Soil moisture determination method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a soil moisture determination method and device, electronic equipment and a storage medium, and relates to the field of data processing, and the method comprises the steps: obtaining to-be-processed observation data with a time sequence, decomposing the to-be-processed observation data through a single-channel signal decomposition algorithm, and obtaining the to-be-processed observation data; a plurality of modal function components are obtained, a false-removed virtual multi-dimensional matrix is constructed based on the plurality of modal function components, the false-removed virtual multi-dimensional matrix is decomposed by using a multi-channel blind source decomposition algorithm to obtain a plurality of blind source signals, and a soil moisture trend signal is determined from the plurality of blind source signals according to soil moisture prior information. And based on the soil moisture prior information, performing normalization processing on the soil moisture trend signal to obtain estimated soil moisture information corresponding to the to-be-processed observation data. According to the method, the morbidity of soil moisture remote sensing inversion can be reduced, and the defect that soil moisture microwave remote sensing quantitative inversion is limited by variables such as surface roughness and vegetation coverage and is difficult to accurately quantify is overcome.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a method, device, electronic device and storage medium for determining soil moisture. Background Art

[0002] Soil moisture is one of the core variables of the earth's surface system and plays an important role in practical applications such as watershed / regional agricultural management, flood forecasting and numerical weather forecasting. The development of remote sensing technology provides an effective means for obtaining soil moisture information at regional and global scales. Remote sensing monitoring can reflect the temporal changes and spatial distribution of soil moisture. It has the advantages of wide monitoring range, high efficiency, low cost, and periodic monitoring, and has become the focus of current research.

[0003] At present, the methods for determining soil moisture are usually as follows: the empirical method based on statistical numerical fitting generally establishes a linear / nonlinear relationship between remote sensing observation brightness temperature and soil moisture to invert soil moisture through statistics and analysis of a series of observation data, the current soil moisture is inverted based on the forward model numerical inversion algorithm, and the soil moisture estimation method based on machine learning.

[0004] The above methods of determining soil moisture all have certain problems and are not effective. Summary of the invention

[0005] The object of the present invention is to provide a soil moisture determination method, device, electronic device and storage medium, which can improve the accuracy of determining soil moisture.

[0006] In order to achieve the above purpose, the technical solution adopted in the embodiment of the present application is as follows:

[0007] In a first aspect, an embodiment of the present application provides a method for determining soil moisture, the method comprising:

[0008] Obtain the observation data to be processed with time series;

[0009] Decomposing the observation data to be processed by using a single-channel signal decomposition algorithm to obtain a plurality of modal function components, and constructing a false-free virtual multidimensional matrix based on the plurality of modal function components;

[0010] Decomposing the de-false virtual multi-dimensional matrix using a multi-channel blind source decomposition algorithm to obtain multiple blind source signals;

[0011] Determining a soil moisture trend signal from a plurality of blind source signals according to soil moisture prior information;

[0012] Based on the soil moisture prior information, the soil moisture trend signal is normalized to obtain the estimated soil moisture information corresponding to the observation data to be processed.

[0013] In an optional embodiment, the step of constructing a false-free virtual multidimensional matrix based on the multiple modal function components includes:

[0014] Processing the plurality of modal function components in sequence based on a whitening algorithm and a de-averaging algorithm to obtain a plurality of standardized modal function components;

[0015] Processing the plurality of standardized modal function components based on Akaike information criterion and / or average profile width algorithm to obtain the number of blind sources of the observation data to be processed;

[0016] respectively calculating the correlations between the observed data to be processed and the multiple modal function components;

[0017] Obtaining multiple target modal function components from multiple modal function components based on the number of blind sources;

[0018] A false-free virtual multidimensional matrix is ​​constructed based on the observation data to be processed and a plurality of target modal function components.

[0019] In an optional implementation, the step of decomposing the false-free virtual multidimensional matrix using a multi-channel blind source decomposition algorithm to obtain a plurality of blind source signals includes:

[0020] The de-false virtual multi-dimensional matrix is ​​decomposed based on an independent component analysis algorithm and / or a non-negative matrix decomposition algorithm to obtain a plurality of blind source signals.

[0021] In an optional implementation, the step of obtaining the observation data to be processed having a time series includes:

[0022] Obtain real ground observation data;

[0023] Inputting the real ground observation data into a microwave radiation transmission model to output a microwave brightness temperature simulating a rough / smooth land surface;

[0024] The microwave brightness temperature is taken as observation data to be processed with a time series.

[0025] In an optional implementation, the step of obtaining the observation data to be processed having a time series includes:

[0026] Obtain SMAP brightness temperature data with time series;

[0027] Processing the outliers of the SMAP brightness temperature data with time series based on a sliding average convolution algorithm to obtain target SMAP brightness temperature data;

[0028] The target SMAP brightness temperature data is used as observation data to be processed with a time series.

[0029] In an optional implementation, the step of normalizing the soil moisture trend signal based on the soil moisture prior information to obtain the estimated soil moisture information corresponding to the observation data to be processed includes:

[0030] Determine the maximum value and the minimum value of the soil moisture prior information;

[0031] The soil moisture trend signal is normalized based on the maximum change value and the minimum change value to obtain estimated soil moisture information corresponding to the observation data to be processed.

