A soil moisture estimation method, medium and program product

Abstract

The present invention relates to a soil moisture estimation method, medium and program product, and belongs to the technical field of soil moisture measurement. The method first inputs green band data G, red band data R, near infrared band data NIR, slope data SLOPE and normalized vegetation index NDVI into a random forest model to obtain an estimated soil moisture. SM RF ; Secondly, calculate the soil moisture estimated based on the environmental information model SM NM ; Finally, based on soil moisture SM RF and soil moisture SM NM The final soil moisture SM is obtained. The present invention reduces the dependence on the accuracy of the random forest model while ensuring the accuracy of soil moisture estimation.

Description

Technical Field

[0001] The present invention relates to the technical field of soil moisture measurement, and to the measurement and estimation technology of soil moisture, and in particular to a soil moisture estimation method combining a random forest model, a model based on environmental information and a drying measurement method, as well as a computer program product and a readable storage medium for implementing the method. Background Art

[0002] Soil moisture is an important parameter in many fields such as farmland management, water resources management and environmental protection. Timely and accurate acquisition of soil moisture information is crucial for scientific management and optimization of agricultural production. Traditional soil moisture measurement methods include drying method and time domain reflectometry (TDR). Although the drying method has high accuracy, it is complex to operate, time-consuming and highly destructive; although the TDR method is fast and non-destructive, the equipment is expensive and highly dependent on the soil environment. In addition, these methods are difficult to implement in large-area soil moisture monitoring and cannot meet the needs of real-time and wide-area measurement.

[0003] With the promotion and popularization of remote sensing technology and artificial intelligence, artificial intelligence models based on remote sensing image information have been applied to the field of soil moisture measurement. For example, CN116482128A uses a radar mounted on a drone to detect the soil, extracts soil moisture features from the collected radar signals, inputs the features into a trained deep neural network model, and outputs soil moisture estimation results; JP2024031096A proposes a low-cost technology for estimating soil moisture content, which uses an estimation model that uses RGB image data obtained by photographing the soil and measurement data obtained by measuring soil moisture to learn the correlation between RGB image data and measurement data.

[0004] Although these soil moisture measurement methods have solved the limitations of traditional methods to a certain extent, their accuracy is highly dependent on the accuracy of the model, and the training of high-precision models requires a large number of tests and large sample data sets. In addition, these methods also face technical problems in data collection, processing, and model deployment in practical applications. For example, the acquisition of radar signals requires specialized hardware equipment and complex signal processing algorithms, and the acquisition of RGB image data is greatly affected by factors such as lighting conditions and shooting angles, which increases the difficulty of model training and practical application. Summary of the invention

[0005] In view of the defects and deficiencies of the existing soil moisture estimation methods in terms of model accuracy, data requirements and practical applications, in order to solve at least one of the above and other technical problems in the prior art, the present invention aims to provide a soil moisture estimation method that combines a random forest model, a model based on environmental information and a drying measurement method. By integrating the random forest model with the model based on environmental information and combining the high-precision characteristics of the drying measurement method, it is achieved that while ensuring the accuracy of soil moisture estimation, the dependence on the accuracy of a single random forest model is reduced, the generalization ability and adaptability of the model are improved, and thus the accuracy and practicality of soil moisture estimation are improved.

[0006] The first invention object of the present invention is to propose a soil moisture estimation method, which estimates soil moisture by integrating a random forest model, an environmental information model and a drying measurement method. The method comprises the following steps when implemented:

[0007] Step S1: input the green band data G, the red band data R, the near infrared band data NIR, the slope data SLOPE and the normalized difference vegetation index NDVI into the random forest model to obtain the estimated soil moisture. SM RF ;

[0008] Step S2: Calculate the soil moisture estimated based on the environmental information model according to the following formula: SM NM ,

[0009]

[0010] Where NR is the net radiation, AH is the air humidity, LST is the surface temperature, SM NM is the soil moisture estimated by the model based on environmental information, a, b, c, d, e are coefficients;

[0011] Step S3, based on soil moisture SM RF and soil moisture SM NM Get the final soil moisture SM,

[0012]

[0013] Among them, w1 and w2 are SM RF and SM NM The corresponding weight.

[0014] Furthermore, step S1 specifically includes the following sub-steps during implementation:

[0015] Sub-step S11, configuring a random forest model;

[0016] Sub-step S12, training and testing the random forest model;

[0017] Sub-step S13, based on the multispectral imager, obtain green band data G, red band data R, near infrared band data NIR, slope data SLOPE and normalized vegetation index NDVI, and standardize them. The standardized green band data G, red band data R, near infrared band data NIR, slope data SLOPE and normalized vegetation index NDVI are input into the random forest model obtained in the above sub-step S12 to obtain the estimated soil moisture. SM RF .

