CN114581791A - Inversion method and system for atmospheric water vapor content based on MODIS data - Google Patents
Inversion method and system for atmospheric water vapor content based on MODIS data Download PDFInfo
- Publication number
- CN114581791A CN114581791A CN202210202976.1A CN202210202976A CN114581791A CN 114581791 A CN114581791 A CN 114581791A CN 202210202976 A CN202210202976 A CN 202210202976A CN 114581791 A CN114581791 A CN 114581791A
- Authority
- CN
- China
- Prior art keywords
- water vapor
- temperature
- vapor content
- data
- channel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Chemical compound O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 102
- 238000000034 method Methods 0.000 title claims abstract description 57
- 230000005855 radiation Effects 0.000 claims description 83
- 230000005540 biological transmission Effects 0.000 claims description 51
- 238000013135 deep learning Methods 0.000 claims description 47
- 238000003062 neural network model Methods 0.000 claims description 30
- 230000008569 process Effects 0.000 claims description 17
- 238000012545 processing Methods 0.000 claims description 17
- 238000002834 transmittance Methods 0.000 claims description 14
- 230000006870 function Effects 0.000 claims description 11
- 238000007781 pre-processing Methods 0.000 claims description 10
- 230000009471 action Effects 0.000 claims description 7
- 238000012937 correction Methods 0.000 claims description 7
- 238000000862 absorption spectrum Methods 0.000 claims description 6
- 238000002310 reflectometry Methods 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims 1
- 238000004422 calculation algorithm Methods 0.000 description 15
- 238000013528 artificial neural network Methods 0.000 description 14
- 238000004088 simulation Methods 0.000 description 11
- 238000012549 training Methods 0.000 description 10
- 238000012795 verification Methods 0.000 description 10
- 230000000694 effects Effects 0.000 description 9
- 238000010521 absorption reaction Methods 0.000 description 7
- 230000008859 change Effects 0.000 description 7
- 238000013136 deep learning model Methods 0.000 description 7
- 239000007789 gas Substances 0.000 description 7
- 238000012360 testing method Methods 0.000 description 6
- 238000005457 optimization Methods 0.000 description 5
- 239000000047 product Substances 0.000 description 5
- 229910002092 carbon dioxide Inorganic materials 0.000 description 4
- 238000011160 research Methods 0.000 description 3
- 230000003595 spectral effect Effects 0.000 description 3
- CURLTUGMZLYLDI-UHFFFAOYSA-N Carbon dioxide Chemical compound O=C=O CURLTUGMZLYLDI-UHFFFAOYSA-N 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 239000006185 dispersion Substances 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000013178 mathematical model Methods 0.000 description 2
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 description 2
- 239000011435 rock Substances 0.000 description 2
- 239000002689 soil Substances 0.000 description 2
- 241000132092 Aster Species 0.000 description 1
- CBENFWSGALASAD-UHFFFAOYSA-N Ozone Chemical compound [O-][O+]=O CBENFWSGALASAD-UHFFFAOYSA-N 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 1
- 239000001569 carbon dioxide Substances 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 230000005494 condensation Effects 0.000 description 1
- 238000009833 condensation Methods 0.000 description 1
- 239000000470 constituent Substances 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000001704 evaporation Methods 0.000 description 1
- 230000008020 evaporation Effects 0.000 description 1
- 230000004907 flux Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 229910052760 oxygen Inorganic materials 0.000 description 1
- 239000001301 oxygen Substances 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
- -1 vegetation Substances 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/11—Complex mathematical operations for solving equations, e.g. nonlinear equations, general mathematical optimization problems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Abstract
The invention discloses an atmospheric water vapor content inversion method and system based on MODIS data. Through the technical scheme, the inversion accuracy of the atmospheric water vapor content can be effectively improved.
Description
Technical Field
The invention relates to the technical field of atmospheric data inversion, in particular to an atmospheric water vapor content inversion method and system based on MODIS data.
Background
The Atmospheric Water Vapor content (Atmospheric Water Vapor), which is also called as the amount of Atmospheric degradable Water, is relatively small in the atmosphere, but represents an active Atmospheric component, which is one of the main driving factors of weather changes and plays an important role in the weather changes. The atmospheric water vapor content space-time change is directly related to a plurality of weather processes and weather changes, is the variable which has the greatest influence on solar radiation in the atmosphere, is changed greatly along with time and place, and is the most unstable and difficult data to accurately acquire in atmospheric molecules. The change of water vapor content affects global heat circulation, water circulation and radiation energy balance, under the condition of clear sky, the water vapor accounts for about 70% of the total radiation absorption amount of the atmosphere, the water vapor can strongly absorb long-wave radiation emitted by the earth surface and also can emit long-wave radiation, the evaporation and condensation of the water vapor can absorb and emit latent heat, the temperature of the ground and the temperature of the air are directly affected, and the method is also important in various researches such as drought monitoring, climate change, temperature inversion and the like. The inversion method of the atmospheric water vapor content adopts light-near infrared inversion and infrared inversion, although the application of the light-near infrared inversion atmospheric water vapor content is mature, the influence of weather conditions and wave band characteristics is large, a scanning radiometer can only image in the daytime, and a water vapor product cannot meet the requirements of many applications on time scale.
Compared with near-infrared inversion, the infrared inversion has lower precision and spatial resolution, can be used all day long, and is sensitive to cloud cover. There are two representative algorithms for the thermal infrared band: an infrared statistical regression method and a split window method based on multiple bands. The statistical regression method is used for obtaining the atmospheric humidity profile, and due to the complexity of profile inversion, the accuracy cannot be high on the basis of the physical algorithm. The split window method is a water vapor inversion method with clear physical significance, and generally constructs the relation between the brightness temperature and the atmospheric water vapor content based on a radiation transmission equation model. The thermal infrared split window method is applied to atmospheric water vapor content inversion aiming at AVHRR data, and the relationship between the bright temperature difference of the water vapor weak sensitive channel and the atmospheric water vapor content is established based on water vapor differential absorption in the split window channel. In the past decades, various windowing algorithms have been proposed for inversion of atmospheric water vapor content and surface temperature, and although the accuracy of the algorithms is high, the algorithms still need to make some assumptions about prior knowledge of the atmosphere, so that simplification inevitably exists in the calculation process, which has a certain influence on inversion accuracy, and high-accuracy acquisition usually comes at the cost of increasing fitting parameters, but the nonlinear relation between the parameters and the interaction factors are difficult to describe clearly, and the acquisition of each parameter cannot be completely described by a mathematical formula and is not generalized.
