Fine precipitation measurement method, system, electronic device, and medium

Abstract

The present disclosure relates to the technical field of precipitation measurement, and particularly relates to a fine precipitation measurement method, system, electronic device and medium. The method comprises: creating a high and low spatial resolution water vapor index, constructing a two-dimensional space with the low spatial resolution water vapor index as the horizontal axis and the precipitation index as the vertical axis, determining four vertices of the per-pixel two-dimensional space extreme precipitation index, obtaining the high spatial resolution precipitation index corresponding to the high spatial resolution water vapor index, enhancing remote sensing precipitation according to the high spatial resolution precipitation index, and obtaining the corrected remote sensing fine precipitation measurement. The present disclosure proposes a water vapor index for representing the water vapor content of the cloud layer based on the normalized temperature of the cloud layer and the reflection of the cirrus cloud, uses the normalized precipitation index of the non-normal distribution precipitation to realize the normal law expression of the spatial heterogeneity precipitation, constructs a new two-dimensional space composed of the water vapor index and the precipitation index, linearizes the nonlinear change of the precipitation, and can realize the fine measurement of the precipitation.

Description

Technical Field

[0001] This disclosure relates to the field of precipitation remote sensing measurement technology, and in particular to a method, system, electronic device and medium for fine precipitation measurement. Background Technology

[0002] Precipitation is a crucial component of the land-atmosphere-ocean material cycle, a vital link in regional and global water cycles, and an important indicator for understanding global climate change. As an indispensable parameter in research related to optimal water resource allocation, analysis of land-sea water vapor fluxes, and the driving forces of eco-hydrological cycles, the spatial resolution of precipitation data directly determines the accuracy of analytical results. High spatial resolution precipitation data helps to deeply understand the interaction mechanisms between the water system, ecosystem, surface system, and climate system based on watershed eco-hydrological process models, and contributes to building a more rational decision-making platform for assessing watershed water resource carrying capacity. Therefore, it is essential to conduct fine-grained measurements of spatially heterogeneous precipitation to support the sustainable socio-economic development of watersheds, regions, and even the world.

[0003] Currently, precipitation measurement methods at different spatial scales, including watershed, regional, and global, mainly include spatial interpolation, model simulation, and remote sensing observation. Spatial interpolation uses mathematical interpolation methods to spatially densify precipitation observations from meteorological stations based on surface environmental factors to cover data-free areas without meteorological stations. For areas with a relatively dense concentration of precipitation meteorological stations, spatial interpolation methods can obtain relatively reliable results; however, the interpolation results for sparsely populated areas are less than ideal. Model simulation is based on the physical processes of precipitation generation, development, and dissipation, using environmental factors such as temperature, water vapor, and clouds to drive real-time forecasts of precipitation spatial distribution. However, due to the complexity of the physical relationship between precipitation and environmental factors, the simplification of the climate system in model simulation, and the uncertainty of model input parameters, model simulation of precipitation still has many imperfections. Remote sensing observation can obtain the spatial distribution of precipitation over large areas and has good spatial coverage capabilities; however, the spatial resolution of remote sensing observation is usually low, often failing to meet the needs of detailed simulation in watershed hydrological models, which greatly limits the widespread application of remote sensing precipitation data.

[0004] The spatial heterogeneity and widespread nature of precipitation distribution are the main reasons why it is difficult to conduct precise precipitation observations. Clarifying the spatial variation patterns of precipitation is an effective way to improve the accuracy of precipitation observations. In order to fully explore the coupling relationship between the probability law of precipitation and its spatial variation, to effectively reveal the spatial topological relationship between precipitation heterogeneity and environmental factors, and to explore effective connection technologies between different precipitation measurement methods, this paper provides a precise precipitation measurement method, system, electronic equipment, and medium. Summary of the Invention

[0005] To address the shortcomings of the aforementioned technologies, a method, system, electronic equipment, and medium for fine precipitation measurement are provided.

[0006] In a first aspect, embodiments of this disclosure provide a method for fine measurement of precipitation, comprising the following steps:

[0007] S1: Create water vapor indices with high and low spatial resolution;

[0008] S2: Construct a pixel-by-pixel two-dimensional space with the low spatial resolution water vapor index as the horizontal axis and the precipitation index as the vertical axis;

[0009] S3: Determine the four vertices of the pixel-by-pixel two-dimensional spatial extreme precipitation index;

[0010] S4: Calculate the high spatial resolution precipitation index corresponding to the high spatial resolution water vapor index;

[0011] S5: Enhance remote sensing precipitation based on high spatial resolution precipitation index;

[0012] S6: Obtain the corrected, fine-grained remote sensing precipitation measurements.

[0013] In one embodiment, S1: creating high and low spatial resolution water vapor indices includes the following steps:

[0014] S11: Create a high spatial resolution water vapor index using remote sensing imagery, specifically:

[0015] WVI = (TIR - CIR) / (TIR + CIR)

[0016] In the formula, TIR represents the normalized temperature of the thermal infrared band of the remote sensing image, CIR represents the on-board reflectance of the cirrus cloud band of the remote sensing image, and WVI represents the water vapor index.

[0017] S12: Spatially upscaled WVI to obtain the water vapor index with low spatial resolution, specifically:

[0018] wvi = UP(WVI)

[0019] In the formula, UP represents the spatial upscaling model, which is used to upscale the spatial resolution of WVI to the spatial resolution of remote sensing precipitation data, and wvi represents the low spatial resolution water vapor index obtained by upscaling WVI using UP.

[0020] In one embodiment, S2: constructing a pixel-by-pixel two-dimensional space with the low spatial resolution water vapor index as the horizontal axis and the precipitation index as the vertical axis includes the following steps:

[0021] S21: Calculate the logarithmic intermediate using remote sensing precipitation data, specifically:

[0022]

[0023]

[0024] In the formula, (x, y) represents the spatial location in the remote sensing precipitation data, P(x, y) represents the remote sensing precipitation at location (x, y), c and r represent the number of columns and rows of the remote sensing precipitation data, lg represents the logarithm to the base 10, and O represents the intermediate logarithm calculated using the remote sensing precipitation data.

[0025] S22: Estimating the shape parameter b using remotely sensed precipitation data, specifically:

[0026]

[0027] In the formula, b represents the shape parameter estimated using remote sensing precipitation data;

[0028] S23: Estimate the scale parameter a using remote sensing precipitation data, specifically:

[0029]

[0030] In the formula, a represents the scale parameter estimated using remote sensing precipitation data;

[0031] S24: Convert remote sensing precipitation data into precipitation probability, including:

[0032] a) When the remotely sensed precipitation is 0, the specific probability of precipitation is as follows:

[0033] F = M / c / r

[0034] In the formula, M represents the number of pixels with zero precipitation, and F represents the probability of zero precipitation as perceived by remote sensing.

[0035] b) When the remotely sensed precipitation is greater than 0, the specific probability of precipitation is as follows:

[0036]

[0037]

[0038] In the formula, Γ represents the gamma distribution, F represents the precipitation probability when the remotely sensed precipitation is P(x,y), and f(z) represents the probability distribution model used to describe daily precipitation.

[0039] S25: Calculate the precipitation index from remotely sensed precipitation data, specifically:

[0040]

[0041] In the formula, when F>0.5, S=1, F=1-F, when F≤0.5, S=-1, PI(x,y) represents the precipitation index of P(x,y), 1≤x≤c, 1≤y≤r, and PI represents the precipitation index of the remote sensing precipitation data obtained.

