All-sky cloud imaging real-time simulation method of three-dimensional structure

Abstract

The invention discloses an all-sky cloud imaging real-time simulation method of a three-dimensional structure, which is used for realizing real-time rapid simulation of a cloud image with set spatial resolution per hour, and comprises the following steps: obtaining basic geographic information data of a research area; carrying out background field data fusion; interpolating the ground meteorological data collected in real time to the grid points of the research area by adopting an interpolation method; obtaining surface meteorological analysis and background field data, and realizing three-dimensional temperature analysis according to a static relation; obtaining an initial guessing field of three-dimensional cloud analysis, and obtaining a three-dimensional cloud analysis field in combination with three-dimensional temperature analysis; an initial three-dimensional water vapor analysis field is obtained according to the background field data, and three-dimensional water vapor analysis is realized by performing optimization iteration solution on a target function; obtaining soil humidity analysis according to the surface meteorological analysis and the three-dimensional cloud analysis field; according to the surface meteorological analysis, the three-dimensional temperature analysis, the three-dimensional cloud analysis field and the three-dimensional water vapor analysis, derivative product analysis is achieved; and simulating to obtain all-sky imaging data.

Description

Technical Field

[0001] The invention relates to the field of meteorological science and technology, and in particular to a real-time simulation method for all-sky cloud imaging with a three-dimensional structure. Background Art

[0002] Since clouds are a mixture of small droplets or small ice crystals condensed from water vapor floating in the air, they play a role in regulating the balance of radiation and promoting global water and energy cycles in the earth-atmosphere environment system. Cloud observation is of great significance to weather forecasts, climate patterns, aviation and navigation, military activities, industrial and agricultural production, and people's lives. At present, cloud observations are still completed by observers at various meteorological stations through regular visual observations. Affected by the recognition ability and observation level of each observer, cloud observation data is more subjective and less continuous. In addition, since the cloud itself is constantly moving and changing and has not formed a complete and regular geometric shape, it brings certain difficulties to observers. Therefore, quantitative and accurate observation of cloud data is a weak link in atmospheric detection.

[0003] In recent years, with the rapid development of hardware technologies such as charge-coupled devices and the continuous advancement of digital image processing technology, meteorological researchers from all over the world have worked hard to transform manual visual observation of clouds into automatic instrument observation, resulting in ground-based all-sky imagers. For example, the all-sky imager WSI of ARM, the total sky imager TSI of Yankee Environmental Systems, and the all-sky imager ASI of the Spanish Center for Atmospheric Physics Research; in China, there are the ASI, the automatic observation system of the visible light all-sky imager developed by the Institute of Atmospheric Physics of the Chinese Academy of Sciences, the total sky imager TCI developed by the Chinese Academy of Meteorological Sciences, and the ASC100 all-sky imager developed by the Anhui Institute of Optics and Precision Mechanics of the Chinese Academy of Sciences. In order to obtain all-sky images, these air imagers use the ultra-wide-angle fisheye lens as the main optical device. However, the fisheye lens will cause a certain degree of distortion in the image when projecting the image, and there is a certain degree of distortion in reflecting the shape and quantity distribution of the sky clouds compared with the results observed by the human eye; in addition, in order to avoid direct sunlight, these imagers are equipped with sun shielding equipment, and the collected all-sky images have partial missing phenomena. Due to the high cost of the instrument, all-sky imagers have not yet been widely promoted and used in the meteorological industry.

[0004] Some Chinese utility model invention patents disclose the design ideas of all-sky imagers, such as the Chinese utility model "An outdoor all-sky imaging device" with application number 2019222214379, the Chinese utility model "Imaging system for all-sky cloud mapper" with application number 202011116680, the Chinese utility model "A ground-based all-sky imaging device" with application number 2010205083741 and the Chinese utility model "All-sky imager" with application number 2013205584414. Although they have solved the problem of all-sky cloud image obstruction, they cannot effectively solve the bottleneck of quantitative cloud observation in the meteorological industry. Summary of the invention

[0005] The purpose of the present invention is to overcome the defects of the prior art and propose a real-time simulation method for all-sky cloud imaging with a three-dimensional structure.

[0006] In order to achieve the above-mentioned object, the present invention proposes a real-time simulation method for all-sky cloud imaging with a three-dimensional structure, which is used to realize real-time and rapid simulation of cloud images with set spatial resolution hour by hour, and is implemented based on the LAPS mode, including:

[0007] Step 1) Set the map projection and spatial resolution of the study area to obtain basic geographic information data of the study area;

[0008] Step 2) performing background field data fusion to generate a three-dimensional space fusion field and a ground two-dimensional fusion field;

[0009] Step 3) Using the interpolation method, the discrete ground meteorological data collected in real time are interpolated to the grid points in the study area;

[0010] Step 4) Obtain surface meteorological analysis and background field data through the LAPS model, and realize three-dimensional temperature analysis based on the static relationship;

[0011] Step 5) Obtain the initial guess field of 3D cloud analysis through LAPS mode, and obtain the 3D cloud analysis field by combining 3D temperature analysis;

[0012] Step 6) obtaining an initial three-dimensional water vapor analysis field based on the background field data, combining the surface meteorological analysis, the three-dimensional temperature analysis and the three-dimensional cloud analysis field, and modifying the initial three-dimensional water vapor analysis field by optimizing the objective function through iterative solution to achieve three-dimensional water vapor analysis;

[0013] Step 7) obtaining soil moisture analysis through the LAPS model based on the surface meteorological analysis and the three-dimensional cloud analysis field;

[0014] Step 8) Based on the surface meteorological analysis, 3D temperature analysis, 3D cloud analysis field and 3D water vapor analysis, the derivative product analysis is realized through the LAPS model;

[0015] Step 9) simulates and obtains full sky imaging data in three-dimensional space.

