, August 18, 2019 25℃ 13℃ clear Southeast wind, level 2 37 Excellent Monday, August 19, 2019 25℃ 14℃ clear Southeast wind, level 2 34 Excellent Tuesday, August 20, 2019 21℃ 9℃ Sunny to partly cloudy Southeast wind, level 2 28 Excellent Wednesday, August 21, 2019 21℃ 10℃ Sunny to partly cloudy Southeast wind, level 2 26 Excellent Thursday, August 22, 2019 23℃ 12℃ Overcast to light rain Southeast wind, level 2 32 Excellent

[0112] 4.2 Satellite Overpass Analysis

[0113] The synchronous observation experiment focuses on operational meteorological satellites FY-3C, FY-3D, and FY-4A, while also taking into account Terra / MODIS, Aqua / MODIS, and Sentinel-3A. This is to evaluate the operational status and calibration coefficient deviation of the Fengyun series satellites, and to conduct cross-comparisons with similar international satellites to ensure the rationality and accuracy of our work.

[0114] Using FY-3D and MODIS as examples, the satellite observation overpass conditions are shown respectively. Figure 7 The image shows the atmospheric conditions over Qinghai Lake as captured by MODIS from August 16th to August 20th. The image reveals that the best atmospheric conditions occurred between August 16th and 18th, while cloud cover was higher on August 19th and 20th. Comparing Terra and Aqua data, cloud cover over Qinghai Lake changed rapidly, with significant differences between morning and afternoon.

[0115] Unlike the Terra / Aqua platform's two daily revisits, FY-3D requires a longer revisit period, typically around 5.5 days. Therefore, based on the revisit period, we selected August 18th-20th for the synchronous test of the FY-3D satellite, and examined real-time changes in atmospheric conditions around Qinghai Lake using FY-4A satellite cloud imagery. The imaging conditions at the actual FY-3D revisit time (…) Figure 8 As can be seen, most of the lake surface was observed by satellite on August 18th and August 20th, which had little impact on our designated flight path area. However, on August 19th, the lake surface was largely covered by clouds, with only the northeastern region remaining unaffected. This area happened to be within the observation range, making it suitable for conducting synchronous experiments. Other satellite synchronous experiments were also designed and carried out using this scheme.

[0116] 5. Calibration data processing and analysis

[0117] 5.1 Extraction of satellite entrance pupil brightness and temperature information

[0118] To ensure the rationality of the calibration results, we simultaneously conducted calibration work on multiple satellites, focusing on FY-3C, FY-3D, FY-4A, and Terra / Aqua. The experiment used imaging data from the satellites' thermal infrared payloads or thermal infrared channels, and calculated the corresponding entrance pupil brightness temperature information based on the satellites' radiometric calibration coefficients for ground-based observation comparison. The radiometric and geometric calibration parameters for the FY series satellites were provided by the National Satellite Meteorological Center's operational department, while the infrared channel geometric and radiometric calibration parameters for the Terra / Aqua satellites were obtained from NASA's data documentation.

[0119] In the experiment, the spectral radiance observed by the satellite can be converted into a brightness temperature value according to Planck's formula:

[0120]

[0121] Where, the formula C1 = 1.191 will -12 W / (cm 2 · / (cm1 male) -1 ) 4 C2 = 1.439 K·cm is the first radiation constant; C2 = 1.439 K·cm is the second radiation constant; σ is the wavenumber (cm⁻¹), T is the temperature (K), and L(σ) is the spectral radiance (W / (cm²)). 2 ·sr·cm -1 By performing an inverse transformation on the above formula, the temperature information can be obtained.