Disclosure of Invention
In order to solve the problems of insufficient inversion accuracy and the like in the prior art, the invention provides an atmospheric water vapor content inversion method and system based on MODIS data, which can effectively improve the inversion accuracy of the atmospheric water vapor content and reduce the use of fitting parameters.
In order to achieve the technical purpose, the invention provides the following technical scheme:
an atmospheric water vapor content inversion method based on MODIS data comprises the following steps:
the method comprises the steps of obtaining remote sensing data, constructing a radiation transmission equation, calculating the remote sensing data through the radiation transmission equation to obtain channel brightness and temperature, and processing the channel brightness and temperature through a deep learning neural network model to obtain an atmospheric water vapor content inversion result.
Optionally, the process of obtaining the remote sensing data includes:
obtaining MODIS image data, and preprocessing the MODIS image data to obtain remote sensing data, wherein the preprocessing comprises radiometric calibration, geometric correction and reflectivity calibration.
Optionally, the building process of the radiation transmission equation includes:
constructing a radiation transmission equation based on the Planck function, wherein the radiation transmission equation is shown as the following formula:
Bi(Ti)＝εiBi(Ts)τi(θ)+(1+(1-εi))τi(θ)(1-τi(θ))Bi(Ta)，
wherein, TiIs the brightness temperature on the satellite for channel i, Bi(Ti) Is the channel brightness temperature received by the sensor; bi(Ts) For surface radiation, wherein TsIs the surface temperature; tau isi(θ) is the atmospheric transmittance of channel i at an observation angle θ; epsiloniIs the surface emissivity of the i channel, Bi(Ta) Is the average radiation of the atmosphere, wherein TaIs the effective average action temperature of the atmosphere.
Optionally, the processing of the channel brightness temperature through the deep learning neural network model includes:
and acquiring the surface temperature and the surface emissivity, in the process of inputting the channel brightness temperature as a deep learning neural network model, simultaneously inputting the surface temperature and the surface emissivity as the deep learning neural network model, and processing the channel brightness temperature, the surface temperature and the surface emissivity through the deep learning neural network model to obtain an inversion result of the atmospheric water vapor content.
Optionally, before processing the channel brightness temperature through the deep learning neural network model, the method includes:
acquiring an optimal infrared band combination, and processing the channel brightness temperature, the surface temperature and the surface emissivity under the optimal infrared band combination through a deep learning neural network model; the optimal infrared band combination is obtained based on the accuracy of an atmospheric water vapor content inversion result under a preset infrared band combination, and the preset infrared band combination is obtained based on a gas absorption spectrum.
In order to better achieve the technical purpose, the invention also provides an atmospheric water vapor content inversion system based on MODIS data, which comprises: an acquisition module and a generation module;
the acquisition module is used for acquiring remote sensing data, constructing a radiation transmission equation, and calculating the remote sensing data through the radiation transmission equation to obtain the channel brightness temperature;
the generation module is used for processing the channel brightness temperature through the deep learning neural network model to obtain an inversion result of the atmospheric water vapor content.
Optionally, the obtaining module includes a first obtaining module;
the first acquisition module is used for acquiring MODIS image data and preprocessing the MODIS image data to obtain remote sensing data, wherein the preprocessing comprises radiometric calibration, geometric correction and reflectivity calibration.
Optionally, the obtaining module includes a second obtaining module;
the second obtaining module is configured to construct a radiation transmission equation based on the planck function, where the radiation transmission equation is expressed as follows:
Bi(Ti)＝εiBi(Ts)τi(θ)+(1+(1-εi))τi(θ)(1-τi(θ))Bi(Ta)，
wherein, TiIs the brightness temperature on the satellite of channel i, Bi(Ti) Is the channel luminance temperature received by the sensor; b isi(Ts) For surface radiation, wherein TsIs the surface temperature; tau.i(θ) is the atmospheric transmittance of channel i at an observation angle θ; epsiloniIs the surface emissivity of the i channel, Bi(Ta) Is the average radiation of the atmosphere, wherein TaIs the effective average action temperature of the atmosphere.
Optionally, the generating module includes a first generating module;
the first generation module is used for acquiring surface temperature and surface emissivity, and in the process of inputting the channel brightness temperature as a deep learning neural network model, the surface temperature and the surface emissivity are simultaneously used as the input of the deep learning neural network model, and the channel brightness temperature, the surface temperature and the surface emissivity are processed through the deep learning neural network model to obtain an atmospheric water vapor content inversion result.
Optionally, the generating module further includes a second generating module;
the second generation module is used for acquiring an optimal infrared band combination and processing the channel brightness temperature, the surface temperature and the surface emissivity under the optimal infrared band combination through a deep learning neural network model; the optimal infrared band combination is obtained based on the accuracy of an atmospheric water vapor content inversion result under a preset infrared band combination, and the preset infrared band combination is obtained based on a gas absorption spectrum.
The invention has the following technical effects:
the invention provides a method for combining a radiation transmission equation with a deep learning neural network, wherein the radiation transmission equation is used as a specific model, the deep learning neural network is used as an optimization algorithm, an optimal solution is obtained, LST and LSE which play an important role in radiation transmission are used as empirical knowledge, and the inversion accuracy of AWV is improved. Meanwhile, the relation among all the parameters of the geophysical parameters is fitted through a deep learning neural network, the number of unknowns in a radiation transmission equation can be reduced, and the inversion accuracy can be further improved due to the existence of priori knowledge.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.
Fig. 1 is a schematic flow chart of a method according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example one
As shown in fig. 1, the invention provides an inversion method of atmospheric water vapor content based on MODIS data, which comprises: the method comprises the steps of obtaining remote sensing data, constructing a radiation transmission equation, calculating the remote sensing data through the radiation transmission equation to obtain channel brightness and temperature, and processing the channel brightness and temperature through a deep learning neural network model to obtain an atmospheric water vapor content inversion result.
In order to verify the applicability and popularization of the deep learning neural network method, firstly, MODTRAN is utilized to simulate the MODIS infrared band, and therefore the theoretical precision of the algorithm is obtained. Selecting MODIS data and ERA5 data to construct a training database, evaluating an algorithm, and performing site verification by using GPS reference data. In practical applications, outliers and non-representative data should be excluded to ensure that the selected data reflects all physical conditions.