[0042] S26: Construct a pixel-by-pixel two-dimensional space with the low spatial resolution water vapor index (wvi) as the horizontal axis and the precipitation index (PI) as the vertical axis, including the following steps:

[0043] S261: Determine the maximum and minimum values ​​of the water vapor index per pixel spatial range for remotely sensed precipitation on WVI, specifically:

[0044]

[0045] In the formula, WVI_max represents the pixel (X,Y) with any given high spatial resolution water vapor index within the coverage area of ​​the low spatial resolution pixel (x,y), WVI_min represents the maximum value of the high spatial resolution water vapor index within the coverage area of ​​the pixel (x,y), and WVI_min represents the minimum value of the high spatial resolution water vapor index within the coverage area of ​​the pixel (x,y).

[0046] S262: Determine the set of pixel locations on the WVI where the water vapor index falls between WVI_min and WVI_max, specifically:

[0047] Ω (x ,y)={(m,n)|WVI_min≤wvi(m,n)| m∈[1 [,c],n∈[1,r]≤WVI_max}

[0048] In the formula, (m,n) represents the pixel position on wvi where the water vapor index is between WVI_min and WVI_max, and Ω (x (x, y) represents the set consisting of {(m, n)} at pixel (x, y);

[0049] S263: Construct a pixel-by-pixel two-dimensional space with the low spatial resolution water vapor index (wvi) as the horizontal axis and the precipitation index (PI) as the vertical axis, specifically:

[0050]

[0051] In the formula, Indicated by Ω (x,y) The space consisting of low spatial resolution water vapor index and precipitation index at all pixel locations, 1≤x≤c, 1≤y≤r.

[0052] In one embodiment, S3: determining the four vertices of the pixel-by-pixel two-dimensional spatial extreme precipitation index includes the following steps:

[0053] S31: The water vapor index along the horizontal axis will be pixel by pixel. Divide into several subspaces;

[0054] S32: Determines the four vertices of the pixel-by-pixel two-dimensional spatial extreme precipitation index, including:

[0055] The maximum value of the precipitation index within the minimum interval of the water vapor index is determined and denoted as PI_1(wvi min PI max1 ), which represents the maximum possible precipitation when the water vapor index is at its minimum, indicating the maximum value of precipitation variation when the water vapor index is relatively small;

[0056] The minimum precipitation index within the minimum water vapor index range is determined and denoted as PI_2(wvi min PI min2 ), which represents the minimum possible precipitation when the water vapor index is extremely low, indicating the minimum value of precipitation variation when the water vapor index is low;

[0057] The minimum precipitation index within the maximum water vapor index range is determined and denoted as PI_3(wvi max PI min3 ), which represents the minimum possible precipitation when the water vapor index is at its maximum, indicating the minimum value of precipitation variation when the water vapor index is relatively large;

[0058] The maximum value of the precipitation index within the maximum range of the water vapor index is determined and denoted as PI_4(wvi max PI max4 The value of α represents the maximum possible precipitation when the water vapor index is at its maximum, indicating the maximum value of precipitation variation when the water vapor index is relatively large.

[0059] In one embodiment, S4: Calculating the high spatial resolution precipitation index corresponding to the high spatial resolution water vapor index includes the following steps:

[0060] S41: The equation for calculating the variation of the extreme precipitation index at each pixel includes:

[0061] The equation for calculating the variation of the maximum precipitation index at each pixel is as follows:

[0062] ku=(PI max4 -PI max1 ) / (wvi max -wvi min )

[0063] PIU(v)=ku*v+(PI max1 wvi max -PI max4 wvi min ) / (wvi max-wvi min )

[0064] In the formula, v represents the water vapor index independent variable, and PIU(v) represents the equation for the change of the maximum precipitation index at pixel (x,y).

[0065] The equation for calculating the variation of the minimum precipitation index at each pixel is as follows:

[0066] kl=(PI min3 -PI min2 ) / (wvi max -wvi min )

[0067] PIL(v) = kl*v + (PI min2 wvi max -PI min3 wvi min ) / (wvi max -wvi min )

[0068] In the formula, PIL(v) represents the equation for the change of the minimum precipitation index at pixel (x,y);

[0069] S42: The equation for calculating the variation of the precipitation index at each pixel is as follows:

[0070] km=kl+(ku-kl) / (PIU(wvi(m,n))-PIL(wvi(m,n)))*(PI(m,n)-PIL(wvi(m,n)))

[0071] PIM(v)=km*v+PI(m,n)-km*wvi(m,n)

[0072] In the formula, PIM(v) represents the equation for the change of precipitation index at each pixel (x,y);

[0073] S43: Calculate the high spatial resolution precipitation index corresponding to the high spatial resolution water vapor index, specifically:

[0074]

[0075] In the formula, WVI(X,Y) represents the water vapor index at any high spatial resolution location (X,Y) within the coverage area of ​​the low spatial resolution pixel (x,y), PIH(WVI(X,Y)) represents the precipitation index at any high spatial resolution pixel location (X,Y) within the coverage area of ​​the low spatial resolution pixel (x,y), and PIH represents the high spatial resolution precipitation index of the corresponding remote sensing precipitation coverage area.

[0076] In one embodiment, S5: enhancing remote sensing precipitation based on the high spatial resolution precipitation index includes the following steps:

[0077] S51: Calculate the high spatial resolution precipitation probability corresponding to the high spatial resolution precipitation index, including:

[0078] J d+1 =J d -G(J d ) / g(J d )

[0079] G(J)=0.0013J 2 +(0.19-0.0013PIH)J 1.5 +(1.4227-0.19PIH)J+(0.197-1.433PIH)J 0.5 -2.516-PIH

[0080] g(J)=0.0026J+(0.285-0.00195PIH)J 0.5 +(1.4227-0.19PIH)+(0.0985-0.7165PIH)J -0.5

[0081] In the formula, the subscript d represents the iteration number, when J d+1 With J d When the absolute deviation is less than a preset threshold, the iteration stops. When, then J d This is what we seek: to achieve a high spatial resolution precipitation probability. Otherwise, discard J. d Solve using the following formula:

[0082] J d+1 =J d -G(J d ) / g(J d )

[0083] G(J)=0.0013J 2 +(0.19+0.0013PIH)J 1.5 +(1.4227+0.19PIH)J+(0.197+1.433PIH)J 0.5 -2.516+PIH

[0084] g(J)=0.0026J+(0.285+0.00195PIH)J 0.5 +(1.4227+0.19PIH)+(0.0985+0.7165PIH)J -0.5

[0085] When J d+1 With J d When the absolute deviation is less than a preset threshold, the iteration stops. When, then J d This is what we seek: to achieve a high spatial resolution precipitation probability.

[0086] S52: Calculate the high spatial resolution remote sensing precipitation R using the high spatial resolution precipitation probability FL, specifically:

[0087]

[0088] In the formula, the subscript d represents the iteration number, when R d+1 With R d When the absolute deviation is less than a preset threshold, the iteration stops, then R... d This is the desired high spatial resolution remote sensing precipitation R.

[0089] In one embodiment, S6: acquiring the corrected fine measurement of remote sensing precipitation specifically includes:

[0090] P R =R*(P / R) ↑ )

[0091] In the formula, R ↑ This represents the result of upscaling R space to P space resolution, where * indicates R and P / R. ↑ Multiplying corresponding spatial ranges, P R This indicates the corrected, refined remote sensing precipitation measurement.

[0092] Secondly, this disclosure provides a fine precipitation measurement system for performing a fine precipitation measurement method as described in the first aspect above, including an input module, a water vapor index module, a two-dimensional space module, a four-vertex module, a precipitation index module, an enhancement module, a correction module, and an output module.

[0093] The input module is used to collect time-synchronized remote sensing images and remote sensing precipitation observations;

[0094] The water vapor index module is used to perform the creation of high and low spatial resolution water vapor indices as described in the first aspect above.

