Intelligent water quality monitoring method based on multi-source spectral data fusion of unmanned aerial vehicle

Abstract

The invention relates to the technical field of intelligent water quality monitoring, in particular to an intelligent water quality monitoring method based on unmanned aerial vehicle multi-source spectral data fusion, the method is realized based on an unmanned aerial vehicle system and a detection platform system, and the method comprises the following steps: unmanned aerial vehicle flight positioning and data acquisition preparation; carrying out multi-source spectral data acquisition; carrying out data real-time transmission and edge calculation; performing multi-source data fusion and feature extraction; performing water quality parameter modeling and dynamic analysis; real-time early warning and visual output are carried out; cloud collaboration and system management are constructed; and establishing adaptive optimization and feedback. Multi-source spectral data are effectively fused, monitoring information dimensions are enriched, monitoring accuracy is improved, an advanced artificial intelligence technology is adopted, intelligent recognition and dynamic prediction of complex water quality changes, real-time dynamic monitoring and early warning functions are achieved, the response speed and the refinement level of water quality management are remarkably improved, the system is designed in a modular mode, and cost is reduced. And popularization, application and later expansion are facilitated.

Description

Technical Field

[0001] This invention relates to the field of intelligent water quality monitoring technology, specifically to an intelligent water quality monitoring method based on the fusion of multi-source spectral data from unmanned aerial vehicles (UAVs). Background Art

[0002] In recent years, remote sensing and spectral imaging technologies have provided new technical approaches for water quality monitoring. Among them, multispectral imaging can obtain the reflection characteristics of water bodies to electromagnetic waves of different bands, reflecting the optical characteristics of pollutants, algae, suspended particles and other substances in the water body.

[0003] However, single spectral sensors are limited by their spectral range and observation angle, making it difficult to comprehensively reflect the complex pollution status of water bodies. Furthermore, the acquired spectral information is often unstable and prone to errors due to environmental factors such as weather, sunlight, and water surface disturbance.

[0004] In recent years, multi-source spectral fusion technology has emerged, attempting to integrate and process remote sensing data from different bands and platforms to improve information dimensionality and monitoring accuracy. Meanwhile, artificial intelligence technology, especially deep learning algorithms, has demonstrated powerful capabilities in image recognition, spectral feature extraction, and water quality parameter modeling, driving the development of intelligent water quality monitoring.

[0005] Against this backdrop, unmanned aerial vehicle (UAV) platforms equipped with multispectral remote sensing devices are gradually becoming important data acquisition carriers for water quality monitoring due to their advantages such as maneuverability, wide operating range, and high resolution. Multi-source spectral data acquired by UAVs, combined with AI algorithms, enables dynamic monitoring of the water environment with higher precision, wider coverage, and greater timeliness.

[0006] However, some existing technical solutions in the field of water quality monitoring are as follows, but they still have the following limitations:

[0007] 1. Single-spectrum monitoring methods: For example, using visible or near-infrared spectroscopy to invert and extrapolate the concentration of certain pollutants in water bodies, but due to the lack of auxiliary information in other bands, the monitoring results are easily interfered with and the accuracy is insufficient.

[0008] 2. Multispectral combined monitoring scheme: Some schemes attempt to use sensors of multiple bands to obtain multidimensional spectral data of water bodies, but they are often deployed on fixed platforms or high-altitude satellites, which have limited spatial resolution and response speed, and usually lack efficient fusion algorithms, resulting in unsatisfactory fusion results.

[0009] 3. Modeling and analysis based on shallow machine learning: Existing models mostly use traditional regression or shallow neural networks to process single-band data, which have problems such as weak feature expression ability and poor model generalization, making it difficult to adapt to complex water quality changes.

[0010] 4. Insufficient real-time dynamic monitoring capabilities: Traditional water quality monitoring mostly involves manual sampling at fixed times and locations, combined with laboratory testing for data analysis. This process is not only time-consuming but also covers a limited area, making it difficult to respond quickly to sudden pollution events and lacking proactive early warning capabilities.

[0011] 5. Lack of integrated solutions for aerial mobile monitoring platforms: Currently, there is a lack of system solutions that deeply integrate UAV platforms with multi-source spectral fusion and AI intelligent analysis technologies, and an integrated, efficient, and intelligent dynamic water quality monitoring system has not yet been formed. Summary of the Invention

[0012] The purpose of this invention is to provide a water quality intelligent monitoring method based on the fusion of multi-source spectral data from unmanned aerial vehicles (UAVs), in order to solve the problems mentioned in the background art, such as the need for workers to enter the unloading platform of building construction projects to retrieve materials, resulting in low work efficiency, and the easy generation of dust during the use of the unloading platform, which has an impact on the site environment.

[0013] To achieve the above objectives, the present invention provides the following technical solution:

[0014] This invention proposes a smart water quality monitoring method based on the fusion of multi-source spectral data from unmanned aerial vehicles (UAVs). The method is implemented based on an UAV system and a detection platform system, and includes the following steps:

[0015] S1. Preparation for UAV flight positioning and data acquisition;

[0016] S2. Perform multi-source spectral data acquisition;

[0017] S3. Perform real-time data transmission and edge computing;

[0018] S4. Perform multi-source data fusion and feature extraction.

