Method and system for generating surface soil humidity based on time-space fusion of remote sensing data

Abstract

The invention discloses a method and a system for generating surface soil humidity based on time-space fusion of remote sensing data, and relates to the technical field of soil humidity processing, and the method comprises the following steps: obtaining a vegetation parameter image, an active microwave remote sensing image and a passive microwave remote sensing image of a required date, and superposing the vegetation parameter image and the active microwave remote sensing image to obtain a vegetation parameter image; obtaining a high-resolution parameter image; resampling the parameter image to low resolution, and performing spatial downscaling on the passive microwave remote sensing image through the low-resolution parameter image; the method comprises the following steps: taking a time sequence of each pixel in a passive microwave remote sensing image as a signal, obtaining phase raster data and high-frequency amplitude raster data of different frequencies of the signal, carrying out spatial interpolation, and fitting a spatial interpolation result and a spatial downscaling result in a time dimension to obtain a daily scale surface soil humidity image. According to the method, fitting of the spatial resolution and the temporal resolution is achieved, the soil humidity image with the fine temporal-spatial resolution is obtained, and the method has high application value.

Description

Technical Field

[0001] The present invention relates to the technical field of soil moisture processing, and in particular to a method and system for generating surface soil moisture based on spatiotemporal fusion of remote sensing data. Background Art

[0002] Soil moisture is a key factor in the energy and water cycles of the Earth system and an important parameter in many fields such as environmental science and agricultural science. On a global scale, soil moisture is extremely valuable for weather forecasting, climate change, and monitoring extreme events such as floods; on a regional scale, soil moisture information is an indispensable part of local agriculture and water resources management. Remote sensing technology has the advantages of fast timeliness, wide monitoring range, long-term dynamic monitoring, and spatial continuity, which overcomes the shortcomings of traditional methods that are time-consuming, labor-intensive, and have a small detection range, providing an effective means for obtaining soil moisture information over a large area.

[0003] Soil moisture inversion methods based on remote sensing can basically be divided into three categories, namely, inversion based on optical, thermal infrared and microwave radar data. The emission and scattering characteristics of soil in the microwave band are affected by its dielectric properties, and the soil dielectric constant is directly related to the amount of soil moisture, which is the physical basis for microwave remote sensing inversion of soil moisture. Microwave remote sensing is not restricted by weather conditions such as light, clouds and fog. It has a strong ability to penetrate clouds and can detect soil information under vegetation cover. It has all-weather simulation capabilities and high sensitivity to soil moisture changes. In terms of the continuous availability of large-scale data products, soil moisture products based on microwave technology are much more advanced than those based on optical and thermal infrared.

[0004] In the existing methods, one or two pairs of coarse and fine resolution images are usually required as reference images, as well as a large amount of difficult-to-obtain remote sensing and auxiliary data, supplemented by a coarse resolution image of the target date to generate a fine resolution image. This results in a large demand for computing power in the existing technology, and the efficiency of generating fine resolution images is low. In addition, due to the contradiction between the temporal resolution and spatial resolution of microwave remote sensing, it is impossible to directly obtain soil moisture products with fine temporal and spatial resolution, resulting in the obtained soil moisture products having insufficient temporal and spatial resolution. Summary of the invention

[0005] The purpose of the present invention is to address the deficiencies of the above-mentioned prior art and to provide a method and system for generating surface soil moisture based on spatiotemporal fusion of remote sensing data, so as to solve the problem in the prior art that soil moisture products with fine spatiotemporal resolution cannot be directly obtained due to the contradiction between the temporal resolution and spatial resolution of microwave remote sensing, resulting in the obtained soil moisture products having insufficient spatiotemporal resolution.

[0006] The present invention specifically provides the following technical solutions:

[0007] A method for generating surface soil moisture based on spatiotemporal fusion of remote sensing data, comprising:

[0008] Obtain vegetation parameter images, active microwave remote sensing images and passive microwave remote sensing images on the required date, and superimpose the vegetation parameter images with the active microwave remote sensing images to obtain high-resolution parameter images; wherein the spatial resolution of the high-resolution image is higher than a set threshold, and the spatial resolution of the low-resolution image is lower than the set threshold;

[0009] The spatial resolution of the parameter image is resampled to a low resolution, and the passive microwave remote sensing image is spatially downscaled by the low-resolution parameter image to obtain a time-discrete initial surface soil moisture image;

[0010] The time series of each pixel in the passive microwave remote sensing image is used as a signal, and the phase grid data and high-frequency amplitude grid data of the signal at different frequencies are obtained. The phase grid data and the high-frequency amplitude grid data are spatially interpolated, and the spatial interpolation result is fitted with the initial surface soil moisture image in the time dimension to obtain a daily-scale surface soil moisture image.

[0011] Preferably, obtaining a vegetation parameter image on a required date includes:

[0012] Get different combinations of vegetation parameters;

[0013] Input different vegetation parameter combinations into the radiation transfer model, simulate and obtain a spectral curve, and convert the spectral curve into atmospheric apparent reflectance to obtain the reflectance of the band captured by the model;

[0014] The band reflectance is input into the random forest model for inversion to obtain the vegetation parameter image of the required date; wherein the vegetation parameter image includes the leaf area index LAI, the canopy water content CWC and the average leaf inclination angle ALA image.