[0095] The two-dimensional space module is used to perform the above-mentioned first aspect of constructing a pixel-by-pixel two-dimensional space with the low spatial resolution water vapor index as the horizontal axis and the precipitation index as the vertical axis.

[0096] The four-vertex module is used to perform the determination of the four vertices of the pixel-by-pixel two-dimensional spatial extreme precipitation index as described in the first aspect above;

[0097] The precipitation index module is used to perform the above-mentioned first aspect to obtain the high spatial resolution precipitation index corresponding to the high spatial resolution water vapor index.

[0098] The enhancement module is used to perform the enhancement of remote sensing precipitation based on the high spatial resolution precipitation index as described in the first aspect above.

[0099] The correction module is used to perform the above-mentioned first aspect of obtaining the corrected fine measurement of remote sensing precipitation;

[0100] The output module is used to output the process and results of fine precipitation measurement.

[0101] Thirdly, embodiments of this disclosure provide an electronic device for fine precipitation measurement, including a sensor, a memory, and a processor. The sensor is used to acquire remote sensing images and remote sensing precipitation. The memory stores a computer program and the remote sensing images and remote sensing precipitation acquired by the sensor. When the processor executes the computer program to process the remote sensing images and remote sensing precipitation acquired by the sensor, it implements the steps of the fine precipitation measurement method described in the first aspect.

[0102] Fourthly, embodiments of this disclosure provide a computer-readable storage medium for fine precipitation measurement, on which a computer program is stored, which, when executed by a processor, implements the steps of the fine precipitation measurement method described in the first aspect.

[0103] Compared with the prior art, the beneficial effects of this disclosure are as follows:

[0104] 1) Based on cloud normalized temperature and cirrus reflectance, a new index for characterizing cloud water vapor content is proposed, which can easily and quickly quantify cloud water vapor content.

[0105] 2) By normalizing non-normally distributed precipitation, the normality of spatially heterogeneous precipitation was expressed using the precipitation index, which is beneficial for further analysis of changes in the spatial distribution of precipitation;

[0106] 3) A novel two-dimensional space composed of water vapor index and precipitation index is proposed to analyze the spatial variation of precipitation, which can linearize the nonlinear variation of precipitation.

[0107] 4) It can organically combine the advantages of heterogeneous remote sensing images and precipitation data. Through spatial enhancement of remote sensing precipitation, the spatial resolution of remote sensing precipitation data can be greatly improved, and the accuracy of precipitation observation can be significantly improved. Attached Figure Description

[0108] The accompanying drawings illustrate embodiments consistent with this disclosure and are used together with the specification to illustrate the principles of this disclosure. To more clearly explain the technical solutions in the embodiments of this disclosure or the prior art, the drawings used in specific embodiments will be briefly described below. It should be noted that, without creative effort, those skilled in the art can obtain other drawings based on these drawings.

[0109] Figure 1 A flowchart illustrating a fine precipitation measurement method provided in this embodiment of the disclosure;

[0110] Figure 2 A flowchart illustrating another fine precipitation measurement method provided in this embodiment of the disclosure;

[0111] Figure 3 A flowchart illustrating another fine precipitation measurement method provided in this embodiment of the disclosure;

[0112] Figure 4 A flowchart illustrating another fine precipitation measurement method provided in this embodiment of the disclosure;

[0113] Figure 5 A flowchart illustrating another fine precipitation measurement method provided in this embodiment of the disclosure;

[0114] Figure 6 A flowchart illustrating another fine precipitation measurement method provided in this embodiment of the disclosure;

[0115] Figure 7 A flowchart illustrating another fine precipitation measurement method provided in this embodiment of the disclosure;

[0116] Figure 8 This is a schematic diagram of a fine precipitation measurement system provided in an embodiment of the present disclosure. Detailed Implementation

[0117] To more clearly illustrate the beneficial effects and implementation steps of this disclosure, the technical solution of this disclosure will be further described below. It is worth noting that the embodiments and steps within the embodiments of this disclosure can be adjusted or rearranged. This disclosure can also be implemented in ways different from those described herein; therefore, the embodiments shown in the specification are only some embodiments of this disclosure, and not all embodiments.

[0118] In one embodiment, such as Figure 1 As shown, Figure 1 A flowchart illustrating a fine precipitation measurement method provided in this disclosure includes the following steps:

[0119] S1: Create water vapor indices with high and low spatial resolution;

[0120] S2: Construct a pixel-by-pixel two-dimensional space with the low spatial resolution water vapor index as the horizontal axis and the precipitation index as the vertical axis;

[0121] S3: Determine the four vertices of the pixel-by-pixel two-dimensional spatial extreme precipitation index;

[0122] S4: Calculate the high spatial resolution precipitation index corresponding to the high spatial resolution water vapor index;

[0123] S5: Enhance remote sensing precipitation based on high spatial resolution precipitation index;

[0124] S6: Obtain the corrected, fine-grained remote sensing precipitation measurements.

[0125] Based on the above embodiments, in the embodiments of this disclosure, further, such as Figure 2 As shown, S1: Creating water vapor indices with high and low spatial resolution, one possible implementation includes the following steps:

[0126] S11: Create a high spatial resolution water vapor index using remote sensing imagery, specifically:

[0127] WVI = (TIR - CIR) / (TIR + CIR)

[0128] In the formula, TIR represents the normalized cloud temperature in the thermal infrared band of the remote sensing image, CIR represents the on-board reflectance in the cirrus band of the remote sensing image, and WVI represents the water vapor index.

[0129] Water vapor in clouds exhibits strong absorption of electromagnetic waves in the wavelength range of 1.36–1.40 μm, forming absorption troughs within this range. This wavelength range corresponds to the cirrus cloud band in remote sensing imagery and can be represented by the reflectance of the corresponding remote sensing image band. Conversely, water vapor in clouds exhibits strong radiation in the electromagnetic wave wavelength range of 10.6–11.2 μm, including strong infrared radiation within this range. This wavelength range corresponds to the thermal infrared band in remote sensing imagery and can be represented by the thermal infrared temperature of the remote sensing imagery. In satellite imagery, cloud areas often appear white, and areas with thicker cloud cover are generally rainy. Therefore, a water vapor index reflecting the water vapor content of clouds can be constructed using the thermal infrared temperature of clouds and the cirrus cloud reflectance.

[0130] The thermal infrared (TIR) ​​cloud temperature of remote sensing imagery can be obtained by downloading existing remote sensing cloud temperature products from the internet, or by inverting the TIR band of the selected remote sensing imagery. For example, the remote sensing TIR cloud temperature can be selected from the ECO2LSTEv001 product, covering a global range from 50°N to 50°S latitude, with a spatial resolution of 70 meters and a temporal resolution of days, covering the period from July 9, 2018 to the present. The data can be downloaded in HDF5 format from https: / / lpdaac.usgs.gov / products / eco2lstev001 / , and preprocessing such as resampling, cropping, and projection conversion can be performed using dedicated remote sensing image processing software such as ENVI. The examples given in this disclosure are only to illustrate that the TIR cloud temperature of remote sensing imagery can be downloaded from the internet, but are not limited to specific products and websites; those skilled in the art can choose according to the actual situation.