[0122] Below, we compare the entrance pupil brightness temperature information obtained from the thermal infrared channels of each satellite, analyze the spatial distribution of brightness temperature during the synchronous test period in the Qinghai Lake area, and the degree of influence of cloud and weather conditions on the lake surface. Based on the actual situation, a suitable synchronization area is selected, as follows:

[0123] (1) FY-3D

[0124] Figure 9 The spatial distribution of entrance pupil brightness temperature information obtained from the inversion of the thermal infrared image of channel 24 of the FY-3D satellite during the synchronous observation period is presented. The brightness temperature information is indicated by warm and cool hues, with cloud-covered areas typically having much lower temperatures than normal surface areas and exhibiting a distinctly cool hue. Therefore, it can be seen from the figure that the lake surface was affected by clouds to varying degrees during the synchronous observation period. Figure 8 Comparative analysis also shows that the brightness temperature of the cloud-affected area is significantly lower than that of other areas of the lake surface. The cloud brightness temperature in the affected area is about 230K, while the brightness temperature of other areas of the lake surface reaches about 287K.

[0125] Based on the cloud-affected area and the synchronization area, lake surfaces suitable for calibration experiments can be selected. The selection criteria are twofold: firstly, to avoid the influence of clouds and cloud edges; and secondly, to choose lake areas with uniform temperature variations. Furthermore, the synchronization area must be within the observation track range. If the track range is affected by clouds, a suitable nearby area is selected. The spatial distribution of brightness temperature obtained from the FY-3D synchronization observation images on August 18th and August 20th shows that most lake surfaces were under clear skies, basically meeting the requirements for accurate synchronization. Figure 9 On August 19th, due to greater cloud cover, it was advisable to choose a location closer to the synchronous test. Analysis of the brightness temperature at the same location during the three time periods revealed that the brightness temperatures of the precisely positioned satellite pixels reached 287K and 288K on August 19th and 20th, respectively, while the brightness temperature on August 18th was close to 285K. This indicates that despite differences in observation time, the water temperature of Qinghai Lake remained relatively stable over the days. Figure 10 We used a 5×5 grid for statistical analysis on the selected synchronization regions and found that the standard deviation of the pixel brightness temperature in the synchronization regions was less than 0.1K.

[0126] (2) FY-3C

[0127] Figure 11 This is the fourth channel brightness temperature information acquired by FY-3C during the synchronous observation experiment at Qinghai Lake. Based on the difference in brightness temperature between clouds and water, it can be seen that the lake surface was less affected by clouds on the 18th and 20th, making it suitable for synchronous calibration experiments. On the 19th, cloud cover had a greater impact, making it impossible to obtain accurate and suitable synchronous radiance information for the lake surface. The results from the 18th and 20th both show that the eastern region of Qinghai Lake had the least cloud impact; this area was also the main region where we conducted synchronous experiments using unmanned surface vessels, and it fully met the requirements for accurate pixel matching.

[0128] Analysis of the brightness temperature information obtained using a 5×5 grid window revealed that the brightness temperatures observed by the satellite remained relatively stable at 285K and 286K, with regional brightness temperature deviations of less than 0.2K. The brightness temperature observation results of FY-3C were very close to those of FY-3D, with only a very small temperature difference between them due to the different transit times.

[0129] (3) FY-4A

[0130] Unlike the polar-orbiting satellites FY-3C and FY-3D, FY-4A is a new generation of geostationary satellite. The thermal infrared channel spatial resolution (4km) of the AGRI sensor is much lower than that of polar-orbiting satellites. Moreover, FY-4A can observe the same observation location once per hour. For field calibration work, this high-frequency observation can provide sufficient data for calibration synchronization test screening, and has a good advantage in data availability.

[0131] FY-4A is a geostationary meteorological satellite with relatively coarse spatial resolution. Therefore, the selection of pixels in the geosynchronous region was more stringent. To avoid the influence of clouds, we selected the observation periods with relatively few clouds during the transit times on the 18th and 20th for our analysis. Figure 12 Brightness temperature information obtained during synchronous observation of Qinghai Lake by FY-4A is presented. Comparative analysis shows that the brightness temperature changes very little over the preceding and following days. Furthermore, brightness temperature statistics using a 2×2 grid are found to be basically stable at around 285K.