[0016] Preferably, the step 3) comprises:

[0017] Step 3-1) for the discrete ground meteorological data collected in real time, the discrete station values ​​are interpolated to the analysis grid points of the study area by the distance weight method of the Gaussian weight function; the ground meteorological data include: ground pressure, ground air temperature, ground dew point temperature and ground wind;

[0018] Step 3-2) For the interpolated surface pressure field, take the average altitude of each observation station in the study area as the reference height, perform two-dimensional variational analysis on the surface pressure field and surface wind field at the reference height, adjust the wind pressure relationship, correct the processed pressure to the ground height of each analysis grid point to obtain the surface pressure; correct it to the sea level to obtain the sea level pressure.

[0019] Preferably, the step 3-1) comprises:

[0020] The distance weighted interpolation method is used to obtain the initial field of discrete ground meteorological data collected in real time, and the value of the observation station position is extracted from the initial field. Satisfy the following formula:

[0021]

[0022] Among them, B (xk,yk) is the kth observation value around the interpolation station, x k ,y k are the observation station coordinates, is the weight function, c is the filter constant, r k is the distance between observation sites, n represents the total number of observations;

[0023] The interpolated value Compared with the observed value B (xk,yk) Subtract and calculate the interpolation error E of each observation value (xk,yk) ;

[0024] Gaussian distance weighted interpolation is performed on the error values ​​of each station, and the interpolated error field is added to the initial interpolation field to obtain the new grid point value. Satisfy the following formula;

[0025]

[0026] in, is the modified weight function, and G is the filtering constant.

[0027] Preferably, the step 4) comprises:

[0028] The surface temperature field, pressure field data and model background field data are obtained through the LAPS model. After quality control of all temperature observations, the influence radius and weight are set, and the Barnes objective analysis method with successive correction is used to interpolate to the analysis grid points in step 3);

[0029] According to the static relationship, the height field is calculated using the atmospheric vertical profile temperature field; then the boundary layer temperature is adjusted according to the temperature field obtained by ground analysis, as well as the potential temperature at different atmospheric heights to ensure that the high-level potential temperature is greater than or equal to the low-level potential temperature;

[0030] Calculate the pressure and height at the top of the boundary layer.

[0031] Preferably, the step 5) comprises:

[0032] Analyze the vertical cloud information reported by the cloud layer in the horizontal direction, obtain the initial guess field of three-dimensional cloud analysis, give the vertical distribution of clouds, and the approximate horizontal analysis field;

[0033] By performing a three-dimensional temperature analysis on the satellite cloud top temperature field, the cloud top height field is calculated, and then the cloud top height field is interpolated to the initial cloud analysis field to determine the cloud top height and the three-dimensional cloud analysis field, and statistical methods are used to eliminate the differences between cloud reports and satellite data;

[0034] Use 3D radar reflectivity to supplement details and obtain cloud physical quantities;

[0035] Cloud physical quantities are corrected using satellite visible light data.

[0036] Preferably, the step 6) comprises:

[0037] The initial three-dimensional water vapor analysis field is obtained based on the background field data;

[0038] Combining the surface meteorological analysis, three-dimensional temperature analysis and three-dimensional cloud analysis field, the initial three-dimensional water vapor analysis field is corrected by optimizing the objective function J through iterative solution to satisfy the following formula:

[0039]

[0040] Among them, S is the switch coefficient for judging whether the data enters the mode calculation, E is the error function describing the observation value or background error, i is the serial number of the atmospheric vertical layer, N is the total number of vertical layers of the atmospheric profile, subscript GEOS is the geostationary satellite imager data, subscript BACK is the initial water vapor analysis field, subscript GPS is the ground-based GPS water vapor data, subscript GVAP is j It is the jth layer of the geostationary satellite layered water vapor product. There are 3 layers in total. j = 1 represents the lowest layer, j = 2 represents the middle layer, and j = 3 represents the highest layer. The subscript CLD is the LAPS cloud analysis field, the subscript SND is the high-altitude sounding data, and ci is the coefficient vector used to adjust the water vapor field; n is the number of satellite channels, L is the spatial weight of each input data; g is the sign function of the vertical layer cloud component; R is the radiance calculated by the forward model; t is the temperature profile; q i is the specific humidity profile; O 3 is the ozone profile; R 0 is the radiance of the satellite water vapor channel; Q GPS is the total water vapor derived from ground-based GPS data; GT(), GC() and G() are cloud control functions; P is the function that converts pressure coordinates to σ coordinates; Q GVAP is the amount of water vapor in the layer at the σ coordinate derived from the geostationary satellite data, q c is the specific humidity when saturated with water vapor; mr is the water vapor mixing ratio;

[0041] The assimilated water vapor data are obtained, including: three-dimensional specific humidity field, relative humidity field and total precipitable water field.

