GNSS-R deep soil moisture inversion method considering precipitation information

Abstract

The present application belongs to a kind of ground remote sensing monitoring technology, and it relates to a kind of GNSS-R deep soil moisture inversion method considering precipitation information, comprising: S1, build ground GNSS-R dual-antenna monitoring platform;S2, utilize signal processing module to GNSS signal processing;S3, carry out preliminary inversion of soil moisture;S4, step-by-step correction eliminates the influence of precipitation;S5, deep soil moisture acquisition.The GNSS-R deep soil moisture inversion method considering precipitation information of the present application adopts multi-step correction method, considers from change rate and precipitation amount in many aspects, and corrects the obtained surface soil moisture;Finally, the statistical result obtained from a large amount of data further obtains deeper soil moisture.The present application greatly reduces the time required in the calculation process, realizes the real-time inversion of deeper soil moisture, and meets the agricultural application demand.

Description

Technical Field

[0001] This invention relates to a remote sensing monitoring technology for Earth, and more particularly to a GNSS-R deep soil moisture inversion method that takes precipitation information into account. Background Technology

[0002] GNSS-R is an emerging method for soil moisture monitoring, offering advantages over traditional technologies, including low cost and high spatiotemporal resolution. Its low cost allows more agricultural producers and research institutions to utilize GNSS-R technology for large-scale soil moisture monitoring and research. The high spatiotemporal resolution makes the monitoring data more accurate, reflecting changes in soil moisture over time and space.

[0003] However, due to the relatively weak GNSS-R signal, with an L-band wavelength of approximately 20 cm, its ability to penetrate soil layers is poor. Higher surface soil moisture content leads to stronger signal attenuation, resulting in shallower GNSS-R signal penetration. Therefore, when using GNSS-R technology to retrieve soil moisture, the effectiveness is often limited by the surface soil moisture content. For important crops like wheat, during the crucial sowing-emergence stage, the main root system is distributed at a depth of 15-30 cm; therefore, agricultural production focuses more on soil moisture at depths above 15 cm. Rainfall, a common weather phenomenon, causes significant variations in soil moisture, exhibiting complex characteristics depending on the intensity and duration of rainfall. For example, with low rainfall, only surface soil moisture changes, while deeper soil moisture remains largely unaffected. With high rainfall, not only does surface soil moisture increase rapidly, but it also gradually penetrates deeper over time, influencing deeper soil layers and resulting in a more consistent trend in soil moisture across different layers.

[0004] Therefore, how to use GNSS-R technology to avoid errors caused by precipitation and accurately obtain soil moisture at a depth of about 15cm is one of the technical problems that needs to be solved. Summary of the Invention

[0005] The purpose of this invention is to solve the above-mentioned problems and provide a GNSS-R deep soil inversion method that takes into account precipitation information.

[0006] To achieve the above objectives, this invention provides a GNSS-R deep soil inversion method that considers precipitation information, comprising:

[0007] S1. Construct a ground-based GNSS-R dual-antenna monitoring platform;

[0008] S2. Process GNSS signals using the signal processing module;

[0009] S3. Conduct preliminary soil moisture inversion;

[0010] S4. Step-by-step correction to eliminate the impact of precipitation;

[0011] S5. Obtaining moisture from deep soil layers.

[0012] Further, S1 includes:

[0013] A liftable platform is erected, and a waterproof enclosure is installed below the platform. The signal acquisition unit and workstation are installed inside the waterproof enclosure. The enclosure is equipped with a right-hand circularly polarized antenna and a left-hand circularly polarized antenna. The right-hand circularly polarized antenna is angled upwards at 45° to receive direct GNSS signals, and the left-hand circularly polarized antenna is angled downwards at 45° to receive reflected GNSS signals. The right-hand circularly polarized antenna and the left-hand circularly polarized antenna are connected to the corresponding channels of the signal acquisition unit.

[0014] Furthermore, S2 employs a CPU+GPU architecture for GNSS signal processing:

[0015] The signal acquisition unit outputs the frequency-converted intermediate frequency digital signal, which is transmitted from the CPU to the GPU. The GPU performs coarse and fine signal acquisition to obtain the peak power of direct and reflected signals. The results are then transmitted to the CPU to obtain the surface reflectivity and output it.

