Multispectral remote sensing-based mountain original red soil nutrient content inversion method and system
By combining drone multispectral remote sensing with artificial neural networks, the problem of time-consuming and labor-intensive traditional methods of monitoring the nutrient content of mountain red soil has been solved, efficient and accurate nutrient monitoring has been achieved, and precision fertilization in agriculture has been supported.
Patent Information
- Application Number
- CN202510983479.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-16
- Publication Date
- 2025-10-21
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing technologies make it difficult to efficiently and accurately monitor the main nutrient content of mountain red soil. Traditional methods are time-consuming and labor-intensive, and difficult to cover large areas and monitor in real time.
By combining UAV multispectral remote sensing with artificial neural network algorithm, we collected and preprocessed multispectral remote sensing image data, established an inversion model for the nutrient content of mountain red soil, and used the artificial neural network algorithm to invert the soil nutrient content.
It has achieved efficient and accurate inversion of the main nutrient content of mountain red soil, improved data acquisition efficiency, expanded the monitoring range, reduced costs and time investment, and provided a strong basis for precise fertilization in agriculture.
Smart Images

Figure CN120823508A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of soil nutrient monitoring, and more particularly to an inversion method and system for mountain red soil nutrient content based on multispectral remote sensing. Background Art
[0002] Mountain red soil is a soil type of great agricultural value, and its nutrient content is directly related to the sustainable development of local agricultural production. Traditional soil nutrient monitoring methods typically rely on manual sampling and laboratory analysis, which are time-consuming and labor-intensive, and difficult to cover large areas and monitor in real time. However, research on the inversion of the main nutrient content of mountain red soil using drone-based multispectral remote sensing is still in its exploratory stage, lacking a systematic and efficient method.
[0003] Therefore, designing a method and system for inverting the nutrient content of mountain red soil based on multispectral remote sensing, which can efficiently and accurately invert the main nutrient content of mountain red soil, is an urgent problem that needs to be solved by technical personnel in this field. Summary of the Invention
[0004] In view of this, the present invention provides a method and system for inverting the nutrient content of mountain red soil based on multispectral remote sensing. By combining unmanned aerial vehicle multispectral remote sensing with artificial neural network algorithm, it can efficiently and accurately invert the main nutrient content of mountain red soil, providing a new and rapid means for soil nutrient monitoring and promoting the development of precision agriculture.
[0005] To achieve the above objectives, the present invention adopts the following technical solution: a method for inverting the nutrient content of mountain red soil based on multispectral remote sensing, comprising:
[0006] Step 1: Collect multispectral remote sensing image data of the mountain red soil area;
[0007] Step 2: Preprocess the multispectral remote sensing image data;
[0008] Step 3: Extract soil nutrient content data from the preprocessed multispectral remote sensing image;
[0009] Step 4: Use artificial neural network algorithm to establish the inversion model of nutrient content of Shanyuan red soil;
[0010] Step 5: Use the inversion model to process the soil nutrient content data to obtain the soil nutrient content distribution map, and evaluate and analyze the soil nutrient content distribution map.
[0011] Preferably, the multispectral remote sensing image data includes a visible light band and a near-infrared band, the visible light band represents soil color information, and the near-infrared band is used to represent soil vegetation coverage.
[0012] Preferably, the pre-processing process in step 2 includes performing radiometric calibration and geometric correction on the visible light band and the near-infrared band, so that the pixel positions of the multispectral remote sensing image data correspond to actual geographic coordinates.
[0013] Preferably, the actual geographic coordinates are imported into ArcGIS software, and the reflectance of each band and the synthetic spectral index corresponding to the actual geographic coordinates are extracted based on the pre-processed multispectral remote sensing image data.
[0014] Preferably, the spectral characteristics of the soil are obtained according to the reflectance of each band, and a regression model is constructed using the spectral characteristic bands to determine different types of soil nutrients and their contents.
[0015] Preferably, the artificial neural network algorithm includes neurons in an input layer, a hidden layer and an output layer. The number of neurons in the input layer is determined according to the number of spectral features, and the number of neurons in the output layer is determined according to the type of soil nutrient content. The hidden layer is optimized and trained so that the artificial neural network algorithm can obtain the nonlinear relationship between multispectral remote sensing image data and soil nutrient content.