There are two infrared atmospheric window ranges: 3.5-5.5 μm and 8-14 μm, we select the water vapor absorption band in the atmospheric window region, which is not overlapped with other gas absorption spectrum as far as possible, i.e. 20, 22, 23, 29, 31, 32 band, 33 band although it is overlapped with CO2The absorption bands coincide, but can also be used as an alternative channel for water vapor inversion. The MODIS L1B product was used in this study to calculate the radiance of the desired channel, the atmospheric moisture content of MOD/YD 05, the surface temperature, and the emissivity (MOD/MYD 11). In the algorithm, the brightness temperature of an MODIS infrared water vapor channel is used as an input parameter, AWV is used as an output parameter, and the surface temperature and the surface emissivity are used as empirical parameters to be input.
MODTRAN is a medium spectral resolution atmospheric transmittance and radiation transmission algorithm and a calculation model, and is one of the most widely applied radiation transmission models with high precision at present. Standard atmosphere is usually used in the simulation process, and parameters are modified according to actual conditions. However, in reality, the weather conditions are very complex, and MODTRAN simulation is very limited, so that the simulation data is only suitable for theoretical analysis and verification. The infrared band of the MODIS is simulated, the maximum spatial fraction is 0.05 degrees, the earth surface approximation is regarded as four types (soil, vegetation, water and rock), parameter setting is carried out, and finally, a brightness temperature result is obtained and used for constructing a deep learning training and testing database.
When electromagnetic waves are transmitted in the atmosphere, attenuation can be generated under the action of absorption, scattering and the like of the atmosphere, and the attenuation can be described by an atmosphere radiation transmission equation. It builds a mathematical model of the transmission and reception of solar radiation in the atmosphere-surface system and then solves the inverse problem to determine the parameters sought. In general, the earth surface is not a black body, and the relation between a general radiator and the black body is usually expressed based on kirchhoff's law, and the value of the relation is the same as the emissivity of the earth surface, so the emissivity of the earth surface itself needs to be considered. In the satellite thermal infrared remote sensing inversion, the radiation value detected by a satellite at the top of the atmosphere is generally the coupling result among the earth surface emissivity, the earth surface temperature and the atmospheric radiation, and the radiation received by an on-satellite sensor mainly comprises three parts: a) the radiation reaching the sensor after the surface radiation passes through the atmosphere, b) the upward radiation of the atmosphere, c) the energy reflected after the downward radiation of the atmosphere reaches the ground.
In the earth atmosphere system, energy exchange is continuously carried out in the earth-gas system, the density of the atmosphere in the vertical direction is not uniform, upward transmission and downward transmission are different, atmospheric radiation is usually expressed into an upward form and a downward form, and in the specific application based on a radiation transmission equation algorithm, the core part is to accurately obtain values of atmospheric upward radiation, atmospheric downward radiation and atmospheric transmittance. Under clear sky conditions, the radiance value at a certain angle of the sensor can be expressed as:
Tiis the on-satellite luminance temperature of channel i, Bi(Ti) The radiation received by the sensor is Planck function, the channel i emits emergent radiation brightness when the brightness temperature is T, and the intensity of the emergent radiation brightness is uniform; b isi(Ts) Surface radiation of earth, wherein TsIs the surface temperature in K; tau isi(θ) is the atmospheric transmittance of channel i at an observation angle θ; epsiloniIs the surface emissivity of the i channel;
and
respectively, atmospheric up-going and down-going long wave radiation. Planck linearization allows us to relate the radiation to the luminance temperature in the corresponding channel. Thus, the atmospheric up or down going radiation can be expressed as:
Rλ(θ)＝(1-τi(θ))Bi(Tc) (2)
Tcatmospheric up and down profile temperatures. While emissivity is generally high in the thermal infrared channel, according to the ASTER spectral data, we find that emissivity of most earth materials is higher than 0.65. According to the radiation transmission equation, the contribution of the atmospheric uplink radiation brightness to the radiation transmission equation is extremely small, and TcUsable atmospheric mean operating temperature TaTo show that:
Bi(Ti)＝εiBi(Ts)τi(θ)+(1+(1-εi))τi(θ)(1-τi(θ))Bi(Ta) (3)
from the above equation, it can be seen that the radiation transmission formula for each channel has four unknowns (transmittance, surface temperature, mean atmospheric temperature, surface emissivity), and if there are N radiation transmission equations, there are 2N +2 unknowns, which is a typical ill-conditioned problem, but these parameters are not independent of each other, but are interrelated. To solve for the atmospheric moisture content more accurately, we need to analyze the relationship between the parameters.
The coupling of the earth surface and the atmospheric radiation means that the brightness temperature contains information of the earth surface and the atmosphere, the earth surface temperature and the emissivity are direct driving forces of the earth surface-atmospheric system for long-wave radiation and latent heat flux exchange, the emissivity of the earth surface is a function of the wavelength, and for the same ground object, the emissivity of a wave band has a local linear relation, as shown in a formula (4). The surface temperature influences the inversion precision of the atmospheric water vapor content, and the 1K temperature inversion error needs 0.6g/cm2The data support of the atmospheric water vapor content, in the transmission process of the thermal infrared radiation, the inversion of the surface temperature can be regarded as a function of emissivity and transmittance:
ε＝f(εi) (4)
LST＝f(εi,τi(θ)) (5)
the atmospheric temperature and the vertical distribution of water vapor are closely related, the influence of the atmospheric temperature and the water vapor is entangled, and according to the research definition of Sobrino and Caselles, the effective average action temperature (T) of the atmospherea) Can be expressed as:
wherein, w is the total content of atmospheric water vapor from the ground, namely the position with the height of 0 to the height Z, and the unit is cm; dw (Z, Z) is the atmospheric water vapor content at height Z; t iszIs the air temperature at height z, in K. In fact, the effective average atmospheric temperature is mainly determined by the water vapor content at 2m height and the near-surface air temperature, the brightness temperature on the satellite is not very sensitive to the near-surface air temperature, but the near-surface air temperature is greatly influenced by the surface temperature, and the two have a linear relation (as shown in formula (7)), so the T isaThere is a strong linear relationship with the luminance temperature on the satellite, and the expression is described as:
T0＝A1+B1Ts (7)
Ta＝A2+B2T0 (8)
Ta≈A3+B3Ti (9)
T0is the air temperature, T, at a height of 2m above the groundiBrightness temperature on satellite, A1，A2，A3Is a constant, B1，B2，B3Is a coefficient, which is different for different regions and seasons.
Many atmospheric constituents, e.g. water vapour, O3、CO2And other gases, all affect the atmospheric transmission rate. However, O3、CO2And other gases are relatively stable in the atmosphere, so their effect on the atmosphere can be assumed to be constant and we can simulate through a standard atmospheric profile. In contrast, the water vapor content varies greatly. Thus, the change in the atmospheric transmittance depends to a large extent on the dynamic change in the moisture content in the profile.