[0042] Preferably, the soil moisture analysis in step 7) obtains three layers of soil moisture data including 0-15 cm, 15-30 cm and 30-90 cm, as well as two-dimensional snow cover and associated soil moisture data.

[0043] Preferably, the derivative product analysis in step 8) includes obtaining a three-dimensional cloud water content product, and the three-dimensional cloud water content product includes six data sets of three-dimensional water clouds, ice clouds, snow accumulation, precipitation, precipitable ice and hydrometeor content.

[0044] Preferably, the step 9) comprises:

[0045] Step 9-1) Calculate the atmospheric light parameter L' obtained by combining the process simulation in three-dimensional space with the light emitted by the sun, moon and city lights, which shines on the clear sky, water vapor condensation nuclei, clouds and aerosol particles, after single or multiple scattering, according to the ray tracing method λ,clear :

[0046]

[0047] Where θ is the scattering angle, P(θ) is the phase function, and τ s is the optical thickness of the incident light from the light source to the scattering point, τ 0 is the optical thickness of the scattered light from the scattering point to the observer;

[0048] Step 9-2) simulates and calculates the radiance L observed in the red, green, and blue bands, then converts the radiance of the three bands into the CIE-xyz color space, and finally converts the CIE-xyz color space into the RBG color space through the conversion matrix to obtain a true color image of normal vision.

[0049] Preferably, the step 9-2) comprises:

[0050] Convolve the spectral radiance with the CIE tristimulus color matching response function to generate the color perception under normal human photopic vision. Each pixel of the image specifies the perceived color in the xyz color space, where chromaticity is represented by x and y values, brightness is represented by z value, and x+y+z=1;

[0051] Apply a 3×3 conversion matrix to convert the xyz image into RGB color space, and the corresponding r, g, and b values ​​satisfy the following formula:

[0052]

[0053] By using γ correction, full-sky imaging data in three-dimensional space is obtained.

[0054] Compared with the prior art, the advantages of the present invention are:

[0055] Based on meteorological observation data, geostationary satellite data and model reanalysis data, the present invention invents a simulation method for all-sky imaging with the help of the LAPS (Local Analysis and Prediction System) model, which can achieve fast simulation of hourly, 500-meter spatial resolution cloud images, solve the problems of large subjectivity of manual visual observation, ultra-wide-angle fisheye lens imaging projection distortion of ground-based all-sky imagers, missing images due to sun occlusion, expensive instruments and slow promotion and use, and provide a new technology for cloud observation in meteorological departments. The simulated cloud images can achieve application benefits in the fields of meteorological disaster prevention and mitigation, weather forecasting, meteorological services, etc., and better serve people's production and life. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This is the technical process of the three-dimensional structure of the full sky cloud imaging simulation of the present invention;

[0057] Figure 2 is the temperature distribution map of the typical pressure layer in the study area;

[0058] Figure 3 is the distribution map of boundary layer thickness in the study area;

[0059] Figure 4 This is the distribution map of the maximum cloud coverage in the study area;

[0060] Figure 5 is the cloud coverage map of the typical pressure layer in the study area;

[0061] Figure 6 It is the spatial distribution map of total water vapor in the study area;

[0062] Figure 7 It is the specific humidity distribution map of the typical pressure layer in the study area;

[0063] Figure 8 is the relative humidity distribution map of the typical pressure layer in the study area;

[0064] Fig. 9 It is the soil moisture distribution map of the study area from top to bottom in three layers;

[0065] Fig.10 It is the two-dimensional soil moisture distribution map of the study area;

[0066] Fig.11 This is a typical spatial distribution map of ice clouds in the atmospheric pressure layer;

[0067] Fig.12 This is a typical spatial distribution map of water clouds in the pressure layer;

[0068] Fig.13 It is a full disk image generated by simulating the all-sky imager;

[0069] Fig.14 It is a full-width image generated by simulating the all-sky imager. DETAILED DESCRIPTION

[0070] Since numerical model forecasting and simulation is a commonly used method in the meteorological industry, the present invention will, with the help of the LAPS model of the Regional Analysis and Forecasting System, utilize a variety of observation data from reanalysis data, satellites, radars, conventional ground observations, and high-altitude detection to invent a three-dimensional all-sky cloud imaging simulation technology to solve the problems of the current cloud observation, such as the high subjectivity of manual visual observation, the projection distortion of ultra-wide-angle fisheye lens imaging of ground-based all-sky imagers, the missing of solar obstruction images, and the expensive and slow promotion of instruments, thus providing a new technology for cloud observation in the meteorological department.

[0071] In view of the problems existing in current cloud observation, the present invention proposes a real-time simulation method for all-sky cloud imaging with a three-dimensional structure, which can overcome the shortcomings of the above-mentioned existing observation methods.

[0072] like Figure 1 As shown, the following steps are included:

[0073] (1) Regionalization of the study area: Setting map projection and spatial resolution, and producing basic geographic information data of the study area.

[0074] (2) Background field data fusion: The large-scale low-spatial-resolution background field data of the model reanalysis is interpolated to the fine grid size of the localization using the Barnes objective analysis method, using bilinear interpolation in the horizontal direction and linear interpolation of the height coordinate in the vertical direction. Point-to-point linear interpolation of grid points is used in time, that is, when the forecast fields of the two previous and subsequent time periods output by the model are interpolated with the background fields at the time between them, the values ​​on the grid points corresponding to the two forecast fields are linearly interpolated in time to obtain the grid point values ​​at the required time. Finally, a three-dimensional spatial fusion field and a two-dimensional ground fusion field in Netcdf format are generated.

