Soil component detection method and system based on spectral analysis

Abstract

The invention relates to the technical field of soil component detection, in particular to a soil component detection method and system based on spectral analysis, and the system comprises an intelligent mobile platform and a multifunctional integrated spectrum acquisition module. In the advancing process, soil is continuously scanned through the spectrum collection module, spectrum information is obtained in real time, data are processed through the spectrum feature extraction unit, the soil component type and distribution proportion are obtained, and a soil component distribution diagram and quantitative data are generated. According to the method, the detection efficiency can be remarkably improved, the problem of data missing of a traditional fixed-point sampling mode is avoided, meanwhile, the adaptability to complex terrains and diversified soil types is enhanced, and high precision and wide applicability of a detection result are ensured.

Description

Technical Field

[0001] The present invention belongs to the technical field of soil component detection, and specifically relates to a soil component detection method and system based on spectral analysis. Background Art

[0002] With the widespread application of spectral analysis technology in soil testing, soil composition detection methods and systems based on spectral analysis have gradually become a research hotspot. This type of technology can obtain soil composition information in a non-destructive and rapid manner, and is of great significance in fields such as agriculture, environmental monitoring, and resource management. However, existing related technical solutions still have shortcomings in terms of detection efficiency, accuracy, adaptability, and system integration, making it difficult to meet the diverse needs of complex scenarios. For example, patent publication number CN118549356B proposes a device for detecting the composition of saline-alkali soil. This device achieves efficient detection through automatic collection, transportation, and spectral analysis, significantly reducing the time required for sample collection, transportation, and laboratory analysis compared to traditional methods. However, this technical solution is mainly targeted at the specific application scenario of saline-alkali soil, and its scope of application is relatively limited, making it difficult to extend to other types of soil or detection under complex substrate conditions. In addition, the device relies on mechanical structures (such as collection wheels and extrusion rods) to collect and process soil samples. The mechanical wear and tear caused by long-term use may lead to a decrease in detection accuracy. At the same time, it has poor adaptability to soils with uneven hardness, further limiting its practical application.

[0003] On the other hand, the patent with publication number CN114813709B proposes a soil composition detection method, equipment and system based on a hybrid model, which improves the detection accuracy by decomposing complex spectral data into linear and nonlinear parts, and combining multiple linear regression and neural network models to predict soil component content. However, this technical solution has high requirements for the preprocessing of spectral data in practical applications, and requires a pre-trained hybrid model, which may lead to insufficient model generalization ability when facing unknown types of soil or complex substrates, affecting the accuracy of the detection results. In addition, the system does not fully consider the real-time and automation requirements in large-scale field detection scenarios, and lacks the ability to comprehensively process multi-source data (such as hyperspectral images and laser-induced breakdown spectroscopy data), which limits its application effect at the regional scale.

[0004] The above issues indicate that existing spectral analysis-based soil composition detection technologies still have certain shortcomings in terms of scope of application, detection accuracy, system integration, and adaptability to complex scenarios. Therefore, a new spectral analysis-based soil composition detection method and system is urgently needed. By optimizing spectral data processing algorithms, enhancing the system's adaptability and multi-source data fusion capabilities, the detection efficiency and accuracy can be improved. At the same time, the method can be expanded to cover different soil types and complex substrate conditions to meet the needs of modern agriculture and environmental monitoring for efficient and accurate soil composition detection. Summary of the Invention

[0005] The embodiments of the present application provide a soil composition detection method and system based on spectral analysis to improve the efficiency, accuracy and adaptability of soil composition detection.

[0006] To achieve the above-mentioned purpose, the present invention provides a soil composition detection method based on spectral analysis, which is characterized in that it is applied to an intelligent mobile platform, and a multifunctional integrated spectral acquisition module is provided on the intelligent mobile platform, and the multifunctional integrated spectral acquisition module includes a light source unit, a spectral sensor unit and a data transmission interface; the method includes: during the movement of the intelligent mobile platform, the soil at different positions is continuously scanned by the multifunctional integrated spectral acquisition module, and multiple groups of soil spectral information are obtained in real time; wherein, each group of spectral information corresponds to a specific scanning area; for any group of spectral information, the spectral information is processed by a spectral feature extraction unit to obtain the main component types of the soil in the scanning area and their relative distribution ratios; and the soil composition distribution map and quantitative data thereof within the coverage area of ​​the intelligent mobile platform are generated by combining the processing results of each group of spectral information.

[0007] Optionally, the spectral feature extraction unit includes a preprocessing step, a feature extraction step and a classification modeling step; the preprocessing step is used to denoise and smooth the original spectral data; the feature extraction step is used to extract key feature variables from the preprocessed spectral data; the classification modeling step is used to input the extracted feature variables into a pre-trained machine learning model to predict the main component types of the soil and their relative distribution ratios.