[0016] Furthermore, coarse and fine acquisition of GNSS-R signals includes:

[0017] 1) Calculate the visible satellite range in advance based on the latitude and longitude of the experimental location;

[0018] 2) Input parameters to the CPU, including the visual satellite PRN, coarse acquisition coherent integration duration, fine acquisition coherent and incoherent integration duration, and Doppler frequency search range and search step size;

[0019] 3) Next, use the GPU for coarse capture;

[0020] 4) Based on the code phase and frequency information obtained from coarse acquisition, complete the fine acquisition of the signal and obtain the peak power of the direct and reflected signals.

[0021] Further, S3 includes:

[0022] Pre-establish a lookup table for reflectance and soil moisture:

[0023] When a GNSS signal reaches a reflecting surface, it undergoes complete reflection. The Fresnel reflection coefficient is used to describe the quantitative relationship between the reflected energy and the incident energy during the reflection process of an electromagnetic wave.

[0024]

[0025] In the formula, ε is the complex permittivity of the soil surface layer; θ is the satellite elevation angle; Γ is the polarization coefficient; v and h represent vertical and horizontal polarization, respectively; and l and r represent left-hand circularly polarized antenna and right-hand circularly polarized antenna, respectively.

[0026] The reflectivity corresponding to different polarization modes can be calculated using the Fresnel reflection coefficients mentioned above.

[0027]

[0028] Where R rl R represents the reflectivity of a left-handed circularly polarized signal. rr Indicates the reflectivity of a right-hand circularly polarized signal;

[0029] To quantify the reflection characteristics of GNSS satellite signals on the Earth's surface, the ratio of reflected signal power to direct signal power is calculated, and this ratio is used as a representation of surface reflectivity.

[0030]

[0031] In the formula, P r It is the power of the reflected signal received by the receiver; P d The power of the direct signal received by the receiver;

[0032] The soil dielectric constant ε is obtained from the relationship between reflectivity R and dielectric constant:

[0033]

[0034] Calculating soil moisture using the Topp model:

[0035]

[0036] Where, m v Reflectance represents soil moisture; a lookup table of reflectance and soil moisture is established based on the results.

[0037] Then, based on the reflectance obtained in step S2, the soil moisture value is directly estimated using the nearest neighbor interpolation method.

[0038] Furthermore, S4 includes:

[0039] Correction using the rate of change:

[0040] Calculate the difference value (diff) for the inversion results of each hour:

[0041] diff i =SM i -SM i-1

[0042] When the diff value is greater than 0.05 or less than -0.05, all soil moisture values ​​within this time period are uniformly multiplied by a coefficient K = 0.85 to correct the soil moisture value.

[0043] SM i =0.85*SM i diff i <-0.05 or>0.05

[0044] Among them SM i This represents the soil moisture value measured in the i-th hour.

[0045] Furthermore, S4 also includes:

[0046] Correction using a piecewise model:

[0047] For soil moisture corrected by the rate of change, a soil moisture value of 0.3m was used. 3 / m 3 With 0.6m 3 / m 3 Using the boundary as a guideline, and based on different intervals, precise calculations are performed using corresponding correction formulas:

[0048]

[0049] Furthermore, S4 also includes:

[0050] Correction using precipitation:

[0051] For the soil moisture results that have been corrected using the piecewise model, when a specific time t occurs... n After verification, no precipitation occurred at that time or in the preceding two hours. If the soil moisture at this time is compared to the adjacent previous time t... n-1 However, if the trend is upward, then the abnormal data should be corrected, that is, t n-1 The soil moisture value at time t is used as t n-1 The result of the moment is used to replace the originally unreasonable t. n Values ​​are inverted at any given time.

[0052] Furthermore, in S5:

[0053] Calculate the soil moisture at a depth of 15cm using the following formula:

[0054] SM deep =SM+0.03.

