Soil nickel concentration inversion method and device, storage medium and computer equipment

Abstract

The application provides a soil nickel concentration inversion method, device, storage medium and computer equipment, and the soil nickel concentration inversion method comprises the following steps: acquiring real-time hyperspectral images of a target area; performing feature extraction on the real-time hyperspectral images to obtain corresponding real-time multiscale spectral features; performing feature dimension reduction on the real-time multiscale spectral features to obtain real-time feature value distribution data of the real-time multiscale spectral features in a low-dimensional space; obtaining corresponding real-time band combinations of soil nickel concentration from the real-time hyperspectral images according to a pre-constructed spectral index model and a Pearson significance test algorithm; inputting the real-time feature value distribution data and the corresponding real-time band combinations of soil nickel concentration into a pre-trained soil nickel concentration inversion model to obtain soil nickel concentration of the target area predicted by the soil nickel concentration inversion model. The application can improve the detection efficiency of soil nickel concentration.

Description

Technical Field

[0001] This application relates to the technical field of soil data acquisition, specifically to a method, apparatus, storage medium, and computer equipment for inverting soil nickel concentration. Background Technology

[0002] Nickel is one of the eight major heavy metals in soil, and it can severely pollute soil. Traditional methods for detecting nickel concentration in soil involve sampling soil samples in the field, which is inefficient and time-consuming. Summary of the Invention

[0003] The purpose of this application is to overcome the shortcomings and deficiencies of the prior art and provide a method, apparatus, storage medium and computer equipment for inverting soil nickel concentration, which can improve the detection efficiency of soil nickel concentration.

[0004] The first aspect of this application provides a method for inverting soil nickel concentration, including:

[0005] Acquire real-time hyperspectral images of the target area;

[0006] Feature extraction is performed on the real-time hyperspectral image to obtain the corresponding real-time multi-scale spectral features;

[0007] The real-time multi-scale spectral features are subjected to feature dimensionality reduction to obtain the real-time feature value distribution data of the real-time multi-scale spectral features in the low-dimensional space;

[0008] Based on the pre-constructed spectral index model and Pearson significance test algorithm, the corresponding real-time band combination of soil nickel concentration is obtained from the real-time hyperspectral image.

[0009] The real-time feature value distribution data and the corresponding real-time band combination of the soil nickel concentration are input into the pre-trained soil nickel concentration inversion model to obtain the soil nickel concentration of the target area predicted by the soil nickel concentration inversion model.

[0010] A second aspect of this application provides a soil nickel concentration inversion device, comprising:

[0011] The real-time hyperspectral image acquisition module is used to acquire real-time hyperspectral images of the target area.

[0012] The spectral feature acquisition module is used to extract features from the real-time hyperspectral image to obtain the corresponding real-time multi-scale spectral features.

[0013] The feature dimensionality reduction module is used to perform feature dimensionality reduction on the real-time multi-scale spectral features to obtain the real-time feature value distribution data of the real-time multi-scale spectral features in a low-dimensional space.

[0014] The band combination acquisition module is used to acquire the corresponding real-time band combination of soil nickel concentration from the real-time hyperspectral image based on the pre-constructed spectral index model and Pearson significance test algorithm.

[0015] The soil nickel concentration acquisition module is used to input the real-time feature value distribution data and the corresponding real-time band combination of the soil nickel concentration into a pre-trained soil nickel concentration inversion model to obtain the soil nickel concentration of the target area predicted by the soil nickel concentration inversion model.

[0016] A third aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the soil nickel concentration inversion method as described above.

[0017] A fourth aspect of this application provides a computer device including a storage device, a processor, and a computer program stored in the storage device and executable by the processor, wherein the processor executes the computer program to implement the steps of the soil nickel concentration inversion method as described above.

[0018] Compared with related technologies, this application extracts and reduces the dimensions of real-time hyperspectral images of the target area to obtain real-time feature value distribution data in low-dimensional space. Then, the real-time feature value distribution data and the corresponding real-time bands of soil nickel concentration are combined and input into a pre-trained soil nickel concentration inversion model to obtain the soil nickel concentration of the target area predicted by the soil nickel concentration inversion model. This method can predict the soil nickel concentration of multiple target areas at the same time, thus improving the efficiency of obtaining soil nickel concentration.