A large number of studies have shown that the water vapor content is related to the observed bright temperature difference value of the channel. Therefore, the transmittance and the temperature difference of the channel luminance are necessarily in a certain functional relationship.
AWV＝f(ΔBT) (10)
τi(θ)＝f(AWV,x) (11)
τi(θ)＝g(ΔBT,x) (12)
x is a constant and represents the influence of other atmospheric absorption components (carbon dioxide, ozone, methane, oxygen and the like) on the atmospheric transmittance; AWV is the atmospheric moisture content. From the above formula, we can see that emissivity, surface temperature and AWV are entangled with each other, and there is a strong correlation.
In a classical atmospheric water vapor content inversion method, the strong correlation between water vapor and atmospheric transmittance is generally utilized to process, and the relationship needs to be simulated by using an atmospheric radiation transmission model. The method needs to obtain various physical parameters from remote sensing data, and due to limited observation conditions, the information of the parameters is difficult to obtain only from satellite signals, and the factors of nonlinear relation and interaction among the parameters are difficult to describe clearly, and the traditional method causes the inversion accuracy to be reduced in a simplified process. In contrast, the invention finds that the DNN can learn strong correlation among parameters, does not need any special adaptation, can realize high-precision convergence by only a few simple parameters, and changes the fitting of the problem on the nonlinear function into the problem of searching the nonlinear multi-function by the DNN. Although it is not clear whether this is advantageous in a strict mathematical sense, it is the task that DNN is good at in various practical applications. The invention hopes to introduce the deep learning neural network into the radiation transmission field, and provides possibility for solving each parameter of the equation by utilizing the remarkable progress obtained in the deep learning aspect in the past decade.
The number of computations required to optimize the deep-learning neural network parameters is quite large compared to conventional approaches, which has the advantage that the deep-learning neural network does not need to be modified if new data is included, because the functional form is very flexible. The accuracy of deep learning neural networks is limited only by the accuracy of the training data, and a fixed network structure cannot be used to predict results for different system sizes, because the optimization weights are only valid for a fixed number of input nodes, and furthermore, fine tuning of certain parameters in the DNN may even result in significant changes in the decision results, which may affect the practical application of making the correct decisions. It is noted that when there is a small change in the training set, and there is also a small change in the result of the web learning, the robustness of the deep learning may make the result fraudulent.
The number of unknowns in the radiation transmission equation can be reduced by the connection among all the parameters of the geophysical parameters, and the inversion accuracy can be further improved due to the existence of the prior knowledge. Deep learning neural networks, as a mathematical or computational model that mimics the structure and function of biological deep learning neural networks, are generally capable of fitting complex nonlinear systems. The method is used as an optimization algorithm, the optimal solution of the problem is continuously searched according to the probability iteration, but a black box function exists in the optimization algorithm, and how to establish the abstract problem into a concrete model is a difficult problem. Based on the method, a method for combining a radiation transmission equation with a deep learning neural network is provided, the radiation transmission equation is used as a specific model, the deep learning neural network is used as an optimization algorithm, an optimal solution is obtained, LST and LSE which play an important role in radiation transmission are used as empirical knowledge, and the inversion accuracy of AWV is improved.
The MODIS has 36 channels, although the existing method for inverting the content of atmospheric water vapor by using visible light-near infrared channels 2, 5, 17, 18 and 19 is more, but only suitable for being used in daytime, and the MODTRAN simulation process needs to consider the observation angles of the sun and the satellite, the situation is complex, and the simulation result may be greatly influenced, so that the research only considers the water vapor absorption channels of an atmospheric window within the range of 3-14 μm, wherein 33 channels (13.185-13.485 μm) are located at the boundary of the water vapor window and carry partial water vapor information, and as alternative channels, the infrared band of 3-8 μm is easily influenced by the sun, and therefore 20, 22 and 23 channels are considered for later data inversion. First, the MODIS/MODTRAN simulation data are divided into three groups: 1) suitable for day and night: BTs for TIR bands 29, 31 and 32 (8-12.5 μm); 2) is suitable for the daytime: BTs for TIR bands 29, 31, 32 and 33 (8-13.5 μm); 3) suitable for night: TIR bands 29, 31, 32 and MWIR bands 20, 22 and 23 (3-5 μm). And (4) verifying the theoretical accuracy of water vapor inversion on three sets of data sets simulated by MODTRAN, and finally performing example verification. In order to obtain more accurate spectral information, the MODIS image data is preprocessed by radiometric calibration, geometric correction, atmospheric correction, reflectivity calibration and the like to obtain BTs, AWV, LST and LSE data. LST is surface temperature, LSE is ground emissivity, BTs is channel brightness temperature, and AWV is atmospheric moisture content.
Firstly, LST, LSE, angle and atmospheric water vapor content range of each infrared band are set, an MODTRAN model is used for simulating an MODIS band, brightness temperature results of each band of the MODIS are output, theoretical precision verification is carried out on the model by the aid of the obtained brightness temperature, and the method comprises three groups of data: and the data is suitable for day/night, day and night, and the prior knowledge (LST and LSE) is added and the data without the prior knowledge is compared to analyze the result. And finally, verifying the actual precision by using the preprocessed MODIS image, performing cross verification on MOD/MYD05 atmospheric water vapor content products, and verifying field data by using GPS (global positioning system) site data as a theoretical true value. The inversion of atmospheric water vapor content using the DNN algorithm can be divided into four basic steps. The method is suitable for theoretical precision analysis and precision verification.
A reliable training and testing database (AWV, BTs for infrared band, with/without LST, LSE) was obtained for theoretical accuracy analysis and practical application. The acquisition times of the individual images used for the accuracy verification data (MODIS, ERA5) must be matched and, if necessary, resampled to ensure matching of the different images. Training and testing of DNNs. The MODIS data is used to calculate BTs. AWV was inverted using DNN.
Specific settings as shown below, for MODIS simulation data from MODTRAN, we set the training data (126720 sets) and the test data (59400 sets) as follows: LST (280-330K), WVC (0.1-4.5 g/cm)2) And LSE (vegetation, water, rock, soil), the biggest visual angle of MODIS sensor is 65, can cover all ranges for the simulation data, and simulation angle setting range is0 ~ 65.