[0015] Preferably, the vegetation parameter image is superimposed with the active microwave remote sensing image to obtain a high-resolution parameter image, specifically:

[0016] By superimposing the active microwave remote sensing image with the leaf area index LAI, canopy water content CWC and average leaf inclination angle ALA images in the vegetation parameter image, a set of images containing seven bands including normalized vegetation index NDVI, leaf area index LAI, canopy water content CWC, average leaf inclination angle ALA, vertical polarization VV, cross polarization VH and radar wave incident angle θ was obtained;

[0017] The seven-band images are used as high-resolution parameter images.

[0018] Preferably, the spatial downscaling of the passive microwave remote sensing image by using the low-resolution parameter image to obtain the time-discrete initial surface soil moisture image includes:

[0019] The spatial resolution of the high-resolution parameter image is resampled to a low resolution corresponding to the pixels of the passive microwave remote sensing image, and the pixels are used as a training set to train the surface soil moisture inversion model;

[0020] The high-resolution parameter image is input into the trained surface soil moisture inversion model to obtain a high-resolution initial surface soil moisture image under time discrete conditions.

[0021] Preferably, the time series of each pixel in the passive microwave remote sensing image is used as a signal, and phase grid data and high-frequency amplitude grid data of different frequencies of the signal are obtained, including:

[0022] Performing a fast Fourier transform on the signal to obtain the amplitude and phase of the signal at different frequencies;

[0023] The time series SSM at different spatial resolutions is represented by the amplitude and phase of different frequencies, specifically:

[0024]

[0025]

[0026] Where n is the number of signal components, i is the frequency of the signal, and A i is the amplitude at low resolution, x is a variable, ψ i is the phase at low resolution, t is the total duration of the signal, a i is the amplitude at high resolution, is the phase at high resolution, I is the low resolution, and U is the high resolution.

[0027] Preferably, the spatial interpolation of the phase grid data and the high-frequency amplitude grid data is specifically as follows:

[0028] Eliminate the low-frequency part in the amplitude grid data, and use bilinear interpolation to perform spatial interpolation on the high-frequency amplitude grid data to obtain high-frequency amplitude grid data; wherein the high frequency and low frequency are divided by a frequency threshold;

[0029] The phase grid data of all frequencies are spatially interpolated using the nearest neighbor method to obtain high-resolution phase grid data.

[0030] Preferably, fitting the spatial interpolation result with the initial surface soil moisture image in the time dimension to obtain a daily-scale surface soil moisture image includes:

[0031] Based on the high-resolution phase grid data and high-frequency amplitude grid data, combined with the soil moisture contribution of the initial surface soil moisture image in the time discrete situation to the grid data in the low-frequency component, downscaling fitting is performed in the time dimension to obtain a high-resolution surface soil moisture image at the daily scale; wherein the low frequency is a frequency below a set threshold;

[0032] According to the measured soil moisture at the site, the high-resolution surface soil moisture image at the daily scale is corrected by coupling inverse Fourier transform and random forest.

[0033] The present invention also provides a system for generating surface soil moisture based on spatiotemporal fusion of remote sensing data, comprising:

[0034] The acquisition module is used to obtain vegetation parameter images, active microwave remote sensing images and passive microwave remote sensing images of the required date, and to superimpose the vegetation parameter images with the active microwave remote sensing images to obtain high-resolution parameter images; wherein the spatial resolution of the high-resolution image is higher than a set threshold, and the spatial resolution of the low-resolution image is lower than the set threshold;

[0035] A downscaling module is used to resample the spatial resolution of the parameter image to a low resolution, and to spatially downscale the passive microwave remote sensing image through the low-resolution parameter image to obtain a time-discrete initial surface soil moisture image;

[0036] The fusion module is used to take the time series of each pixel in the passive microwave remote sensing image as a signal, obtain the phase grid data and high-frequency amplitude grid data of the signal at different frequencies, and spatially interpolate the phase grid data and the high-frequency amplitude grid data, fit the spatial interpolation result with the initial surface soil moisture image in the time dimension, and obtain a daily-scale surface soil moisture image.

[0037] The present invention provides a computer device, including a memory and a processor, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the processor executes the steps of the above-mentioned method for generating surface soil moisture based on spatiotemporal fusion of remote sensing data.

[0038] The present invention provides a storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned method for generating surface soil moisture based on spatiotemporal fusion of remote sensing data are implemented.