[0131] Cloud temperature (TIR) ​​in the thermal infrared band of remote sensing imagery can also be obtained through inversion using the thermal infrared band of the remote sensing imagery. For example, satellite remote sensing imagery can be selected from the Landsat series, such as images acquired by the Landsat 8 Thermal Infrared Sensor (TIRS). The wavelength range of Landsat 8 TIRS band 10TIRS1 is 10.60–11.19 μm, which falls precisely in the region of strong electromagnetic radiation from cloud water vapor. This imagery can be downloaded free of charge from http: / / glovis.usgs.gov / . The TIR inversion algorithm can be either a single-channel or multi-channel algorithm. The cloud emissivity used in the inversion algorithm can be the MOD11A1 product, which can be downloaded from https: / / ladsweb.modaps.eosdis.nasa.gov / . It is raster data organized in a tiled manner, with global coverage, spatiotemporal resolution of days and 1km, and a starting time of 2000. Preprocessing such as format conversion, reprojection, cropping, and stitching can be completed using the dedicated processing tool HEG of the MODIS product. The examples provided in this disclosure are only used to illustrate that cloud temperature TIR in the thermal infrared band of remote sensing images can be inverted from the thermal infrared band of remote sensing images, and are not intended to specify the selected remote sensing image and / or temperature inversion algorithm. Those skilled in the art can make the selection according to the specific circumstances.

[0132] The CIR (Cirrus Reflectance Index) of remote sensing image cirrus bands can be obtained by downloading existing remote sensing cirrus band reflectance products from the internet, or by inverting the cirrus band of the selected remote sensing image. For example, the remote sensing cirrus band reflectance product can be the MOD021KM product, which covers the globe, has a spatial resolution of 1000 meters, a temporal resolution of 5 minutes, and a time range from 2002 to the present. The data can be downloaded in HDF5 format from https: / / ladsweb.modaps.eosdis.nasa.gov / search / . Preprocessing such as data format conversion, resampling, cropping, and projection conversion can be performed using dedicated remote sensing image processing software such as ENVI or MODIS product-specific processing software HEG. The examples given in this disclosure are only used to illustrate that the CIR of remote sensing image cirrus bands can be downloaded from the internet, and are not intended to limit the selection of remote sensing image cirrus band reflectance products. Those skilled in the art can choose the appropriate product based on the specific circumstances.

[0133] The reflectance CIR of the cirrus band in remote sensing images can also be obtained through inversion using the cirrus band of remote sensing images. For example, the satellite remote sensing image can be a Landsat series image, such as an image acquired by the Landsat 8 Operational Land Imager (OLI). The wavelength range of the Landsat 8 OLI band 9 Cirrus (cirrus band) is 1.360–1.390 μm, which falls precisely in the region where cloud water vapor strongly absorbs electromagnetic waves. This image data can be downloaded free of charge from http: / / glovis.usgs.gov / . The Landsat 8 OLI band 9 Cirrus reflectance is obtained through radiometric calibration, projection transformation, image cropping, and mosaicking of the band 9 Cirrus digital values. The examples given in this disclosure are only used to illustrate that the reflectance CIR of the cirrus band in remote sensing images can be inverted using the cirrus band of remote sensing images, and are not intended to limit the selection of remote sensing images. Those skilled in the art can select other remote sensing images as needed.

[0134] Before calculating the water vapor index, the thermal infrared (TIR) ​​cloud temperature of the remote sensing image needs to be normalized to match the numerical range (0-1) of the cirrus cloud reflectance (CIR) of the remote sensing image. The normalization algorithm can be an extremum adjustment method, which involves subtracting the minimum TIR value from the TIR value and then dividing by the difference between the maximum and minimum TIR values. Because the spatial resolution of the cirrus cloud reflectance (CIR) is usually higher than that of the TIR value, it is necessary to spatially upscale the CIR value to match the spatial resolution of the TIR value. For example, the upscaling method can be a simple averaging method, empirical regression method, geostatistical method, Bayesian method, or others. This disclosure does not specifically specify any particular method, and those skilled in the art can choose according to the actual situation. Then, the water vapor index can be calculated using the normalized TIR value of cloud temperature in the thermal infrared band of the remote sensing image and the CIR reflectance in the cirrus band of the remote sensing image.

[0135] The Water Vapor Index (WVI) can also be obtained using an inversion algorithm. For example, the reflectance ratio of bands 2 and 19 of MOD021KM can be used to estimate the WVI, specifically:

[0136] WVI={[0.02-ln(ρ 19 / ρ2)] / 0.651} 2

[0137] In the formula, ρ 19 ρ2 represents the apparent reflectance of bands 19 and 2 of MOD021KM, respectively.

[0138] The water vapor index (WVI) can also be obtained by inverting CloudSat satellite imagery. CloudSat provides vertical profiles of cloud liquid water and ice water content, while the 2B2CWC2RO product provides information on condensate water content and effective radius on radar-detected profiles. The vertical resolution is 300m, the cloud thickness is 30km, and the spatial resolution is 1.4×2.5km, spanning from 2006 to the present. This information can be downloaded from http: / / CloudSat.cira.colostate.edu / data-ICDlist.php. Based on cloud cover information from the 2B2GEOPROF product, cloud-containing profiles are identified. The 2B2CLDCLASS product is used to determine the cloud type on the profile. Assuming the droplet distribution of liquid or ice phase particles in each cloud mass, the water vapor index is retrieved using assumed prior values ​​and radar observations from 2B2GEOPROF.

[0139] Liquid phase:

[0140] Solid phase:

[0141] In the formula, N T r represents the total particle density. g Or D g σ represents the average geometric radius of liquid water or solid ice. l ρ represents the natural logarithm of the geometric standard deviation. w or ρ i This indicates the density of water or ice.

[0142] The water vapor index can also be obtained using other remote sensing images / products, which will not be listed in this disclosure. Those skilled in the art can choose according to the actual situation.

[0143] S12: Spatially upscaled WVI to obtain the water vapor index with low spatial resolution, specifically:

[0144] wvi = UP(WVI)

[0145] In the formula, UP represents the spatial upscaling model, which is used to upscale the spatial resolution of WVI to the spatial resolution of remote sensing precipitation data, and wvi represents the low spatial resolution water vapor index obtained by upscaling WVI using UP.

[0146] For example, the spatial upscaling model UP can be selected from simple averaging, empirical regression, geostatistical methods, Bayesian methods or others. This disclosure does not specifically specify any, and those skilled in the art can choose according to the actual situation.

[0147] Based on the above embodiments, in some embodiments of this disclosure, further, such as Figure 3 As shown, constructing a pixel-by-pixel two-dimensional space with the low spatial resolution water vapor index as the horizontal axis and the precipitation index as the vertical axis, one possible implementation includes the following steps:

[0148] S21: Calculate the logarithmic intermediate using remote sensing precipitation data, specifically:

[0149]

[0150]

[0151] In the formula, (x, y) represents the spatial location in the remote sensing precipitation data, P(x, y) represents the remote sensing precipitation at location (x, y), c and r represent the number of columns and rows of the remote sensing precipitation data, lg represents the logarithm to the base 10, and O represents the intermediate logarithm calculated using the remote sensing precipitation data.

[0152] For example, remote sensing precipitation data can be selected from the PERSIANN-CCS product of the GEO series satellites, with a spatial resolution of 4 km, a temporal resolution of 30 minutes, a spatial coverage of 60°S-60°N, a time range of 2006 to the present, and units of mm / 30m, which can be downloaded from ftp: / / persiann.eng.uci.edu / pub / GC CS. Alternatively, remote sensing precipitation data can be selected from the CMORPH product, which is a fusion of precipitation data from TMI, SSM / I, AMSU-B, AMSR-E, etc. satellites, with a spatial resolution of 8 km, a temporal resolution of 30 minutes, a time range of 1998 to the present, and units of mm / 30m, which can be downloaded from http: / / rda.ucar.edu / datasets / ds502.0 / . Remote sensing precipitation data can also be acquired using other remote sensing images / products; this disclosure will not list them all, and those skilled in the art can choose according to the actual situation.