Selecting Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) as measures of deviation of the result from the true value, Standard Deviation (SD) as measures of degree of dispersion of a set of numbers themselves, and correlation coefficient (R) as measures of degree of correlation between data2) And analyzing the inversion result as an index. The variation range of the water vapor content of the atmospheric column is 0.3-73 kg/m2(0.03～7.3g/cm2) However, when the WVC content exceeds 5g/cm2For some TIR wave bands, the transmissivity is lower than 0.4, water vapor in air is contained more, the water vapor is easy to condense into small water drops to form clouds, the observation is not suitable for the sensor, and the abnormal values are removed in order to guarantee the training accuracy.
LST and LSE are used as priori knowledge, test data are divided into three groups to verify the accuracy of the deep learning neural network, and MODIS data are subjected to theoretical precision analysis. Specifically, the first group uses 29, 31, 32TIR bands for day/night water vapor inversion; a second set 29, 31, 32, 33 of TIR bands is used for daytime water vapor inversion; a third set of 20, 22, 23, 29, 31, 32TIR bands is used for late vapor inversion, each set inputting a priori knowledge data.
When the number of hidden nodes is 500(MAE 0.1396, RMSE 0.2350, SD 0.9992, R0.9738), the accuracy is highest, and when the number of hidden nodes is 7 and 700, the accuracy is highest (MAE 0.1011, RMSE 0.1551, SD 1.0159, R0.9887); when the number of hidden layers is 10 and the number of hidden nodes is 600, the accuracy is the highest (MAE — 0.0937, RMSE — 0.1304, SD — 1.0171, R — 0.9920). Therefore, according to the preliminary analysis, the more the band information, the better the inversion effect, but the more unstable the inversion result.
For the model without prior knowledge, the test data is also divided into three groups, and the models with the wave bands having prior knowledge are selected to be the same. When the number of hidden layers is 7 and the number of hidden nodes is 600 (MAE: 0.4110, RMSE: 0.5577, SD: 0.8655, and R: 0.8417), the accuracy is highest; when the number of hidden layers is 7 and the number of hidden nodes is 700, the accuracy is the highest (MAE: 0.3663, RMSE: 0.4973, SD: 0.8930, R: 0.8765); when the number of hidden layers is 8 and the number of hidden nodes is 800, the accuracy is the highest (MAE — 0.1711, RMSE — 0.2472, SD — 0.9964, and R — 0.9710). Similarly, it can be seen from the preliminary analysis that the more the band information, the better the inversion effect, and the worse the stability of the inversion result. Compared with the model with prior knowledge, the inversion precision is obviously reduced.
The optimal results of the two models are compared, and the two sets of data meet the rule that the error is reduced along with the increase of the number of the wave bands, but the degree of self dispersion is increased, the inversion result is more unstable, and the fact that the rule is met no matter the inversion result is transverse comparison or longitudinal comparison can be seen from table 1, and the reason that the result is unstable can be caused by the fact that the structure of the deep learning neural network is unstable along with the increase of the information quantity. On the premise that LST and LSE are known, the accuracy of the inversion result is improved with the increase of the number of the wave bands, but the effect is not obvious, and the radiation transmission equation can be solved well by the brightness temperature of 3 wave bands; compared with data without prior knowledge, the inversion result is greatly improved by increasing the number of the wave bands, particularly under the condition of 6 wave band information, the inversion result is similar to the result of the known prior knowledge, but the accuracy is improved at the cost of increasing the information quantity, so that the inversion accuracy can be obviously improved by the prior knowledge. Table 1 shows the inversion error comparison for the optimal results of the model.
TABLE 1
Under the condition that prior knowledge exists, the deep learning model based on the radiation transmission model is used for inverting the atmospheric water vapor content, data are divided into three groups, the deep learning model is suitable for inversion of MODIS in different time periods, the deep learning model is suitable for inversion all day and comprises 3 infrared band information (29, 31 and 32), the deep learning model is suitable for inversion in day and comprises 4 infrared band information (29, 31, 32 and 33), and the deep learning model is suitable for inversion at night and comprises 6 infrared band information (20, 22, 23, 29, 31 and 32). With known a priori knowledge, we further study the band information for the night, here we supplement the inversion of the combination of 4 infrared band information (23, 29, 31, 32) and the combination of 5 infrared band information (20, 22, 29, 31, 32).
When the number of infrared bands is 4, the MAE with the highest inversion accuracy is 0.0803, RMSE is 0.1208, SD is 1.0275, and R is 0.9931, when the number of infrared bands is 5, the MAE with the highest accuracy is 0.0765, RMSE is 0.1113, SD is 1.0253, and R is 0.9942, but the inversion accuracy is not very different, and the accuracy is not improved for increasing the bands. When the number of the infrared bands is 4, compared with the bands suitable for being inverted in the daytime, the accuracy is improved, the MAE is improved by 0.02, the RMSE is improved by 0.04, the 23 bands are more suitable for inverting the atmospheric water vapor content than the 33 bands, the band quality of the visible inversion atmospheric water vapor content can have certain influence on the inversion result, and the degree is not very large due to the strong fitting capability and fault tolerance of the deep learning model. Data redundancy may be brought about due to the increase of the number of the wave bands, and inversion accuracy is affected. The inversion of the MODIS nighttime data may use a combination of 4 bands or 5 bands.
With MAE, RMSE and R2And evaluating three sets of water vapor inversion results of the MODIS. The results of the validation of the three sets of data showed no significant difference in MAE and RMSE, with 0.26g/cm for the day and night combinations of MAE and RMSE, respectively2And 0.31g/cm2The MAE and RMSE were 0.2636g/cm in the daytime, respectively2And 0.3441g/cm2The MAE and RMSE at night were 0.2475g/cm respectively2And 0.31g/cm2. However, for R2In other words, the inversion result at night is superior to the other two combinations, the inversion result all day is the worst, and the inversion errors of the three groups of data are all 0.5g/cm2On the left and right, the atmospheric moisture content is underestimated as a whole. This result may have two reasons. Firstly, for MOD05 water vapor products, the estimation of the measured value is significantly higher, and secondly, the AWV image data is mostly concentrated in 0-2 g/cm2According to theoretical analysis, the underestimation phenomenon of the data exists.
To further compare the three AWV inversion results, we divided the verification results into 3 partitions (total AWV, v,>2g/cm2、<2g/cm2) Wherein the inversion result suitable for day or night and the inversion result suitable for day are>2g/cm2And the accuracy is poor, and the inversion accuracy of the whole section is not very different only when the method is suitable for night.