[0132] (4) Terra / Aqua MODIS

[0133] The experiment primarily utilized the 31st and 32nd thermal infrared channels of the MODIS sensor on the Terra / Aqua satellite platform, using synchronous lake surface observations to evaluate the rationality of MODIS calibration. The spatial distribution of brightness temperature in MODIS channel 31 is shown below. Figure 13As shown, the synchronous observation period, consistent with that of the FY series satellites, was selected from August 18th to August 20th. The satellite radiation brightness temperature results in the figure show that there were large clear-sky areas over Qinghai Lake on August 18th and 19th, suitable for synchronous testing and analysis. On August 19th, the Aqua satellite could observe part of the lake surface, but cloud cover had a significant impact; therefore, the synchronous calibration experiment focused on data from the 18th and 19th. We also analyzed the brightness temperature of the selected synchronous observation area using a 5×5 grid. The results showed that the brightness temperatures obtained from the Aqua satellite were generally around 287K and 288K, slightly higher than the radiation brightness temperature results obtained from the Fengyun series satellites during certain periods. Considering the differences in time and spectrum, the actual observation information from both shows good consistency.

[0134] We further analyzed the brightness temperature of the lake surface observations using different regions of interest (ROIs). The analysis involved sampling the entire lake surface in different regions to distinguish the brightness temperature differences between the eastern and western parts of Qinghai Lake. Overall, the difference in radiation brightness temperature between different affected areas was less than 1 K, with the smallest difference within the ROI reaching 0.1 K. ROIs 1 and 2 are both located in the western waters of Qinghai Lake, while ROIs 3 and 4 are located in the eastern waters. Comparing different lake surfaces revealed that the temperature in the western waters was slightly lower than in the eastern waters, by approximately 0.2 K. Additionally, the standard deviation of brightness temperature in the western waters was smaller than that in the eastern waters. Given that Qinghai Lake spans an area of ​​over 100 km², the fact that the brightness temperature deviation remains within 1 K indicates that the lake water is clean and stable, and that observations of adjacent areas can be used as references for calibration pixels.

[0135] 5.2 Extraction of Surface Radiation Information

[0136] Lake water radiance was observed using the CLIMEL CE312, a handheld low-noise thermal infrared radiometer operating in an 8-14 μm atmospheric window. Its design encompasses four different thermal infrared channels: Band-1: 8-14 μm; Band-2: 11.5-12.5 μm; Band-3: 10.3-11.3 μm; and Band-4: 8.2-9.2 μm. Bands 2 and 3 showed good overlap and matching with the thermal infrared channels of the FY series satellites and MODIS sensors. Furthermore, CE312-Band 1 can be matched with most satellite infrared absorption channels. To ensure the accuracy of the experimental results, the spectral response functions of different payloads were processed and matched. Additionally, the CE312 was blackbody calibrated before and after the experiment, and the measured parameters were converted to obtain the spectral radiance at specific wavelengths.

[0137] Figure 14The brightness temperature of the lake surface observed during the synchronous test period is presented, including brightness temperature information of both lakeshore features and lake water. It is evident that the brightness temperature of lakeshore features exceeded 30℃, while the lake water temperature remained stable at around 16-17℃. After excluding observations from the pre-launch and recovery phases of the unmanned surface vessel (USV), the lake surface brightness temperature information obtained from August 18th to 21st was very stable, further demonstrating the rationale for using Qinghai Lake as a calibration feature in the thermal infrared band. Taking August 18th as an example, the synchronous observation status of the lake surface at the satellite's transit time is further analyzed. Figure 15 On the 18th, the brightness temperature observed on the lake surface by CE312 showed very stable conditions in both the 12μm and 10.8μm thermal infrared channels. The brightness temperature in the 10.8μm channel was slightly higher than that in the 12μm channel, by approximately 0.17K. The fluctuation range of the brightness temperature in both thermal infrared channels was also very small. The standard deviation of brightness temperature change during the observation period was 0.043 for the 10.8μm channel and 0.032 for the 12μm channel, both less than 0.1K, fully meeting the accuracy requirements for satellite calibration and synchronous observation. Further analysis of the synchronous observation period revealed that the times when multiple satellites passed over Qinghai Lake (MODIS / Aqua, FY3D / MERSI, and FY4A / AGRI) were also evenly distributed within stable synchronous observation periods, fully meeting the synchronous observation requirements for half an hour before and after the passing time. The number of data samples from the unmanned surface vessel's synchronous observations also far exceeded those from the large ship's synchronous observations, which is beneficial for data processing, analysis, and evaluation.