[0019] To provide a clearer understanding of this application, the specific embodiments of this application will be described below in conjunction with the accompanying drawings. Attached Figure Description

[0020] Figure 1 This is a flowchart of a soil nickel concentration inversion method according to an embodiment of this application.

[0021] Figure 2 This is a schematic diagram of the spectral index model of a soil nickel concentration inversion method according to an embodiment of this application.

[0022] Figure 3 This is a schematic diagram of the band combination related to soil nickel concentration in an embodiment of the soil nickel concentration inversion method of this application.

[0023] Figure 4 This is a schematic diagram of the module connections of a soil nickel concentration inversion device according to an embodiment of this application.

[0024] 100. Soil nickel concentration inversion device; 101. Real-time hyperspectral image acquisition module; 102. Spectral feature acquisition module; 103. Feature dimensionality reduction module; 104. Band combination acquisition module; 105. Soil nickel concentration acquisition module. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0026] It should be understood that the described embodiments are merely some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of the embodiments of this application.

[0027] In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. In the description of this application, it should be understood that the terms "first," "second," "third," etc., are used only to distinguish similar objects and are not necessarily used to describe a specific order or sequence, nor should they be construed as indicating or implying relative importance. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances. The singular forms "a," "the," and "the" used in this application and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise. The word "if" as used herein can be interpreted as "when," "when," or "in response to determination."

[0028] Furthermore, in the description of this application, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0029] Please see Figure 1 This is a flowchart of the soil nickel concentration inversion method according to the first embodiment of this application. The soil nickel concentration inversion method includes:

[0030] S1: Acquire real-time hyperspectral images of the target area.

[0031] Real-time hyperspectral imagery refers to hyperspectral images acquired in real time. Hyperspectral imagery refers to images with a spectral resolution of 10⁻⁶. -2 Spectral images in the order of λ.

[0032] S2: Extract features from the real-time hyperspectral image to obtain the corresponding real-time multi-scale spectral features.

[0033] Among them, the real-time multi-scale spectral features include the low-frequency components of the approximate spectrum and the high-frequency components that characterize the detailed phases.

[0034] S3: Perform feature dimensionality reduction on the real-time multi-scale spectral features to obtain the real-time feature value distribution data of the real-time multi-scale spectral features in the low-dimensional space.

[0035] S4: Based on the pre-constructed spectral index model and Pearson significance test algorithm, obtain the corresponding real-time band combination of soil nickel concentration from the real-time hyperspectral image.

[0036] The pre-built spectral index models include the difference model (DI), ratio model (RI), soil model (SI), product model (PI), and normalized difference index models (NPDI and NDSI). These pre-built models are constructed using all possible combinations of available bands in the full hyperspectral imagery.

[0037] S5: Input the real-time feature value distribution data and the corresponding real-time band combination of the soil nickel concentration into the pre-trained soil nickel concentration inversion model to obtain the soil nickel concentration of the target area predicted by the soil nickel concentration inversion model.

[0038] In order to stabilize and improve the generalization ability of the soil nickel concentration inversion model, the pre-trained soil nickel concentration inversion model will be trained using the dimensionality-reduced multi-scale spectral features. Therefore, when predicting soil nickel concentration in real time, the real-time multi-scale spectral features are also dimensionality reduced, which can improve the accuracy of the prediction.

[0039] The pre-trained soil nickel concentration inversion model can be any of the following prediction models: Random Forest (RF), eXtreme Gradient Boosting (XGBoost), Back Propagation Neural Networks (BPNN), Support Vector Machine Regression (SVMR), and Gaussian Process Regression (GPR).

[0040] Compared with related technologies, this application extracts and reduces the dimensions of real-time hyperspectral images of the target area to obtain real-time feature value distribution data in low-dimensional space. Then, the real-time feature value distribution data and the corresponding real-time bands of soil nickel concentration are combined and input into a pre-trained soil nickel concentration inversion model to obtain the soil nickel concentration of the target area predicted by the soil nickel concentration inversion model. This method can predict the soil nickel concentration of multiple target areas at the same time, thus improving the efficiency of obtaining soil nickel concentration.