[0153] Remote sensing precipitation data and the water vapor index (WVI) acquired by S11 need to be synchronized in time. This time synchronization means that the time of acquisition of the remote sensing imagery or product for the water vapor index should be as close as possible to the time of acquisition of the remote sensing precipitation data. It does not require the acquisition times to be the same. As long as the difference in acquisition time between the selected multi-source remote sensing data is within a suitable time range, it is considered time-synchronized data. For example, if the difference between the acquisition time of the remote sensing precipitation data and the acquisition time of the remote sensing imagery or product for the water vapor index is within one day—that is, if the remote sensing precipitation data is collected on July 15, 2015, and the acquired remote sensing imagery or product falls between 00:00 and 23:59 on July 15, 2015—then the remote sensing precipitation data and the remote sensing imagery or product for the water vapor index are considered time-synchronized data.

[0154] It should be noted that the spatial resolution of the remote sensing imagery or products used to obtain the water vapor index (WVI) must be higher than that of the remote sensing precipitation data. In other words, the high spatial resolution WVI should have higher observational accuracy than the remote sensing precipitation data. Furthermore, the remote sensing precipitation data and the remote sensing imagery or products used to obtain the WVI should cover the same geographic spatial area.

[0155] S22: Estimating shape parameters using remotely sensed precipitation data, specifically:

[0156]

[0157] In the formula, b represents the shape parameter estimated using remote sensing precipitation data;

[0158] S23: Estimating scale parameters using remotely sensed precipitation data, specifically:

[0159]

[0160] In the formula, a represents the scale parameter estimated using remote sensing precipitation data;

[0161] S24: Convert remote sensing precipitation data into precipitation probability, including:

[0162] a) When the remotely sensed precipitation is 0, the specific probability of precipitation is as follows:

[0163] F = M / c / r

[0164] In the formula, M represents the number of pixels with zero precipitation, and F represents the probability of zero precipitation as perceived by remote sensing.

[0165] b) When the remotely sensed precipitation is greater than 0, the specific probability of precipitation is as follows:

[0166]

[0167]

[0168] In the formula, Γ represents the gamma distribution, F represents the precipitation probability when the remotely sensed precipitation is P(x,y), and f(z) represents the probability distribution model used to describe the precipitation.

[0169] S25: Calculate the precipitation index from remotely sensed precipitation data, specifically:

[0170]

[0171] In the formula, when F>0.5, S=1, F=1-F, when F≤0.5, S=-1, PI(x,y) represents the precipitation index of P(x,y), 1≤x≤c, 1≤y≤r, and PI represents the precipitation index of the remote sensing precipitation data obtained.

[0172] S26: Construct a pixel-by-pixel two-dimensional space with the low spatial resolution water vapor index wvi as the horizontal axis and the precipitation index PI as the vertical axis.

[0173] Based on the above embodiments, in some embodiments of this disclosure, further, such as Figure 4 As shown, one possible implementation of S26 includes the following steps:

[0174] S261: Determine the maximum and minimum values ​​of the water vapor index per pixel spatial range for remotely sensed precipitation on WVI, specifically:

[0175]

[0176] In the formula, WVI_max represents the pixel (X,Y) with any given high spatial resolution water vapor index within the coverage area of ​​the low spatial resolution pixel (x,y), WVI_min represents the maximum value of the high spatial resolution water vapor index within the coverage area of ​​the pixel (x,y), and WVI_min represents the minimum value of the high spatial resolution water vapor index within the coverage area of ​​the pixel (x,y).

[0177] In the formula, (x,y) represents the geographic spatial range covered by each pixel of remote sensing precipitation data. WVI(X,Y) represents the pixel location of the high spatial resolution water vapor index (WVI) within the geographic space covered by (x,y). WVI(X,Y) represents the water vapor index value at any given high spatial resolution pixel location (X,Y) within the geographic space covered by (x,y). WVI_max≥WVI(X,Y) means that the water vapor index value at any given high spatial resolution pixel location (X,Y) within the geographic space covered by (x,y) is less than or equal to WVI_max, i.e., WVI_max represents the maximum value of the high spatial resolution water vapor index (WVI) within the geographic space covered by (x,y). WVI_min≤WVI(X,Y) means that the water vapor index value at any given high spatial resolution pixel location (X,Y) within the geographic space covered by (x,y) is greater than or equal to WVI_min, i.e., WVI_min represents the minimum value of the high spatial resolution water vapor index (WVI) within the geographic space covered by (x,y).

[0178] S262: Determine the set of pixel locations on the WVI where the water vapor index falls between WVI_min and WVI_max, specifically:

[0179] Ω (x,y) ={(m,n)|WVI_min≤wvi(m,n)| m∈[1 [,c],n∈[1,r]≤WVI_max}

[0180] In the formula, (m,n) represents the pixel position on wvi where the water vapor index is between WVI_min and WVI_max, and Ω (x,y) Let {x,y} represent the set consisting of {(m,n)} at pixel (x,y);

[0181] In the formula, c represents the number of columns of the low spatial resolution water vapor index wvi, r represents the number of rows of wvi, the subscript m∈[1,c] represents the m-th column of wvi, the subscript n∈[1,r] represents the n-th column of wvi, and WVI_min≤wvi(m,n)| m∈[1,c],n∈[1,r]≤WVI_max represents the pixel position (m,n) in wvi where the water vapor index is between WVI_min and WVI_max, Ω (x,y) This represents the set of all pixel positions (m,n) in wvi for each pixel (x,y) of remotely sensed precipitation;

[0182] S263: Construct a pixel-by-pixel two-dimensional space with the low spatial resolution water vapor index (wvi) as the horizontal axis and the precipitation index (PI) as the vertical axis, specifically:

[0183]

[0184] In the formula, Indicated by Ω (x,y) The space consisting of low spatial resolution water vapor index and precipitation index at all pixel locations, 1≤x≤c, 1≤y≤r;

[0185] In the formula, (m,n)∈Ω (x,y) Represents the set of pixels Ω (x,y) In the context of pixel positions, [wvi(m,n),PI(m,n)] represents a data pair consisting of the wvi value and PI value at position (m,n). Represents the set of cells Ω (x,y) The space for cell location (x,y) is formed by data pairs consisting of the wvi and PI values ​​at all cell locations.

[0186] Based on the above embodiments, in some embodiments of this disclosure, further, such as Figure 5 As shown, to determine the four vertices of the pixel-by-pixel two-dimensional spatial extreme precipitation index, one possible approach includes the following steps:

[0187] S31: The water vapor index along the horizontal axis will be pixel by pixel. Divide into several subspaces;

[0188] Based on The WVI_min and WVI_max of the water vapor index in space divide the horizontal water vapor index into several equal intervals. The water vapor index falling in each interval and its corresponding precipitation index constitute the subspace corresponding to the water vapor index sub-interval.

[0189] S32: Determines the four vertices of the pixel-by-pixel two-dimensional spatial extreme precipitation index, including:

[0190] The maximum value of the precipitation index within the minimum interval of the water vapor index is determined and denoted as PI_1(wvi min PI max1 ), which represents the maximum possible precipitation when the water vapor index is at its minimum, indicating the maximum value of precipitation variation when the water vapor index is relatively small;

[0191] The minimum precipitation index within the minimum water vapor index range is determined and denoted as PI_2(wvi min PI min2 ), which represents the minimum possible precipitation when the water vapor index is extremely low, indicating the minimum value of precipitation variation when the water vapor index is low;

[0192] The minimum precipitation index within the maximum water vapor index range is determined and denoted as PI_3(wvi max PI min3 ), which represents the minimum possible precipitation when the water vapor index is at its maximum, indicating the minimum value of precipitation variation when the water vapor index is relatively large;

[0193] The maximum value of the precipitation index within the maximum range of the water vapor index is determined and denoted as PI_4(wvi max PI max4 The value of α represents the maximum possible precipitation when the water vapor index is at its maximum, indicating the maximum value of precipitation variation when the water vapor index is relatively large.