[0039] Compared with the prior art, the present invention has the following significant advantages:

[0040] The present invention realizes spatial downscaling of passive microwave remote sensing images by resampling the spatial resolution of the acquired parameter image. It only needs to resample the spatial resolution, which reduces the demand for computing power, further reduces the spatial downscaling cost, and improves the downscaling efficiency of objects with large scale differences. The pixel time series of the passive microwave remote sensing image constitutes spatial interpolation of signal phase grid data and high-frequency amplitude grid data, and the spatial interpolation result is fitted with the initial surface soil moisture image in the time dimension, thus realizing the fitting of spatial resolution and temporal resolution, and directly obtaining soil moisture products with fine spatiotemporal resolution, which has high application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 The present invention provides a distribution map of soil moisture observation points and land cover types in the Jiefangzha irrigation area;

[0042] Figure 2 It is the technical roadmap for generating a spatiotemporally continuous 20m×20m soil moisture provided by the present invention;

[0043] Figure 3 It is a framework diagram of the spatiotemporal fusion algorithm provided by the present invention;

[0044] Figure 4 is a performance evaluation diagram of the random forest replacing the PROSAIL model provided by the present invention; wherein, Figure 4 (a) is the performance evaluation diagram of ALA parameters. Figure 4 (b) is the performance evaluation diagram of LAI parameters. Figure 4 (c) is the performance evaluation diagram of CWC parameters;

[0045] Figure 5 It is the accuracy assessment of the surface soil moisture inversion model provided by the present invention; Figure 5 (a) The inversion accuracy of the model for SMAP SSM at 9 km resolution. Figure 5 (b) Model performance at 20m resolution;

[0046] Figure 6 It is the soil moisture with a spatial resolution of 20m from July 7 to 18, 2018 obtained by two spatiotemporal fusion algorithms provided by the present invention; Figure 6 (a) Surface soil moisture obtained by ESTARFM algorithm, the reference image dates are July-8, 2018 and July-20, 2018, Figure 6 (b) Surface soil moisture obtained by STFFT algorithm (without ground data correction);

[0047] Figure 7The present invention provides the in-situ observations of the Hangjinhou Banner Station in 2018 and 2019 and the time series changes of downscaled surface soil moisture; Figure 7 (a) and Figure 7 (b) are the results of 2018 and 2019 respectively;

[0048] Figure 8 It is the verification of downscaling surface soil moisture at Hangjinhouqi Station provided by the present invention; Figure 8 (a) and Figure 8 (b) are the results of 2018 and 2019 respectively;

[0049] Fig. 9 It is the surface soil moisture map of the research area with a daily scale and a spatial resolution of 20m provided by the present invention; wherein, Fig. 9 (a) Fig. 9 (b) Fig. 9 (c) Fig. 9 (d) and Fig. 9 (e) are the surface soil moisture maps from July 17, 2018 to July 21, 2018, Fig. 9 (f), Fig. 9 (g), Fig. 9 of (h), Fig. 9 (i) and Fig. 9 (j) are the surface soil moisture maps from May 16, 2019 to May 20, 2019. DETAILED DESCRIPTION

[0050] The following is a clear and complete description of the technical solutions of the embodiments of the present invention in conjunction with the drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0051] The research area of ​​this invention is Jiefangzha Irrigation Area ( Figure 1 )(106°43'~107°27'E,40°34'~41°14'N), with an altitude of 1032m~1050m, covering an area of ​​about 2.157*105 hectares, including 1.543*105 hectares of arable land, and the soil is mainly silt loam, loam and clay loam.

[0052] The study area is located in arid, semi-arid and semi-desert steppe areas with sufficient sunshine, an average annual sunshine duration of about 3180 hours, and an average frost-free period of 130 to 168 days. The climate here is mainly temperate continental, with cold and little snow in winter and hot and dry summers. The average annual temperature is about 6.3℃ to 7.7℃, the average precipitation is 138.2mm, and the average evaporation is 2096.4mm. Regional agricultural production is highly dependent on irrigation. Irrigation in the irrigation area is mainly based on the Yellow River water, with an annual water diversion volume of 1 billion to 1.2 billion cubic meters. Ten-meter-level daily soil moisture products will have a very important guiding significance for the efficient water resources management of local agriculture.

[0053] Four types of datasets will be used in the process of soil moisture temporal and spatial downscaling (Table 1). They are: (1) Sentinel-1GRD IW mode active microwave remote sensing products (images), (2) Sentinel-2L1C atmospheric apparent reflectance products (images), (3) Soil Moisture Active Passive Remote Sensing Satellite SMAP (The Soil Moisture Active Passive) Level-4SSM surface soil moisture products, that is, passive microwave remote sensing image data, and (4) in situ observed surface soil moisture data.

[0054] Table 1 Datasets used to develop the soil moisture reduction method

[0055] Dataset Variable Spatialresolution Temporal resolution Sentinel-2L2A Surfacereflectance 20m×20m 2-3days Sentinel-1GRD VV / VH / Angle 5m×20m 12days SMAPLevel-4 SSM 9km×9km 3h MeteorologicalStation SSM Point Daily OrdinaryObservationPoints SSM Point Monthly

[0056] Sentinel-1GRD IW mode active microwave remote sensing products:

[0057] Sentinel-1 consists of two twin satellites, Sentinel-1A and Sentinel-1B. The revisit period of a single satellite is 12 days, and the complementarity of the two satellites can be reduced to 6 days, and satellite images with a resolution of 5 meters to 40 meters can be obtained under all-day and all-weather conditions. The Sentinel-1 satellite is equipped with a C-band SAR sensor with an operating frequency of 5.4GHz, multi-polarization imaging capability and four imaging modes. The Sentinel-1 image used in the present invention is the descending orbit data of the interferometric wide-band mode. The corresponding ground resolution in this mode is 5m×20m, the data level is Level-1, and the product data is in the ground range format (Ground Range Detected, GRD). Dual polarization is cross-polarization VH (Vertical horizontal) and vertical polarization VV (Verticalvertical). The required data time span is from April 1, 2018 to September 30, 2018 and the same period of 2019, and the time resolution in the study area is 12 days. The data were obtained from the Google Earth Engine platform and processed with Refined Lee filtering, and then exported with a spatial resolution of 20 m.