[0075] (3) Fusion of surface meteorological data: The Barnes interpolation method is used, that is, the discrete station values ​​are interpolated to the analysis grid points through the distance weight method of the Gaussian weight function. It is specifically divided into two steps: first, a distance weight interpolation method is used to obtain a relatively smooth initial field from the data of the discrete meteorological stations, and the value of the observation station location is extracted from the initial field. The interpolation method is as follows:

[0076]

[0077] Among them, B (xk,yk) is the kth observation value around the interpolation station, x k ,y k are the observation station coordinates, is the weight function, c is the filter constant, r k The distance between the observation sites. Then the interpolated value Compared with the observed value B (xk,yk) Subtract and calculate the interpolation error of each observation The error values ​​of each station are interpolated with Gaussian distance weights, and the interpolated error field is added to the initial interpolation field to obtain the new grid point value.

[0078]

[0079] in: is the weight function for correction, and G is the filter constant. The interpolated surface pressure field needs to be corrected to a reference height, which is generally the average altitude of each observation station in the study area; then the pressure field and surface wind field at the reference height are analyzed by two-dimensional variation to adjust the wind-pressure relationship, and finally the processed pressure is corrected to the ground height of each grid point to obtain the surface pressure; corrected to the sea level to obtain the sea level pressure. The ground data analysis results will produce physical quantities such as surface pressure, sea level pressure, reference height pressure, surface air temperature, surface dew point temperature, surface wind, vorticity, divergence, and potential temperature at the LAPS ground grid point.

[0080] (4) Simulation analysis of three-dimensional temperature

[0081] The three-dimensional temperature analysis algorithm is mainly carried out in the following three steps:

[0082] The first step is to obtain the ground temperature field, pressure field data and model background field data obtained by the ground analysis of the LAPS model. After quality control of all temperature observations, reasonable influence radius and weight are set for various observations, and the Barnes objective analysis method with successive correction is used to interpolate to the fine grid size during localization.

[0083] The second step is to calculate the height field based on the static relationship using the vertical profile temperature field of the atmosphere obtained from the previous analysis; then adjust the boundary layer temperature based on the temperature field obtained from the ground analysis, as well as the potential temperature at different atmospheric heights to ensure that the high-level potential temperature is greater than or equal to the low-level potential temperature.

[0084] The third step is to calculate the pressure and height of the top of the boundary layer.

[0085] (5) Simulation analysis of three-dimensional clouds

[0086] For the simulation analysis of three-dimensional clouds, firstly, the vertical cloud information of the cloud layer report is analyzed in the horizontal direction to obtain the preliminary guess field of the three-dimensional cloud analysis, and the vertical distribution of the cloud and the approximate horizontal analysis field are given; secondly, the cloud top height field is calculated by performing a three-dimensional temperature analysis on the satellite cloud top temperature field, and then the cloud top height field is interpolated to the preliminary guess cloud analysis field to determine the cloud top height and the three-dimensional cloud analysis field, and the statistical method is used to eliminate the difference between the cloud layer report and the satellite data; finally, the three-dimensional radar reflectivity is used to supplement the details and obtain the cloud physical quantity.

[0087] After calculating the preliminary cloud amount field, it is necessary to conduct a three-dimensional cloud amount field analysis, which mainly includes: fusion analysis of conventional observation data, satellite data fusion analysis, radar data fusion analysis, satellite visible light data fusion analysis and corresponding quality control steps.

[0088] In the first step, based on the initial cloud amount and cloud base information provided by conventional observations, a cloud thickness is assumed to obtain the cloud top height and cloud profile. It is usually assumed that the cloud thickness is 1000m, with a buffer zone of 500m above and below. Then, using the Barnes objective analysis method, the weight is the distance function from the grid to the observation point, and the original observations with sparse data are interpolated to the localized grid to produce a continuous cloud field. At this time, the rationality of the cloud existence is judged according to the environmental stability and corrections are made. In this way, the three-dimensional cloud structure of the first stage can be obtained.

[0089] The second step is to use the brightness temperature observed by the satellite 11μm channel for adjustment. First, based on the cloud field obtained by three-dimensional temperature analysis and Barnes interpolation, an expected brightness temperature is calculated; then, based on the difference between the expected brightness temperature and the actual brightness temperature observed by the satellite 11μm channel, the three-dimensional cloud field obtained in the first stage is adjusted. If the brightness temperature observed by the satellite is lower than the expected brightness temperature, the cloud top height and satellite cloud amount are calculated using satellite channel data, and then the satellite calculation results are compared with the cloud bottom height and cloud top height calculated by ground analysis, and the three-dimensional cloud amount is corrected. The area of ​​high clouds is based on satellite observation data, while the area of ​​low clouds mainly depends on sounding and ground observation data. Finally, a brightness temperature value is calculated using the analyzed cloud field, which is compared with the brightness temperature value of the 11μm channel, and the cloud amount is further adjusted to further balance the two.

[0090] The third step is to use radar reflectivity to correct cloud cover. The radar reflectivity observed by multiple radars is projected onto localized grid points, and the reflectivity of multiple radars at each grid point is obtained by distance-weighted average to obtain a reflectivity value. If there is a cloud base in the previous analysis, and the radar echo exceeds a certain threshold above the cloud base, the cloud cover at this grid point is set to 1.