[0032] In an optional embodiment, the method further comprises:

[0033] Determining actual soil moisture information corresponding to the estimated soil moisture information;

[0034] Calculating the difference between the estimated soil moisture information and the actual soil moisture information;

[0035] The accuracy of the estimated soil moisture information is evaluated based on the difference.

[0036] In a second aspect, an embodiment of the present application provides a soil moisture determination device, the device comprising:

[0037] An acquisition module, used to acquire observation data to be processed with a time series;

[0038] A signal decomposition module is used to decompose the observation data to be processed using a single-channel signal decomposition algorithm to obtain multiple modal function components and construct a false-free virtual multidimensional matrix based on the multiple modal function components; decompose the false-free virtual multidimensional matrix using a multi-channel blind source decomposition algorithm to obtain multiple blind source signals;

[0039] The determination module is used to determine the soil moisture trend signal from the multiple blind source signals according to the soil moisture prior information; based on the soil moisture prior information, the soil moisture trend signal is normalized to obtain the estimated soil moisture information corresponding to the observation data to be processed.

[0040] In a third aspect, an embodiment of the present application provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the soil moisture determination method when executing the computer program.

[0041] In a fourth aspect, an embodiment of the present application provides a storage medium having a computer program stored thereon, which implements the steps of the soil moisture determination method when executed by a processor.

[0042] This application has the following beneficial effects:

[0043] This application obtains the observation data to be processed with a time series, decomposes the observation data to be processed using a single-channel signal decomposition algorithm, obtains multiple modal function components, and constructs a false virtual multidimensional matrix based on the multiple modal function components, decomposes the false virtual multidimensional matrix using a multi-channel blind source decomposition algorithm to obtain multiple blind source signals, determines the soil moisture trend signal from the multiple blind source signals based on the soil moisture prior information, and normalizes the soil moisture trend signal based on the soil moisture prior information to obtain the estimated soil moisture information corresponding to the observation data to be processed. This application estimates soil moisture based on blind source decomposition. Considering the autocorrelation of information before and after microwave remote sensing instantaneous observations, providing time dimension constraints for blind source decomposition estimation of soil moisture, combining single-channel with multi-channel blind source decomposition methods to increase the prior information of remote sensing estimation of soil moisture, can reduce the pathological nature of remote sensing inversion of soil moisture, and make up for the deficiency that quantitative inversion of microwave remote sensing of soil moisture is limited by variables such as surface roughness and vegetation coverage, which are difficult to accurately quantify. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0045] Figure 1 A block diagram of an electronic device provided by an embodiment of the present invention;

[0046] Figure 2 One of the flow charts of a method for determining soil moisture provided by an embodiment of the present invention;

[0047] Figure 3 A second flow chart of a method for determining soil moisture provided by an embodiment of the present invention;

[0048] Figure 4 A third flow chart of a method for determining soil moisture provided by an embodiment of the present invention;

[0049] Figure 5 A fourth flowchart of a method for determining soil moisture provided by an embodiment of the present invention;

[0050] Figure 6 A fifth flow chart of a method for determining soil moisture provided by an embodiment of the present invention;

[0051] Figure 7A sixth flow chart of a method for determining soil moisture provided by an embodiment of the present invention;

[0052] Figure 8 A structural block diagram of a soil moisture determination device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0054] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0056] In the description of the present invention, it should be noted that if the terms "upper", "lower", "inside", "outside", etc. appear to indicate an orientation or position relationship, they are based on the orientation or position relationship shown in the accompanying drawings, or are the orientation or position relationship in which the product of the invention is usually placed when used. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0057] In addition, the terms “first”, “second”, etc., if used, are merely used to distinguish between the descriptions and should not be understood as indicating or implying relative importance.

[0058] In the description of this application, it should also be noted that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0059] After extensive research, the inventors found that the current soil moisture passive microwave remote sensing inversion algorithm has experienced a statistical method with incomplete theoretical basis and lack of understanding of physical mechanisms, a forward model numerical inversion algorithm that models the relationship between characteristic parameters and sensor received signals and requires the input of a large number of surface parameters and has difficulty in inverting surface characteristic parameters, and a machine learning algorithm that requires sample training.

[0060] The empirical method based on statistical numerical fitting generally inverts soil moisture by statistically analyzing a series of observation data and establishing a linear / nonlinear relationship between remote sensing observation brightness temperature and soil moisture. Although the inversion process is simple and the accuracy of inverting soil moisture in a specific area is high, due to the incomplete theoretical basis, lack of understanding and recognition of physical mechanisms, and lack of logical relationship between parameters, the inversion algorithm has poor universality in different regions.

