An intelligent unmanned aerial vehicle remote sensing monitoring system for soil heavy metal content

Abstract

The application discloses an intelligent soil heavy metal content unmanned aerial vehicle remote sensing monitoring system, and belongs to the technical field of soil environment monitoring. The system comprises an unmanned aerial vehicle flight platform, a multi-source remote sensing load module, an on-board data processing unit, a ground control station and a cloud intelligent analysis platform. The multi-source remote sensing load synchronously collects hyperspectral, multispectral and terrain elevation data; the on-board FPGA chip realizes real-time data preprocessing and abnormality elimination; the cloud is internally provided with a multi-algorithm inversion model library such as PLSR, SVM and CNN, and can invert the distribution of various heavy metal contents, automatically divide pollution grades and give early warnings, and can also analyze the temporal and spatial variation trend and predict pollution diffusion. The application solves the problems of long cycle, high cost and limited coverage range of the traditional monitoring method, realizes rapid, large-area and high-precision real-time monitoring of soil heavy metals, and provides a scientific basis for soil pollution prevention and control decision-making.

Description

Technical Field

[0001] This invention relates to the field of soil environmental monitoring technology, and in particular to an intelligent unmanned aerial vehicle (UAV) remote sensing monitoring system for soil heavy metal content. Background Technology

[0002] Heavy metal pollution in soil is one of the most significant environmental problems facing the world today. Heavy metals are highly toxic, difficult to degrade, and prone to accumulation, affecting not only soil quality and crop growth but also entering the human body through the food chain, posing a threat to human health. Therefore, rapid, accurate, and large-scale monitoring of heavy metal content in soil is of great importance for soil pollution prevention and control and ecological environmental protection.

[0003] Traditional methods for monitoring heavy metals in soil mainly involve field sampling combined with laboratory chemical analysis. While these methods offer high accuracy, they suffer from drawbacks such as long sampling cycles, high costs, limited coverage, and difficulty in achieving real-time dynamic monitoring. In recent years, with the development of remote sensing technology, unmanned aerial vehicle (UAV) remote sensing has been widely applied in the field of soil environmental monitoring due to its advantages such as maneuverability, low cost, high resolution, and repeatable observations.

[0004] However, most existing UAV-based remote sensing monitoring systems for heavy metals in soil suffer from the following problems: First, they lack data processing capabilities, requiring the raw data to be brought back to the ground for post-processing, making real-time monitoring impossible; second, the inversion models are not highly accurate, mostly employing a single statistical model that is difficult to adapt to the complex conditions of different regions and soil types; and third, they have low levels of intelligence, lacking automatic early warning and spatiotemporal analysis functions, thus failing to provide timely and effective support for soil pollution prevention and control decisions. Summary of the Invention

[0005] The purpose of this invention is to propose an intelligent unmanned aerial vehicle (UAV) remote sensing monitoring system for soil heavy metal content, which can achieve rapid, accurate, and large-area real-time monitoring of soil heavy metal content, improve monitoring efficiency and accuracy, and provide a scientific basis for soil pollution prevention and control.

[0006] To achieve the above objectives, the present invention provides an intelligent unmanned aerial vehicle (UAV) remote sensing monitoring system for soil heavy metal content, comprising: an UAV flight platform, a multi-source remote sensing payload module, an airborne data processing unit, a ground control station, and a cloud-based intelligent analysis platform; The multi-source remote sensing payload module is mounted on an unmanned aerial vehicle (UAV) flight platform and is used to collect hyperspectral image data, multispectral image data, and terrain elevation data of the target area. The airborne data processing unit is connected to the multi-source remote sensing payload module for real-time preprocessing of the acquired raw data and removal of abnormal data. The ground control station communicates bidirectionally with the UAV flight platform and airborne data processing unit to plan flight paths, send control commands, and receive pre-processed remote sensing data. The cloud-based intelligent analysis platform communicates with the ground control station and has a built-in soil heavy metal content inversion model library and machine learning algorithm module. It is used to perform in-depth analysis on preprocessed remote sensing data, invert the content distribution of various heavy metal elements in the soil, and generate pollution level distribution maps and early warning reports.