[0091] The fourth step is to use satellite visible light data to correct cloud cover. Since visible light satellite data has a high spatial resolution, it can enhance the description of low clouds and see thin and sparse clouds. Therefore, the visible light data is interpolated onto a localized grid, and then the cloud brightness value is normalized and corrected according to the sunlight and satellite observation geometry to obtain cloud cover data with uniform brightness. If the cloud corrected by radar reflectivity in the third step is higher than the cloud corrected by satellite visible light, the cloud corrected by satellite visible light data shall prevail.

[0092] (6) Three-dimensional water vapor simulation analysis

[0093] The main process of three-dimensional water vapor analysis is to first calculate an initial three-dimensional water vapor analysis field using the model background field data; then a one-dimensional variational assimilation method is used to integrate high-altitude data, satellite water vapor products, ground-based GPS total water vapor, LAPS cloud analysis results, and satellite imager visible-infrared channel data. According to the following formula, the initial water vapor field is corrected by optimizing the iterative minimum value of the objective function J;

[0094]

[0095] Where i is the number of the vertical layer of the atmosphere, N is the total number of vertical layers of the atmospheric profile, with a maximum of 40; subscript GEOS is the geostationary satellite imager data, subscript BACK is the initial water vapor analysis field, subscript GPS is the ground-based GPS water vapor data, subscript GVAP is the geostationary satellite layered water vapor product, with a total of 3 layers, subscript CLD is the LAPS cloud analysis field, subscript SND is the high-altitude sounding data, E is the error function that describes the observation value or background error, and c i is the coefficient vector used to adjust the water vapor field; S is the switch coefficient for judging whether the data enters the model calculation, n is the number of satellite channels, L is the spatial weight of each input data; g is the sign function of the vertical layer cloud component; R is the radiance calculated by the forward model; t is the temperature profile; q is the specific humidity profile; O 3 is the ozone profile; R 0 is the radiance of the satellite water vapor channel; Q GPS is the total water vapor derived from ground-based GPS data; j is the number of layers of geostationary satellite water vapor products, with a maximum of 3, with 1 representing the lowest layer, 2 the middle layer, and 3 the highest layer; GT, GC, and G are cloud control functions; P is the function that converts pressure coordinates into σ coordinates; Q GVAP is the amount of water vapor in the layer at the σ coordinate derived from the geostationary satellite data, q c is the specific humidity when saturated with water vapor; mr is the water vapor mixing ratio. Finally, the assimilated water vapor data are obtained, including: three-dimensional specific humidity field, relative humidity field and atmospheric total precipitable water field, which can be used as input parameters for all-sky imager data simulation.

[0096] (7) Simulation analysis of soil moisture

[0097] The evaporation of surface soil moisture will form atmospheric water vapor, which in turn affects cloud formation. Using the results of surface analysis and cloud analysis, the soil moisture analysis program is run in the LAPS mode to obtain three layers of soil moisture data including 0-15cm, 15-30cm and 30-90cm, as well as two-dimensional snow cover and attached soil moisture data.

[0098] (8) Analysis of derivative products

[0099] Based on the results of surface analysis, 3D temperature analysis, 3D cloud analysis and 3D water vapor analysis, many products related to cloud, wind, atmospheric stability and fire weather can be derived by running the derivative product program in the LAPS mode. One of the products involved in the simulation of all-sky imaging data is the 3D cloud water content product. The 3D cloud water content product includes 6 data sets: 3D water cloud, ice cloud, snow accumulation, precipitation, precipitable ice and hydrometeor content.

[0100] (9) Simulation of all-sky imaging data

[0101] The key technology of simulating the data of the all-sky imager is to calculate in three-dimensional space based on the ray tracing method, tracing the light emitted by light sources such as the sun, moon and city lights, irradiating on particles such as clear sky, water vapor condensation nuclei, clouds and aerosols, and combining the atmospheric light parameters obtained by the process simulation after single or multiple scattering, such as optical thickness τ, attenuation coefficient β, scattering coefficient b, single scattering albedo α, scattering asymmetry factor g, phase function P, etc. The radiance L observed by the imager in the red, green and blue bands is simulated and calculated, and then the radiance of the three bands is converted into the CIE-xyz color space, and finally the CIE-xyz color space is converted into the RBG color space through the conversion matrix to obtain a true color image of normal vision.

[0102] In the clear sky ray tracing process, the effects of Rayleigh single scattering and stratospheric aerosols are mainly considered. Therefore, the calculation of sky radiance can be carried out using a relatively coarse angular grid model. The tracing algorithm from the light radiated by the sun to the observer or imager can be expressed as:

[0103]

[0104] Where: θ is the scattering angle, P(θ) is the phase function, τ s is the optical thickness of the incident light from the light source to the scattering point, τ 0 It is the optical thickness of scattered light from the scattering point to the observer. The clear sky radiance calculated here includes the scattering of the entire atmospheric molecules and the scattering of aerosols at the top of the troposphere.

[0105] According to the previous algorithm, we can calculate the sky radiance image. In order to obtain an image that can be displayed by human vision or computer, the following three steps are required to generate an RGB true color image.