[0061] The forward model numerical inversion algorithm is the mainstream means of soil moisture inversion at present. The microwave radiation transmission model describes the radiation transmission process in which the microwave signal radiated from the natural surface reaches the top of the atmosphere after being attenuated by vegetation, atmosphere and other media during the remote sensing process and is received by the sensor. The input of the forward model is surface variables such as soil moisture, soil temperature and surface roughness, and the output is remote sensing observations. The so-called inversion process is to obtain the input variables of the forward model through the sensor observations, and find the model input variables with the smallest difference between remote sensing observations and forward model output as the inversion result. Therefore, the inversion based on the forward model generally requires two steps: forward model selection and optimization algorithm selection. At present, the most commonly used forward models are based on the τ-ω zero-order radiation transmission model, and single-channel method, dual-channel method, L-band microwave radiation model of the biosphere, surface parameter inversion model, etc. have been developed; in addition, a lookup table inversion method has been constructed for AMSR2 microwave soil moisture products. Although this type of method has a clear physical basis, the algorithm itself is relatively complex and involves many surface parameters. In addition, variables such as surface roughness and vegetation cover are difficult to accurately quantify at the microwave pixel scale, which inevitably leads to errors in soil moisture inversion.

[0062] The soil moisture estimation method based on machine learning has characteristics that other methods do not have. The biggest advantage is that it can theoretically approximate any complex nonlinear relationship without the need to design a complex algorithm and does not rely on a complex physical forward model. Assuming that there is a linear / nonlinear relationship between surface parameters and microwave remote sensing observations, the surface parameters and remote sensing observations are regarded as the input data set of the model, and the real soil moisture observations are regarded as the output value of the model. By continuously training and testing the network, the nonlinear relationship between soil moisture and remote sensing observations is obtained, and then the soil moisture is estimated based on the nonlinear relationship. At present, machine learning algorithms have been widely used in the study of passive microwave remote sensing soil moisture estimation. Although machine learning algorithms can improve the accuracy of soil moisture estimation, this method often relies on the selection of training samples. If the training samples are not representative enough, the estimation results are poor. In addition, since machine learning training is a "black box" non-parametric process, it is difficult to see the internal mechanism of the association between the input and output data sets. The poor interpretability of the physical mechanism of soil moisture estimation limits its application.

[0063] Through the above review and summary, the passive microwave soil moisture remote sensing inversion / estimation method has been continuously improved, mainly reflected in the continuous improvement of the physical mechanism of the forward model and the continuous optimization of the soil moisture inversion / estimation method. However, due to the characteristics of the satellite itself and the limitations of the surface and atmospheric observation parameters and methods, there is room for further exploration and improvement in the soil moisture inversion / estimation method. Surface roughness and vegetation cover are two key variables that affect whether soil moisture can be accurately inverted / estimated. How to reduce the influence of these two variables and invert soil moisture more quantitatively is the difficulty of current research. At present, the soil moisture inversion algorithm does not consider the intrinsic information before and after the instantaneous remote sensing observation, and time series remote sensing observations have autocorrelation characteristics. Therefore, in the process of soil moisture inversion, considering the information before and after the instantaneous value of remote sensing observations can impose certain constraints on the inversion of soil moisture.

[0064] In view of the discovery of the above problems, the present embodiment provides a soil moisture determination method, device, electronic device and storage medium, which can obtain the observation data to be processed with a time series, decompose the observation data to be processed using a single-channel signal decomposition algorithm to obtain multiple modal function components, and construct a false-free virtual multidimensional matrix based on the multiple modal function components, decompose the false-free virtual multidimensional matrix using a multi-channel blind source decomposition algorithm to obtain multiple blind source signals, determine the soil moisture trend signal from the multiple blind source signals according to the soil moisture / brightness temperature prior information, and normalize the soil moisture trend signal based on the prior soil moisture information to obtain the estimated soil moisture information corresponding to the observation data to be processed. This application decomposes the time series microwave brightness temperature based on the blind source decomposition algorithm, considers the autocorrelation of information before and after microwave remote sensing instantaneous observation, provides time dimension constraints for soil moisture blind source decomposition estimation, combines single-channel and multi-channel blind source decomposition methods to improve the accuracy of soil moisture remote sensing estimation, reduces the pathological nature of soil moisture remote sensing inversion, and makes up for the deficiency that soil moisture microwave remote sensing quantitative inversion is difficult to accurately quantify due to the limitation of variables such as surface roughness and vegetation cover. The solution provided in this embodiment is elaborated in detail below.

[0065] This embodiment provides an electronic device that can determine soil moisture. In a possible implementation, the electronic device can be a user terminal, for example, the electronic device can be, but is not limited to, a server, a smart phone, a personal computer (PC), a tablet computer, a personal digital assistant (PDA), a mobile Internet device (MID), etc.

[0066] Please refer to Figure 1 , Figure 1 1 is a schematic diagram of the structure of the electronic device 100 provided in the embodiment of the present application. The electronic device 100 may also include Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown. Figure 1 Each component shown in the figure can be implemented by hardware, software or a combination thereof.

[0067] The electronic device 100 includes a soil moisture determination device 110 , a memory 120 and a processor 130 .