[0018] Furthermore, step S2 specifically includes the following sub-steps during implementation:

[0019] Sub-step S21, establishing a model based on environmental information, wherein the model based on environmental information is represented by the following formula:

[0020]

[0021] Where NR is the net radiation, AH is the air humidity, LST is the surface temperature, SM NM is the soil moisture estimated by the model based on environmental information, a, b, c, d, e are coefficients;

[0022] Sub-step S22, calculating coefficients a, b, c, d, e of the model based on the environmental information, and then obtaining the model based on the environmental information.

[0023] Sub-step S23, collecting net radiation NR, air humidity AH, and ground surface temperature LST, and inputting the net radiation NR, air humidity AH, and ground surface temperature LST into the model based on environmental information obtained in the above sub-step S22 to obtain estimated soil moisture SM NM .

[0024] Furthermore, step S3 specifically includes the following sub-steps during implementation:

[0025] Sub-step S31, establish a fusion calculation model based on soil moisture SM RF and soil moisture SM NM To obtain the final soil moisture SM, the fusion calculation model is expressed by the following formula:

[0026]

[0027] Among them, w1 and w2 are SM RF andSM NM The corresponding weights;

[0028] Sub-step S32, calculate in combination with the drying measurement method SM RF and SM NM The corresponding weights w1 and w2 obtain the fusion calculation model.

[0029] Sub-step S33, the soil moisture obtained in step S1 is SM RF and the soil moisture obtained in step S2 SM NM The fusion calculation model obtained in the above sub-step S32 is input to obtain the soil moisture SM.

[0030] The present invention also includes a computer program product, which includes a computer program, and when the computer program is executed by a processor, the soil moisture estimation method can be implemented.

[0031] The present invention also includes a readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the soil moisture estimation method can be implemented.

[0032] Compared with the prior art, the present invention integrates the random forest model, the model based on environmental information and the drying measurement method, while ensuring the accuracy of soil moisture estimation, it reduces the dependence on the accuracy of the random forest model. Through the standardized processing of multi-source data and the application of the fusion calculation model, the accuracy and reliability of soil moisture estimation are significantly improved. The random forest model uses multi-band data to estimate soil moisture, which enhances the robustness of the model; the model based on environmental information combines key environmental parameters to achieve a more comprehensive soil moisture estimation; the drying measurement method provides high-precision benchmark data for optimizing the weights of the fusion calculation model. This method not only improves the estimation accuracy, but also enhances the generalization ability and adaptability of the model. It is suitable for large-scale, real-time soil moisture monitoring, and provides important technical support for agricultural management and water resource optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a schematic diagram of the soil moisture estimation method;

[0034] Figure 2 Soil moisture SM RF Schematic diagram of the calculation method;

[0035] Figure 3 Soil moisture SM NM Schematic diagram of the calculation method;

[0036] Figure 4 Schematic diagram of the final soil moisture SM calculation method. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solution and advantages of the present invention clearer, the embodiments of the present invention will be further described in detail in conjunction with the accompanying drawings. The present invention provides a soil moisture estimation method, and the accompanying drawings will show the specific operation flow and data processing process of each step, including the configuration and training of the random forest model, the establishment and calculation based on the environmental information model, and the application of the fusion calculation model. Through these detailed illustrations and descriptions, the technical solution of the present invention will be more intuitive and easy to understand, and the significant technical effects of the present invention in improving the accuracy and applicability of soil moisture estimation will be further explained.

[0038] The embodiment of the present invention provides a high-precision soil moisture estimation method, which combines a random forest model, a model based on environmental information, and a traditional drying measurement method, aiming to improve the accuracy and practicality of soil moisture estimation. The method reduces the dependence on the accuracy of a single model through multi-source data fusion and model optimization, enhances the generalization ability and adaptability of the model, and can provide more reliable soil moisture measurement results in practical applications. Specifically, Figure 1 As shown, the soil moisture estimation method of the present invention comprises the following steps:

[0039] Step S1: input the green band data G, the red band data R, the near infrared band data NIR, the slope data SLOPE and the normalized difference vegetation index NDVI into the random forest model to obtain the estimated soil moisture. SM RF ;

[0040] Step S2: Calculate the soil moisture estimated based on the environmental information model according to the following formula: SM NM ,

[0041]

[0042] Where NR is the net radiation, AH is the air humidity, LST is the surface temperature, SM NM is the soil moisture estimated by the model based on environmental information, a, b, c, d, e are coefficients;

[0043] Step S3, based on soil moisture SM RF and soil moisture SM NM Get the final soil moisture SM,

[0044]

[0045] Among them, w1 and w2 are SM RF and SM NM The corresponding weight.