[0007] Preferably, the multi-source remote sensing payload module includes: a hyperspectral imager, a multispectral camera, a lidar, and a GPS / BeiDou dual-mode positioning unit; The hyperspectral imager has a spectral range of 400-2500nm, a spectral resolution of ≤5nm, and a spatial resolution of ≤0.1m; the multispectral camera contains several spectral bands, covering the visible, near-infrared, and short-wave infrared regions; the lidar is used to acquire digital elevation model (DEM) data of the target area, with an elevation measurement accuracy of ≤5cm.

[0008] Preferably, the data processing unit includes: a data acquisition card, an FPGA preprocessing chip, a storage module, and a wireless transmission module; The FPGA preprocessing chip is used to perform the following operations: radiometric correction, geometric correction and bad pixel repair for hyperspectral images; dehazing and image enhancement for multispectral images; noise reduction and filtering for lidar point cloud data; the storage module uses an industrial-grade solid-state drive with a capacity of ≥1TB for temporary storage of raw data and preprocessed data.

[0009] Preferably, the ground control station includes: a flight control module, a data receiving module, a display module, and a data export module; The flight control module supports both automatic path planning and manual control modes. In automatic path planning, it can generate the optimal flight path based on the shape of the target area and the monitoring accuracy requirements, setting the overlap to 60%-80% and the flight altitude to 50-200m. The display module is used to display the UAV's flight status, payload working status, and pre-processed remote sensing images in real time.

[0010] Preferably, the soil heavy metal content inversion model library of the cloud-based intelligent analysis platform contains several inversion models based on different algorithms, including partial least squares regression (PLSR) model, support vector machine (SVM) model, random forest (RF) model, and convolutional neural network (CNN) model; the machine learning algorithm module is used to train and optimize the inversion model based on measured soil sample data to improve the inversion accuracy.

[0011] Preferably, the input to the convolutional neural network inversion model is preprocessed hyperspectral image data, and the output is the content values ​​of five heavy metal elements in the soil: cadmium (Cd), lead (Pb), chromium (Cr), copper (Cu), and zinc (Zn). The following loss function is used for training: ; in, The value of the loss function. The number of training samples. For the first The measured heavy metal content values ​​of each sample For the first The model prediction value for each sample. The regularization coefficient is . The number of model weight parameters, For the first Each weight parameter.

[0012] Preferably, the cloud-based intelligent analysis platform also includes a pollution level classification module and an early warning module; The pollution level classification module divides soil heavy metal pollution into five levels: clean, relatively clean, lightly polluted, moderately polluted, and heavily polluted; the early warning module is used to automatically issue early warnings for areas with moderate or higher pollution levels, marking the pollution location, the type of pollutant element, and the multiple of exceedance.

[0013] Preferably, the cloud-based intelligent analysis platform also includes a spatiotemporal analysis module, used to compare and analyze changes in soil heavy metal content in the same area at different times, generate trend maps, and predict the direction and speed of pollution diffusion; the prediction adopts the following grey prediction model: ; in, For the first The predicted value for the period, These are the measured values ​​at the initial moment. For development coefficient, For gray action quantity, and The solution is obtained using the least squares method.

[0014] Preferably, it also includes a mobile terminal APP, which communicates and connects with the cloud-based intelligent analysis platform to view monitoring results, pollution distribution maps and early warning information anytime and anywhere, and supports one-click sharing and report export functions.

[0015] Therefore, the present invention employs the above-mentioned intelligent soil heavy metal content UAV remote sensing monitoring system, which has the following advantages: (1) The present invention uses a multi-source remote sensing payload module, which can simultaneously acquire hyperspectral, multispectral and topographic elevation data, making full use of the advantages of different data sources and improving the accuracy of soil heavy metal content inversion; (2) The present invention is equipped with an airborne data processing unit, which can perform real-time preprocessing of raw data and removal of abnormal data, reduce data transmission volume, improve monitoring efficiency, and realize near real-time monitoring; (3) The cloud-based intelligent analysis platform of the present invention has multiple inversion models and machine learning algorithms built in, which can continuously optimize the model based on measured data, adapt to the complex situations of different regions and different soil types, and improve the inversion accuracy and versatility; (4) This invention has the functions of pollution level classification, automatic early warning and spatiotemporal analysis, which can provide timely, comprehensive and scientific support for soil pollution prevention and control decision-making; (5) The present invention is equipped with a mobile terminal APP, which makes it convenient for managers to view monitoring results and early warning information anytime and anywhere, thereby improving management efficiency.