[0058] Sentinel-2L1C atmospheric apparent reflectance product:

[0059] Like Sentinel-1, Sentinel-2 satellite is also composed of two binary satellites, Sentinel-2A and Sentinel-2B. The revisit period of a single satellite is 10 days, which can be reduced to 5 days by the complementarity of two satellites. The revisit period in the area of ​​the present invention is 2-3 days. Sentinel-2 satellite carries a multispectral imager with 13 spectral bands, a width of 290km, and a ground resolution ranging from 10m to 60m depending on the band. Since the Sentinel-2L2A (S2L2A) surface reflectance product in the area of ​​the present invention in 2018 is missing on the GEE platform, the Sentinel-2L1C (S2L1C) product is selected and atmospheric correction is performed on it with reference to the research results of Yin et al. to obtain the S2L2A product. In order to match the Sentinel-1 data, the date of S2L2A is selected to be consistent with the date of Sentinel-1 passing. Due to the high temporal resolution of Sentine-2, it can meet the matching requirements for most of the time. The dates that could not be matched were filled by linear interpolation and SG filtering of S2L2A data. This method was also used to fill the local blanks of the image after cloud removal using the QA60 band. The present invention selected 10 bands (B02-B8A, B11 and B12) covering the spectral region from visible light to short-wave infrared at 10m and 20m spatial resolutions, and resampled them all to a spatial resolution of 20m.

[0060] SMAP Level-4SSM surface soil moisture product:

[0061] The SMAP Level-4 surface soil moisture product uses data assimilation technology to integrate satellite-borne L-band brightness temperature measurements and precipitation observations into the basin land surface process model to generate surface soil moisture (0-5 cm vertical average), root zone soil moisture (0-100 cm vertical average) and other research products (unverified), including surface meteorological forcing variables, soil temperature, evapotranspiration and net radiation. The present invention uses the SMAP Level-4 surface soil moisture product obtained from the GEE platform as the background field for SSM downscaling, with a spatial resolution of 9 km and a temporal resolution set to output on a daily basis.

[0062] In situ surface soil moisture observations:

[0063] Two long-term observation stations, Hangjinhouqi Station (40°51′00″N, 107°07′00″E) and Uradhouqi Station (41°04′00″N, 107°03′00″E), can provide daily soil volumetric moisture content at different depths of 10, 40, and 100 cm. The present invention extracts the SSM automatically measured by the station from April 30 to September 30, 2018 and the same period in 2019. Taking into account that the SMAP product has non-negligible errors in the study area, the SSM observations of Uradhouqi Station are used for the correction of the downscaled products in the spatiotemporal fusion stage, while the observations of Hangjinhouqi Station are used for the verification of the final surface soil moisture downscaled products. This part of the data is obtained from the China Meteorological Data Service Center (http: / / data.cma.cn).

[0064] In addition to the in-situ data of soil surface moisture continuously observed at fixed points, a total of 14 soil volume moisture content measurements were carried out at 19 common observation points in the study area from April to September in 2018 and 2019. These ground observations will be used to evaluate the performance of the soil moisture inversion model at a resolution of 20 m.

[0065] The results of many studies have shown that when there is no vegetation coverage on the surface, there is a good linear relationship between the radar backscatter coefficient and soil moisture. However, during the vigorous growth period of vegetation, due to the influence of the vegetation canopy, the accuracy of directly using the backscatter coefficient to invert soil moisture is poor. At this time, it is necessary to strip the soil backscatter coefficient from the echo signal collected by the SAR, and then invert the soil moisture information from it. Referring to the idea of ​​the water-cloud model, the present invention believes that the total backscatter of the vegetation-covered surface is composed of two parts, one is the volume scattering term directly reflected back by the vegetation canopy, and the other is the ground backscatter term after double attenuation by the vegetation canopy. Both of these items are related to the water content of vegetation (VWC), the type of vegetation (V), and the soil moisture (M).S ), radar wave wavelength (λ) and radar wave incident angle (θ). Therefore, the total backscatter coefficient (σ) can be expressed as:

[0066] σ=f(VWC,V,M S ,λ,θ).

[0067] The vegetation water content is considered to be a function of the canopy water content (CWC):

[0068] VWC=f(CWC).

[0069] Canopy water content is defined as the product of leaf area index (LAI) and leaf water content (LWC):

[0070] CWC=LAI×LWC.

[0071] In addition, the present invention believes that the influence of vegetation types on backscattering mainly comes from the differences in the canopy structures of different vegetation. The canopy structure can be simply characterized by the leaf area index and the average leaf inclination angle (ALA). In addition, NDVI is used to describe the difference in vegetation coverage. Therefore, the vegetation category can be expressed as:

[0072] V=f(LAI,ALA,NDVI).

[0073] Combined formula:

[0074] σ=f(LAI,CWC,ALA,NDVI,M S ,λ,θ).