TABLE 2
The inversion results were verified in the field using GPS PWV and we compared MOD/MYD05 with GPS and DNN inversion results with GPS, the results are shown in Table 2. Table 2 shows inversion error tables in different AWV ranges, which are suitable for combination in day and night among three band combinations of MODIS, and have the worst inversion effect, where MAE and RMSE are 0.4391 and 0.4903, respectively, and are the best inversion effect in day, where MAE and RMSE are 0.2451 and 0.2932, respectively, and the accuracy of the inversion result has a large relationship with the accuracy of the training sample. The data of the MODIS05 has an overestimation phenomenon, the data of the high AWV has an underestimation phenomenon at night, the cloud detection at night is worse than that at day time due to the fact that short-wave observation is not carried out at night, the MODIS cannot identify all cloud pixels and cannot completely identify whether the pixels around the ground GPS station are completely clear, the reason that the data are underestimated under the condition of the high AWV at night can be partially explained, and although the stability of the model is poor, the overestimation or underestimation phenomenon can be well reduced. In addition, synchronous measurement and space difference exist among data, water vapor in the atmosphere is active, errors exist in the acquisition time of various water vapor, the GPS mainly obtains the water vapor content of the near-low altitude, and the MODIS obtains the water vapor on the inclined path.
At present, a radiation transmission equation is used as an important means for inverting AWV, but in the solving process, enough physical parameters cannot be obtained often, inversion is often called 'sick' inversion, theoretically, some relation exists among the physical parameters, and how to find the relation is a difficult problem at present.
The invention discusses the physical mechanism of parameter inversion, the internal relation among physical parameters and how to combine wave bands to realize high-precision inversion in detail, and well solves the problem of 'ill-condition' in the inversion. On the basis, the influence of the prior knowledge on AWV inversion is researched, MODTRAN is used for simulating MODIS data, theoretical analysis is carried out, and results show that under the condition that the prior knowledge is known, inversion accuracy can be greatly improved, the inversion accuracy can also be improved when the number of wave bands is increased, fluctuation of inversion results is reduced, and for the same information quantity, the prior knowledge has higher accuracy than the inversion with the increased number of wave bands. High content of (>2g/cm2) Underestimation of the presence of water vapor, low content of (C:<2g/cm2) Overestimating of water vapor and simulation of water vapor on two sidesThe effect is poor. In addition, the inversion result is influenced by the quality of the wave band for inverting the atmospheric water vapor content, but the degree is not very high due to the strong fitting capability and fault tolerance of the deep learning model, and it is worth noting that the inversion result is influenced by the data redundancy caused by the increase of the number of the wave bands.
Based on the verification of simulation data, the minimum MAE obtained by the optimal waveband combination is less than 0.1g/cm2，RMSE＝0.1304g/cm2. The result of the practical application of the MODIS image data compared with the MODIS05 water vapor product shows that the minimum MAE obtained by the optimal waveband combination is about 0.2475g/cm2RMSE of about 0.31g/cm2。Seemann[35]And King[67]MOD07 was verified to have a Root Mean Square (RMS) of about 4kg/m2While MOD05 verified that the displayed root mean square was 1.16kg/m2[29]The accuracy is improved. Increasing the number of infrared bands of the atmospheric window can improve the inversion accuracy of the AWV, and increasing the priori knowledge can obtain a better inversion effect, but in practical application, the accuracy and effectiveness of the training data should be ensured. Furthermore, the existing GPS water vapor data is used for field verification, but due to the fact that mismatching possibly exists in time and space and the influence of night cloud pixels, an inversion result is poor, and the phenomenon of overestimation or underestimation of AWV can be well reduced by the aid of the advanced learning model based on the priori knowledge.
Example two
The invention also provides an atmospheric water vapor content inversion system based on MODIS data, which comprises the following steps: an acquisition module and a generation module; the acquisition module is used for acquiring remote sensing data, constructing a radiation transmission equation, and calculating the remote sensing data through the radiation transmission equation to obtain the channel brightness temperature; the generation module is used for processing the channel brightness temperature through the deep learning neural network model to obtain an inversion result of the atmospheric water vapor content.
Optionally, the obtaining module includes a first obtaining module; the first acquisition module is used for acquiring MODIS image data and preprocessing the MODIS image data to obtain remote sensing data, wherein the preprocessing comprises radiometric calibration, geometric correction and reflectivity calibration.
Optionally, the obtaining module includes a second obtaining module; the second obtaining module is configured to construct a radiation transmission equation based on the planck function, where the radiation transmission equation is expressed as follows: b isi(Ti)＝εiBi(Ts)τi(θ)+(1+(1-εi))τi(θ)(1-τi(θ))Bi(Ta) Wherein, TiIs the brightness temperature on the satellite of channel i, Bi(Ti) Is the channel luminance temperature received by the sensor; bi(Ts) For surface radiation, wherein TsIs the surface temperature; tau isi(θ) is the atmospheric transmittance of channel i at an observation angle θ; epsiloniIs the surface emissivity of the i channel, Bi(Ta) Is the average radiation of the atmosphere, wherein TaIs the effective average action temperature of the atmosphere.
Optionally, the generating module includes a first generating module; the first generation module is used for acquiring surface temperature and surface emissivity, and in the process of inputting the channel brightness temperature as a deep learning neural network model, the surface temperature and the surface emissivity are simultaneously used as the input of the deep learning neural network model, and the channel brightness temperature, the surface temperature and the surface emissivity are processed through the deep learning neural network model to obtain an atmospheric water vapor content inversion result.
Optionally, the generating module further includes a second generating module; the second generation module is used for acquiring an optimal infrared band combination and processing the channel brightness temperature, the surface temperature and the surface emissivity under the optimal infrared band combination through a deep learning neural network model; the optimal infrared band combination is obtained based on the accuracy of an atmospheric water vapor content inversion result under a preset infrared band combination, and the preset infrared band combination is obtained based on a gas absorption spectrum. The system and method content in the invention are corresponding, and are not described in detail.
The foregoing illustrates and describes the principles, general features, and advantages of the present invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are described in the specification and illustrated only to illustrate the principle of the present invention, but that various changes and modifications may be made therein without departing from the spirit and scope of the present invention, which fall within the scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (10)
1. An atmospheric water vapor content inversion method based on MODIS data is characterized by comprising the following steps:
the method comprises the steps of obtaining remote sensing data, constructing a radiation transmission equation, calculating the remote sensing data through the radiation transmission equation to obtain channel brightness and temperature, and processing the channel brightness and temperature through a deep learning neural network model to obtain an atmospheric water vapor content inversion result.