[0055] The beneficial effects of this invention are as follows:

[0056] The GNSS-R deep soil moisture inversion method of this invention, which considers precipitation information, has significant advantages over traditional monitoring methods such as contact measurement and optical remote sensing. In terms of economic cost, it uses freely available GNSS signals; it achieves a relatively high level of spatiotemporal resolution and enables continuous monitoring around the clock, making it an extremely effective means of soil moisture monitoring. This invention employs a CPU+GPU program architecture, significantly improving program efficiency; it establishes a lookup table for soil moisture ranging from 0 to 1, directly obtaining soil moisture values ​​based on surface reflectivity; it uses a multi-step correction method, considering factors such as the rate of change and precipitation, to correct the obtained surface soil moisture; and finally, it uses statistical results obtained from a large amount of data to further derive deeper levels of soil moisture. These innovations greatly reduce the time required for computation, enabling real-time inversion of deeper soil moisture levels and meeting the needs of agricultural applications. Attached Figure Description

[0057] Figure 1 This schematic diagram illustrates the data processing flow of the GNSS-R deep soil moisture inversion method that takes precipitation information into account according to the present invention.

[0058] Figure 2 A schematic diagram illustrating the reflectance and soil moisture lookup representation of the GNSS-R deep soil moisture inversion method considering precipitation information according to the present invention.

[0059] Figure 3 This diagram illustrates the verification results of the GNSS-R deep soil moisture inversion method based on precipitation information according to the present invention. Detailed Implementation

[0060] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly described below. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0061] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. The embodiments cannot be described in detail here, but the embodiments of the present invention are not limited to the following embodiments.

[0062] Combination Figure 1As shown, the present invention provides a GNSS-R deep soil moisture inversion method that takes precipitation information into account, including: S1, building a ground-based GNSS-R dual-antenna monitoring platform; S2, processing GNSS signals using a signal processing module; S3, performing preliminary soil moisture inversion; S4, stepwise correction to remove the influence of precipitation; S5, acquiring deep soil moisture.

[0063] In S1, assuming a lifting platform approximately 15m high, a waterproof enclosure is installed beneath the platform. Inside the enclosure, a 62MHz signal acquisition unit is housed, and a right-hand circularly polarized antenna and a left-hand circularly polarized antenna are fixed to the workstation enclosure. The right-hand circularly polarized antenna faces upwards at a 45° angle to receive direct GNSS signals, while the left-hand circularly polarized antenna faces downwards at a 45° angle to receive reflected GNSS signals. Both antennas are connected to their respective channels on the signal acquisition unit. Then, the remote control device on the workstation opens the acquisition unit control software, modifies the sampling duration as needed, and performs the reception and storage of GNSS intermediate frequency data.

[0064] Then, a CPU+GPU architecture is used in S2 for GNSS signal processing:

[0065] The signal acquisition unit outputs a frequency-converted intermediate frequency digital signal, which is transmitted from the CPU to the GPU. The GPU performs coarse and fine acquisition of the signal to obtain the peak power of the direct and reflected signals. The results are then transmitted back to the CPU to obtain the surface reflectivity and output it. The coarse and fine acquisition of the GNSS-R signal includes: 1) calculating the visible satellite range based on the latitude and longitude of the experimental location; 2) inputting parameters to the CPU, including the visible satellite PRN, the coarse acquisition coherent integration time, the fine acquisition coherent and incoherent integration time, and the Doppler frequency search range and search step size; 3) then performing coarse acquisition using the GPU; and 4) based on the code phase and frequency information obtained from the coarse acquisition, completing the fine acquisition of the signal to obtain the peak power of the direct and reflected signals.

[0066] Then, in S3, the obtained surface reflectance is used to obtain the corresponding soil moisture value according to a pre-established reflectance and soil moisture lookup table. The reflectance and soil moisture lookup table includes:

[0067] Ideally, GNSS signals undergo complete reflection upon reaching the reflecting surface. The Fresnel reflection coefficient is used to describe the quantitative relationship between the reflected energy and the incident energy during the reflection process of electromagnetic waves.

[0068]

[0069] In the formula, ε is the complex permittivity of the soil surface layer; θ is the satellite elevation angle; Γ is the polarization coefficient; v and h represent vertical and horizontal polarization, respectively; and l and r represent left-hand circularly polarized antenna and right-hand circularly polarized antenna, respectively.