[0068] The memory 120 and the processor 130 are electrically connected to each other directly or indirectly to achieve data transmission or interaction. For example, these elements can be electrically connected to each other through one or more communication buses or signal lines. The soil moisture determination device 110 includes at least one software function module that can be stored in the memory 120 in the form of software or firmware or fixed in the operating system (OS) of the electronic device 100. The processor 130 is used to execute the executable modules stored in the memory 120, such as the software function modules and computer programs included in the soil moisture determination device 110.

[0069] The memory 120 may be, but is not limited to, a random access memory (RAM), a read only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable read-only memory (EEPROM), etc. The memory 120 is used to store a program, and the processor 130 executes the program after receiving an execution instruction.

[0070] Please refer to Figure 2 , Figure 2 For application Figure 1 A flow chart of a soil moisture determination method of the electronic device 100 is shown, and the method including each step is described in detail below.

[0071] S201: Obtain observation data to be processed with a time series.

[0072] S202: Decomposing the observation data to be processed by using a single-channel signal decomposition algorithm to obtain a plurality of modal function components, and constructing a false-free virtual multidimensional matrix based on the plurality of modal function components.

[0073] S203: Decompose the false virtual multi-dimensional matrix using a multi-channel blind source decomposition algorithm to obtain multiple blind source signals.

[0074] S204: Determine a soil moisture trend signal from multiple blind source signals based on soil moisture prior information.

[0075] S205: Based on the soil moisture prior information, the soil moisture trend signal is normalized to obtain estimated soil moisture information corresponding to the observation data to be processed.

[0076] As a cutting-edge method for extracting effective signals from complex time series signals, blind source decomposition continues to be active in many fields such as multi-user communication, speech processing, biomedical engineering, pollution source extraction, and image processing. Blind source decomposition originated from the "cocktail party" problem, which can be simply described as: at a noisy cocktail party where many people are talking at the same time and surrounded by background music and other noises, people can hear the topics they are interested in and ignore other interference. In this case, each sound is a source signal, the human ear is equivalent to a sensor, the human brain is equivalent to a speech processor, and the captured sound is equivalent to a blind source.

[0077] Therefore, blind source decomposition can be simply summarized as: multiple source signals are mixed in an unknown mixing manner and then received by one or more sensors, and then the multiple source signals are judged and obtained through a separation system based on the sensor received signals. The remarkable feature of blind source decomposition is that it does not require a forward model, and can be decomposed using a single or multiple observation signals; another feature is "blind", that is, the source signal is unobservable, the mixing method of the source signal is unknown, it is difficult to establish an accurate mathematical model between the transmitter and receiver, or the prior knowledge of the transmission channel is difficult to obtain. Blind source decomposition can be used to mine different source signals without the need for forward simulation, assuming that the source signals are independent of each other, in the context of little prior information about the blind source signal and its mixing process. Passive microwave remote sensing inversion of soil moisture can be summarized as: microwave signals radiated from the natural surface are attenuated by vegetation, atmosphere and other media before reaching the top of the atmosphere and being received by the sensor. Soil moisture is inverted by establishing the relationship between the sensor received signal and surface variables such as soil moisture, soil temperature, surface roughness, and vegetation cover.

[0078] The study of blind source signal decomposition originated from array signal processing technology. Taking linear mixed signal decomposition as an example, its basic mathematical model is x(t) = As(t) + n(t), where s(t) is the source signal vector, A is the unknown mixing matrix, x(t) is the observed signal, and n(t) is the noise. Blind separation of signals is actually to use the observed signal x(t) to obtain a separation matrix W, and restore the source signal through the matrix and the observed signal. Let y(t) be the output source signal estimate, then the separation system can be expressed as y(t) = g(x(t)). Then the separability of the system is to use the correlation on the mixed signal time series to find an unmixing system g(·) using the optimization function, so that y(t) is the optimal estimate or approximation of s(t), that is, it satisfies y(t) = Wx(t)≈s(t). The separability of complex signals is guaranteed by assuming that the blind source signals are independent of each other.

[0079] In recent years, blind source decomposition algorithms have been gradually introduced into the field of remote sensing for preliminary exploration, especially in the decomposition of mixed pixels in remote sensing images, with good application effects. For example, in view of the difficulty of current quantitative remote sensing - the problem of unmixing multi-component pixels in hyperspectral images, the blind source decomposition of hyperspectral information can accurately obtain the component spectra and their weight information. The blind source decomposition method based on ICA (independent component decomposition) technology is applied to the quantitative decomposition of remote sensing mixed pixels, which solves the problem of amplitude uncertainty and realizes the simultaneous acquisition of quantitative component spectrum information and component weight information from hyperspectral data. In addition, blind source decomposition has also been initially explored in the remote sensing estimation of surface variables. By analyzing 189 groups of field-measured surface visible-near infrared reflectance spectra of planting areas with different vegetation coverage and different salinization degrees, the results of comparing and evaluating the prediction of soil salt content based on the original spectrum and the spectrum after blind source decomposition show that the blind source decomposition algorithm can effectively decompose the mixed spectrum of vegetation and soil, and improve the accuracy of soil salt inversion based on visible-near infrared reflectance spectra under vegetation coverage. Therefore, the characteristic of blind source decomposition that can separate complex signals has reference value and application potential for how to separate soil moisture signals from microwave remote sensing observation signals.