[0138] 5.3 Processing of Surface Atmospheric Observation Data

[0139] (1) Atmospheric aerosol observation

[0140] Atmospheric aerosol parameters were retrieved using the CLIMEL CE318 automatic scanning and tracking solar photometer. The CE318 has eight observation channels and can automatically track the sun to measure direct solar radiation, which can be used to retrieve and calculate atmospheric transmittance, extinction optical thickness, aerosol optical thickness, total atmospheric water vapor column, and total ozone. During the experiment, the CE318 instrument was leveled and placed near the ground observation site for all-weather automatic observation. The Langley method was used to obtain the relationship between aerosol optical thickness and wavelength under a specific distribution. Using the measured aerosol optical thickness at two wavelengths, the following calculations were performed:

[0141]

[0142] Using the measured α and β values, the optical thickness of aerosols at the desired wavelength can be retrieved. During the synchronous experiment, CE318 conducted all-weather observations of direct solar radiation and sky radiation. Its 10-band observation data were processed to obtain aerosol optical thickness information for different bands and time periods. After cloud filtering, valid aerosol observation information was obtained. Then, using the results from the 443nm and 865nm channels combined with the aerosol Junge parameters, the optical thickness of aerosols at 550nm was calculated and retrieved.

[0143] Figure 16 The effective aerosol optical thickness (AOD) at 550 nm was obtained during the synchronous experiment from August 18th to August 20th. The effective results show that the aerosol attenuation effect on light in the Qinghai Lake area is minimal, with an AOD of approximately 0.2, indicating relatively ideal atmospheric transmittance in this region. The highest AOD value during the synchronous experiment did not exceed 0.25, further demonstrating that the atmosphere at Qinghai Lake is relatively clean and has good transmittance, making it highly suitable for conducting synchronous calibration experiments in the thermal infrared channel.

[0144] (2) Acquisition of atmospheric profile layering information

[0145] Based on the satellite's transit time and the atmospheric cloud conditions during the synchronous test, a decision is made regarding whether to release the sounding balloon, which is then released within half an hour of the synchronous transit. The sounding balloon's onboard radiosonde tracks its changing position and corresponding observation parameters, such as temperature, humidity, air pressure, wind direction, and wind speed, obtaining real-time profiles of these parameters. Figure 17 This data represents the atmospheric profile changes observed on August 18, 2019, primarily focusing on temperature and water vapor. Based on the balloon's altitude, it passed through different pressure layers, eventually entering the stratosphere at an altitude of approximately 20,000 meters. A clear decreasing trend in tropospheric temperature was observed, while the trend reversed in the stratosphere. Water vapor content also showed a significant decreasing trend in the troposphere, reaching its lowest point at the tropopause. The profile data, after manual review and stratified sampling, was used for the MODTRAN model.

[0146] 5.4 Atmospheric Radiation Simulation Analysis

[0147] Atmospheric radiation simulation was performed using the MODTRAN 4.3 atmospheric radiative transfer model to obtain a quantitative understanding of the atmospheric influence on the upward radiation over the lake surface. The atmospheric radiation simulation considered factors such as the location of the observation field, satellite altitude, satellite observation angle, aerosol optical thickness, visible distance, atmospheric temperature and water vapor profile variations, spectral range, and corresponding increments. This information was input into MODTRAN to calculate the transmittance and path radiation characteristics for each spectral band. Figure 18 and Figure 19Information on atmospheric transmittance and path radiation obtained from the simultaneous tests conducted by FY-3D and MODIS at the Qinghai Lake test site on August 18, 2019, is presented.

[0148] We further calculated the transmittance and path radiance information for different channels by convolving the transmittance information with the channel spectral response functions (RSF files) of different satellite infrared payloads. Based on this information, we can quantify the influence of the atmosphere on the Earth's surface thermal infrared radiation. Table 2 compares the atmospheric transmittance and path radiance information of FY-3D and MODIS on August 18th. The table shows that the atmospheric transmittance reached 0.97 and 0.98 in the infrared radiation channels of FY-3D, while the atmospheric transmittance obtained by MODIS was slightly higher than that of FY-3D, exceeding 0.98. The path radiance of FY-3D in channels 24 and 25 was relatively low, approximately 1.3E-07, while that of MODIS in channels 31 and 32 was slightly lower than that of FY-3D.