[0075] Change the form:

[0076] M S =f(LAI,CWC,ALA,NDVI,VV,VH,θ).

[0077] Therefore, the first step of the method proposed in the present invention is to use the S2L2A product to calculate NDVI and invert vegetation LAI, CWC and ALA through the PROSAIL radiation transfer model. The second step is to fuse NDVI, LAI, CWC, ALA and Sentinel-1GRD products, and then use the random forest algorithm to spatially downscale SMAP. The third step is to develop a new spatiotemporal fusion algorithm and perform spatiotemporal fusion of soil moisture.

[0078] like Figure 2 As shown, a method for generating surface soil moisture based on spatiotemporal fusion of remote sensing data provided by an embodiment of the present invention is described below, including the following steps:

[0079] Step S1: Obtain vegetation parameter images, active microwave remote sensing images and passive microwave remote sensing images of a required date, and superimpose the vegetation parameter images with the active microwave remote sensing images to obtain high-resolution parameter images; wherein the spatial resolution of the high-resolution is higher than a set threshold, and the spatial resolution of the low-resolution is lower than the set threshold.

[0080] In step S1, the vegetation parameter image of the required date is obtained, including:

[0081] Different combinations of vegetation parameters are constructed and input into the radiation transfer model to simulate and generate spectral curves. The spectral curves are then converted into the reflectance of the bands that can be captured by the atmospheric apparent reflectance acquisition model (atmospheric apparent reflectance product - Sentinel-2).

[0082] GEE (Google Earth Engine-cloud geographic information processing platform) has a built-in calculation function for the normalized vegetation index NDVI, so here we focus on how to use the S2L2A (Sentinel-2-L2A) product to invert vegetation parameters such as LAI, CWC and ALA. The present invention first uses the radiation transfer PROSAIL model to simulate 5000 groups of different vegetation parameter combinations (Table 2) and their corresponding spectral curves, and converts the spectral curves into the band reflectance that Sentinel-2 can capture through the spectral response function of Sentinel-2.

[0083] Table 2 Different vegetation parameter combinations based on the PROSAIL model

[0084] Parameters Abbreviation valuerange Leafchlorophyllcontent <![CDATA[C ab (μg / cm 2 )]]> 0-100 Leaf structure parameter N (unitless) 1-3 Leafcarotenoidcontent <![CDATA[C ar (μg / cm 2 )]]> <![CDATA[C ab / 7]]> Brownpigmentcontent <![CDATA[C b ]]> 0 Drymattercontent <![CDATA[C m (g / cm 2 )]]> 0.002-0.020 Leafwatercontent <![CDATA[LWC(g / cm 2 )]]> 0.005-0.050 Averageleafangle ALA(°) 40-80 Leafareaindex LAI 0.1-5 Hotspot parameter hspot 0.1

[0085] The band reflectance is input into the random forest model for inversion to obtain the vegetation parameter image of the required date; the vegetation parameter image includes the leaf area index LAI, canopy water content CWC and average leaf inclination angle ALA image.

[0086] The simulated data set obtained by PROSAIL model simulation is used as the training set of random forest, in which the band reflectance is used as the input variable of random forest, and the leaf area index LAI, canopy water content CWC and average leaf inclination angle ALA vegetation parameters are used as output variables. Then the trained random forest model is applied pixel by pixel to Sentinel-2 imagery to obtain the LAI, CWC and ALA images of the required date.

[0087] In addition to using the PROSAIL model to simulate the data set, the training of the random forest model and the inversion of vegetation parameters are all performed on the GEE platform. The advantage of this is that only the final vegetation parameter image needs to be downloaded, rather than the original Sentinel-2 image. In addition, the inversion process of the PROSAIL model is not easy to deploy on the GEE platform, but the random forest model trained with the simulated data set can easily complete the inversion of parameters such as LAI on the GEE platform.

[0088] In step S1, the vegetation parameter image is superimposed with the active microwave remote sensing image to obtain a parameter image with a spatial resolution of Ukm, specifically:

[0089] Acquire active microwave remote sensing images (Sentinel-1 images).

[0090] By superimposing the active microwave remote sensing image with the leaf area index LAI, canopy water content CWC and average leaf inclination angle ALA images in the vegetation parameter image, a set of images including the normalized vegetation index NDVI, leaf area index LAI, canopy water content CWC, average leaf inclination angle ALA, vertical polarization VV, cross polarization VH and radar wave incident angle θ in seven bands are obtained. Among them, the active microwave remote sensing image here is the Sentinel-1 image after Refined Lee filtering.

[0091] The seven-band images are used as high-resolution parameter images.

[0092] Step S2: resample the spatial resolution of the parameter image to a low resolution, and spatially downscale the passive microwave remote sensing image using the low-resolution parameter image to obtain a time-discrete initial surface soil moisture image.

[0093] Step S2 specifically includes:

[0094] The spatial resolution of the high-resolution parameter image is resampled to a low resolution using mean aggregation. In this embodiment, the low resolution is 9 km, which corresponds one-to-one to the pixels of the passive microwave remote sensing image (the image obtained by the soil moisture active-passive remote sensing satellite SMAP), and the pixels are used as a training set to train the surface soil moisture inversion model.