[0041] In a feasible embodiment, step S2: extracting features from the real-time hyperspectral image to obtain corresponding real-time multi-scale spectral features includes:

[0042] S21: Obtain spectral reflectance from the real-time hyperspectral image.

[0043] S22: Perform discrete wavelet transform on the first derivative of the spectral reflectance to obtain low-frequency and high-frequency components.

[0044] The maximum decomposition pool for discrete wavelet transform processing is 5.

[0045] S23: Based on the low-frequency component and the high-frequency component, the real-time multi-scale spectral characteristics are obtained.

[0046] Specifically, the real-time multi-scale spectral features are obtained using the following formula:

[0047]

[0048] Where S represents the real-time multi-scale spectral features, and A H H represents the low-frequency component of the preset final scale, where H is the preset final scale and D is the low-frequency component of the preset final scale. h For the high-frequency components at scale h.

[0049] In this embodiment, by performing discrete wavelet transform processing on the spectral reflectance of real-time hyperspectral images, two different scales of spectral features, namely low-frequency components and high-frequency components, which reflect spectral characteristic information, can be obtained.

[0050] In a feasible embodiment, step S3: performing feature dimensionality reduction on the real-time multi-scale spectral features to obtain real-time feature value distribution data of the real-time multi-scale spectral features in a low-dimensional space, includes:

[0051] The real-time multi-scale spectral features are reduced in dimensionality using a nonlinear manifold learning dimensionality reduction algorithm. The formula for the nonlinear manifold learning dimensionality reduction algorithm is as follows:

[0052]

[0053] Where C represents the eigenvalue distribution of real-time multi-scale spectral features in low-dimensional space; P ij Q represents the joint probability distribution of data point i and data point j in the feature space before dimensionality reduction in real-time multi-scale spectral features; ij Let represent the joint probability distribution of data point i and data point j in the low-dimensional space in real-time multi-scale spectral features.

[0054] In this embodiment, the nonlinear manifold learning dimensionality reduction algorithm used is T-SNE, which is a dimensionality reduction algorithm that uses Kullback-Leibler (KL) divergence to measure the difference in probability distributions of data points i and j and to perform optimization.

[0055] In this embodiment, the multi-scale spectral features are reduced in dimensionality by using a nonlinear manifold learning dimensionality reduction algorithm, which can obtain feature value distribution data with high stability in low-dimensional space.

[0056] In one feasible embodiment, before step S4: obtaining the real-time band combination related to soil nickel concentration from the real-time hyperspectral image based on the pre-constructed spectral index model and Pearson significance test algorithm, the step includes:

[0057] S41: Obtain the spectral reflectance values ​​of each band from the real-time hyperspectral image.

[0058] S42: Construct the spectral index model based on the spectral reflectance values ​​of each band.

[0059] Among them, the spectral index models of the difference model (DI), ratio model (RI), soil model (SI), product model (PI), and normalized difference index models (NPDI and NDSI) are as follows: Figure 2 As shown:

[0060] Figure 2 In this context, i and j are wavelengths measured in nanometers (nm), and i ≠ j. R is the spectral reflectance value for the corresponding wavelength.

[0061] S43: Obtain real-time band combinations related to soil nickel concentration using the Pearson significance test algorithm.

[0062] The Pearson significance test algorithm can be used to determine the most effective and sensitive indicator of nickel based on the spectral index model, namely, the band combination related to soil nickel concentration. The results are as follows: Figure 3 As shown:

[0063] Figure 3 In this context, |R| represents the absolute value of the correlation coefficient, * indicates a significant correlation at the p<0.05 level, ** indicates a significant correlation at the p<0.01 level, and p is the output value of the Pearson significance test algorithm.

[0064] In this embodiment, through steps S41-S43, the real-time band combinations related to soil nickel concentration corresponding to each spectral index model can be obtained respectively. This is beneficial for users to select different real-time band combinations related to soil nickel concentration as inputs to the pre-trained soil nickel concentration inversion model according to different spectral index models, so as to accurately predict the soil nickel concentration in the target area.