[0008] Optionally, the machine learning model includes a support vector machine, a random forest, or a deep neural network.

[0009] Optionally, the multifunctional integrated spectrum acquisition module further includes a filter component, and the filter component is used to selectively filter out background light signals that are not related to the target band.

[0010] Optionally, the intelligent mobile platform comprises a chassis, a driving motor, a navigation module and a power management module; the chassis is made of lightweight material; the navigation module integrates a GPS positioning unit and an inertial measurement unit.

[0011] To solve the above problems, the application further provides a terminal device comprising a processor and a memory, wherein the memory stores a computer program capable of running on the processor, and when the processor executes the computer program, the method for detecting soil composition based on spectrum analysis is realized.

[0012] To solve the above problems, the application further provides a system for detecting soil composition based on spectrum analysis, which comprises an intelligent mobile platform, a multifunctional integrated spectrum acquisition module, a spectrum feature extraction unit and a data processing and visualization module; the multifunctional integrated spectrum acquisition module is arranged on the intelligent mobile platform and comprises a light source unit, a spectrum sensor unit and a data transmission interface; the spectrum feature extraction unit is used for processing the spectrum information acquired by the multifunctional integrated spectrum acquisition module; and the data processing and visualization module is used for generating a soil composition distribution map and its quantitative data.

[0013] Optionally, the spectrum feature extraction unit comprises a high-performance processor, a memory and a communication interface; the high-performance processor is used for executing the pre-processing, feature extraction and classification modeling tasks of spectrum data; the memory is used for caching original spectrum data and intermediate calculation results; and the communication interface is used for data interaction with the multifunctional integrated spectrum acquisition module.

[0014] Optionally, the multifunctional integrated spectrum acquisition module further comprises a filter assembly, and the filter assembly realizes rapid acquisition of multi-band spectrum through rotation switching.

[0015] Optionally, the data processing and visualization module runs on a remote server or a cloud platform and supports export of multiple data formats.

[0016] This method is applied to an intelligent mobile platform equipped with a multifunctional integrated spectral acquisition module. The method includes: while the intelligent mobile platform is moving, the multifunctional integrated spectral acquisition module is used to continuously scan the soil at different locations and obtain multiple sets of soil spectral information in real time. Each set of spectral information corresponds to a specific scanning area. For each set of spectral information, the spectral information is processed using a spectral feature extraction unit to obtain the main component types and relative distribution ratios of the soil in the scanning area. Combining the processing results of each set of spectral information, a soil component distribution map and quantitative data within the coverage area of ​​the intelligent mobile platform are generated. Because the intelligent mobile platform can complete the continuous collection and processing of spectral information while moving, it avoids the data loss problem caused by the limited point selection in traditional fixed-point sampling methods, significantly improving detection efficiency. At the same time, the design of the multifunctional integrated spectral acquisition module enhances the adaptability to complex terrain and diverse soil types, ensuring the high accuracy and wide applicability of the detection results. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flow chart of a soil composition detection method based on spectral analysis provided in one embodiment of the present application;

[0018] Figure 2 Schematic diagram of the structure of a soil composition detection system based on spectral analysis provided in one embodiment of the present application;

[0019] Figure 3 It is a structural diagram of the intelligent mobile platform provided in one embodiment of the present application. DETAILED DESCRIPTION

[0020] The present invention provides a method and system for detecting soil components based on spectral analysis, which can significantly improve the efficiency, accuracy and adaptability of soil component detection. Figure 1 To the attached Figure 3 Specific embodiments of the present invention are described in detail.

[0021] First reference Figure 1, which is a flow chart of a soil composition detection method based on spectral analysis provided in one embodiment of the present application. In practical applications, the method mainly relies on an intelligent mobile platform equipped with a multifunctional integrated spectral acquisition module. The intelligent mobile platform can be a wheeled or tracked unmanned vehicle, or other forms of automated equipment such as drones, and its design must meet the requirements of stable operation under complex terrain conditions. The multifunctional integrated spectral acquisition module includes a light source unit, a spectral sensor unit, and a data transmission interface, wherein the light source unit is used to emit light signals in a specific wavelength range to the soil surface, and the spectral sensor unit is responsible for receiving the reflected or scattered light signals and converting them into digital spectral information. The design of this module enables it to maintain stable performance under different lighting conditions, while having certain waterproof and dustproof capabilities to adapt to outdoor environments.