[0095] The high-resolution parameter image is input into the trained surface soil moisture inversion model to obtain an initial surface soil moisture image with high spatial resolution under time discrete conditions.

[0096] The following is an example:

[0097] The surface soil moisture inversion model is applied to the parameter image with a spatial resolution of 20m before resampling to obtain the surface soil moisture with a spatial resolution of 20m, thus realizing the spatial downscaling of SMAP. The 20m here is a practical application of high resolution. This method is widely used in the spatial downscaling of soil moisture products. It should be noted that at this time, only the SMAP SSM spatial downscaling product of the Sentinel-1 overpass date can be obtained, which is discontinuous in time ( Figure 3 (g)), that is, a set of surface soil moisture images with high spatial resolution and low temporal resolution is obtained ( Figure 3 (g)).

[0098] Step S3: Take the time series of each pixel in the passive microwave remote sensing image as a signal, obtain the phase grid data and high-frequency amplitude grid data of the signal at different frequencies, where the high frequency is the frequency above the set threshold, and perform spatial interpolation on the phase grid data and the high-frequency amplitude grid data, fit the spatial interpolation result with the initial surface soil moisture image in the time dimension, and obtain a daily-scale surface soil moisture image. The frequency represents the event that affects soil moisture, the amplitude represents the degree of influence of the event on soil moisture, and the phase represents the time when the event starts.

[0099] SMAP itself has the characteristics of high temporal resolution and low spatial resolution. The advantages of both are combined through the spatiotemporal fusion algorithm to obtain soil moisture products with high spatiotemporal resolution. The time of the reference image and the variability of soil moisture both affect the application of traditional spatiotemporal fusion algorithms on soil moisture products to a certain extent. The present invention proposes a soil moisture spatiotemporal fusion algorithm (STFFT) based on Fourier transform to solve the above problems.

[0100] Soil moisture is affected by many factors, including precipitation, evapotranspiration, infiltration, groundwater recharge, and irrigation. This algorithm divides these factors into high-frequency and low-frequency information in the time series, among which irrigation, precipitation, and evapotranspiration are considered high-frequency information, while infiltration and groundwater recharge are considered low-frequency information. Interestingly, in terms of spatial distribution, high-frequency information in the time series becomes low-frequency information, and vice versa. In terms of atmospheric phenomena that control precipitation and evapotranspiration, changes in boundary layer vortices, cumulonimbus clouds, fronts, and squall lines are all on the scale of kilometers. These atmospheric phenomena are almost all daily events, that is, they change day by day in time. Another intuitive feeling is that the soil texture that affects water infiltration is different within a farmland, but it basically does not change in the time series. This phenomenon of opposite spatiotemporal frequencies of the same information provides a hypothetical basis for our spatiotemporal fusion algorithm.

[0101] In step S3, the time series of each pixel in the passive microwave remote sensing image is used as a signal to obtain phase grid data and high-frequency amplitude grid data of the signal at different frequencies, including:

[0102] Each pixel of SMAP SSM can be regarded as a signal in the time series ( Figure 3 (a) shows the temporal variation of soil moisture at a low resolution (9 km is actually used here).

[0103] Perform fast Fourier transform on the signal to obtain its amplitude and phase at different frequencies ( Figure 3 (b)).

[0104] The time series SSM at different spatial resolutions is represented by the amplitude and phase of different frequencies, specifically:

[0105]

[0106]

[0107] Where n is the number of signal components, i is the frequency of the signal, and A i is the amplitude at low resolution, x is a variable, ψ i is the phase at low resolution, t is the total duration of the signal, a i is the amplitude at high resolution, is the phase at high resolution, I is the low resolution, and U is the high resolution.

[0108] The same operation is performed on all pixels of SMAP SSM to obtain a set of amplitude grid data and phase grid data of different frequencies ( Figure 3 (c), Figure 3 (d)).

[0109] In step S3, the phase grid data and the high-frequency amplitude grid data are spatially interpolated, specifically:

[0110] The low-frequency part of the amplitude grid data is removed, and the high-frequency amplitude grid data is spatially interpolated using bilinear interpolation to obtain high-frequency amplitude grid data, where high frequency and low frequency are divided by frequency threshold.

[0111] The phase grid data of all frequencies are spatially interpolated using the nearest neighbor method to obtain high-resolution phase grid data.

[0112] In step S3, the spatial interpolation result is fitted with the initial surface soil moisture image in the time dimension to obtain a daily-scale surface soil moisture image, including:

[0113] Based on high-resolution phase grid data and high-frequency amplitude grid data, combined with the soil moisture contribution of the initial surface soil moisture image in the time discrete condition to the grid data in the low-frequency component, downscaling fitting is performed in the time dimension to obtain a high-resolution surface soil moisture image on the daily scale.

[0114] According to the measured soil moisture at the site, the high-resolution surface soil moisture image at the daily scale is corrected by coupling inverse Fourier transform and random forest.