[0016] Preferably, the soil nutrient content data is processed using an inversion model to obtain a soil nutrient content distribution map, and the soil nutrient content distribution map is verified and evaluated based on the soil nutrient data sampled in the field to determine the accuracy and reliability of the soil nutrient content distribution map.
[0017] Preferably, a system for inverting nutrient content of mountain red soil based on multispectral remote sensing comprises:
[0018] Acquisition module: collects multispectral remote sensing image data of mountain and red soil areas;
[0019] Preprocessing module: preprocess multispectral remote sensing image data;
[0020] Extraction module: extract soil nutrient content data from pre-processed multispectral remote sensing images;
[0021] Model building module: Use artificial neural network algorithm to establish the inversion model of nutrient content of mountain red soil;
[0022] Analysis and evaluation module: Use the inversion model to process the soil nutrient content data to obtain the soil nutrient content distribution map, and evaluate and analyze the soil nutrient content distribution map.
[0023] As can be seen from the above technical solutions, compared to existing technologies, the present invention provides a method and system for inverting the nutrient content of mountainous red soil based on multispectral remote sensing. By using an artificial neural network algorithm to establish an inversion model for the main nutrient content of mountainous red soil, this method improves inversion accuracy and reliability. It also enables more accurate prediction of soil nutrient content, providing a powerful basis for precision fertilization in agriculture. Furthermore, the use of drone multispectral remote sensing technology improves data acquisition efficiency, expands the monitoring range, and reduces costs and time investment, demonstrating its promising application prospects. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.
[0025] Figure 1 The present invention provides a flow chart of the method. DETAILED DESCRIPTION
[0026] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.
[0027] like Figure 1 As shown, the embodiment of the present invention discloses a method for inverting the nutrient content of mountain red soil based on multispectral remote sensing, comprising:
[0028] Step 1: Collect multispectral remote sensing image data of the mountain red soil area;
[0029] Step 2: Preprocess the multispectral remote sensing image data;
[0030] Step 3: Extract soil nutrient content data from the preprocessed multispectral remote sensing image;
[0031] Step 4: Use artificial neural network algorithm to establish the inversion model of nutrient content of Shanyuan red soil;
[0032] Step 5: Use the inversion model to process the soil nutrient content data to obtain the soil nutrient content distribution map, and evaluate and analyze the soil nutrient content distribution map.
[0033] Specifically, the multispectral remote sensing image data includes a visible light band and a near-infrared band. The visible light band represents soil color information, and the near-infrared band is used to represent soil vegetation coverage.
[0034] In one specific embodiment of the present invention, an unmanned aerial vehicle (UAV) equipped with a multispectral remote sensing sensor is used to fly over the study area to acquire multispectral remote sensing image data. Appropriate flight altitude, flight speed, and sensor parameters are selected. In this embodiment, the appropriate multispectral remote sensing sensor is selected to ensure that the acquired remote sensing images have high spatial and spectral resolution, accurately reflecting information such as soil and vegetation cover.
[0035] Specifically, the preprocessing process in step 2 includes performing radiometric calibration and geometric correction on the visible light band and the near-infrared band, so that the pixel positions of the multispectral remote sensing image data correspond to actual geographic coordinates.
[0036] In one embodiment of the present invention, the reflectivity of the near-infrared band is related to vegetation growth, which is in turn constrained by soil nutrients, thus indirectly reflecting soil nutrient information. Wavebands sensitive to soil nutrients are selected from different spectral bands. Visible light bands, such as red, green, and blue, are used to capture soil color information, while the near-infrared band is used to reflect soil vegetation cover.
[0037] Furthermore, radiometric calibration involves converting the sensor's digital signals into actual physical radiance values to eliminate radiometric variations between multispectral remote sensing sensors and between data collected at different times. Geometric correction involves selecting appropriate calibration models and ground control points to ensure that pixel positions in the remote sensing image accurately correspond to actual geographic coordinates, thereby improving the image's geometric accuracy. Furthermore, atmospheric correction is required to remove the effects of atmospheric scattering and absorption on remote sensing images, restore the true reflectance information of the ground objects, and provide accurate and reliable data for subsequent soil nutrient content extraction.