[0022] As the intelligent mobile platform moves, the multifunctional integrated spectral acquisition module continuously scans the soil at different locations and acquires multiple sets of soil spectral information in real time. Each set of spectral information corresponds to a specific scanning area, and its coverage is determined by the field of view of the spectral sensor and the travel speed of the intelligent mobile platform. To ensure the accuracy and completeness of the spectral information, the spectral acquisition module needs to dynamically adjust the light source intensity and exposure time according to the reflective characteristics of the soil surface. For example, in areas with high reflectivity of sandy soil, the light source intensity is appropriately reduced to avoid signal saturation; while in areas with low reflectivity of clay, the light source intensity needs to be increased to ensure a sufficient signal-to-noise ratio. In addition, to reduce interference from external ambient light, the spectral acquisition module is also equipped with a filter assembly that can selectively filter out background light signals that are not related to the target band.

[0023] For any set of spectral information, the spectral feature extraction unit processes it to determine the primary soil component types and their relative distribution ratios within the scanned area. The core algorithm of the spectral feature extraction unit consists of a preprocessing step, a feature extraction step, and a classification modeling step. In the preprocessing step, the raw spectral data is first denoised and smoothed to eliminate outliers caused by sensor noise or environmental interference. Common denoising methods include wavelet transform and Savitzky-Golay filtering, while smoothing can be achieved using a moving average method. Next, in the feature extraction step, key characteristic variables are extracted from the preprocessed spectral data. These characteristic variables typically include the position, width, and depth of spectral absorption peaks, as well as reflectance trends within specific wavelength bands. For example, soils with high organic matter content will have a distinct absorption peak in the near-infrared band (700-1300nm), a characteristic that can serve as an important basis for distinguishing between high and low organic matter content. Finally, in the classification modeling step, the extracted characteristic variables are input into a pretrained machine learning model to predict the primary soil component types and their relative distribution ratios. Common machine learning models include support vector machines (SVM), random forests (RF), and deep neural networks (DNN). Taking random forests as an example, its classification process can be expressed as the following formula:

[0024]

[0025] Among them, C represents the final classification result, k is the category label, N is the number of decision trees, f(x i ) is the output of the i-th decision tree, and I(·) is the indicator function. By integrating the prediction results of multiple decision trees, random forests can effectively reduce the risk of overfitting of a single model, thereby improving classification accuracy.

[0026] Combining the processing results of each set of spectral information, a soil composition distribution map and its quantitative data within the coverage area of ​​the intelligent mobile platform are generated. Specifically, the spectral processing results of each scanned area are first mapped to the geographic coordinate system to form a discretized soil composition distribution matrix. Then, the discrete data is spatially expanded using an interpolation algorithm to generate a continuous soil composition distribution map. Commonly used interpolation algorithms include Kriging interpolation and inverse distance weighted interpolation. Taking Kriging interpolation as an example, its calculation formula is as follows:

[0027]

[0028] Among them, Z(s0) represents the soil composition value of the point to be estimated, z(s i ) is the soil composition value of the known sampling point, λ iwhere w is the weight coefficient, n is the number of known sampling points. The determination of the weight coefficient needs to consider the spatial correlation between the sampling points, which is usually fitted by a variogram model. The finally generated soil composition distribution map not only can intuitively show the differences of soil composition in different regions, but also can further analyze the soil fertility, pollution degree and other indicators through quantitative data.

[0029] Reference Figure 2 which is a structural schematic diagram of a soil composition detection system based on spectral analysis provided by an embodiment of the present application. The system mainly includes an intelligent mobile platform, a multifunctional integrated spectral acquisition module, a spectral feature extraction unit, and a data processing and visualization module. The main structure of the intelligent mobile platform includes a chassis, a drive motor, a navigation module, and a power management module. The chassis is made of lightweight materials to reduce the overall weight and improve the endurance; the drive motor is responsible for controlling the direction and speed of travel, and its parameters can be dynamically adjusted according to the terrain conditions; the navigation module integrates a GPS positioning unit and an inertial measurement unit, which is used to obtain the geographical position and attitude information of the platform in real time; the power management module is responsible for providing stable power supply for each sub-module. The multifunctional integrated spectral acquisition module is installed on the top of the chassis, including a light source unit, a spectral sensor unit, a filter assembly, and a data transmission interface. The light source unit adopts LED array design, which can cover a wide spectral range from visible light to near-infrared band; the spectral sensor unit selects high-sensitivity CCD or CMOS sensors to ensure high resolution and high accuracy of spectral data; the filter assembly realizes fast acquisition of multi-band spectrum through rotary switching; the data transmission interface uses high-speed USB or wireless communication protocol to transmit the collected spectral information to the spectral feature extraction unit in real time.

[0030] The spectral feature extraction unit is deployed in the central processing unit of the intelligent mobile platform, and its hardware architecture includes a high-performance processor, a memory, and a communication interface. The high-performance processor is responsible for performing the tasks of preprocessing, feature extraction, and classification modeling of spectral data; the memory is used to cache raw spectral data and intermediate calculation results; the communication interface is responsible for data interaction with the multifunctional integrated spectral acquisition module and other external devices. The data processing and visualization module runs on a remote server or cloud platform, and its main function is to summarize and analyze the results output by the spectral feature extraction unit, and generate the final soil composition distribution map and its quantitative data. This module supports the export of multiple data formats, which is convenient for users to carry out subsequent research or decision support.