[0080] When the present application obtains the observation data to be processed, the observation data to be processed is data obtained by mixing multiple blind source signals. The observation data to be processed is first decomposed using a single-channel signal decomposition algorithm to obtain multiple modal function components containing false signals.

[0081] The observation data to be processed may be simulated brightness temperature data or satellite observed brightness temperature data of the study area. For example, the time resolution of the observation data to be processed may be set to 30 minutes, 60 minutes, 90 minutes, 1 day, etc. The observation time range of the observation data to be processed may be 1 week, 10 months, 1 year, 2 years, etc., and the embodiment of the present application does not impose any specific limitation on this.

[0082] Specifically, the microwave radiation transmission model can be driven by the wireless network sensor network observation data in area A to obtain the time series observation data (simulation) to be processed. It is also possible to directly download the microwave brightness temperature data of time series such as SMAP, AMSR2 and AMSR-E as the observation data to be processed.

[0083] Based on a single-channel signal decomposition algorithm, the observation data to be processed with a time series is decomposed to obtain multiple modal function components. The single-channel signal decomposition algorithm can be a CEEMDAN (Complete Ensemble Empirical Mode Decomposition with Adaptive Noise) algorithm, an EMD (Empirical Mode Decomposition) algorithm, and a CEEMD (Complete Ensemble Empirical Mode Decomposition) algorithm, which decomposes the observation data to be processed to obtain multiple modal function components.

[0084] Based on multiple modal function components, a virtual multidimensional matrix is ​​constructed to remove false information. The virtual multidimensional matrix is ​​decomposed to obtain multiple blind source signals using a multi-channel blind source decomposition algorithm. The blind source decomposition algorithm based on single-channel and multi-channel coupling can solve the morbidity of single signal decomposition. Based on soil moisture prior information, the soil moisture signal is determined from multiple blind source signals, and the soil moisture is estimated based on the soil moisture prior information.

[0085] When decomposing a single-channel signal, adaptive white noise is added at each stage of the decomposition to overcome the non-zero reconstruction error problem of the improved empirical mode decomposition method, ensure the integrity of the decomposition, and reduce the false components and modal aliasing. However, the signal decomposition algorithm still cannot solve the underdetermination problem in the blind source decomposition process, and the decomposed source signal is far from the actual signal. Based on this problem, the effective signal decomposed from the single channel is virtualized into multiple channels, and the multi-channel blind source decomposition algorithm is used to convert the underdetermination problem into a positive solution, thereby improving the accuracy of determining soil moisture.

[0086] In another implementation method of determining the estimated moisture information corresponding to the observation data to be processed, the observation data to be processed in the first area with a time series can be obtained, and the observation data to be processed in the first area can be decomposed using a single-channel signal decomposition algorithm to obtain multiple modal function components, and a virtual multidimensional matrix that eliminates false information is constructed based on the multiple modal function components. The virtual multidimensional matrix that eliminates false information is decomposed using a multi-channel blind source decomposition algorithm to obtain multiple blind source signals, and according to soil moisture prior information, a soil moisture trend signal is determined from the multiple blind source signals, and a corresponding relationship between the soil moisture prior information in the first area and the soil moisture trend signal in the first area is established.

[0087] When determining the soil moisture information of the second region under the same watershed and climatic conditions, the observation data to be processed in the second region with a time series is obtained, and the observation data to be processed in the second region is decomposed using a single-channel signal decomposition algorithm to obtain multiple modal function components, and a virtual multidimensional matrix is ​​constructed based on the multiple modal function components to remove false information. The virtual multidimensional matrix is ​​decomposed to remove false information using a multi-channel blind source decomposition algorithm to obtain multiple blind source signals. According to the soil moisture prior information, the soil moisture trend signal of the second region is determined from the multiple blind source signals. Based on the corresponding relationship between the soil moisture prior information of the first region and the soil moisture trend signal of the first region and the soil moisture trend signal of the second region, the soil moisture information of the second region is determined.

[0088] The above soil moisture determination method can reduce the amount of calculation, thereby improving the efficiency of determining soil moisture. Example:

[0089] Assume that the soil moisture trend in area A is: Trend (a1,a2,…a n )

[0090] Assume that the soil moisture prior information of area A is: SM Obs (sm1,sm2,…sm n )

[0091] Then the mapping relationship is: Y = f (A Trend , S.M. Obs )

[0092] Assume that the soil moisture trend in area B is: Trend (b1,b2,…b n )

[0093] Then the estimated soil moisture in area B is: SM = Y(B Trend )

[0094] There are many ways to implement the construction of a virtual multidimensional matrix based on multiple modal function components. In one implementation, Figure 3 As shown, the following steps are included:

[0095] S301: sequentially processing multiple modal function components based on a whitening algorithm and a de-meaning algorithm to obtain multiple standardized modal function components.

[0096] S302: Processing a plurality of standardized modal function components based on the Akaike Information Criterion and / or the mean profile width algorithm to obtain the number of blind sources of the observation data to be processed.

[0097] S303: Calculate the correlations between the observation data to be processed and the multiple modal function components respectively.