[0194] Based on the above embodiments, in some embodiments of this disclosure, further, such as Figure 6 As shown, one possible way to obtain the high spatial resolution precipitation index corresponding to the high spatial resolution water vapor index includes the following steps:

[0195] S41: The equation for calculating the variation of the extreme precipitation index at each pixel includes:

[0196] The equation for calculating the variation of the maximum precipitation index at each pixel is as follows:

[0197] ku=(PI max4 -PI max1 ) / (wvi max -wvi min )

[0198] PIU(v)=ku*v+(PI max1 wvi max -PI max4 wvi min ) / (wvi max -wvi min )

[0199] In the formula, v represents the water vapor index independent variable, and PIU(v) represents the equation for the change of the maximum precipitation index at pixel (x,y).

[0200] The equation for calculating the variation of the minimum precipitation index at each pixel is as follows:

[0201] kl=(PI min3 -PI min2 ) / (wvi max -wvimin )

[0202] PIL(v) = kl*v + (PI min2 wvi max -PI min3 wvi min ) / (wvi max -wvi min )

[0203] In the formula, PIL(v) represents the equation for the change of the minimum precipitation index at pixel (x,y);

[0204] S42: The equation for calculating the variation of the precipitation index at each pixel is as follows:

[0205] km=kl+(ku-kl) / (PIU(wvi(m,n))-PIL(wvi(m,n)))*(PI(m,n)-PIL(wvi(m,n)))

[0206] PIM(v)=km*v+PI(m,n)-km*wvi(m,n)

[0207] In the formula, PIM(v) represents the equation for the change of precipitation index at each pixel (x,y);

[0208] S43: Calculate the high spatial resolution precipitation index corresponding to the high spatial resolution water vapor index, specifically:

[0209]

[0210] In the formula, WVI(X,Y) represents the water vapor index at any high spatial resolution location (X,Y) within the coverage area of ​​the low spatial resolution pixel (x,y), PIH(WVI(X,Y)) represents the precipitation index at any high spatial resolution pixel location (X,Y) within the coverage area of ​​the low spatial resolution pixel (x,y), and PIH represents the high spatial resolution precipitation index within the corresponding remote sensing precipitation coverage area.

[0211] Based on the above embodiments, in some embodiments of this disclosure, further, such as Figure 7 As shown, one possible way to enhance remote sensing precipitation based on the high spatial resolution precipitation index includes the following steps:

[0212] S51: Calculate the high spatial resolution precipitation probability corresponding to the high spatial resolution precipitation index, including:

[0213] J d+1 =J d -G(J d ) / g(J d )

[0214] G(J)=0.0013J 2 +(0.19-0.0013PIH)J 1.5 +(1.4227-0.19PIH)J+(0.197-1.433PIH)J 0.5 -2.516-PIH

[0215] g(J)=0.0026J+(0.285-0.00195PIH)J 0.5 +(1.4227-0.19PIH)+(0.0985-0.7165PIH)J -0.5

[0216] In the formula, the subscript d represents the iteration number, when J d+1 With J d When the absolute deviation is less than a preset threshold, the iteration stops. When, then J d This is what we seek: to achieve a high spatial resolution precipitation probability. Otherwise, discard J. d Solve using the following formula:

[0217] J d+1 =J d -G(J d ) / g(J d )

[0218] G(J)=0.0013J 2 +(0.19+0.0013PIH)J 1.5 +(1.4227+0.19PIH)J+(0.197+1.433PIH)J 0.5 -2.516+PIH

[0219] g(J)=0.0026J+(0.285+0.00195PIH)J 0.5 +(1.4227+0.19PIH)+(0.0985+0.7165PIH)J -0.5

[0220] When J d+1 With J d When the absolute deviation is less than a preset threshold, the iteration stops. When, then J d This is what we seek: to achieve a high spatial resolution precipitation probability.

[0221] In the formula, for example, one method for determining the initial value J0 is: find all pixels with the same precipitation index as the target pixel PIH on the low spatial resolution precipitation index PI, and assign the average value of the remote sensing precipitation occurrence probability corresponding to all pixels to J0.

[0222] In the formula, for example, one method for determining the preset threshold is: to find all pixels with the same precipitation index as the target pixel PIH on the low spatial resolution precipitation index PI, and to assign the standard deviation of the remote sensing precipitation occurrence probability corresponding to all pixels to the preset threshold.

[0223] The determination of J0 and the preset threshold is not limited to this; there may be other methods for determination, and those skilled in the art can set them according to the actual situation.

[0224] S52: Utilizing high spatial resolution precipitation probability to obtain high spatial resolution remote sensing precipitation, specifically:

[0225]

[0226] In the formula, FL represents the high spatial resolution precipitation probability, and the subscript d represents the iteration number. When R d+1 With R d When the absolute deviation is less than a preset threshold, the iteration stops, then R... d This is the desired high spatial resolution remote sensing precipitation R.

[0227] In the formula, for example, one method for determining the initial value R0 is as follows: determine the occurrence probability of each precipitation amount in remote sensing precipitation, determine all low spatial resolution pixels corresponding to the occurrence probability equal to the precipitation probability at the target pixel of high spatial resolution precipitation probability FL, and assign the mean value of remote sensing precipitation corresponding to all low spatial resolution pixels to R0.

[0228] In the formula, for example, one method for determining the preset threshold is as follows: determine the occurrence probability of each precipitation amount in remote sensing precipitation, determine all low spatial resolution pixels corresponding to the occurrence probability equal to the precipitation probability at the target pixel of high spatial resolution precipitation probability FL, and assign the mean square error of the remote sensing precipitation corresponding to all low spatial resolution pixels to the preset threshold.

[0229] The determination of R0 and the preset threshold is not limited to this; there may be other methods for determination, and those skilled in the art can set them according to the actual situation.

[0230] Based on the above embodiments, in some embodiments of this disclosure, further, such as Figure 7 As shown, one possible way to obtain the corrected, fine-grained remote sensing precipitation measurements is as follows:

[0231] P R =R*(P / R)↑ )

[0232] In the formula, R ↑ This represents the result of upscaling R space to P space resolution, where * indicates R and P / R. ↑ Multiplying corresponding spatial ranges, P R This indicates the corrected, refined remote sensing precipitation measurement;

[0233] In the formula, we get R ↑ The spatial upscaling model can be selected from simple averaging, empirical regression, geostatistical methods, Bayesian methods or others. This disclosure does not specify any particular model, and those skilled in the art can choose according to the actual situation.

[0234] In the formula, * represents P / R ↑ Each low spatial resolution pixel value is multiplied by the high spatial resolution pixel value in the geographic space R it covers.

[0235] It should be pointed out that, although Figures 1-7 The steps in each flowchart are linked by arrows in sequence, but this does not mean that all steps must be executed in the order of the arrows. Unless explicitly stated otherwise, these steps do not need to be executed in a strict order and can be executed in other orders. Furthermore, Figures 1-7 At least some steps in the process contain multiple sub-steps or parts. These sub-steps or parts do not need to be completed simultaneously; they can achieve the desired effect by being executed at different times. These sub-steps or parts do not need to be executed sequentially; rather, some steps or sub-steps are executed alternately with other steps or sub-steps.

[0236] In one embodiment, such as Figure 8 As shown, a fine precipitation measurement system is provided for performing a fine precipitation measurement method described in the above embodiments, including: an input module 1, a water vapor index module 2, a two-dimensional space module 3, a four-vertex module 4, a precipitation index module 5, an enhancement module 6, a correction module 7, and an output module 8.