[0038] Specifically, the actual geographic coordinates are imported into ArcGIS software, and the reflectance and synthetic spectral index of each band corresponding to the actual geographic coordinates are extracted based on the preprocessed multispectral remote sensing image data.
[0039] In a specific embodiment of the present invention, the reflectance of each band and the synthetic spectral index each play an important role in inverting the nutrient content of mountain red soil. The reflectance of each band can reflect the spectral characteristics of the soil. The reflectance of the red light band is negatively correlated with the soil organic matter and total nitrogen content, and the reflectance of the near-infrared band is positively correlated with the available potassium. These characteristics are related to the soil nutrient content and can be used as basic data for inversion. The synthetic spectral index highlights the relationship between soil nutrient content and spectral characteristics by combining the reflectance of different bands. The vegetation index can indirectly reflect the organic matter content and help extract the total nitrogen content information of the soil. It is used to construct a model related to the soil nutrient content, thereby improving the accuracy and reliability of the inversion.
[0040] Furthermore, this example describes the impact of soil organic matter on spectral reflectance, primarily manifesting as absorption in the visible and near-infrared bands. In the visible band, higher organic matter content leads to stronger soil absorption and lower reflectance. In the near-infrared band, organic matter absorption is relatively weaker, but changes in its content still affect reflectance. Furthermore, soil organic matter content is correlated with multiple spectral indices.
[0041] Specifically, the spectral characteristics of the soil are obtained according to the reflectance of each band, and a regression model is constructed using the spectral characteristic bands to determine different types of soil nutrients and their contents.
[0042] Specifically, the artificial neural network algorithm includes neurons in an input layer, a hidden layer, and an output layer. The number of neurons in the input layer is determined according to the number of spectral features, and the number of neurons in the output layer is determined according to the type of soil nutrient content. The hidden layer is optimized and trained so that the artificial neural network algorithm can obtain a nonlinear relationship between multispectral remote sensing image data and soil nutrient content.
[0043] In a specific embodiment of the present invention, the specific process of establishing an inversion model for the nutrient content of mountain red soil using an artificial neural network algorithm includes:
[0044] A large amount of mountain red soil sample data was collected, including multispectral remote sensing image data and corresponding field measurement soil nutrient content data, and divided into training sets and test sets.
[0045] Determine the structure of the artificial neural network, including the number of neurons in the input layer, hidden layer, and output layer. The number of neurons in the input layer is determined according to the number of spectral features of the selected multispectral remote sensing image, and the number of neurons in the output layer is determined according to the type of soil nutrient content (such as organic matter content, total nitrogen content, total phosphorus content, etc.). The number of neurons in the hidden layer is optimized through experiments to achieve the best model performance. The artificial neural network is trained using the training set data, and by adjusting the network weights and thresholds, the network is able to learn the complex nonlinear relationship between multispectral remote sensing image data and soil nutrient content. During the training process, appropriate optimization algorithms (such as gradient descent, genetic algorithm, etc.) and evaluation indicators (such as root mean square error, coefficient of determination, etc.) are used to evaluate and optimize the model until the model meets the predetermined accuracy requirements.
[0046] Specifically, the soil nutrient content data is processed using the inversion model to obtain a soil nutrient content distribution map. The soil nutrient content distribution map is verified and evaluated based on the soil nutrient data sampled in the field to determine the accuracy and reliability of the soil nutrient content distribution map.
[0047] In a specific embodiment of the present invention, sampling points are evenly distributed in the study area (such as a grid method), soil samples are collected, and the nutrient content (such as organic matter, total nitrogen, etc.) is measured to form a set of measured values.
[0048] Furthermore, the coordinates of the sampling points were matched to the nutrient distribution map, and the predicted values of the corresponding positions were extracted. The mean absolute error (MAE) formula was used to directly reflect the average deviation between the predicted values and the measured values. The unit was consistent with the nutrient content, and the smaller the value, the higher the accuracy.
[0049] If the MAE is less than 10% of the measured nutrient content (e.g., an organic matter MAE of 3%), the distribution map is generally considered to have good accuracy. If the MAE is greater than 10% of the measured nutrient content, further analysis of the causes of high-error areas, such as complex terrain and sparse sampling points, is necessary, taking into account the spatial distribution characteristics of the study area, to provide a basis for model optimization.