[0098] S304: obtaining a plurality of target modal function components from a plurality of modal function components based on the number of blind sources;

[0099] S305: constructing a false-free virtual multidimensional matrix based on the observation data to be processed and the multiple target modal function components.

[0100] In view of the false components and modal aliasing of the decomposed modal function, the whitening algorithm and the demeaning algorithm are used to process each modal function component to obtain multiple standardized modal function components to reduce the correlation between each modal function component. Each standard modal function component is processed based on the Akaike information criterion and / or the average contour width algorithm to determine the final number of blind sources of the observed data to be processed.

[0101] The correlation between the observed data signal and each modal function component or each standard modal function component may be calculated using methods such as the Pearson correlation coefficient, the Spearman rank correlation coefficient, or the Kendall rank correlation coefficient.

[0102] Exemplarily, when the number of blind sources is 3, each modal function component can be sorted from large to small according to the correlation, and two first-ranked first target modal function components and second target modal function components are obtained. Then, a virtual multidimensional matrix for removing false information is constructed based on the first target modal function component, the second target modal function component and the observation data to be processed.

[0103] In addition to processing multiple standard modal function components based on the Akaike information criterion and / or the average profile width algorithm to determine the optimal number of blind sources, the present application can also determine the number of blind sources based on the Bayesian information criterion. Or based on the MH parameter estimation algorithm, the number of source signals is determined by constructing a likelihood function.

[0104] Akaike Information Criterion AIC is a standard for measuring the goodness of model fit based on information theory. Usually, AIC is a weighted function of fitting accuracy and the number of unknown parameters, and is defined as:

[0105] AIC = 2k-2ln(L);

[0106] Where k is the number of unknown parameters in the model, and L is the maximum value of the model likelihood function. When selecting the best model from a set of available models, the model with the smallest AIC is usually selected. The fewer the parameters, the smaller the AIC value, and the better the model.

[0107] The calculation formula of Bayesian Information Criterion BIC is as follows:

[0108] BIC = 2kln(n)-2ln(L);

[0109] Among them, k and L have the same meaning as AIC, and n represents the total number of nodes in the data set. Generally, when the complexity of the sample data increases, the likelihood function L will also increase, thereby increasing the BIC value. However, when this value is too large, the model will be too complex and cause overfitting.

[0110] In one example, when determining the number of blind sources, AIC and BIC can be used in collaboration to construct a standard matrix and a weight matrix for estimating the number of source signals using the two methods to further determine the number of source signals.

[0111] The average contour width algorithm ASW is an intuitive and simple data clustering quality detection method that does not rely on statistical model assumptions. Its calculation formula is as follows:

[0112]

[0113] Among them, i is the number of vectors of the multidimensional matrix after single-channel blind source decomposition, n is the number of all vectors, and a i To reconstruct the average distance between a vector and other vectors in a multidimensional matrix, b i It is the minimum average distance between a vector and all the most similar vectors in the reconstructed multidimensional matrix.

[0114] When the number of blind sources has been determined, the false components of the reconstructed multi-channel signal need to be removed according to the determined number of blind sources. It is proposed to take the observed signal to be processed and the first n-1 modal function components with the best correlation with brightness temperature as the virtual multidimensional matrix (n) corresponding to the source signal, and perform de-meaning and whitening processing on the n-dimensional matrix to further eliminate the modal aliasing phenomenon of each component, and finally obtain a virtual multidimensional matrix without false components.

[0115] There are many ways to use a multi-channel blind source decomposition algorithm to decompose and remove the false virtual multidimensional matrix to obtain multiple blind source signals. In one implementation, the false virtual multidimensional matrix is ​​decomposed based on an independent component analysis algorithm (ICA) and / or a non-negative matrix factorization algorithm (NMF) to obtain multiple blind source signals.

[0116] Exemplarily, the virtual multidimensional matrix decomposed by independent component analysis algorithm and / or non-negative matrix decomposition algorithm is decomposed to obtain multiple blind source signals. Under the assumption that the virtual multi-channel signals are independent of each other, the process of solving the independent component analysis algorithm is converted into the selection problem of cost function and optimization method. In order to effectively decompose, the mutual information between the components of the estimated signal is used as the cost function, and the stochastic gradient method is used to solve the minimum value of the cost function.

[0117] The non-negative matrix decomposition algorithm requires that the virtual multi-channel signal is non-negative. This application intends to construct a basis matrix of the separation function and the source signal, and to stabilize the non-negative matrix decomposition algorithm by continuously iteratively updating the separation function and the blind source signal through multiplication, and finally obtain multiple blind source signals.

[0118] There are many ways to obtain the observation data to be processed with time series. In one implementation, for example, Figure 4 As shown, the following steps are included:

[0119] S401: Acquire real ground observation data.

[0120] S402: Input the real ground observation data into the microwave radiation transmission model to output the microwave brightness temperature of the simulated rough / smooth land surface.

[0121] S403: Use microwave brightness temperature as observation data to be processed with a time series.