[0106] ①Convolve the spectral radiance with the CIE tristimulus color matching response function to generate the color perception under normal human photopic vision. Each pixel of the image specifies the perceived color in the xyz color space, the chromaticity is represented by the xy value, the brightness is represented by the z value, and x+y+z=1, and the xyz chromaticity value represents the perception of the three primary colors. For example:

[0107] Pure Rayleigh scattered blue sky, chromaticity value x is 0.235, y is 0.235.

[0108] ② Apply a 3×3 transformation matrix to convert the xyz image into RGB color space

[0109]

[0110] ③ Gamma correction of color space

[0111] In order to make the generated RGB image match the nonlinear brightness ratio of the display, a γ correction of 2.2 (approximately an exponential function) can be used. The brightness of the corrected image can be directly proportional to the brightness of the natural scene, obtaining a true contrast and avoiding unrealistic color saturation. Without correction, both the contrast and brightness are wrong.

[0112] The technical solution of the present invention is described in detail below with reference to the accompanying drawings and embodiments.

[0113] Example

[0114] An embodiment of the present invention provides a real-time simulation method for all-sky cloud imaging of a three-dimensional structure, which specifically includes:

[0115] (1) Regionalization of the study area

[0116] The Beijing South Suburbs Meteorological Observation Station (39.8N and 116.47E) was taken as the study area. When regionalizing, the map projection mode was set to Lambert, the spatial resolution was 500m, and the input files such as surface, topography, ground temperature, and albedo were imported. The localization script file was run in the LAPS mode. After regionalizing the study area, information including the scope of the study area, surface utilization type, surface albedo, topography, soil type, soil temperature, slope, etc. was generated, laying the foundation for other data fusion and simulation analysis in the LAPS mode.

[0117] (2) Background field data fusion

[0118] The American NCEP / GFS forecast data with a spatial resolution of 0.25° and a time interval of 6 hours was used as the background field data. The background field fusion program was run in the LAPS mode to complete the fusion of the background field data in the study area, and generate hourly three-dimensional data fusion fields and ground data fusion fields in netcdf format. The three-dimensional data fusion field contains 43 layers of background field height, temperature, specific humidity, u-wind speed, v-wind speed, and m-wind speed data; the ground data fusion field contains the background field ground u-wind speed, v-wind speed, temperature, dew point temperature, sea level pressure, ground pressure, specific humidity, and precipitation data.

[0119] (3) Analysis of surface meteorological data

[0120] Using ground meteorological data and sounding data, the ground analysis program is run in the LAPS mode to correct the ground data lgb fused with background field and the three-dimensional data lga fused with background field. The generated surface analysis result file lsx contains data such as uv wind speed, surface temperature, surface dew point, relative humidity, sea level pressure, temperature advection, potential temperature, surface pressure, vorticity, mixing ratio, convergence, divergence, visibility, etc., which can be used for other analyses in the LAPS mode.

[0121] (4) Simulation analysis of three-dimensional temperature

[0122] When performing 3D temperature analysis in LAPS mode, the input parameters mainly include: background field fusion data, surface temperature data for surface analysis, radiosonde wind temperature data, etc.; run the 3D temperature analysis program temp.exe, and the results will generate a 3D temperature grid file lt1, a 2D boundary layer thickness file pbl, and a temperature observation file tmg for internal drawing, which can be used as input parameters for 3D cloud and water vapor analysis. Figure 2 Shown is the temperature distribution of a typical pressure layer in the study area.

[0123] (5) Simulation analysis of three-dimensional clouds

[0124] The results of three-dimensional cloud simulation analysis are important input parameters for all-sky imager data simulation. When performing three-dimensional cloud analysis in the LAPS mode, the input parameters mainly include: background field fusion results lga and lgb, conventional observation data fusion results lso, satellite observation data fusion results lvd, including visible light channel data, 3.9μm and 11μm channel brightness temperature data, three-dimensional radar reflectivity, three-dimensional temperature field and height field. Three-dimensional cloud simulation analysis can obtain three-dimensional cloud amount field, cloud classification, cloud water content product lc3, three-dimensional radar reflectivity product lps, two-dimensional cloud / snow cover product lcv, and cloud base and cloud top height products lcb, which are input parameters for three-dimensional water vapor analysis and all-sky imager data simulation. Figure 3 Shown is the distribution map of boundary layer thickness in the study area; Figure 4 This is the distribution map of the maximum cloud coverage in the study area; Figure 5 This is a map of the cloud coverage of a typical pressure layer in the study area.

[0125] (6) Three-dimensional water vapor simulation analysis

[0126] Three-dimensional water vapor analysis is to assimilate and analyze background field fusion data, ground meteorological data, satellite brightness temperature and water vapor data, ground-based GPS water vapor and other information to obtain a three-dimensional water vapor field. The data input in the LAPS water vapor analysis include: the specific humidity field of the background model, the three-dimensional pressure-temperature data as a result of temperature analysis, the three-dimensional cloud amount field as a result of cloud analysis, the pressure, temperature and dew point field as a result of ground analysis, the satellite's atmospheric precipitable water product, the satellite's visible light channel reflectivity, 3.9μm and 11μm channel brightness temperature, GPS's total atmospheric precipitable water, sounding data, etc. The results of the three-dimensional water vapor analysis include: three-dimensional specific humidity lq3, relative humidity field lh3 and atmospheric total precipitable water field lh4, which can be used as input parameters for all-sky imager data simulation. Such as Figure 6 The figure shows the spatial distribution of total water vapor in the study area; Figure 7 The following is the specific humidity distribution diagram of the typical pressure layer in the study area; Figure 8 Shown is the relative humidity distribution of a typical pressure layer in the study area.