[0065] In one feasible embodiment, the training steps of the soil nickel concentration inversion model include:

[0066] S51: Acquire training soil data for the target area, and training hyperspectral images that spatiotemporally correspond to the training soil data; the training soil data includes the nickel concentration of the training soil.

[0067] S52: Extract features from the training hyperspectral image to obtain the corresponding training multi-scale spectral features.

[0068] The training soil data can be historical soil data, and the training hyperspectral image is a historical hyperspectral image that corresponds to the training soil data in time and space.

[0069] S53: Perform feature dimensionality reduction on the trained multi-scale spectral features to obtain the feature value distribution of the trained multi-scale spectral features in a low-dimensional space.

[0070] S54: Based on the pre-constructed spectral index model and Pearson significance test algorithm, obtain the training band combination related to soil nickel concentration from the training hyperspectral image.

[0071] S55: The initial prediction model is trained by taking the training multi-scale spectral features and the training bands related to soil nickel concentration as inputs and the training soil nickel concentration as outputs, to obtain the soil nickel concentration inversion model.

[0072] In this embodiment, a soil nickel concentration inversion model with high prediction accuracy can be trained by using historical data as training data.

[0073] In a feasible embodiment, soil data samples and hyperspectral image samples can also be obtained from historical soil data and historical hyperspectral images of the target area. The hyperspectral image samples are then input into the soil nickel concentration inversion model to obtain soil nickel concentration samples. Based on the soil nickel concentration samples and soil data samples, the relative analysis error, coefficient of determination, and ratio of explained total variance to actual total variance of the soil nickel concentration inversion model are obtained. Finally, based on the relative analysis error, the coefficient of determination, and the ratio of explained total variance to actual total variance, the predictive ability level of the soil nickel concentration inversion model is obtained.

[0074] Wherein, the steps of obtaining the relative prediction deviation, coefficient of determination of the soil nickel concentration inversion model, and the ratio of the explainable total variance to the actual total variance comprise:

[0075] Obtain the relative prediction deviation of the soil nickel concentration inversion model through the following formula:

[0076]

[0077] Wherein, RPD is the relative prediction deviation, m is the number of soil nickel concentration samples, y n is the true value of the n-th soil nickel concentration sample, f(x n ) is the predicted value of the n-th soil nickel concentration sample output by the soil nickel concentration inversion model, is the average true value of the soil nickel concentration samples;

[0078] Obtain the coefficient of determination of the soil nickel concentration inversion model through the following formula:

[0079]

[0080] Wherein, R 2 is the coefficient of determination;

[0081] Obtain the ratio of the explainable total variance to the actual total variance of the soil nickel concentration inversion model through the following formula:

[0082]

[0083] Wherein, SSR / SST is the ratio of the explainable total variance to the actual total variance.

[0084] Wherein, to ensure the stability of the model, on the premise that SSR / SST > 0.5 is satisfied, the prediction ability of the model is divided into three categories according to the values of RPD and R 2 : Category A has good prediction ability (RPD>2.0 and R 2 >0.6), Category B has medium prediction ability (1.4<RPD<2.0 and 0.4<R 2 <0.6), and Category C has poor prediction ability (RPD<1.4).

[0085] In this embodiment, through the soil data samples and hyperspectral image samples, users can understand the prediction ability and prediction accuracy of the soil nickel concentration inversion model, so as to select the soil nickel concentration inversion model with strong prediction ability and high prediction accuracy according to different regions and time.

[0086] Please refer to Figure 4 , a second embodiment of the present application provides a soil nickel concentration inversion apparatus 100, comprising:

[0087] The real-time hyperspectral image acquisition module 101 is used to acquire real-time hyperspectral images of the target area.

[0088] The spectral feature acquisition module 102 is used to extract features from the real-time hyperspectral image to obtain the corresponding real-time multi-scale spectral features.

[0089] The feature dimensionality reduction module 103 is used to perform feature dimensionality reduction on the real-time multi-scale spectral features to obtain the real-time feature value distribution data of the real-time multi-scale spectral features in the low-dimensional space.