[0115] The following is a detailed description of step S3:

[0116] One thing that needs to be made clear is that according to the sampling law, interpolation or fitting of limited sampling points performs better on low-frequency information than high-frequency information. The algorithm removes the low-frequency part of the amplitude grid data ( Figure 3 (c) and use bilinear interpolation to calculate the high-frequency amplitude grid data ( Figure 3 The blue part in (c) is used for spatial interpolation, and the phase grid data of all frequencies are interpolated using the nearest neighbor method to obtain the phase grid data with a spatial resolution of 20m ( Figure 3 (f)) and high-frequency amplitude grid data ( Figure 3 (e)). This approach is based on the fact that the spatial changes of high-frequency information such as irrigation, precipitation, and evapotranspiration are low-frequency. Therefore, the influence of these factors on soil moisture can be downscaled in the spatial dimension using interpolation methods, assuming that the initial time of all events within 9 km is the same. At this time, only {a0,…,a n} etc. There are n+1 unknowns that need to be solved. If there are more than n+1 equations, all the unknowns can be solved ( Figure 3 (h)) thus achieving the spatiotemporal downscaling of SSM. The establishment of the solution equation depends on the spatial downscaling results of the second step SMAP SSM ( Figure 3 (g)), that is, the SSM downscaling result for each day, an equation can be established. This approach is based on the fact that the high-frequency information in space, such as infiltration and groundwater recharge, varies at a low frequency in time, and the contribution of these factors to soil moisture can be downscaled in the time dimension.

[0117] However, since the second step of constructing the surface soil moisture inversion model with SMAP SSM as the dependent variable, it is believed that the accuracy of the SMAP SSM product is high enough, which is undoubtedly an ideal state. In fact, the accuracy of SMAP SSM products in some areas lacking ground data needs to be improved. Taking the area of ​​the present invention as an example, soil moisture is seriously underestimated, so it is necessary to correct the downscaled product. The specific correction method is to use a random forest model instead of the original polynomial sum during the inverse Fourier transform, and the SSM observations of the Urad Rear Banner Station are used to train the random forest model.

[0118] The above steps can be used to downscale the SMAP SSM to obtain a surface soil moisture product with a daily spatial resolution of 20 m ( Figure 3 (j)).

[0119] Based on the above method, the present invention provides a system for generating surface soil moisture based on spatiotemporal fusion of remote sensing data, including: an acquisition module, a downscaling module and a fusion module.

[0120] Among them, the acquisition module is used to obtain the vegetation parameter image, active microwave remote sensing image and passive microwave remote sensing image of the required date, and superimpose the vegetation parameter image with the active microwave remote sensing image to obtain a high-resolution parameter image; the spatial resolution of the high resolution is higher than the set threshold, and the spatial resolution of the low resolution is lower than the set threshold; the downscaling module is used to resample the spatial resolution of the parameter image to a low resolution, and spatially downscale the passive microwave remote sensing image through the low-resolution parameter image to obtain a time-discrete initial surface soil moisture image; the fusion module is used to use the time series of each pixel in the passive microwave remote sensing image as a signal, obtain the phase grid data and high-frequency amplitude grid data of the signal at different frequencies, and perform spatial interpolation on the phase grid data and the high-frequency amplitude grid data, fit the spatial interpolation result with the initial surface soil moisture image in the time dimension, and obtain a daily-scale surface soil moisture image.

[0121] The present invention provides a computer device, comprising a memory and a processor. A computer program is stored in the memory. When the computer program is executed by the processor, the processor executes the steps of a method for generating surface soil moisture based on spatiotemporal fusion of remote sensing data.

[0122] In accordance with the disclosed embodiments, a computing device may communicate with one or more external devices (e.g., keyboards, pointing devices, Bluetooth communications, etc.), or with any device (e.g., routers, modems, etc.) that enables a computing device to communicate with one or more other computing devices.

[0123] The present invention provides a storage medium on which a computer program is stored. When the computer program is executed by a processor, steps of a method for generating surface soil moisture based on time-space fusion of remote sensing data are implemented.

[0124] According to the disclosed embodiments, the storage medium may be a non-volatile computer-readable storage medium, such as but not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, the storage medium may be any tangible medium containing or storing a program that may be used by or in conjunction with an instruction execution system, apparatus, or device.

[0125] The above content is a further detailed description of the present invention in combination with a specific preferred embodiment. For technicians in the technical field to which the present invention belongs, several simple deductions or substitutions can be made without departing from the concept of the present invention, which should be regarded as belonging to the protection scope of the present invention.

Claims

1. A method for generating surface soil moisture based on spatiotemporal fusion of remote sensing data, characterized in that: include: Obtain vegetation parameter images, active microwave remote sensing images and passive microwave remote sensing images on the required date, and superimpose the vegetation parameter images with the active microwave remote sensing images to obtain high-resolution parameter images; wherein the spatial resolution of the high-resolution image is higher than a set threshold, and the spatial resolution of the low-resolution image is lower than the set threshold; The spatial resolution of the parameter image is resampled to a low resolution, and the passive microwave remote sensing image is spatially downscaled by the low-resolution parameter image to obtain a time-discrete initial surface soil moisture image; The time series of each pixel in the passive microwave remote sensing image is used as a signal, and the phase grid data and high-frequency amplitude grid data of the signal at different frequencies are obtained. The phase grid data and the high-frequency amplitude grid data are spatially interpolated, and the spatial interpolation result is fitted with the initial surface soil moisture image in the time dimension to obtain a daily-scale surface soil moisture image.