2. The MODIS data-based atmospheric water vapor content inversion method as claimed in claim 1, wherein:
the process for acquiring the remote sensing data comprises the following steps:
obtaining MODIS image data, and preprocessing the MODIS image data to obtain remote sensing data, wherein the preprocessing comprises radiometric calibration, geometric correction and reflectivity calibration.
3. The MODIS data-based atmospheric water vapor content inversion method as claimed in claim 1, wherein:
the construction process of the radiation transmission equation comprises the following steps:
constructing a radiation transmission equation based on the Planck function, wherein the radiation transmission equation is shown as the following formula:
Bi(Ti)＝εiBi(Ts)τi(θ)+(1+(1-εi))τi(θ)(1-τi(θ))Bi(Ta)，
wherein, TiIs the brightness temperature on the satellite of channel i, Bi(Ti) Is the channel brightness temperature received by the sensor; b isi(Ts) For surface radiation, wherein TsIs the surface temperature; tau isi(θ) is the atmospheric transmittance of channel i at an observation angle θ; epsiloniIs the surface emissivity of the i channel, Bi(Ta) Is the average radiation of the atmosphere, wherein TaIs the effective average action temperature of the atmosphere.
4. The MODIS data-based atmospheric water vapor content inversion method as claimed in claim 1, wherein:
the process of processing the channel brightness temperature through the deep learning neural network model comprises the following steps:
and acquiring the earth surface temperature and the earth surface emissivity, taking the earth surface temperature and the earth surface emissivity as the input of the deep learning neural network model in the process of inputting the channel brightness temperature as the deep learning neural network model, and processing the channel brightness temperature, the earth surface temperature and the earth surface emissivity through the deep learning neural network model to obtain an inversion result of the atmospheric water vapor content.
5. The MODIS data-based atmospheric water vapor content inversion method as claimed in claim 4, wherein:
the method comprises the following steps of processing the channel brightness temperature through a deep learning neural network model:
acquiring an optimal infrared band combination, and processing the channel brightness temperature, the surface temperature and the surface emissivity under the optimal infrared band combination through a deep learning neural network model; the optimal infrared band combination is obtained based on accuracy of an inversion result of the atmospheric water vapor content under a preset infrared band combination, and the preset infrared band combination is obtained based on gas absorption spectrum.
6. The inversion system based on the MODIS data atmospheric water vapor content inversion method according to any one of claims 1 to 5, comprising:
the acquisition module generates a module;
the acquisition module is used for acquiring remote sensing data, constructing a radiation transmission equation, and calculating the remote sensing data through the radiation transmission equation to obtain the channel brightness temperature;
the generation module is used for processing the channel brightness temperature through a deep learning neural network model to obtain an atmospheric water vapor content inversion result.
7. The MODIS data-based atmospheric water vapor content inversion system of claim 6, wherein:
the acquisition module comprises a first acquisition module;
the first acquisition module is used for acquiring MODIS image data and preprocessing the MODIS image data to obtain remote sensing data, wherein the preprocessing comprises radiometric calibration, geometric correction and reflectivity calibration.
8. The MODIS data-based atmospheric water vapor content inversion system of claim 6, wherein:
the acquisition module comprises a second acquisition module;
the second obtaining module is configured to construct a radiation transmission equation based on the planck function, where the radiation transmission equation is expressed as follows:
Bi(Ti)＝εiBi(Ts)τi(θ)+(1+(1-εi))τi(θ)(1-τi(θ))Bi(Ta)，
wherein, TiIs the brightness temperature on the satellite for channel i, Bi(Ti) Is the channel brightness temperature received by the sensor; bi(Ts) For surface radiation, wherein TsIs the surface temperature; tau isi(θ) is the atmospheric transmittance of channel i at an observation angle θ; epsiloniIs the surface emissivity of the i channel, Bi(Ta) Is the average radiation of the atmosphere, wherein TaIs the effective average action temperature of the atmosphere.
9. The MODIS data-based atmospheric water vapor content inversion system of claim 6, wherein:
the generating module comprises a first generating module;
the first generation module is used for acquiring surface temperature and surface emissivity, and in the process of inputting the channel brightness temperature as a deep learning neural network model, the surface temperature and the surface emissivity are simultaneously used as the input of the deep learning neural network model, and the channel brightness temperature, the surface temperature and the surface emissivity are processed through the deep learning neural network model to obtain an atmospheric water vapor content inversion result.
10. The MODIS data-based atmospheric water vapor content inversion system of claim 9, wherein:
the generation module further comprises a second generation module;
the second generation module is used for acquiring an optimal infrared band combination and processing the channel brightness temperature, the surface temperature and the surface emissivity under the optimal infrared band combination through a deep learning neural network model; the optimal infrared band combination is obtained based on the accuracy of an atmospheric water vapor content inversion result under a preset infrared band combination, and the preset infrared band combination is obtained based on a gas absorption spectrum.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210202976.1A CN114581791A (en) | 2022-03-02 | 2022-03-02 | Inversion method and system for atmospheric water vapor content based on MODIS data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210202976.1A CN114581791A (en) | 2022-03-02 | 2022-03-02 | Inversion method and system for atmospheric water vapor content based on MODIS data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114581791A true CN114581791A (en) | 2022-06-03 |
Family
ID=81771257
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210202976.1A Pending CN114581791A (en) | 2022-03-02 | 2022-03-02 | Inversion method and system for atmospheric water vapor content based on MODIS data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114581791A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115422703A (en) * | 2022-07-19 | 2022-12-02 | 南京航空航天大学 | Surface thermal infrared emissivity inversion method based on MODIS data and Transformer network |
| CN115795781A (en) * | 2022-09-23 | 2023-03-14 | 北京大学 | Atmospheric water vapor content estimation method and system based on ground infrared radiometer |