[0122] Based on the microwave radiation transmission model AIEM, the real ground observation data are used as the input parameters of the model, and then the microwave brightness temperature of the simulated rough / smooth bare earth surface is output as the observation data to be processed.

[0123] The simulated time series microwave brightness temperature is decomposed by signal decomposition algorithm to construct a false virtual multidimensional matrix. The false virtual multidimensional matrix is ​​decomposed by multi-channel blind source decomposition algorithm to obtain multiple blind source signals. The decomposed blind source signals are trend-screened according to soil moisture prior information to obtain soil moisture signal. The trend is normalized with the maximum and minimum values ​​of moisture change of soil moisture prior information to obtain the estimated soil moisture information corresponding to the observed data to be processed.

[0124] In another implementation method of obtaining observation data to be processed with a time series, such as Figure 5 As shown, the following steps are included:

[0125] S501: Acquire SMAP brightness temperature data with time series.

[0126] S502: Processing the outliers of the SMAP brightness temperature data with time series based on the sliding average convolution algorithm to obtain the target SMAP brightness temperature data.

[0127] S503: Use the target SMAP brightness temperature data as observation data to be processed with a time series.

[0128] The target SMAP brightness temperature data is decomposed by a single-channel signal decomposition algorithm to construct a false virtual multidimensional matrix. The false virtual multidimensional matrix is ​​decomposed by a multi-channel blind source decomposition algorithm to obtain multiple blind source signals. The decomposed blind source signals are trend-screened according to the soil moisture prior information to obtain the soil moisture signal. The trend is normalized with the maximum and minimum values ​​of moisture change based on the soil moisture prior information to obtain the estimated soil moisture information corresponding to the observed data to be processed.

[0129] Based on the soil moisture prior information, the soil moisture signal is normalized to obtain the estimated soil moisture information corresponding to the observed data to be processed. There are many ways to achieve this. In one implementation, Figure 6 As shown, the following steps are included:

[0130] S601: Determine the maximum and minimum changes in soil moisture prior information.

[0131] S602: normalizing the soil moisture signal based on the maximum value and the minimum value of the change to obtain estimated soil moisture information corresponding to the observation data to be processed.

[0132] Exemplarily, the soil moisture signal is as follows:

[0133] {0.06,0.12,0.18,0.24,0.30,0.36,0.42};

[0134] The maximum value of the known soil moisture prior information is 0.42 and the minimum value is 0.06.

[0135] The normalized data is:

[0136]

[0137] The final calculation results are {0.00, 0.14, 0.29, 0.43, 0.57, 0.71, 1.00}, which is used as the estimated soil moisture information.

[0138] There are several ways to determine the accuracy of the estimated soil moisture information. In one implementation, Figure 7 As shown, the following steps are included:

[0139] S701: Determine actual soil moisture information corresponding to the estimated soil moisture information.

[0140] S702: Calculate the difference between the estimated soil moisture information and the actual soil moisture information.

[0141] S703: Evaluate the accuracy of the estimated soil moisture information based on the difference.

[0142] The real soil moisture information can also be compared with the long-term soil moisture product data of SMAP, SMOS, CCI and AMSR2, and their accuracy is used to evaluate the accuracy of the estimated soil moisture information.

[0143] Another way to determine the accuracy of the estimated moisture information is to calculate evaluation indicators based on the estimated soil moisture information and the actual soil moisture information. The evaluation indicators may include RMSE (Root Mean Squared Error), ubRMSE (Unbiased RMSE), R (Pearson Correlation Coefficient) and KGE (Kling-Gupta Efficiency, a comprehensive evaluation indicator).

[0144] Specifically, RMSE is calculated by the following formula:

[0145]

[0146] θ true represents the real soil moisture information, θ results Represents estimated soil moisture information; Represents the average value of the real soil moisture information, is to estimate the average value of soil moisture information, and N is the observation time.

[0147] Exemplarily, the observation duration may be days, months, hours, etc., and this embodiment of the present application does not impose any specific limitation on this.

[0148] Calculate ubRMSE by the following formula:

[0149]

[0150] θ true represents the real soil moisture information, θ results Represents estimated soil moisture information; Represents the average value of the real soil moisture information, is to estimate the average value of soil moisture information, and N is the observation time.

[0151] R is calculated by the following formula:

[0152]

[0153] θ true represents the real soil moisture information, θ results Represents estimated soil moisture information; Represents the average value of the real soil moisture information, is to estimate the average value of soil moisture information, and N is the observation time.

[0154] KGE is calculated by the following formula:

[0155]

[0156] θ true represents the real soil moisture information, θ results Represents estimated soil moisture information; Represents the average value of the real soil moisture information, is to estimate the average value of soil moisture information, and N is the observation time.

[0157] The accuracy of the estimated soil moisture information can be evaluated by any two or more of the above evaluation indicators.