[0050] Specifically, a system for inverting nutrient content of mountain red soil based on multispectral remote sensing includes:
[0051] Acquisition module: collects multispectral remote sensing image data of mountain and red soil areas;
[0052] Preprocessing module: preprocess multispectral remote sensing image data;
[0053] Extraction module: extract soil nutrient content data from pre-processed multispectral remote sensing images;
[0054] Model building module: Use artificial neural network algorithm to establish the inversion model of nutrient content of mountain red soil;
[0055] Analysis and evaluation module: Use the inversion model to process the soil nutrient content data to obtain the soil nutrient content distribution map, and evaluate and analyze the soil nutrient content distribution map.
[0056] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Reference can be made to the common and similar parts between the various embodiments. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method description.
[0057] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for inverting the nutrient content of mountain red soil based on multispectral remote sensing, characterized in that: include: Step 1: Collect multispectral remote sensing image data of the mountain red soil area; Step 2: Preprocess the multispectral remote sensing image data; Step 3: Extract soil nutrient content data from the preprocessed multispectral remote sensing image; Step 4: Use artificial neural network algorithm to establish the inversion model of nutrient content of Shanyuan red soil; Step 5: Use the inversion model to process the soil nutrient content data to obtain the soil nutrient content distribution map, and evaluate and analyze the soil nutrient content distribution map.
2. The method for inverting nutrient content of mountain red soil based on multispectral remote sensing according to claim 1, characterized in that: The multispectral remote sensing image data includes a visible light band and a near-infrared band. The visible light band represents soil color information, and the near-infrared band is used to represent soil vegetation coverage.
3. The method for inverting nutrient content of mountain red soil based on multispectral remote sensing according to claim 2, characterized in that: The pre-processing process in step 2 includes performing radiometric calibration and geometric correction on the visible light band and the near infrared band, so that the pixel positions of the multispectral remote sensing image data correspond to actual geographic coordinates.
4. The method for inverting nutrient content of mountain red soil based on multispectral remote sensing according to claim 3, characterized in that: The actual geographic coordinates were imported into ArcGIS software, and the reflectance of each band and the synthetic spectral index corresponding to the actual geographic coordinates were extracted based on the preprocessed multispectral remote sensing image data.
5. The method for inverting nutrient content of mountain red soil based on multispectral remote sensing according to claim 4, characterized in that: The spectral characteristics of the soil are obtained according to the reflectance of each band, and a regression model is constructed using the spectral characteristic bands to determine different types of soil nutrients and their contents.
6. The method for inverting nutrient content of mountain red soil based on multispectral remote sensing according to claim 4, characterized in that: The artificial neural network algorithm includes neurons in an input layer, a hidden layer, and an output layer. The number of neurons in the input layer is determined according to the number of spectral features, and the number of neurons in the output layer is determined according to the types of soil nutrient content. The hidden layer is optimized and trained so that the artificial neural network algorithm can obtain a nonlinear relationship between multispectral remote sensing image data and soil nutrient content.
7. The method for inverting nutrient content of mountain red soil based on multispectral remote sensing according to claim 4, characterized in that: The soil nutrient content data were processed using the inversion model to obtain a soil nutrient content distribution map. The soil nutrient content distribution map was verified and evaluated based on the soil nutrient data collected in the field to determine the accuracy and reliability of the soil nutrient content distribution map.
8. A system for inverting nutrient content of mountain red soil based on multispectral remote sensing, characterized in that: include: Acquisition module: collects multispectral remote sensing image data of mountain and red soil areas; Preprocessing module: preprocess multispectral remote sensing image data; Extraction module: extract soil nutrient content data from pre-processed multispectral remote sensing images; Model building module: Use artificial neural network algorithm to establish the inversion model of nutrient content of mountain red soil; Analysis and evaluation module: Use the inversion model to process the soil nutrient content data to obtain the soil nutrient content distribution map, and evaluate and analyze the soil nutrient content distribution map.
Citation Information
Cited By
Intelligent monitoring and evaluation method and system for newly-cultivated land
CN121505442A