| CN116029211A (en) * | 2023-01-10 | 2023-04-28 | 中国农业科学院农业资源与农业区划研究所 | Artificial intelligence-based agricultural meteorological remote sensing key parameter joint inversion method |
| CN116295853A (en) * | 2023-03-22 | 2023-06-23 | 大连海事大学 | Multichannel thermal infrared data surface temperature inversion method and device and electronic equipment |
| CN116486931A (en) * | 2023-06-21 | 2023-07-25 | 上海航天空间技术有限公司 | Full-coverage atmospheric methane concentration data production method and system coupled with physical mechanism |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101655564A (en) * | 2009-09-15 | 2010-02-24 | 中国农业科学院农业资源与农业区划研究所 | Method for inversing surface temperature and emissivity from MODIS data |
| CN101936877A (en) * | 2010-03-05 | 2011-01-05 | 中国农业科学院农业资源与农业区划研究所 | Method for inverting atmospheric water vapor content from MODIS (Moderate Resolution Imaging Spectroradiometer) data |
| CN103293522A (en) * | 2013-05-08 | 2013-09-11 | 中国科学院光电研究院 | Intermediate infrared two-channel remote sensing data surface temperature inversion method and device |
| CN108896188A (en) * | 2018-06-06 | 2018-11-27 | 北京空间飞行器总体设计部 | A kind of high rail remote sensing satellite infrared channel atmospheric correction method of Shared aperture |
| CN113486581A (en) * | 2021-07-02 | 2021-10-08 | 宁夏大学 | Machine learning earth surface temperature estimation method, system, storage medium and equipment based on knowledge driving and logical reasoning |
-
2022
- 2022-03-02 CN CN202210202976.1A patent/CN114581791A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101655564A (en) * | 2009-09-15 | 2010-02-24 | 中国农业科学院农业资源与农业区划研究所 | Method for inversing surface temperature and emissivity from MODIS data |
| CN101936877A (en) * | 2010-03-05 | 2011-01-05 | 中国农业科学院农业资源与农业区划研究所 | Method for inverting atmospheric water vapor content from MODIS (Moderate Resolution Imaging Spectroradiometer) data |
| CN103293522A (en) * | 2013-05-08 | 2013-09-11 | 中国科学院光电研究院 | Intermediate infrared two-channel remote sensing data surface temperature inversion method and device |
| CN108896188A (en) * | 2018-06-06 | 2018-11-27 | 北京空间飞行器总体设计部 | A kind of high rail remote sensing satellite infrared channel atmospheric correction method of Shared aperture |
| CN113486581A (en) * | 2021-07-02 | 2021-10-08 | 宁夏大学 | Machine learning earth surface temperature estimation method, system, storage medium and equipment based on knowledge driving and logical reasoning |
Non-Patent Citations (3)
| Title |
|---|
| HAN WANG 等: ""A method for land surface temperature retrieval based on model-data-knowledge-driven and deep learning"", 《REMOTE SENSING OF ENVIRONMENT》 * |
| 陆桂华 等: ""全球水循环研究进展"", 《水科学进展》 * |
| 雷忠腾: ""基于MODIS热红外分裂窗波段反演晴空大气水汽的研究"", 《中国优秀硕士论文全文数据库 基础科学辑》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115422703A (en) * | 2022-07-19 | 2022-12-02 | 南京航空航天大学 | Surface thermal infrared emissivity inversion method based on MODIS data and Transformer network |
| CN115422703B (en) * | 2022-07-19 | 2023-09-19 | 南京航空航天大学 | Surface thermal infrared emissivity inversion method based on MODIS data and transducer network |
| CN115795781A (en) * | 2022-09-23 | 2023-03-14 | 北京大学 | Atmospheric water vapor content estimation method and system based on ground infrared radiometer |
| CN116029211A (en) * | 2023-01-10 | 2023-04-28 | 中国农业科学院农业资源与农业区划研究所 | Artificial intelligence-based agricultural meteorological remote sensing key parameter joint inversion method |
| CN116295853A (en) * | 2023-03-22 | 2023-06-23 | 大连海事大学 | Multichannel thermal infrared data surface temperature inversion method and device and electronic equipment |
| CN116295853B (en) * | 2023-03-22 | 2024-03-15 | 大连海事大学 | Multichannel thermal infrared data surface temperature inversion method and device and electronic equipment |
| CN116486931A (en) * | 2023-06-21 | 2023-07-25 | 上海航天空间技术有限公司 | Full-coverage atmospheric methane concentration data production method and system coupled with physical mechanism |
| CN116486931B (en) * | 2023-06-21 | 2023-08-29 | 上海航天空间技术有限公司 | Full-coverage atmospheric methane concentration data production method and system coupled with physical mechanism |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114581791A (en) | Inversion method and system for atmospheric water vapor content based on MODIS data | |
| Mao et al. | A practical split‐window algorithm for retrieving land‐surface temperature from MODIS data | |
| Li et al. | Satellite-derived land surface temperature: Current status and perspectives | |
| Francois et al. | Analytical parameterization of canopy directional emissivity and directional radiance in the thermal infrared. Application on the retrieval of soil and foliage temperatures using two directional measurements | |
| CN106547840A (en) | A kind of parsing of global three-dimensional atmospheric data and management method | |
| CN105608293B (en) | Merge Biomass retrieval method and system on the woodland of spectrum and textural characteristics | |
| Yao et al. | LAI retrieval and uncertainty evaluations for typical row-planted crops at different growth stages | |
| CN101738620B (en) | Method by utilizing passive microwave remote sensing data AMSR-E (Advanced Microwave Scanning Radiometer-EOS ) to invert surface temperature | |
| CN104483663B (en) | A kind of target in hyperspectral remotely sensed image Atmospheric Correction method and system | |
| Vereecken et al. | Characterization of crop canopies and water stress related phenomena using microwave remote sensing methods: A review | |
| Yao et al. | A simple temperature domain two‐source model for estimating agricultural field surface energy fluxes from Landsat images | |
| CN101295022A (en) | Method for ground surface temperature and emissivity inversion by remote sensing data ASTER | |
| Jia et al. | A practical algorithm to infer soil and foliage component temperatures from bi-angular ATSR-2 data | |
| CN105204024B (en) | Method of the microwave remote sensing surface temperature to thermal infrared remote sensing skin face temperature transition | |
| Cheng et al. | Generating the 30-m land surface temperature product over continental China and USA from landsat 5/7/8 data | |
| Czajkowski et al. | Thermal remote sensing of near-surface water vapor | |
| Tao et al. | Soil moisture retrieval from SAR and optical data using a combined model | |
| CN103743679B (en) | A kind of method accelerating greenhouse gases vertical column retrieving concentration speed | |
| CN113486581A (en) | Machine learning earth surface temperature estimation method, system, storage medium and equipment based on knowledge driving and logical reasoning | |
| Ermida et al. | A multi-sensor approach to retrieve emissivity angular dependence over desert regions | |
| CN101655564A (en) | Method for inversing surface temperature and emissivity from MODIS data | |
| Rautiainen et al. | On the relationship of canopy LAI and photon recollision probability in boreal forests | |
| Czajkowski et al. | Estimating environmental variables using thermal remote sensing | |
| CN115269549A (en) | Atmospheric water vapor inversion method coupling physics-statistics-deep learning | |
| CN109029735A (en) | A kind of Land surface emissivity calculation method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20220603 |