[0149] Table 2 shows the atmospheric transmittance and path radiation information obtained from the simulation.

[0150]

[0151] 5.5 Spectral Matching

[0152] Different sensors may have variations in center wavelength and bandwidth during design and manufacturing, leading to differences in their observations of the same target. A comparison of the spectral response functions between the ground-based observation instrument CE312 and the satellite's thermal infrared channel is presented. Figure 20 As can be seen, the satellite's thermal infrared channel settings take into account the influence of the water vapor absorption window well. Most sensors are set with two channels, 10.8μm and 12μm, respectively. Although the center wavelengths of different sensors are similar, their spectral widths vary greatly. For example, MODIS is narrower than FY-3D.

[0153] To reduce errors in radiometric calibration caused by differences in sensor spectral response functions, spectral matching was performed between the satellite sensor and the ground-based CE312 radiometer. First, the MODTRAN radiative transfer model was used to simulate the top-of-layer spectral emission radiance of 972 atmospheric profiles under 9 different underlying surfaces (cloud tops, deserts, dry grasslands, farmland, forests, new snow, maple forests, oceans, and wet grasslands), 2 observation geometries (vertical and 10°), 6 atmospheric profiles (tropical, mid-latitude summer, mid-latitude winter, subarctic summer, subarctic winter atmosphere, and standard model), and 9 temperatures (278K, 283K, 288K, 293K, 298K, 303K, 308K, 313K, and 320K). Second, the spectral emission radiance curves were convolved with the spectral response functions of the corresponding channels of the satellite sensor and the CE312 radiometer to obtain the spectral radiance of the corresponding channels. Finally, linear regression was performed on the spectral radiance of the corresponding channels of the two sensors at the 972 points to obtain the spectral matching factor of the corresponding channels (e.g., spectral radiance of the satellite sensor and the CE312 radiometer). Figure 21 , 22 (As shown in Figure 23).

[0154] 6. Evaluation of calibration coefficients

[0155] The existing calibration coefficients of the satellite's thermal infrared payload were evaluated using a calibration experiment at Qinghai Lake. Table 3 presents the specific evaluation results, which were based on synchronous test data from August 18th and August 20th. The experiment focused on the thermal infrared split-window channels for surface temperature retrieval, including 10.8 and 12 micrometers. All evaluation results show that:

[0156] Table 3. Field Calibration Evaluation of Multiple Satellite Thermal Infrared Channels

[0157]

[0158] The FY-3D thermal infrared payload is operating at its optimal state, with the smallest difference in brightness temperature between satellite and ground observations reaching 0.26K and the largest difference being less than 0.99K. Launched on November 15, 2017, FY-3D's current calibration results indicate that the infrared payload's on-orbit calibration and operational status are excellent.

[0159] Launched in 2013, about four years earlier than satellite D, FY-3C's on-orbit performance is worse than that of satellite D, according to current test results. During synchronous tests, the best satellite-to-ground difference was found to be 1.16K, while the difference was greater than 2K at other times. Although there are differences between transit times and between observation angles, it is not as ideal as satellite D overall. The thermal infrared channel has a certain degree of attenuation and drift, but the overall accuracy is less than 1.5K, which is within an acceptable range.

[0160] FY-4A is a geostationary satellite, and its observation angle is relatively large for the Qinghai Lake synchronous test site, which affects the accuracy of atmospheric transport calculations. Furthermore, FY-4A has a relatively low spatial resolution of 4km, making its pixels more susceptible to atmospheric and cloud influences, making it difficult to select ideal synchronous observation pixels. In the experiment, we used pixels near the water area as substitutes for calibration coefficient evaluation, which also introduced some error. Therefore, overall, the evaluation results for FY-4A are not ideal, with the best satellite-to-ground contrast difference reaching 0.93K.