[0127] (7) Simulation analysis of soil moisture

[0128] The evaporation of surface soil moisture will form atmospheric water vapor, which in turn affects cloud formation. In the LAPS model, using the surface analysis results lsx and cloud analysis results lcv data as input parameters, by running the soil moisture analysis program, three layers of soil moisture data lm1, 0-15cm, 15-30cm and 30-90cm, as well as two-dimensional snow cover and attached soil moisture data lm2 can be obtained. Fig. 9 The following is the soil moisture distribution map of the study area from top to bottom in three layers; Fig.10 Shown is a two-dimensional soil moisture distribution map of the study area;

[0129] (8) Analysis of derivative products

[0130] Based on the results of surface analysis, 3D temperature analysis, 3D cloud analysis and 3D water vapor analysis, running the derivative product program in the LAPS mode can derive many products related to clouds, winds, atmospheric stability and fire weather. Among them, the 3D cloud water content product lwc is a necessary product for participating in the simulation of all-sky imager data. The lwc product contains 6 data sets: 3D water cloud, ice cloud, snow cover, precipitation, precipitable ice and hydrometeor content. Fig.11 The figure shows the spatial distribution of ice clouds in a typical atmospheric pressure layer; Fig.12 Shown is a typical spatial distribution map of water clouds in the pressure layer.

[0131] (9) Simulation of all-sky imaging data

[0132] Using the results of the previous 8 steps, the full disk image and full width image of the all-sky imager are simulated. Fig.13 , Fig.14shown.

[0133] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention is described in detail with reference to the embodiments, it should be understood by those skilled in the art that any modification or equivalent replacement of the technical solutions of the present invention does not depart from the spirit and scope of the technical solutions of the present invention and should be included in the scope of the claims of the present invention.

Claims

1. A real-time simulation method for all-sky cloud imaging with a three-dimensional structure, which is used to achieve real-time and rapid simulation of cloud images with a set spatial resolution hour by hour, based on the LAPS mode, including: Step 1) Set the map projection and spatial resolution of the study area to obtain basic geographic information data of the study area; Step 2) performing background field data fusion to generate a three-dimensional space fusion field and a ground two-dimensional fusion field; Step 3) Using the interpolation method, the discrete ground meteorological data collected in real time are interpolated to the grid points in the study area; Step 4) Obtain surface meteorological analysis and background field data through the LAPS model, and realize three-dimensional temperature analysis based on the static relationship; Step 5) Obtain the initial guess field of 3D cloud analysis through LAPS mode, and combine it with 3D temperature analysis to obtain the 3D cloud analysis field; Step 6) obtaining an initial three-dimensional water vapor analysis field based on the background field data, combining the surface meteorological analysis, the three-dimensional temperature analysis and the three-dimensional cloud analysis field, and modifying the initial three-dimensional water vapor analysis field by optimizing the objective function through iterative solution to achieve three-dimensional water vapor analysis; Step 7) obtaining soil moisture analysis through the LAPS model based on the surface meteorological analysis and the three-dimensional cloud analysis field; Step 8) Based on the surface meteorological analysis, 3D temperature analysis, 3D cloud analysis field and 3D water vapor analysis, the derivative product analysis is realized through the LAPS model; Step 9) simulates and obtains full sky imaging data in three-dimensional space.

2. The method for real-time simulation of full sky cloud imaging of a three-dimensional structure according to claim 1, characterized in that: The step 3) comprises: Step 3-1) for the discrete ground meteorological data collected in real time, the discrete station values ​​are interpolated to the analysis grid points of the study area by the distance weight method of the Gaussian weight function; the ground meteorological data include: ground pressure, ground air temperature, ground dew point temperature and ground wind; Step 3-2) For the interpolated surface pressure field, take the average altitude of each observation station in the study area as the reference height, perform two-dimensional variational analysis on the surface pressure field and surface wind field at the reference height, adjust the wind pressure relationship, correct the processed pressure to the ground height of each analysis grid point to obtain the surface pressure; correct it to the sea level to obtain the sea level pressure.

3. The method for real-time simulation of full sky cloud imaging of a three-dimensional structure according to claim 2, characterized in that: The step 3-1) comprises: The distance weighted interpolation method is used to obtain the initial field of discrete ground meteorological data collected in real time, and the value of the observation station position is extracted from the initial field. Satisfy the following formula: Among them, B (xk,yk) is the kth observation value around the interpolation station, x k ,y k are the observation station coordinates, is the weight function, c is the filter constant, r k is the distance between observation sites, n represents the total number of observations; The interpolated value Compared with the observed value B (xk,yk) Subtract and calculate the interpolation error E of each observation value (xk,yk) ; Gaussian distance weighted interpolation is performed on the error values ​​of each station, and the interpolated error field is added to the initial interpolation field to obtain the new grid point value. Satisfy the following formula; in, is the modified weight function, and G is the filtering constant.