[0237] The input module 1 is used to collect time-synchronized remote sensing images and remote sensing precipitation observations;

[0238] The water vapor index module 2 is used to perform the creation of high and low spatial resolution water vapor indices as described in the above embodiments.

[0239] The two-dimensional space module 3 is used to perform the above embodiment to construct a pixel-by-pixel two-dimensional space with the low spatial resolution water vapor index as the horizontal axis and the precipitation index as the vertical axis.

[0240] The four-vertex module 4 is used to perform the determination of the four vertices of the pixel-by-pixel two-dimensional spatial extreme precipitation index as described in the above embodiment;

[0241] The precipitation index module 5 is used to perform the above embodiment to obtain the high spatial resolution precipitation index corresponding to the high spatial resolution water vapor index.

[0242] The enhancement module 6 is used to perform the enhanced remote sensing precipitation based on the high spatial resolution precipitation index as described in the above embodiments;

[0243] The correction module 7 is used to perform the above embodiment to obtain the corrected fine measurement of remote sensing precipitation;

[0244] The output module 8 is used to output the process and results of fine precipitation measurement.

[0245] This disclosure provides an electronic device for fine precipitation measurement, including: a sensor, a memory, and a processor. The sensor is used to acquire remote sensing images and remote sensing precipitation data. The memory is used to store a computer program and the remote sensing images and remote sensing precipitation data acquired by the sensor. The processor executes the computer program to process the remote sensing images and remote sensing precipitation data acquired by the sensor. Figures 1 to 7 The technical solutions of any of the method embodiments shown are similar in implementation principle and technical effect, and will not be described again here.

[0246] This disclosure also provides a computer-readable storage medium for fine precipitation measurement, on which a computer program is stored. When executed by a processor, the computer program can implement the fine precipitation measurement method provided in this disclosure. For example, when executed by a processor, the computer program implements... Figures 1 to 7 The technical solutions of any of the method embodiments shown are similar in implementation principle and technical effect, and will not be described again here.

[0247] It will be readily understood by those skilled in the art that all or part of the steps in the embodiments or methods provided in this disclosure can be executed by a computer program controlling related hardware. The computer program can be stored in a computer-readable storage medium, such as a read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage. When the computer program is executed, it can complete the steps of the embodiments of the above methods. Furthermore, other media referenced in the embodiments provided in this disclosure may also include volatile memory, such as random access memory (RAM) or external cache memory.

[0248] Compared with the prior art, the beneficial effects of this disclosure are as follows:

[0249] 1) The proposed water vapor index only uses the normalized cloud temperature in the remote sensing thermal infrared band and the reflectivity in the cirrus band, which can conveniently assess the water vapor content of clouds.

[0250] 2) By converting precipitation into a standardized normal distribution precipitation index, the nonlinear variation of precipitation with environmental factors is expressed, which is beneficial for the analysis of spatial heterogeneity of precipitation.

[0251] 3) The nonlinear variation of precipitation was quantified by using a new two-dimensional space composed of water vapor index and precipitation index, and the spatial details of local precipitation were estimated by the range of precipitation variation.

[0252] 4) By considering the spatial heterogeneity across scales, the spatial distribution information of precipitation with high spatial resolution is downscaled and simulated, which improves the accuracy of precipitation observation.

[0253] It should be noted that the above embodiments are described in detail, but this should not be construed as a limitation on the scope of the present invention. In addition, the technical features of the above embodiments can be combined in other ways. This disclosure does not describe all possible combinations of technical features. As long as these combinations do not contradict each other, they should all be within the scope of this specification. Furthermore, those skilled in the art can modify or improve some steps and technical features in the embodiments of this disclosure without creative effort, but these changes are all within the protection scope of this disclosure.