[0016] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall structure of an intelligent soil heavy metal content UAV remote sensing monitoring system according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the internal structure of the cloud-based intelligent analysis platform in an embodiment of the present invention.

[0018] Reference numerals: 1. Unmanned Aerial Vehicle (UAV) flight platform; 2. Multi-source remote sensing payload module; 21. Hyperspectral imager; 22. Multispectral camera; 23. LiDAR; 24. GPS / BeiDou positioning unit; 3. Airborne data processing unit; 31. Data acquisition card; 32. FPGA preprocessing chip; 33. Storage module; 34. Wireless transmission module; 4. Ground control station; 41. Flight control module; 42. Data receiving module; 43. Display module; 44. Data export module; 5. Cloud-based intelligent analysis platform; 51. Data receiving module; 52. Inversion model library; 53. Machine learning algorithm module; 54. Pollution level classification module; 55. Early warning module; 56. Spatiotemporal analysis module; 57. Report generation module; 6. Mobile terminal APP. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. 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 claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0020] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0021] Example like Figure 1-2 As shown in the figure, this embodiment proposes an intelligent soil heavy metal content UAV remote sensing monitoring system, which specifically includes: UAV flight platform 1, multi-source remote sensing payload module 2, airborne data processing unit 3, ground control station 4, and cloud intelligent analysis platform 5.

[0022] The UAV flight platform 1 adopts a hexacopter UAV, which has the advantages of long endurance, good stability and strong payload capacity. The endurance is ≥60 minutes and the maximum payload is ≥5kg, which can meet the needs of large-area soil monitoring.

[0023] The multi-source remote sensing payload module 2 is mounted beneath the UAV flight platform 1 and includes a hyperspectral imager 21, a multispectral camera 22, a lidar 23, and a GPS / BeiDou dual-mode positioning unit 24. The hyperspectral imager 21 is a Headwall Nano-Hyperspec hyperspectral imager with a spectral range of 400-1000 nm, a spectral resolution of 2.1 nm, and a spatial resolution of 0.05 m. The multispectral camera 22 is a MicaSense RedEdge-MX multispectral camera, containing 10 spectral bands covering the 450-900 nm range. The lidar 23 is a Velodyne Puck lidar, capable of acquiring high-precision terrain elevation data. The GPS / BeiDou dual-mode positioning unit 24 is used to acquire the UAV's real-time location information, with a positioning accuracy ≤0.1 m.

[0024] The airborne data processing unit 3 is installed inside the UAV flight platform 1 and is connected to the multi-source remote sensing payload module 2 via a high-speed data cable. The airborne data processing unit 3 includes a data acquisition card 31, an FPGA preprocessing chip 32, a storage module 33, and a wireless transmission module 34. The data acquisition card 31 is used to acquire the raw data output by the multi-source remote sensing payload module 2. The FPGA preprocessing chip 32 uses a Xilinx Zynq-7000 series chip, which has high-speed parallel processing capabilities and can perform real-time preprocessing of the raw data. The storage module 33 uses a 1TB industrial-grade solid-state drive for temporary data storage. The wireless transmission module 34 uses a 5G communication module, which can transmit the preprocessed data to the ground control station 4 in real time.

[0025] Ground control station 4 is a portable ground station, including a flight control module 41, a data receiving module 42, a display module 43, and a data export module 44. The flight control module 41 is used to plan flight paths and control the UAV's flight, supporting both automatic and manual control modes. The data receiving module 42 receives pre-processed data transmitted from the onboard data processing unit 3. The display module 43 uses a high-definition touchscreen to display the UAV's flight status and remote sensing images in real time. The data export module 44 exports data to a USB flash drive or other storage device.

[0026] like Figure 2 As shown, the cloud-based intelligent analysis platform 5 is deployed on a cloud server and includes a data receiving module 51, an inversion model library 52, a machine learning algorithm module 53, a pollution level classification module 54, an early warning module 55, a spatiotemporal analysis module 56, and a report generation module 57. The data receiving module 51 receives preprocessed data uploaded by the ground control station 4. The inversion model library 52 contains various inversion models such as PLSR, SVM, RF, and CNN. The machine learning algorithm module 53 trains and optimizes the inversion models based on measured soil sample data. The pollution level classification module 54 classifies soil heavy metal pollution according to relevant national standards. The early warning module 55 provides automatic early warnings for areas with moderate or higher pollution levels. The spatiotemporal analysis module 56 analyzes the spatiotemporal trends of soil heavy metal content. The report generation module 57 automatically generates monitoring reports and pollution distribution maps.