[0161] In our experiments, we also compared and analyzed the results using internationally recognized on-orbit calibrated thermal infrared payloads. The comparison between the infrared channels of the MODIS sensor on the Terra and Aqua satellite platforms and the infrared payloads of the FY series satellites showed that we could obtain good results using unmanned surface vessels (USVs) on both Terra and Aqua satellites. The optimal satellite-to-ground contrast deviation for the Aqua / MODIS payload reached 0.09K, while the optimal deviation for Terra / MODIS reached approximately 0.99K. This demonstrates that the observations obtained using USVs are feasible and reasonable, and can be used as a calibration observation method for future field experiments. The differences between Terra and Aqua are twofold: the latter is a later-launched satellite, and its payload stability is superior to the former; the former also experiences some degradation after long-term operation. Furthermore, Terra transits in the morning, and our ground observation test time was not optimally matched with it. Both the CE312 observations and the selection of satellite synchronization pixels involved corresponding substitutions, therefore, a certain degree of deviation is understandable.

[0162] Table 3, comparing calibration results at different times, also shows that the calibration accuracy of the field test was significantly affected by atmospheric conditions and cloud cover. The differences in the observation fields on August 18th and 20th were mainly due to cloud influence. On August 18th, the surrounding area of ​​the entire observation field was mostly clear, while on August 20th, there was heavy cloud cover around the observation field, with only the lake area experiencing less cloud influence at certain times. Overall, increased cloud cover, cloud edge effects, and differences in satellite observation angles all contributed to the deviation in satellite entrance pupil radiance, making it one of the most important calibration deviation factors. Therefore, continuous monitoring by FY-4A at multiple times showed a deviation of approximately 3K between the lake surface brightness temperature on August 18th and August 20th. For Qinghai Lake, continuous observations of the same area over several days generally remained within a deviation range of about 1K. Therefore, this deviation of the same sensor at different times should mainly come from cloud influence, and we need to pay special attention to selecting appropriate weather conditions for our subsequent calibration field tests.

[0163] 7. Summary

[0164] From August 14th to 22nd, 2019, a field calibration experiment for the infrared payload was conducted at Qinghai Lake. The synchronous observation data obtained from this experiment were used to evaluate the radiometric calibration coefficients of operational meteorological satellites and to calculate the alternative calibration coefficients for each payload's thermal infrared channel. The results of the field calibration experiment indicate that the overall on-orbit calibration of my country's operational meteorological satellites is reasonable, especially FY-3D, which achieved the best satellite-to-ground observation entrance pupil brightness temperature deviation of 0.35K. FY-3C and FY-4A showed slightly weaker calibration, reaching levels of 1.7K and 1.6K, respectively. Simultaneously, comparative experiments with the thermal infrared payloads of the Terra and Aqua satellite platforms also revealed relatively good on-orbit calibration, with optimal satellite-to-ground contrast deviations of 0.7K and 1.2K, respectively. This indicates that MODIS is still operating well and that the new calibration observation method using unmanned surface vessels is reasonable and can be used in operational field calibration.

[0165] In summary, by utilizing the technical solution of this invention, the use of an unmanned surface vessel equipped with an infrared radiometer CE312 replaces the traditional field measurement work of collecting water surface radiation information by personnel traveling by boat on the lake for radiation calibration, which can greatly save manpower, material resources, and financial resources. By combining the acquisition of encrypted atmospheric profile data, the calibration frequency can be greatly increased, and the calibration accuracy can be further improved. By comparing the entrance pupil brightness temperature information obtained from the thermal infrared channels of various satellites, the spatial distribution of brightness temperature during the synchronous test period at the calibration site and the degree of influence of cloud and weather conditions on the lake surface can be analyzed, which has great reference value for the selection of the synchronous area.