[0158] Please refer to Figure 8 The present application embodiment also provides a method for applying Figure 1 The soil moisture determination device 110 of the electronic device 100 includes:

[0159] An acquisition module 111 is used to acquire observation data to be processed with a time series;

[0160] The signal decomposition module 112 is used to decompose the observation data to be processed by using a single-channel signal decomposition algorithm to obtain multiple modal function components and construct a virtual multidimensional matrix to remove false signals based on the multiple modal function components; decompose the virtual multidimensional matrix to remove false signals by using a multi-channel blind source decomposition algorithm to obtain multiple blind source signals;

[0161] The determination module 113 is used to determine the soil moisture trend signal from the multiple blind source signals according to the soil moisture prior information; based on the soil moisture prior information, the soil moisture trend signal is normalized to obtain the estimated soil moisture information corresponding to the observation data to be processed.

[0162] The present application also provides an electronic device 100, which includes a processor 130 and a memory 120. The memory 120 stores computer executable instructions, and when the computer executable instructions are executed by the processor 130, the soil moisture determination method is implemented.

[0163] The embodiment of the present application further provides a storage medium, which stores a computer program. When the computer program is executed by the processor 130, the soil moisture determination method is implemented.

[0164] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely schematic. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the devices, methods and computer program products according to the multiple embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of a code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or the flowchart, and the combination of boxes in the block diagram and / or the flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0165] In addition, each functional module in each embodiment of the present application can be integrated together to form an independent part, or each module can exist separately, or two or more modules can be integrated to form an independent part. If the function is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a disk or an optical disk.

[0166] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0167] The above are only various implementations of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for determining soil moisture, characterized in that: The method comprises: Obtain the observation data to be processed with time series; Decomposing the observation data to be processed by using a single-channel signal decomposition algorithm to obtain a plurality of modal function components, and constructing a false-free virtual multidimensional matrix based on the plurality of modal function components; Decomposing the de-false virtual multi-dimensional matrix using a multi-channel blind source decomposition algorithm to obtain multiple blind source signals; Determining a soil moisture trend signal from a plurality of blind source signals according to soil moisture prior information; Based on the soil moisture prior information, the soil moisture trend signal is normalized to obtain the estimated soil moisture information corresponding to the observation data to be processed.

2. The method according to claim 1, characterized in that The step of constructing a virtual multidimensional matrix for removing false information based on the multiple modal function components comprises: Processing the plurality of modal function components in sequence based on a whitening algorithm and a de-averaging algorithm to obtain a plurality of standardized modal function components; Processing the plurality of standardized modal function components based on Akaike information criterion and / or average profile width algorithm to obtain the number of blind sources of the observation data to be processed; respectively calculating the correlations between the observed data to be processed and the multiple modal function components; Obtaining multiple target modal function components from multiple modal function components based on the number of blind sources; A false-free virtual multidimensional matrix is ​​constructed based on the observation data to be processed and a plurality of target modal function components.

3. The method according to claim 1, characterized in that The step of using a multi-channel blind source decomposition algorithm to decompose the false-free virtual multidimensional matrix to obtain multiple blind source signals includes: The de-false virtual multi-dimensional matrix is ​​decomposed based on an independent component analysis algorithm and / or a non-negative matrix decomposition algorithm to obtain a plurality of blind source signals.

4. The method according to claim 1, characterized in that: The step of obtaining the observation data to be processed with a time series includes: Obtain real ground observation data; Inputting the real ground observation data into a microwave radiation transmission model to output a microwave brightness temperature simulating a rough / smooth land surface; The microwave brightness temperature is taken as observation data to be processed with a time series.

5. The method according to claim 1, characterized in that The step of obtaining the observation data to be processed with a time series includes: Obtain SMAP brightness temperature data with time series; Processing the outliers of the SMAP brightness temperature data with time series based on a sliding average convolution algorithm to obtain target SMAP brightness temperature data; The target SMAP brightness temperature data is used as observation data to be processed with a time series.

6. The method according to claim 1, characterized in that The step of normalizing the soil moisture trend signal based on the soil moisture prior information to obtain the estimated soil moisture information corresponding to the observation data to be processed includes: Determine the maximum value and the minimum value of the soil moisture prior information; The soil moisture trend signal is normalized based on the maximum change value and the minimum change value to obtain estimated soil moisture information corresponding to the observation data to be processed.

7. The method according to claim 1, characterized in that The method further comprises: Determining actual soil moisture information corresponding to the estimated soil moisture information; Calculating the difference between the estimated soil moisture information and the actual soil moisture information; The accuracy of the estimated soil moisture information is evaluated based on the difference.

8. A soil moisture determination device, characterized in that: The device comprises: An acquisition module, used to acquire observation data to be processed with a time series; A signal decomposition module is used to decompose the observation data to be processed using a single-channel signal decomposition algorithm to obtain multiple modal function components and construct a false-free virtual multidimensional matrix based on the multiple modal function components; decompose the false-free virtual multidimensional matrix using a multi-channel blind source decomposition algorithm to obtain multiple blind source signals; The determination module is used to determine the soil moisture trend signal from the multiple blind source signals according to the soil moisture prior information; based on the soil moisture prior information, the soil moisture trend signal is normalized to obtain the estimated soil moisture information corresponding to the observation data to be processed.

9. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores a computer program executable by the processor, and the processor can execute the computer program to implement the steps of the method according to 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 the method according to any one of claims 1 to 7 are implemented.

Images