[0027] The input to the Convolutional Neural Network (CNN) inversion model is preprocessed hyperspectral image data, and the output is the content values ​​of five heavy metals in the soil: cadmium (Cd), lead (Pb), chromium (Cr), copper (Cu), and zinc (Zn). The following loss function is used for training: ; in, The value of the loss function. The number of training samples. For the first The measured heavy metal content values ​​of each sample For the first The model prediction value for each sample. The regularization coefficient is . The number of model weight parameters, For the first Each weight parameter.

[0028] The spatiotemporal analysis module 56 is used to compare and analyze the changes in soil heavy metal content in the same area at different times, generate a trend map, and predict the direction and speed of pollution diffusion; the prediction uses the following grey prediction model: ; in, For the first The predicted value for the period, These are the measured values ​​at the initial moment. For development coefficient, For gray action quantity, and The solution is obtained using the least squares method.

[0029] The system also includes a mobile terminal APP6, which communicates with the cloud-based intelligent analysis platform5 to view monitoring results, pollution distribution maps, and early warning information anytime, anywhere, and supports one-click sharing and report export functions.

[0030] The working process of this invention is as follows: (1) Preliminary preparation: Collect an appropriate amount of soil samples in the target area, conduct laboratory chemical analysis, and obtain the measured values ​​of soil heavy metal content for training and verification of the inversion model.

[0031] (2) Flight planning: Input the boundary coordinates of the target area and the monitoring accuracy requirements at the ground control station 4. The flight control module 41 automatically generates the optimal flight route, sets the flight altitude to 100m, the heading overlap to 70%, and the lateral overlap to 60%.

[0032] (3) Data acquisition: The UAV flies along the planned route, and the multi-source remote sensing payload module 2 simultaneously acquires hyperspectral images, multispectral images and lidar point cloud data.

[0033] (4) Airborne preprocessing: The FPGA preprocessing chip 32 performs real-time preprocessing on the collected raw data, including radiometric correction, geometric correction, defogging, noise reduction and other operations, and removes abnormal data.

[0034] (5) Data transmission: The pre-processed data is transmitted to the ground control station 4 in real time via the 5G network, and the ground control station 4 then uploads the data to the cloud intelligent analysis platform 5.

[0035] (6) Cloud-based analysis: The cloud-based intelligent analysis platform 5 calls a suitable inversion model to analyze the data, inverts the distribution of various heavy metal elements in the soil, and then classifies the pollution level to generate a pollution level distribution map. If a moderate or higher pollution area is found, an early warning is automatically issued. At the same time, the spatiotemporal analysis module 56 compares historical data to analyze the pollution change trend and predicts the direction and speed of pollution diffusion.

[0036] (7) Results Output: The report generation module 57 automatically generates a complete monitoring report, including an overview of the monitoring area, data collection details, distribution of heavy metal content, pollution level assessment, trend analysis, and prevention and control recommendations. Users can view the monitoring results and reports through the ground control station 4 or the mobile terminal APP 6.

[0037] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. An intelligent unmanned aerial vehicle (UAV) remote sensing monitoring system for heavy metal content in soil, characterized in that, include: Unmanned aerial vehicle (UAV) flight platform, multi-source remote sensing payload module, airborne data processing unit, ground control station, and cloud-based intelligent analysis platform; The multi-source remote sensing payload module is mounted on an unmanned aerial vehicle (UAV) flight platform and is used to collect hyperspectral image data, multispectral image data, and terrain elevation data of the target area. The airborne data processing unit is connected to the multi-source remote sensing payload module for real-time preprocessing of the acquired raw data and removal of abnormal data. The ground control station communicates bidirectionally with the UAV flight platform and airborne data processing unit to plan flight paths, send control commands, and receive pre-processed remote sensing data. The cloud-based intelligent analysis platform communicates with the ground control station and has a built-in soil heavy metal content inversion model library and machine learning algorithm module. It is used to perform in-depth analysis on preprocessed remote sensing data, invert the content distribution of various heavy metal elements in the soil, and generate pollution level distribution maps and early warning reports.