[0166] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for in-orbit radiometric calibration of a meteorological satellite thermal infrared channel, characterized in that, Comprising the following steps: S1. Selecting appropriate working time for the field calibration test according to satellite overpass information and weather condition of the calibration site; S2. Using unmanned boat to carry infrared radiometer CE312, releasing the unmanned boat from the wharf 2 hours before the satellite overpass, and sailing on the lake according to the pre-set route, carrying out lake surface radiance observation to obtain continuous observation water surface brightness temperature information according to the specified route, and returning to the base according to the pre-set route after completion; S3. Calibration data processing and analysis: S31. Extracting satellite entrance pupil brightness temperature information: selecting satellite thermal infrared load or thermal infrared channel overpass synchronous image data, and calculating the corresponding entrance pupil brightness temperature information according to the radiation calibration coefficient of each satellite, comparing the entrance pupil brightness temperature information obtained by each satellite thermal infrared channel, analyzing the brightness temperature spatial distribution of the calibration site during the synchronous test period and the influence degree of the lake surface by clouds and weather conditions, and selecting appropriate synchronous area according to the actual situation; selecting the lake surface for calibration test according to the cloud influence area and the synchronous area, and selecting the lake surface area according to the avoidance of cloud and cloud edge influence and uniform temperature change, and the synchronous area should be ensured within the observation track range, if the track range is affected by clouds, then select the adjacent suitable area; S32. Extracting ground surface radiance information: processing and matching the spectral response functions of different loads, and carrying out blackbody calibration on infrared radiometer CE312 before and after the test, and obtaining the spectral radiance of a specific wavelength after converting the measured parameters; S33. Processing ground atmospheric observation data: using a solar photometer to calculate and invert the aerosol optical thickness, and using the Langley method to obtain the relationship between the aerosol optical thickness and the wavelength under a specific distribution; Using the sonde carried by the atmospheric sounding balloon to track the balloon position and the corresponding observation parameters to obtain the real-time profile state of these parameters; S34. Atmospheric radiation simulation: using atmospheric radiation transfer mode to simulate atmospheric radiation, obtaining the quantitative influence of atmosphere on the uplink radiation of the lake surface; according to the spectral response function of the thermal infrared channel of different satellite infrared loads, convoluting the transmittance information and the channel spectral response function to obtain the transmittance and path radiation information of different channels, and quantifying the influence of atmosphere on ground surface thermal infrared radiation according to the transmittance and path radiation information; S35. Spectral matching of satellite sensor and ground infrared radiometer CE312.

2. The method according to claim 1, wherein, Calculating the spectral matching factor to reduce the spectral response difference between ground observation instrument and satellite thermal infrared load.

3. The method according to claim 1, wherein, The satellite observation zenith angle during the synchronization period is less than 15 degrees.

4. The method according to claim 1, wherein, The standard deviation between ground observation pixels during the synchronous observation period is less than 10%.

5. The method for on-orbit radiometric calibration of a meteorological satellite thermal infrared channel according to claim 1, characterized in that, In step S31, the spectral radiance observed by the satellite can be converted to brightness temperature value according to the Planck formula: where C1 = 1.191 formula will -12 W / (cm 2 ·sr·cm -1 ) 4 ), is the first radiation constant; C2 = 1.439 K-cm is the second radiation constant; σ is the wave number (cm"1), T is the temperature (K), and L(σ) is the spectral radiance (W / (cm 2 ·sr-cm -1 ).

6. The method for on-orbit radiometric calibration of a meteorological satellite thermal infrared channel according to claim 1, characterized in that, In step S33, the calculation method of aerosol optical thickness is: calculating the aerosol optical thickness of the required wavelength by measuring the aerosol optical thickness of two wavelengths: Then use the measured α and β values to invert the optical thickness of the aerosol of the required wavelength.

7. The method according to claim 1, wherein, In step S33, the corresponding observation parameters include temperature, humidity, air pressure, wind direction, and wind speed.

8. The method for on-orbit radiometric calibration of a meteorological satellite thermal infrared channel according to claim 1, characterized in that, Further, it also comprises: S4 inputs the observation data of the unmanned vehicle, the synchronous image data of the meteorological satellite and the sounding profile data measured during the synchronization into the MODTRAN model to simulate the apparent radiance of the atmospheric top.

9. The method according to claim 8, wherein, Further, the method further comprises: S5 calculates the absolute radiometric calibration coefficient of the channel to be calibrated by linear regression according to the average value of the DN of the meteorological satellite and the apparent radiance of the atmospheric top of the channel to be calibrated simulated.

Images