[0070] The reflectivity corresponding to different polarization modes can be calculated using the Fresnel reflection coefficients mentioned above.

[0071]

[0072] Where R rl R represents the reflectivity of a left-handed circularly polarized signal. rr Indicates the reflectivity of a right-hand circularly polarized signal;

[0073] To quantify the reflection characteristics of GNSS satellite signals on the Earth's surface, the ratio of reflected signal power to direct signal power is calculated, and this ratio is used as a representation of surface reflectivity.

[0074]

[0075] In the formula, P r It is the power of the reflected signal received by the receiver; P d The power of the direct signal received by the receiver;

[0076] The soil dielectric constant ε is obtained from the relationship between reflectivity R and dielectric constant:

[0077]

[0078] Calculating soil moisture using the Topp model:

[0079]

[0080] Where, m v The soil moisture value is represented by the reflectance obtained in step S2, and then the soil moisture value is directly estimated by nearest neighbor interpolation.

[0081] Figure 2 This is a lookup table showing the relationship between soil moisture and reflectivity. However, under conditions of precipitation, the accumulation of water on the soil surface leads to a significant increase in GNSS signal reflectivity, resulting in a large discrepancy between the soil moisture obtained through inversion and the in-situ data. Further corrections are needed in S4, specifically including:

[0082] Correction using the rate of change:

[0083] Calculate the difference value (diff) for the inversion results of each hour:

[0084] diff i =SM i-SM i-1

[0085] When the diff value is greater than 0.05 or less than -0.05, all soil moisture values ​​within this time period are uniformly multiplied by a coefficient K = 0.85 to correct the soil moisture value.

[0086] SM i =0.85*SM i diff i <-0.05 or>0.05

[0087] Among them SM i This represents the soil moisture value measured in the i-th hour. This operation effectively reduces drastic changes in moisture values, making the inversion results closer to the true soil moisture state.

[0088] Subsequently, the soil moisture data obtained in the previous step was corrected using a piecewise model:

[0089] With a soil moisture value of 0.3m 3 / m 3 With 0.6m 3 / m 3 Using the boundary as a guideline, and based on different intervals, precise calculations are performed using corresponding correction formulas:

[0090]

[0091] Furthermore, S4 also includes:

[0092] Correction using precipitation:

[0093] For the soil moisture results that have undergone piecewise model correction, a reasonableness assessment is performed by incorporating precipitation data: when a specific time t occurs... n After verification, no precipitation occurred at that time or in the preceding two hours. If the soil moisture at this time is compared to the adjacent previous time t... n-1 However, if the trend is upward, then the abnormal data should be corrected, that is, t n-1 The soil moisture value at time t is used as t n-1 The result of the moment is used to replace the originally unreasonable t. n Values ​​are inverted at any given time.

[0094] Finally, to obtain deeper soil moisture values, further processing of the results after the aforementioned series of corrections is necessary. Considering that deep soil moisture, due to its unique physical location, is less affected by external environmental factors and typically exhibits a larger value compared to surface soil moisture, this invention employs a method of uniformly adding a constant to the correction results. After systematic analysis and research of a large amount of long-term experimental data, this constant was determined to be 0.03. This operation allows the inversion results to more accurately reflect the actual state of deep soil moisture, effectively improving the usability and scientific rigor of the data. In other words, deep soil moisture is obtained as follows:

[0095] SM deep =SM+0.03

[0096] The present invention will be described below with reference to specific embodiments:

[0097] The experimental data acquisition frequency was set at 120 seconds per hour for a total of 48 hours. The intermediate frequency data sampling rate was 62MHz. Soil moisture monitoring was performed using the QZSS GEO L5 frequency signal, specifically including:

[0098] S1. Relevant parameter settings:

[0099] Based on the GNSS-R test results of this foundation, the following settings were made:

[0100] 1) Pre-store the local code and local phase of L5, thereby reducing the time consumed by each repeated generation;

[0101] 2) The duration of the incoherent integration is 50ms;

[0102] 3) Doppler frequency shift range and search step size settings: The search range for coarse acquisition is [-5000Hz, +5000Hz], with a step size of 200Hz, for a total of 51 search frequency points; the search range for fine acquisition is [-300Hz, 300Hz], with a step size of 25Hz, for a total of 25 search frequency points.