[0046] Through the above steps, the soil moisture estimation method of the present invention can reduce the dependence on the accuracy of a single model while ensuring the accuracy of soil moisture estimation, improve the generalization ability and adaptability of the model, and improve the accuracy and practicality of soil moisture estimation.

[0047] In some preferred embodiments, Figure 2 As shown, the above step S1 may specifically include the following sub-steps when implemented:

[0048] Sub-step S11, configure the random forest model. Specifically, the number of decision trees of the model is selected from one of the number sets {20, 40, 60, 80, 100, 120, 140, 160}, preferably 40; the maximum depth of a single decision tree is selected from one of the number sets {10, 20, 30, 40, 50, 60, 70, 80}, preferably 20; the minimum number of samples required for internal node splitting is one of {1, 2, 3, 4, 5, 6, 7, 8, 9, 10}, preferably 5; the minimum number of samples required for leaf nodes is one of {1, 2, 3, 4, 5, 6, 7, 8, 9, 10}, preferably 4. The model input is the standardized green band data G, red band data R, near infrared band data NIR, slope data SLOPE and normalized vegetation index NDVI; the output layer is the estimated soil moisture SM RF .

[0049] Sub-step S12, training and testing the random forest model. Specifically, N groups of data DATE_a are collected to train and test the random forest model, each group of data DATE_a includes: soil moisture collected by a time domain reflectometer at a certain time, green band data G, red band data R, near infrared band data NIR, slope data SLOPE and normalized vegetation index NDVI obtained at the same time and standardized by the remote sensing data processing module, and each group of data DATE_a is collected at different times.

[0050] Sub-step S13, calculating soil moisture SM RFSpecifically, green band data G, red band data R, near infrared band data NIR, slope data SLOPE and normalized vegetation index NDVI are obtained based on the multispectral imager and standardized. The standardized green band data G, red band data R, near infrared band data NIR, slope data SLOPE and normalized vegetation index NDVI are input into the random forest model obtained in the above sub-step S12 to obtain the estimated soil moisture. SM RF .

[0051] Step S1 ensures the consistency of data input and the accuracy of model output by reasonably configuring the parameters of the random forest model. When configuring the random forest model, the appropriate number of decision trees, depth, and minimum number of samples for splitting and leaf nodes are selected to balance the model between complexity and generalization ability. In addition, this method of combining remote sensing image data and measured data fully utilizes the advantages of different data sources and improves the accuracy and reliability of the model. At the same time, through standardization, the dimensional differences between different features are eliminated, which helps the convergence and generalization ability of the model.

[0052] In some preferred embodiments, Figure 3 As shown, the above step S2 may specifically include the following sub-steps when implemented:

[0053] Sub-step S21, establishing a model based on environmental information, wherein the model based on environmental information is represented by the following formula:

[0054]

[0055] Where NR is the net radiation, AH is the air humidity, LST is the surface temperature, SM NM is the soil moisture estimated by the model based on environmental information, a, b, c, d, e are coefficients;

[0056] Sub-step S22, calculate the coefficients a, b, c, d, e of the model based on the environmental information, and then obtain the model based on the environmental information. Specifically, collect M groups of data DATE_b to calculate the coefficients a, b, c, d, e of the model based on the environmental information. Each group of data DATE_b includes: soil moisture collected by the time domain reflectometer at a certain time, net radiation NR, air humidity AH, and surface temperature LST obtained at the same time. Each group of data DATE_b is collected at different times.

[0057] Sub-step S23, calculating soil moisture SM NMSpecifically, net radiation NR, air humidity AH, and ground surface temperature LST are collected and input into the model based on environmental information obtained in the above sub-step S22 to obtain estimated soil moisture. SM NM .

[0058] In the above sub-steps S12 and S22, the time domain reflectometer belongs to the existing technology. Its principle is that the speed at which high-frequency electromagnetic waves propagate in the soil depends on the dielectric constant of the soil, and soil moisture greatly affects the dielectric constant of the soil. Therefore, the dielectric constant of the soil can be calculated based on the time it takes for the electromagnetic waves to propagate in the soil, and then the soil moisture content can be calculated using the empirical relationship between the soil dielectric constant and the soil volume moisture content.