2. The intelligent soil heavy metal content UAV remote sensing monitoring system according to claim 1, characterized in that: The multi-source remote sensing payload module includes: a hyperspectral imager, a multispectral camera, a lidar, and a GPS / BeiDou dual-mode positioning unit; The hyperspectral imager has a spectral range of 400-2500nm, a spectral resolution of ≤5nm, and a spatial resolution of ≤0.1m; the multispectral camera contains several spectral bands, covering the visible, near-infrared, and short-wave infrared regions; the lidar is used to acquire digital elevation model (DEM) data of the target area, with an elevation measurement accuracy of ≤5cm.

3. The intelligent soil heavy metal content UAV remote sensing monitoring system according to claim 1, characterized in that: The data processing unit includes: a data acquisition card, an FPGA preprocessing chip, a storage module, and a wireless transmission module; The FPGA preprocessing chip is used to perform the following operations: radiometric correction, geometric correction and bad pixel repair for hyperspectral images; dehazing and image enhancement for multispectral images; noise reduction and filtering for lidar point cloud data; the storage module uses an industrial-grade solid-state drive with a capacity of ≥1TB for temporary storage of raw data and preprocessed data.

4. The intelligent soil heavy metal content UAV remote sensing monitoring system according to claim 1, characterized in that: The ground control station includes: a flight control module, a data receiving module, a display module, and a data export module; The flight control module supports both automatic path planning and manual control modes. In automatic path planning, it can generate the optimal flight path based on the shape of the target area and the monitoring accuracy requirements, setting the overlap to 60%-80% and the flight altitude to 50-200m. The display module is used to display the UAV's flight status, payload working status, and pre-processed remote sensing images in real time.

5. The intelligent soil heavy metal content UAV remote sensing monitoring system according to claim 1, characterized in that: The cloud-based intelligent analysis platform's soil heavy metal content inversion model library contains several inversion models based on different algorithms, including the Partial Least Squares Regression (PLSR) model, the Support Vector Machine (SVM) model, the Random Forest (RF) model, and the Convolutional Neural Network (CNN) model. The machine learning algorithm module is used to train and optimize the inversion models based on measured soil sample data to improve inversion accuracy.

6. The intelligent soil heavy metal content UAV remote sensing monitoring system according to claim 5, characterized in that: The input to the convolutional neural network inversion model is preprocessed hyperspectral image data, and the output is the content values ​​of five heavy metal elements in the soil: cadmium (Cd), lead (Pb), chromium (Cr), copper (Cu), and zinc (Zn). The following loss function is used for training: ； in, The value of the loss function. The number of training samples. For the first The measured heavy metal content values ​​of each sample For the first The model prediction value for each sample. The regularization coefficient is . The number of model weight parameters, For the first Each weight parameter.

7. The intelligent soil heavy metal content UAV remote sensing monitoring system according to claim 1, characterized in that: The cloud-based intelligent analysis platform also includes a pollution level classification module and an early warning module; The pollution level classification module divides soil heavy metal pollution into five levels: clean, relatively clean, lightly polluted, moderately polluted, and heavily polluted; the early warning module is used to automatically issue early warnings for areas with moderate or higher pollution levels, marking the pollution location, the type of pollutant element, and the multiple of exceedance.

8. The intelligent soil heavy metal content UAV remote sensing monitoring system according to claim 1, characterized in that: The cloud-based intelligent analysis platform also includes a spatiotemporal analysis module, used to compare and analyze changes in soil heavy metal content in the same area at different times, generate trend maps, and predict the direction and speed of pollution diffusion; the prediction uses the following grey prediction model: ； in, For the first The predicted value for the period, These are the measured values ​​at the initial moment. For development coefficient, For gray action quantity, and The solution is obtained using the least squares method.

9. The intelligent soil heavy metal content UAV remote sensing monitoring system according to claim 1, characterized in that: It also includes a mobile terminal APP that communicates and connects with the cloud-based intelligent analysis platform, allowing users to view monitoring results, pollution distribution maps, and early warning information anytime, anywhere, and supports one-click sharing and report export functions.

Images