[0103] 4) The signal acquisition device stores 120 seconds of data per hour.

[0104] S2. Data Acquisition:

[0105] The signal to be processed every second is read into the CPU and then transmitted from the CPU to the GPU for signal capture. After coarse and fine capture, the capture result is transmitted back to the CPU.

[0106] S3. Data Processing

[0107] The surface reflectance is obtained by calculating the peak ratio of direct and reflected signals on the CPU. The initial soil moisture value is obtained by using the nearest neighbor interpolation method based on a lookup table.

[0108] The obtained soil moisture values ​​were corrected step by step:

[0109] 1) First, calculate the difference value diff for the inversion results every hour. For all soil moisture values ​​that are in the period from diff greater than 0.05 to less than -0.05, i.e. moisture values ​​that change too drastically and do not conform to the natural change law, multiply them by a coefficient of 0.85.

[0110] 2) Correct the piecewise model by substituting the corrected result into the formula:

[0111]

[0112] 3) After correction by the segmented model, precipitation is corrected according to the rules described above;

[0113] 4) Based on long-term observation data from the TDR hygrometer, the difference between the surface soil and the soil at a depth of about 15cm is approximately 0.03. This difference is added to the surface soil moisture after all corrections to obtain the moisture content of the deep soil.

[0114] S4. Result Verification

[0115] like Figure 3 As shown, to verify the correctness of the results in this embodiment, in-situ soil moisture data of the detected area were collected using a TDR soil moisture meter. Comparing the in-situ data at a depth of 15cm with the detected soil moisture, the experimental data accuracy was found to be 0.035m. 3 / m 3 This verifies the correctness of this embodiment.

[0116] This invention presents a GNSS-R deep soil moisture inversion method that incorporates precipitation information. By combining the CPU's superior logic control capabilities with the GPU's unique advantages in parallel processing of massive amounts of data, this program architecture overcomes the problem of low computational efficiency in GNSS-R software. This innovative combination significantly improves program processing speed, enabling the originally time-consuming process to be completed efficiently, achieving real-time calculation of soil moisture inversion, enhancing the practicality of this invention, and providing solid technical support for GNSS-R soil moisture detection on ground-based platforms.

[0117] The GNSS-R deep soil moisture inversion method of the present invention, which takes into account precipitation information, can eliminate the interference caused by precipitation on the soil moisture inversion values ​​by a piecewise correction function. By setting different judgment thresholds and combining precipitation data, the abnormally high soil moisture inversion values ​​due to surface water accumulation are corrected accordingly, thereby obtaining reliable deep soil moisture data.

[0118] Furthermore, a lookup table was established through forward calculation, and the Topp model was used to invert the data obtained from the experiments to obtain the soil moisture value calculated based on the soil dielectric constant. Based on the correspondence between reflectivity and soil moisture value, soil moisture was interpolated with a step size of 0.001. By first interpolating and then extrapolating, a 0-1cm range was established. 3 / cm 3 This method uses a range of soil moisture lookup tables to directly estimate soil moisture content based on reflectivity, simplifying the intermediate parameter of soil dielectric constant. It solves the problem of requiring a large number of inverse calculations in the traditional inversion process and greatly reduces the time required for the original model to solve.

[0119] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A GNSS-R method for inverting deep soil moisture considering precipitation information, characterized in that, include: S1. Construct a ground-based GNSS-R dual-antenna monitoring platform; S2. Process GNSS signals using the signal processing module; S3. Conduct preliminary soil moisture inversion; S4. Step-by-step correction to eliminate the impact of precipitation; S5. Deep soil moisture acquisition; S4 includes: Correction using the rate of change: Calculate the difference value (diff) for the inversion results of each hour: When the diff value is greater than 0.05 or less than -0.05, all soil moisture values ​​within this time period are uniformly multiplied by a coefficient K=0.85 to correct the soil moisture value. in Indicates the first Soil moisture values ​​measured hourly; S4 further includes: Correction using a piecewise model: For soil moisture corrected for the rate of change, a soil moisture value of 0.3 was used. With 0.6 Using the boundary as a guideline, and based on different intervals, precise calculations are performed using corresponding correction formulas: ； S4 further includes: Correction using precipitation: For soil moisture results that have been corrected using a segmented model, when a specific moment occurs... n After verification, no precipitation occurred at that time or in the preceding two hours. If the soil moisture at this time is compared to the immediately preceding time... n-1 However, if the trend is upward, then the abnormal data will be corrected, that is... n-1 Soil moisture value at any given time as n-1 The result of the moment is used to replace the originally unreasonable t. n Values ​​are inverted at any given time.