[0090] The band combination acquisition module 104 is used to acquire the corresponding real-time band combination of soil nickel concentration from the real-time hyperspectral image based on the pre-constructed spectral index model and Pearson significance test algorithm.

[0091] The soil nickel concentration acquisition module 105 is used to input the real-time feature value distribution data and the corresponding real-time band combination of the soil nickel concentration into a pre-trained soil nickel concentration inversion model to obtain the soil nickel concentration of the target area predicted by the soil nickel concentration inversion model.

[0092] The spectral feature acquisition module 102 includes:

[0093] A spectral reflectance acquisition module is used to acquire spectral reflectance from the real-time hyperspectral image;

[0094] The component acquisition module is used to perform discrete wavelet transform processing on the first derivative of the spectral reflectance to obtain low-frequency and high-frequency components.

[0095] The spectral feature acquisition submodule is used to obtain the real-time multi-scale spectral features based on the low-frequency and high-frequency components.

[0096] It should be noted that the soil nickel concentration inversion device 100 provided in the second embodiment of this application is only illustrated by the above-described division of functional modules when performing the soil nickel concentration inversion method. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the soil nickel concentration inversion device 100 provided in the second embodiment of this application and the soil nickel concentration inversion method of the first embodiment of this application belong to the same concept, and its implementation process is detailed in the method embodiment, which will not be repeated here.

[0097] A third aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the soil nickel concentration inversion method as described above.

[0098] A fourth aspect of this application provides a computer device including a storage device, a processor, and a computer program stored in the storage device and executable by the processor, wherein the processor executes the computer program to implement the steps of the soil nickel concentration inversion method as described above.

[0099] The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this application according to actual needs. Those skilled in the art can understand and implement this without any inventive effort.

[0100] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0101] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 The computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function selected in one or more boxes.

[0102] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function selected in one or more boxes.

[0103] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0104] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0105] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0106] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0107] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for inverting soil nickel concentration, characterized in that, include: Acquire real-time hyperspectral images of the target area; Feature extraction is performed on the real-time hyperspectral image to obtain the corresponding real-time multi-scale spectral features, including: Spectral reflectance is obtained from the real-time hyperspectral image; The low-frequency component and the high-frequency component are obtained by performing discrete wavelet transform on the first derivative of the spectral reflectance. The real-time multi-scale spectral features are obtained based on the low-frequency components and the high-frequency components. The real-time multi-scale spectral features are subjected to feature dimensionality reduction to obtain the real-time feature value distribution data of the real-time multi-scale spectral features in the low-dimensional space; Based on the pre-constructed spectral index model and Pearson significance test algorithm, the corresponding real-time band combination of soil nickel concentration is obtained from the real-time hyperspectral image. The real-time feature value distribution data and the corresponding real-time band combination of the soil nickel concentration are input into the pre-trained soil nickel concentration inversion model to obtain the soil nickel concentration of the target area predicted by the soil nickel concentration inversion model. The training steps for the soil nickel concentration inversion model include: Acquire training soil data for the target area, and training hyperspectral images that spatiotemporally correspond to the training soil data; the training soil data includes the nickel concentration of the training soil. Feature extraction is performed on the training hyperspectral images to obtain the corresponding training multi-scale spectral features; The trained multi-scale spectral features are subjected to feature dimensionality reduction to obtain the feature value distribution of the trained multi-scale spectral features in a low-dimensional space. Based on the pre-constructed spectral index model and Pearson significance test algorithm, training band combinations related to soil nickel concentration are obtained from the training hyperspectral images; The initial prediction model is trained by taking the eigenvalue distribution of the trained multi-scale spectral features in low-dimensional space and the training band combination related to soil nickel concentration as input, and the training soil nickel concentration as output, to obtain the soil nickel concentration inversion model.

2. The method for inverting soil nickel concentration according to claim 1, characterized in that, The step of obtaining the real-time multi-scale spectral features based on the low-frequency and high-frequency components includes: The real-time multi-scale spectral features are obtained using the following formula: in, The real-time multi-scale spectral features, The low-frequency components are the preset final scale. The final scale is preset. For scale The high-frequency components.