[0059] Through the above steps, the soil moisture estimation model based on environmental information can effectively utilize key environmental parameters such as net radiation (NR), air humidity (AH) and land surface temperature (LST), combined with soil moisture data obtained by time domain reflectometry, to establish a high-precision estimation model. This model based on environmental information can significantly improve the accuracy of soil moisture estimation.

[0060] In some preferred embodiments, Figure 4 As shown, the above step S3 specifically includes the following sub-steps when implemented:

[0061] Sub-step S31, establish a fusion calculation model based on soil moisture SM RF and soil moisture SM NM To obtain the final soil moisture SM, the fusion calculation model is expressed by the following formula:

[0062]

[0063] Among them, w1 and w2 are SM RF and SM NM The corresponding weights;

[0064] Sub-step S32, calculate SM RF and SM NM The corresponding weights w1 and w2 are used to obtain the fusion calculation model. Specifically, Z groups of data DATE_c are collected to calculate SM RF and SM NM The corresponding weights w1 and w2. Each set of data DATE_c includes: the soil moisture measured by drying the soil samples collected at a certain time SMOD , the standardized green band data G, red band data R, near infrared band data NIR, slope data SLOPE and normalized vegetation index NDVI acquired at the same time, as well as the net radiation NR, air humidity AH and surface temperature LST collected at the same time; based on the random forest model obtained in the above sub-step S12 and the model based on environmental information obtained in the above sub-step S22, the soil moisture of group Z can be obtained from the above group Z data DATE_c SM OD , SM RF and SM NM , the soil moisture SM OD Corresponding to the soil moisture SM, the soil moisture of group Z is SM OD , SM RF and SM NM The fusion calculation model established in the above sub-step S31 is introduced to obtain the values ​​of the Z groups of weights w1 and w2 (that is, each group of weights includes a w1 value and a w2 value), and the average value of the Z groups of weights w1 and w2 is obtained by calculating the average value. SM RF and SM NM The corresponding values ​​of weights w1 and w2 are then used to obtain the fusion calculation model.

[0065] Sub-step S33, the soil moisture obtained in step S1 is SM RF and the soil moisture obtained in step S2 SM NM The fusion calculation model obtained in the above sub-step S32 is input to obtain the soil moisture SM.

[0066] In the above sub-step S32, the soil sample to be tested is weighed to obtain the sample wet weight W WET Then the soil sample to be tested is completely dried in a drying oven to obtain the sample dry weight W DRY , drying measured soil moisture .

[0067] In some preferred embodiments, in the above sub-steps S12, S13, and S32, the method for obtaining the green band data G, the red band data R, the near infrared band data NIR, the slope data SLOPE, and the normalized vegetation index NDVI includes:

[0068] Step a, obtaining blue band (e.g., wavelength 0.450–0.515 μm) data B, green band (e.g., wavelength 0.525–0.600 μm) data G, red band (e.g., wavelength 0.630–0.680 μm) data R, and near infrared band (e.g., wavelength 0.845–0.885 μm) data NIR based on a multispectral imager, the multispectral imager is carried on a UAV or remote sensing satellite, such as Landsat-8;

[0069] Step b, obtaining a digital elevation model (DEM) based on the visible light image of the multispectral imager, and collecting slope data SLOPE based on the digital elevation model;

[0070] Step c: obtaining the Normalized Difference Vegetation Index (NDVI) based on the blue band, green band, and red band of the multispectral imager.

[0071] In the above sub-steps S12, S13, and S32, the standardization refers to standardizing the green band data G, the red band data R, the near infrared band data NIR, the slope data SLOPE, and the normalized vegetation index NDVI by the Z-score standardization method, and the standardization process is based on the following formula:

[0072]

[0073] Where X represents the initial eigenvalue, X Mean represents the average value; X Std Represents the standard deviation of the eigenvalues.

[0074] In the above sub-steps S22, S23, and S32, the net radiation NR, air humidity AH, and surface temperature LST are collected by a net radiation meter, a hygrometer, and a surface temperature inversion module, respectively. Among them, the surface temperature inversion module is preferably based on the inversion calculation of UAV thermal infrared remote sensing data. The UAV thermal infrared remote sensing technology has the advantage of controllable time resolution. Combined with the net radiation and atmospheric humidity that can be collected in a timely manner, the environmental information collection of the present invention has high real-time performance, which is convenient for experimenters to collect environmental information that matches the remote sensing information collection time.

[0075] The soil moisture estimation method of the present invention integrates the random forest model, the model based on environmental information and the drying measurement method, thereby reducing the dependence on the accuracy of the random forest model while ensuring the accuracy of soil moisture estimation.