2. The GNSS-R deep soil moisture inversion method considering precipitation information according to claim 1, characterized in that, S1 includes: A liftable platform is erected, and a waterproof enclosure is installed below the platform. The signal acquisition unit and workstation are installed inside the waterproof enclosure. The enclosure is equipped with a right-hand circularly polarized antenna and a left-hand circularly polarized antenna. The right-hand circularly polarized antenna is angled upwards at 45° to receive direct GNSS signals, and the left-hand circularly polarized antenna is angled downwards at 45° to receive reflected GNSS signals. The right-hand circularly polarized antenna and the left-hand circularly polarized antenna are connected to the corresponding channels of the signal acquisition unit.

3. The GNSS-R deep soil moisture inversion method considering precipitation information according to claim 2, characterized in that, The S2 uses a CPU+GPU architecture for GNSS signal processing. The signal acquisition unit outputs the frequency-converted intermediate frequency digital signal, which is transmitted from the CPU to the GPU. The GPU performs coarse and fine signal acquisition to obtain the peak power of direct and reflected signals. The results are then transmitted to the CPU to obtain the surface reflectivity and output it.

4. The GNSS-R deep soil moisture inversion method considering precipitation information according to claim 3, characterized in that, Coarse and fine acquisition of GNSS-R signals include: 1) Calculate the visible satellite range in advance based on the latitude and longitude of the experimental location; 2) Input parameters to the CPU, including the visual satellite PRN, coarse acquisition coherent integration duration, fine acquisition coherent and incoherent integration duration, and Doppler frequency search range and search step size; 3) Next, use the GPU for coarse capture; 4) Based on the code phase and frequency information obtained from coarse acquisition, complete the fine acquisition of the signal and obtain the peak power of the direct and reflected signals.

5. The GNSS-R deep soil moisture inversion method considering precipitation information according to claim 4, characterized in that, S3 includes: Pre-establish a lookup table for reflectance and soil moisture: When a GNSS signal reaches a reflecting surface, it undergoes complete reflection. The Fresnel reflection coefficient is used to describe the quantitative relationship between the reflected energy and the incident energy during the reflection process of an electromagnetic wave. In the formula The complex permittivity of the soil surface layer; The satellite's elevation angle; The polarization coefficient; , These represent vertical polarization and horizontal polarization, respectively. , These represent left-hand circularly polarized antennas and right-hand circularly polarized antennas, respectively. The reflectivity corresponding to different polarization modes can be calculated using the Fresnel reflection coefficients mentioned above. in Indicates the reflectivity of a left-handed circularly polarized signal; Indicates the reflectivity of a right-hand circularly polarized signal; To quantify the reflection characteristics of GNSS satellite signals on the Earth's surface, the ratio of reflected signal power to direct signal power is calculated, and this ratio is used as a representation of surface reflectivity. In the formula, It is the power of the reflected signal received by the receiver; The power of the direct signal received by the receiver; The soil dielectric constant is obtained from the relationship between reflectivity R and dielectric constant. : Calculating soil moisture using the Topp model: in, This represents soil moisture, and then a lookup table of reflectance and soil moisture is created based on the results. Finally, based on the reflectance obtained in step S2, the soil moisture value is directly estimated using the nearest neighbor interpolation method.

6. The GNSS-R deep soil moisture inversion method considering precipitation information according to claim 5, characterized in that, In S5: Calculate the soil moisture at a depth of 15cm using the following formula: .

Images