2. A method for generating surface soil moisture based on spatiotemporal fusion of remote sensing data as claimed in claim 1, characterized in that: The step of obtaining the vegetation parameter image of the required date includes: Get different combinations of vegetation parameters; Input different vegetation parameter combinations into the radiation transfer model, simulate and obtain a spectral curve, and convert the spectral curve into atmospheric apparent reflectance to obtain the reflectance of the band captured by the model; The band reflectance is input into the random forest model for inversion to obtain the vegetation parameter image of the required date; wherein the vegetation parameter image includes the leaf area index LAI, the canopy water content CWC and the average leaf inclination angle ALA image.

3. The method for generating surface soil moisture based on spatiotemporal fusion of remote sensing data as claimed in claim 1, characterized in that: The vegetation parameter image is superimposed with the active microwave remote sensing image to obtain a high-resolution parameter image, specifically: By superimposing the active microwave remote sensing image with the leaf area index LAI, canopy water content CWC and average leaf inclination angle ALA images in the vegetation parameter image, a set of images containing seven bands including normalized vegetation index NDVI, leaf area index LAI, canopy water content CWC, average leaf inclination angle ALA, vertical polarization VV, cross polarization VH and radar wave incident angle θ was obtained; The seven-band images are used as high-resolution parameter images.

4. The method for generating surface soil moisture based on spatiotemporal fusion of remote sensing data according to claim 1, characterized in that: The method of spatially downscaling the passive microwave remote sensing image by using the low-resolution parameter image to obtain a time-discrete initial surface soil moisture image includes: The spatial resolution of the high-resolution parameter image is resampled to a low resolution corresponding to the pixels of the passive microwave remote sensing image, and the pixels are used as a training set to train the surface soil moisture inversion model; The high-resolution parameter image is input into the trained surface soil moisture inversion model to obtain a high-resolution initial surface soil moisture image under time discrete conditions.

5. The method for generating surface soil moisture based on spatiotemporal fusion of remote sensing data as claimed in claim 1, characterized in that: The time series of each pixel in the passive microwave remote sensing image is used as a signal to obtain phase grid data and high-frequency amplitude grid data of the signal at different frequencies, including: Performing a fast Fourier transform on the signal to obtain the amplitude and phase of the signal at different frequencies; The time series SSM at different spatial resolutions is represented by the amplitude and phase of different frequencies, specifically: Where n is the number of signal components, i is the frequency of the signal, and A i is the amplitude at low resolution, x is a variable, ψ i is the phase at low resolution, t is the total duration of the signal, a i is the amplitude at high resolution, is the phase at high resolution, I is the low resolution, and U is the high resolution.

6. The method for generating surface soil moisture based on spatiotemporal fusion of remote sensing data according to claim 1, characterized in that: The phase grid data and the high-frequency amplitude grid data are spatially interpolated, specifically: Eliminate the low-frequency part in the amplitude grid data, and use bilinear interpolation to perform spatial interpolation on the high-frequency amplitude grid data to obtain high-frequency amplitude grid data; wherein the high frequency and low frequency are divided by a frequency threshold; The phase grid data of all frequencies are spatially interpolated using the nearest neighbor method to obtain high-resolution phase grid data.

7. A method for generating surface soil moisture based on spatiotemporal fusion of remote sensing data as claimed in claim 6, characterized in that: The spatial interpolation result is fitted with the initial surface soil moisture image in the time dimension to obtain a daily-scale surface soil moisture image, including: Based on the high-resolution phase grid data and high-frequency amplitude grid data, combined with the soil moisture contribution of the initial surface soil moisture image in the time discrete situation to the grid data in the low-frequency component, downscaling fitting is performed in the time dimension to obtain a high-resolution surface soil moisture image on a daily scale; According to the measured soil moisture at the site, the high-resolution surface soil moisture image at the daily scale is corrected by coupling inverse Fourier transform and random forest.

8. A system for generating surface soil moisture based on spatiotemporal fusion of remote sensing data, characterized in that: include: The acquisition module is used to obtain vegetation parameter images, active microwave remote sensing images and passive microwave remote sensing images of the required date, and to superimpose the vegetation parameter images with the active microwave remote sensing images to obtain high-resolution parameter images; wherein the spatial resolution of the high-resolution image is higher than a set threshold, and the spatial resolution of the low-resolution image is lower than the set threshold; A downscaling module is used to resample the spatial resolution of the parameter image to a low resolution, and to spatially downscale the passive microwave remote sensing image through the low-resolution parameter image to obtain a time-discrete initial surface soil moisture image; The fusion module is used to take the time series of each pixel in the passive microwave remote sensing image as a signal, obtain the phase grid data and high-frequency amplitude grid data of the signal at different frequencies, and spatially interpolate the phase grid data and the high-frequency amplitude grid data, fit the spatial interpolation result with the initial surface soil moisture image in the time dimension, and obtain a daily-scale surface soil moisture image.

9. A computer device, characterized in that: It includes a memory and a processor, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the processor executes the steps of a method for generating surface soil moisture based on spatiotemporal fusion of remote sensing data as described in any one of claims 1 to 7.

10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a method for generating surface soil moisture based on spatiotemporal fusion of remote sensing data according to any one of claims 1 to 7 are implemented.

Images