3. The method for inverting soil nickel concentration according to claim 1, characterized in that, The step of performing feature dimensionality reduction on the real-time multi-scale spectral features to obtain the real-time feature value distribution data of the real-time multi-scale spectral features in a low-dimensional space includes: The real-time multi-scale spectral features are reduced in dimensionality using a nonlinear manifold learning dimensionality reduction algorithm. The formula for the nonlinear manifold learning dimensionality reduction algorithm is as follows: in, The eigenvalue distribution of real-time multi-scale spectral features in low-dimensional space; Data points in real-time multi-scale spectral features and data points The joint probability distribution in the feature space before dimensionality reduction; Data points in real-time multi-scale spectral features and data points Joint probability distribution in a low-dimensional space.

4. The method for retrieving soil nickel concentration according to claim 1, characterized in that, Before the step of obtaining the real-time band combination related to soil nickel concentration from the real-time hyperspectral image based on the pre-constructed spectral index model and Pearson significance test algorithm, the following steps are included: The spectral reflectance values ​​of each band are obtained from the real-time hyperspectral image; Based on the spectral reflectance values ​​of each band, the spectral index model is constructed. The Pearson significance test algorithm was used to obtain real-time band combinations related to soil nickel concentration.

5. A soil nickel concentration inversion device, characterized in that, include: The real-time hyperspectral image acquisition module is used to acquire real-time hyperspectral images of the target area. The spectral feature acquisition module is used to extract features from the real-time hyperspectral image to obtain corresponding real-time multi-scale spectral features, including: Spectral reflectance is obtained from the real-time hyperspectral image; The low-frequency component and the high-frequency component are obtained by performing discrete wavelet transform on the first derivative of the spectral reflectance. The real-time multi-scale spectral features are obtained based on the low-frequency components and the high-frequency components. The feature dimensionality reduction module is used to perform feature dimensionality reduction on the real-time multi-scale spectral features to obtain the real-time feature value distribution data of the real-time multi-scale spectral features in a low-dimensional space. The band combination acquisition module is used to acquire the corresponding real-time band combination of soil nickel concentration from the real-time hyperspectral image based on the pre-constructed spectral index model and Pearson significance test algorithm. The soil nickel concentration acquisition module is used to input the real-time feature value distribution data and the corresponding real-time band combination of the soil nickel concentration into the pre-trained soil nickel concentration inversion model to obtain the soil nickel concentration of the target area predicted by the soil nickel concentration inversion model. The training steps for the soil nickel concentration inversion model include: Acquire training soil data for the target area, and training hyperspectral images that spatiotemporally correspond to the training soil data; the training soil data includes the nickel concentration of the training soil. Feature extraction is performed on the training hyperspectral images to obtain the corresponding training multi-scale spectral features; The trained multi-scale spectral features are subjected to feature dimensionality reduction to obtain the feature value distribution of the trained multi-scale spectral features in a low-dimensional space. Based on the pre-constructed spectral index model and Pearson significance test algorithm, training band combinations related to soil nickel concentration are obtained from the training hyperspectral images; The initial prediction model is trained by taking the eigenvalue distribution of the trained multi-scale spectral features in low-dimensional space and the training band combination related to soil nickel concentration as input, and the training soil nickel concentration as output, to obtain the soil nickel concentration inversion model.

6. The soil nickel concentration inversion device according to claim 5, characterized in that, The spectral feature acquisition module includes: A spectral reflectance acquisition module is used to acquire spectral reflectance from the real-time hyperspectral image; The component acquisition module is used to perform discrete wavelet transform processing on the first derivative of the spectral reflectance to obtain low-frequency and high-frequency components. The spectral feature acquisition submodule is used to obtain the real-time multi-scale spectral features based on the low-frequency and high-frequency components.

7. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by the processor, it implements the steps of the soil nickel concentration inversion method as described in any one of claims 1 to 4.

8. A computer device, characterized in that: It includes a storage device, a processor, and a computer program stored in the storage device and executable by the processor, wherein the processor executes the computer program to implement the steps of the soil nickel concentration inversion method as described in any one of claims 1 to 4.

Images