[0076] The present invention also includes a computer program product, which includes a computer program, and when the computer program is executed by a processor, the soil moisture estimation method can be implemented.

[0077] The present invention also includes a readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the soil moisture estimation method can be implemented.

[0078] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principle of the present invention should be covered by the protection scope of the present invention.

Claims

1. A soil moisture estimation method, characterized in that: The method comprises at least the following steps when implemented: Step S1: input the green band data G, the red band data R, the near infrared band data NIR, the slope data SLOPE and the normalized difference vegetation index NDVI into the random forest model to obtain the estimated soil moisture. SM RF ; The step S1 specifically includes: Sub-step S11, configure the random forest model, the model input is the standardized green band data G, red band data R, near infrared band data NIR, slope data SLOPE and normalized vegetation index NDVI; the output layer is the estimated soil moisture SM RF ; Sub-step S12, training and testing the random forest model; Sub-step S13, based on the multispectral imager, obtain the green band data G, the red band data R, the near infrared band data NIR, the slope data SLOPE and the normalized vegetation index NDVI, and standardize them, and input the standardized parameters into the random forest model obtained in the above sub-step S12 to obtain the estimated soil moisture SM RF ; In sub-step S12, N groups of data DATE_a are collected to train and test the random forest model, each group of data DATE_a includes: soil moisture collected by a time domain reflectometer at a certain time, green band data G, red band data R, near infrared band data NIR, slope data SLOPE and normalized vegetation index NDVI obtained at the same time and standardized by a remote sensing data processing module; Step S2: Calculate the soil moisture estimated based on the environmental information model according to the following formula: SM NM , Where NR is the net radiation, AH is the air humidity, LST is the surface temperature, SM NM is the soil moisture estimated by the model based on environmental information, a, b, c, d, e are coefficients; The step S2 specifically includes: Sub-step S21, establishing a model based on environmental information, wherein the model based on environmental information is represented by the following formula: Sub-step S22, calculating coefficients a, b, c, d, e of the model based on the environmental information, and then obtaining the model based on the environmental information; Sub-step S23, collecting net radiation NR, air humidity AH, and ground surface temperature LST, and inputting the net radiation NR, air humidity AH, and ground surface temperature LST into the model based on environmental information obtained in the above sub-step S22 to obtain estimated soil moisture SM NM ; In sub-step S22, M groups of data DATE_b are collected to calculate coefficients a, b, c, d, e of the model based on environmental information, and each group of data DATE_b includes: soil moisture collected by a time domain reflectometer at a certain time, net radiation NR, air humidity AH, and surface temperature LST obtained at the same time; Step S3, based on soil moisture SM RF and soil moisture SM NM Get the final soil moisture SM,+, where w1 and w2 are SM RF and SM NM The corresponding weights; The step S3 specifically includes: Sub-step S31, establishing a fusion calculation model, the fusion calculation model is represented by the following formula: Sub-step S32, calculate SM RF and SM NM The corresponding weights w1 and w2 are used to obtain the fusion calculation model; Sub-step S33, the soil moisture obtained in step S1 is SM RF and the soil moisture obtained in step S2 SM NM Input the fusion calculation model obtained in the above sub-step S32 to obtain soil moisture SM; In sub-step S32, Z groups of data DATE_c are collected, each group of data DATE_c includes: soil moisture measured by drying the soil sample collected at a certain time SM OD , the standardized green band data G, red band data R, near infrared band data NIR, slope data SLOPE and normalized vegetation index NDVI acquired at the same time, as well as the net radiation NR, air humidity AH and surface temperature LST collected at the same time; based on the random forest model obtained in the above sub-step S12 and the model based on environmental information obtained in the above sub-step S22, the soil moisture of group Z can be obtained from the above group Z data DATE_c SM OD , SM RF and SM NM , the soil moisture SM OD Corresponding to the soil moisture SM, the soil moisture of group Z is SM OD , SM RF and SM NM The fusion calculation model established in the above sub-step S31 can be used to obtain the values ​​of the Z group weights w1 and w2. The average value of the Z group weights w1 and w2 is calculated to obtain SM RF and SM NM The corresponding values ​​of weights w1 and w2 are then used to obtain the fusion calculation model.

2. The soil moisture estimation method according to claim 1, characterized in that: In sub-step S32, the soil sample to be tested is weighed to obtain the sample wet weight W WET Then the soil sample to be tested is completely dried in a drying oven to obtain the sample dry weight W DRY , drying measured soil moisture .

3. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 2 can be implemented.

4. A readable storage medium, characterized in that: The readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in any one of claims 1 to 2 can be implemented.

Images