Claims

1. A method for precise measurement of precipitation, characterized in that, Includes the following steps: S1: Create water vapor indices with high and low spatial resolutions, including the following steps: S11: Create a high spatial resolution water vapor index using remote sensing imagery, specifically: WVI = (TIR - CIR) / (TIR + CIR) In the formula, TIR represents the normalized temperature of the thermal infrared band of the remote sensing image, CIR represents the on-board reflectance of the cirrus cloud band of the remote sensing image, and WVI represents the water vapor index. S12: Spatially upscaled WVI to obtain the water vapor index with low spatial resolution, specifically: wvi=UP(WVI) In the formula, UP represents the spatial upscaling model, which is used to upscale the spatial resolution of WVI to the spatial resolution of remote sensing precipitation data, and wvi represents the low spatial resolution water vapor index obtained by upscaling WVI using UP. S2: Construct a pixel-by-pixel two-dimensional space with the low spatial resolution water vapor index as the horizontal axis and the precipitation index as the vertical axis, including the following steps: S21: Calculate the logarithmic intermediate using remote sensing precipitation data, specifically: In the formula, (x, y) represents the spatial location in the remote sensing precipitation data, P(x, y) represents the remote sensing precipitation at location (x, y), c and r represent the number of columns and rows of the remote sensing precipitation data, lg represents the logarithm to the base 10, and O represents the intermediate logarithm calculated using the remote sensing precipitation data. S22: Estimating shape parameters using remotely sensed precipitation data, specifically: In the formula, b represents the shape parameter estimated using remote sensing precipitation data; S23: Estimate the scale parameter a using remote sensing precipitation data, specifically: In the formula, a represents the scale parameter estimated using remote sensing precipitation data; S24: Convert remote sensing precipitation data into precipitation probability, including: a) When the remotely sensed precipitation is 0, the specific probability of precipitation is as follows: F=M / (c*r) In the formula, M represents the number of pixels with zero precipitation, and F represents the probability of zero precipitation as perceived by remote sensing. b) When the remotely sensed precipitation is greater than 0, the specific probability of precipitation is as follows: In the formula, Γ represents the gamma distribution, F represents the precipitation probability when the remote sensing precipitation is P(x,y), and f(z) represents the probability distribution model used to describe the daily precipitation. S25: Calculate the precipitation index from remotely sensed precipitation data, specifically: In the formula, when F>0.5, S=1, F=1-F, when F≤0.5, S=-1, PI(x,y) represents the precipitation index of P(x,y), 1≤x≤c, 1≤y≤r, and PI represents the precipitation index of the remote sensing precipitation data obtained. S26: Construct a pixel-by-pixel two-dimensional space with the low spatial resolution water vapor index (wvi) as the horizontal axis and the precipitation index (PI) as the vertical axis, including the following steps: S261: Determine the maximum and minimum values ​​of the water vapor index per pixel spatial range for remotely sensed precipitation on WVI, specifically: WVI_max≥WVI(X, Y) WVI_min≤WVI(X, Y) ∀(X, Y)∈(x, y) In the formula, ∀(X, Y)∈(x, y) represents any given high spatial resolution water vapor index pixel (X, Y) within the coverage area of ​​low spatial resolution pixel (x, y), WVI_max represents the maximum value of high spatial resolution water vapor index within the coverage area of ​​pixel (x, y), and WVI_min represents the minimum value of high spatial resolution water vapor index within the coverage area of ​​pixel (x, y). S262: Determine the set of pixel locations on WVI where the water vapor index falls between WVI_min and WVI_max, specifically: Oh (x,y) ={(m, n)|WVI_min≤wvi(m, n)| m∈[1, c], n∈[1, r] ≤WVI_max} In the formula, (m, n) represents the pixel position on wvi where the water vapor index is between WVI_min and WVI_max, and Ω (x,y) Let {x,y} represent the set of {(m,n)} at pixel (x,y); S263: Construct a pixel-by-pixel two-dimensional space with the low spatial resolution water vapor index (wvi) as the horizontal axis and the precipitation index (PI) as the vertical axis, specifically: Ŝ={[wvi(m, n),PI(m, n)]|(m, n)∈Ω (x,y) } In the formula, Ŝ represents the expression derived from Ω. (x,y) The space consisting of low spatial resolution water vapor index and precipitation index at all pixel locations, 1≤x≤c, 1≤y≤r; S3: Determine the four vertices of the pixel-by-pixel two-dimensional spatial extreme precipitation index, including the following steps: S31: Divide each pixel into several subspaces along the horizontal axis water vapor index; S32: Determines the four vertices of the pixel-by-pixel two-dimensional spatial extreme precipitation index, including: The maximum value of the precipitation index within the minimum interval of the water vapor index is determined and denoted as PI_1 (wvi min PI max1 (), represents the maximum possible precipitation when the water vapor index is at its minimum, indicating the maximum value of precipitation variation when the water vapor index is relatively small; The minimum precipitation index within the minimum water vapor index range is determined and denoted as PI_2 (wvi min PI min2 The value of 1 represents the minimum possible precipitation when the water vapor index is at its minimum, indicating the minimum value of precipitation variation when the water vapor index is low. The minimum precipitation index within the maximum water vapor index range is determined and denoted as PI_3 (wvi max PI min3 The value of 1 represents the minimum possible precipitation when the water vapor index is at its maximum, indicating the minimum value of precipitation variation when the water vapor index is relatively large. The maximum value of the precipitation index within the maximum water vapor index range is determined and denoted as PI_4 (wvi max PI max4 The maximum possible precipitation occurs when the water vapor index is at its maximum, indicating the maximum value of precipitation variation when the water vapor index is relatively large. S4: Calculate the high spatial resolution precipitation index corresponding to the high spatial resolution water vapor index, including the following steps: S41: The equation for calculating the variation of the extreme precipitation index at each pixel includes: The equation for calculating the variation of the maximum precipitation index at each pixel is as follows: standing=(PI max4 -PI max1 ) / (wvi max -wvi min ) PIU(v)=ku*v+(PIU max1 wvi max -PI max4 wvi min ) / (wvi max -wvi min ) In the formula, v represents the water vapor index independent variable, and PIU(v) represents the equation for the change of the maximum precipitation index at pixel (x, y). The equation for calculating the variation of the minimum precipitation index at each pixel is as follows: kl=(PI min3 -PI min2 ) / (wvi max -wvi min ) PIL(v)=kl*v+(PI min2 wvi max -PI min3 wvi min ) / (wvi max -wvi min ) In the formula, PIL(v) represents the equation for the change of the minimum precipitation index at pixel (x, y); S42: The equation for calculating the variation of the precipitation index at each pixel is as follows: km=kl+(ku-kl) / (PIU(wvi(m, n))-PIL(wvi(m, n)))*(PI(m, n)-PIL(wvi(m, n))) PIM(v)=km*v+PI(m, n)-km*wvi(m, n) In the formula, PIM(v) represents the equation for the change of precipitation index at each pixel (x, y); S43: Calculate the high spatial resolution precipitation index corresponding to the high spatial resolution water vapor index, specifically: PIH(WVI(X, Y))=km*WVI(X, Y)+PI(m, n)-km*wvi(m, n) ∀(X, Y)∈(x, y) In the formula, WVI(X, Y) represents the water vapor index at any high spatial resolution location (X, Y) within the coverage area of ​​the low spatial resolution pixel (x, y), PIH(WVI(X, Y)) represents the precipitation index at any high spatial resolution pixel location (X, Y) within the coverage area of ​​the low spatial resolution pixel (x, y), and PIH represents the high spatial resolution precipitation index of the corresponding remote sensing precipitation coverage area; S5: Enhance remote sensing precipitation based on high spatial resolution precipitation index, including the following steps: S51: Calculate the high spatial resolution precipitation probability corresponding to the high spatial resolution precipitation index, including: J d+1 =J d -G(J d ) / g(J d ) G(J)=0.0013J 2 +(0.19-0.0013PIH)J 1.5 +(1.4227-0.19PIH)J+(0.197-1.433PIH)J 0.5 -2.516-PIH g(J)=0.0026J+(0.285-0.00195PIH)J 0.5 +(1.4227-0.19PIH)+(0.0985-0.7165PIH)J -0.5 In the formula, the subscript d represents the iteration number, when J d+1 With J d When the absolute deviation is less than a preset threshold, the iteration stops. When J < 0.5, then d This is what we are looking for, so that the high spatial resolution precipitation probability FL = 1 - Otherwise, discard J. d Solve using the following formula: J d+1 =J d -G(J d ) / g(J d ) G(J)=0.0013J 2 +(0.19+0.0013PIH)J 1.5 +(1.4227+0.19PIH)J+(0.197+1.433PIH)J 0.5 -2.516+PIH g(J)=0.0026J+(0.285+0.00195PIH)J 0.5 +(1.4227+0.19PIH)+(0.0985+0.7165PIH)J -0.5 When J d+1 With J d When the absolute deviation is less than a preset threshold, the iteration stops. When J < 0.5, then d That is, let the high spatial resolution precipitation probability FL = ; S52: Calculate the high spatial resolution remote sensing precipitation R using the high spatial resolution precipitation probability FL, specifically: In the formula, the subscript d represents the iteration number, when R d+1 With R d When the absolute deviation is less than a preset threshold, the iteration stops, then R... d This is the desired high spatial resolution remote sensing precipitation R; S6: Obtain the corrected, fine-grained remote sensing precipitation measurements.

2. The method according to claim 1, characterized in that, Specifically, S6 is: P R =R*(P / R ↑ ) In the formula, R ↑ This represents the result of upscaling R space to P space resolution, where * indicates R and P / R. ↑ Multiplying corresponding spatial ranges, P R This indicates the corrected, refined remote sensing precipitation measurement.

3. A fine precipitation measurement system, characterized in that, The method for performing a fine precipitation measurement as described in claim 1 includes an input module, a water vapor index module, a two-dimensional space module, a four-vertex module, a precipitation index module, an enhancement module, a correction module, and an output module. The input module is used to collect time-synchronized remote sensing images and remote sensing precipitation observations; The water vapor index module is used to perform the method of claim 1 to create water vapor indices with high and low spatial resolution; The two-dimensional space module is used to perform the method of claim 1 to construct a pixel-by-pixel two-dimensional space with the low spatial resolution water vapor index as the horizontal axis and the precipitation index as the vertical axis. The four-vertex module is used to perform the method of claim 1 to determine the four vertices of the pixel-by-pixel two-dimensional spatial extreme precipitation index; The precipitation index module is used to perform the method described in claim 1 to obtain the high spatial resolution precipitation index corresponding to the high spatial resolution water vapor index. The enhancement module is used to perform the method of claim 1 to enhance remote sensing precipitation based on the high spatial resolution precipitation index; The correction module is used to perform the method of claim 2 to obtain the corrected fine measurement of remote sensing precipitation; The output module is used to output the process and results of fine precipitation measurement.

4. An electronic device for fine precipitation measurement, comprising a sensor, a memory, and a processor, wherein the sensor is used for acquiring remote sensing images and remote sensing precipitation, and the memory stores a computer program and the remote sensing images and remote sensing precipitation acquired by the sensor, characterized in that... When the processor executes the computer program to process the remote sensing images and remote sensing precipitation acquired by the sensor, it implements the steps of the precipitation fine measurement method according to any one of claims 1 to 2.

5. A computer-readable storage medium for fine precipitation measurement, having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the fine precipitation measurement method according to any one of claims 1-2.

Images