[0031] In practical application scenarios, the method and system of the present invention can be widely used in fields such as agriculture, environmental protection, and geological exploration. For example, in agricultural production, farmers can use the intelligent mobile platform to quickly obtain soil composition distribution information on their farmland, thereby formulating more scientific and reasonable fertilization plans. In the field of environmental protection, researchers can use the system to monitor soil pollution and assess the spatial distribution characteristics of pollutants. In geological exploration, the system can also help identify potential distribution areas of mineral resources, providing important reference for resource development.

[0032] In summary, the spectral analysis-based soil composition detection method and system provided in the embodiments of this application achieve efficient and accurate collection and processing of soil composition information through the organic combination of an intelligent mobile platform and a multifunctional integrated spectral acquisition module. This method not only overcomes the limitations of traditional fixed-point sampling methods but also significantly improves adaptability to complex terrain conditions, providing strong technical support for soil science research and practical applications.

Claims

1. A soil composition detection method based on spectral analysis, characterized in that: The invention is applied to an intelligent mobile platform, wherein the intelligent mobile platform is provided with a multifunctional integrated spectrum acquisition module, which includes a light source unit, a spectrum sensor unit and a data transmission interface. The method comprises: during the movement of the intelligent mobile platform, continuously scanning soil at different positions by the multifunctional integrated spectrum acquisition module, and acquiring multiple sets of soil spectral information in real time; wherein each set of spectral information corresponds to a specific scanning area; for any set of spectral information, using a spectral feature extraction unit to process the set of spectral information, to obtain the main component types of the soil in the scanning area and their relative distribution ratios; and combining the processing results of each set of spectral information to generate a soil component distribution map and quantitative data thereof within the coverage area of ​​the intelligent mobile platform.

2. The soil component detection method based on spectral analysis according to claim 1, characterized in that: The spectral feature extraction unit includes a preprocessing step, a feature extraction step, and a classification modeling step; the preprocessing step is used to perform denoising and smoothing on the original spectral data; The feature extraction step is used to extract key characteristic variables from the preprocessed spectral data; the classification modeling step is used to input the extracted characteristic variables into a pre-trained machine learning model to predict the main component types of the soil and their relative distribution ratios.

3. The soil component detection method based on spectral analysis according to claim 2, characterized in that: The machine learning model includes a support vector machine, a random forest, or a deep neural network.

4. The soil component detection method based on spectral analysis according to claim 1, characterized in that: The multifunctional integrated spectrum acquisition module further includes a filter component, which is used to selectively filter out background light signals that are irrelevant to the target wavelength band.

5. The soil component detection method based on spectral analysis according to claim 1, characterized in that: The intelligent mobile platform includes a chassis, a drive motor, a navigation module and a power management module; the chassis is made of lightweight materials; and the navigation module integrates a GPS positioning unit and an inertial measurement unit.

6. A soil composition detection system based on spectral analysis, characterized in that: It includes an intelligent mobile platform, a multifunctional integrated spectrum acquisition module, a spectrum feature extraction unit and a data processing and visualization module; the multifunctional integrated spectrum acquisition module is arranged on the intelligent mobile platform, and the multifunctional integrated spectrum acquisition module includes a light source unit, a spectrum sensor unit and a data transmission interface; the spectrum feature extraction unit is used to process the spectrum information obtained by the multifunctional integrated spectrum acquisition module; the data processing and visualization module is used to generate a soil composition distribution map and its quantitative data.

7. The soil composition detection system based on spectral analysis according to claim 6, characterized in that: The spectral feature extraction unit includes a high-performance processor, a memory, and a communication interface; the high-performance processor is used to perform spectral data preprocessing, feature extraction, and classification modeling tasks; The memory is used to cache original spectral data and intermediate calculation results; the communication interface is used to exchange data with the multifunctional integrated spectral acquisition module.

8. The soil composition detection system based on spectral analysis according to claim 6, characterized in that: The multifunctional integrated spectrum acquisition module further includes a filter assembly, which realizes rapid acquisition of multi-band spectra by rotating and switching.

9. The soil composition detection system based on spectral analysis according to claim 6, characterized in that: The data processing and visualization module runs on a remote server or cloud platform and supports exporting data in multiple formats.

10. A terminal device comprising a processor and a memory, wherein the memory stores a computer program that can be run on the processor, characterized in that: When the processor executes the computer program, the soil composition detection method based on spectral analysis according to any one of claims 1 to 5 is implemented.

Images