4. The method for real-time simulation of all-sky cloud imaging of a three-dimensional structure according to claim 1, characterized in that: The step 4) comprises: The surface temperature field, pressure field data and model background field data are obtained through the LAPS model. After quality control of all temperature observations, the influence radius and weight are set, and the Barnes objective analysis method with successive correction is used to interpolate to the analysis grid points in step 3); According to the static relationship, the height field is calculated using the atmospheric vertical profile temperature field; then the boundary layer temperature is adjusted according to the temperature field obtained by ground analysis, as well as the potential temperature at different atmospheric heights to ensure that the high-level potential temperature is greater than or equal to the low-level potential temperature; Calculate the pressure and height at the top of the boundary layer.

5. The method for real-time simulation of all-sky cloud imaging of a three-dimensional structure according to claim 1, characterized in that: The step 5) comprises: Analyze the vertical cloud information reported by the cloud layer in the horizontal direction, obtain the initial guess field of three-dimensional cloud analysis, give the vertical distribution of clouds, and the approximate horizontal analysis field; By performing a three-dimensional temperature analysis on the satellite cloud top temperature field, the cloud top height field is calculated, and then the cloud top height field is interpolated to the initial cloud analysis field to determine the cloud top height and the three-dimensional cloud analysis field, and statistical methods are used to eliminate the differences between cloud reports and satellite data; Use 3D radar reflectivity to supplement details and obtain cloud physical quantities; Cloud physical quantities are corrected using satellite visible light data.

6. The method for real-time simulation of all-sky cloud imaging of a three-dimensional structure according to claim 1, characterized in that: The step 6) comprises: The initial three-dimensional water vapor analysis field is obtained based on the background field data; Combining the surface meteorological analysis, three-dimensional temperature analysis and three-dimensional cloud analysis field, the initial three-dimensional water vapor analysis field is corrected by optimizing the objective function J through iterative solution to satisfy the following formula: Among them, S is the switch coefficient for judging whether the data enters the mode calculation, E is the error function describing the observation value or background error, i is the serial number of the atmospheric vertical layer, N is the total number of vertical layers of the atmospheric profile, subscript GEOS is the geostationary satellite imager data, subscript BACK is the initial water vapor analysis field, subscript GPS is the ground-based GPS water vapor data, subscript GVAP is j It is the jth layer of the geostationary satellite layered water vapor product. There are 3 layers in total. j = 1 represents the lowest layer, j = 2 represents the middle layer, and j = 3 represents the highest layer. The subscript CLD is the LAPS cloud analysis field, the subscript SND is the high-altitude sounding data, and c i is the coefficient vector used to adjust the water vapor field; n is the number of satellite channels, L is the spatial weight of each input data; g is the sign function of the vertical layer cloud component; R is the radiance calculated by the forward model; t is the temperature profile; q i is the specific humidity profile; O3 is the ozone profile; R 0 is the radiance of the satellite water vapor channel; Q GPS is the total water vapor derived from ground-based GPS data; GT(), GC() and G() are cloud control functions; P is the function that converts pressure coordinates to σ coordinates; Q GVAP is the amount of water vapor in the layer at the σ coordinate derived from the geostationary satellite data, q c is the specific humidity when saturated with water vapor; mr is the water vapor mixing ratio; The assimilated water vapor data are obtained, including: three-dimensional specific humidity field, relative humidity field and total precipitable water field.

7. The method for real-time simulation of all-sky cloud imaging of a three-dimensional structure according to claim 1, characterized in that: The soil moisture analysis in step 7) obtains three layers of soil moisture data including 0-15 cm, 15-30 cm and 30-90 cm, as well as two-dimensional snow cover and associated soil moisture data.

8. The method for real-time simulation of all-sky cloud imaging of a three-dimensional structure according to claim 1, characterized in that: The derivative product analysis in step 8) includes obtaining a three-dimensional cloud water content product, and the three-dimensional cloud water content product includes six data sets: three-dimensional water cloud, ice cloud, snow accumulation, precipitation, precipitable ice and hydrometeor content.

9. The method for real-time simulation of all-sky cloud imaging of a three-dimensional structure according to claim 1, characterized in that: The step 9) comprises: Step 9-1) Calculate the atmospheric light parameter L' obtained by combining the process simulation in three-dimensional space with the light emitted by the sun, moon and city lights, which shines on the clear sky, water vapor condensation nuclei, clouds and aerosol particles, after single or multiple scattering, according to the ray tracing method λ,clear : Where θ is the scattering angle, P(θ) is the phase function, and τ s is the optical thickness of the incident light from the light source to the scattering point, and τ0 is the optical thickness of the scattered light from the scattering point to the observer; Step 9-2) simulates and calculates the radiance observed in the red, green, and blue bands, then converts the radiance of the three bands into the CIE-xyz color space, and finally converts the CIE-xyz color space into the RBG color space through the conversion matrix to obtain a true color image of normal vision.

10. The method for real-time simulation of all-sky cloud imaging of a three-dimensional structure according to claim 9, characterized in that: The step 9-2) comprises: Convolve the radiance with the CIE tristimulus color matching response function to generate the color perception under normal human photopic vision. Each pixel of the image specifies the perceived color in the xyz color space, with chromaticity represented by x and y values, brightness represented by z value, and x+y+z=1; Apply a 3×3 conversion matrix to convert the xyz image into RGB color space, and the corresponding r, g, and b values ​​satisfy the following formula: By using γ correction, full-sky imaging data in three-dimensional space is obtained.

Images