[0019] S5. Conduct water quality parameter modeling and dynamic analysis;

[0020] S6. Provide real-time early warnings and visual output;

[0021] S7. Building cloud-based collaboration and system management;

[0022] S8. Establish adaptive optimization and feedback.

[0023] Preferably, the implementation steps of step S1 are as follows: high-precision positioning of the UAV is achieved by combining RTK and PPK with GPS, Beidou and IMU navigation technologies, and a three-axis gimbal stabilized multispectral or hyperspectral detector is used to detect visible light, near infrared and short-wave infrared bands to ensure data acquisition stability.

[0024] Preferably, the implementation steps of step S2 are as follows: the UAV flies along a preset route, simultaneously collects multi-band spectral data of spectral reflectance information of dissolved oxygen, nitrogen and phosphorus indices in the target water area, and obtains 3D point cloud data of the water body through lidar to assist in the construction of an underwater terrain model.

[0025] Preferably, step S3 is implemented by transmitting the collected spectral data and point cloud data to the edge computing node in real time via 5G or wireless network for data preprocessing, including noise reduction, filtering, and removal of interference information.

[0026] Preferably, the implementation steps of step S4 are as follows:

[0027] S41. Employs the TCNet fusion model, i.e., a CNN fusion Transformer architecture:

[0028] S411. Local features are extracted using CNN and global dependencies are constructed using a multi-head self-attention mechanism;

[0029] S412. Integrating multispectral data with spatial distribution information of 3D point clouds to enhance pollutant identification capabilities;

[0030] S42. Combining the YOLO-V11 algorithm, feature extraction is optimized through the C3k2 module, and target association is enhanced under complex backgrounds through the C2PSA module, identifying pollutants, algae, and suspended matter targets and labeling rotated bounding boxes.

[0031] Preferably, the implementation steps of step S5 are as follows: establish a water quality parameter inversion model including pollutant concentration and turbidity based on fused data, combine historical data and real-time data, and predict water quality change trends through AI algorithms.

[0032] Preferably, the implementation steps of step S6 are as follows: constructing an abnormal data triggering early warning mechanism including pollution events, automatically generating reports, integrating data through a GIS platform, and generating a multi-dimensional monitoring map.

[0033] Preferably, the implementation steps of step S7 are as follows: using the K8S cluster management platform to elastically schedule computing resources, perform real-time processing at the edge and in-depth analysis in the cloud, and integrate Prometheus and Grafana to achieve end-to-end monitoring.

[0034] Preferably, the implementation steps of step S8 are as follows: adjust the frequency and route of drone inspections according to the monitoring results, update the AI ​​model threshold, and adapt to the needs of different aquatic environments.

[0035] Preferably, the unmanned aerial vehicle (UAV) system includes a flight positioning module, a data acquisition module, and a pollutant identification module. The flight positioning module is used to provide RTK and PPK technologies, the data acquisition module is used to acquire multispectral and hyperspectral data, and the pollutant identification module is used to provide a YOLO-V11 algorithm and a CTNet fusion model.

[0036] The detection platform system includes a data visualization module and a 3D modeling module. The 3D modeling module is used to provide 3D point cloud processing technology and to acquire corresponding water body graphic models through lidar to display the water body conditions.

[0037] The beneficial effects achieved by this invention are as follows: This invention effectively integrates multi-source spectral data, enriches the dimensions of monitoring information, improves monitoring accuracy, adopts advanced artificial intelligence technology, realizes intelligent identification and dynamic prediction of complex water quality changes, real-time dynamic monitoring and early warning functions, significantly improves the response speed and precision of water quality management, and the modular design of the system facilitates promotion, application and future expansion. Attached Figure Description

[0038] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0039] Figure 1 A flowchart of the invention is shown;

[0040] Figure 2 A functional block diagram of the present invention is shown;

[0041] Figure 3 The schematic diagram of the PPK principle of the present invention is shown;

[0042] Figure 4 A schematic diagram of the framework structure of K8S of the present invention is shown;

[0043] Figure 5 The graph shows the test results of the TCNet model of the present invention relative to other models on the same dataset;

[0044] Figure 6 A visual comparison chart of the feature extraction methods of CNN and TCNet in this invention is shown;

[0045] Figure 7 The following is a line graph showing the test results of part of the YOLO algorithm of this invention on the COCO dataset;

[0046] Figure 8 The image shows a comparison of the lidar of the present invention and the image after 3D point cloud processing. Detailed Implementation

[0047] 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.

[0048] Example 1, please refer to Figures 1 to 8 This invention proposes a smart water quality monitoring method based on the fusion of multi-source spectral data from unmanned aerial vehicles (UAVs). It should be noted that:

[0049] The unmanned aerial vehicle (UAV) system includes a flight positioning module, a data acquisition module, and a pollutant identification module. The flight positioning module is used to provide RTK and PPK technologies, the data acquisition module is used to acquire multispectral and hyperspectral data, and the pollutant identification module is used to provide a YOLO-V11 algorithm and a CTNet fusion model.

[0050] The detection platform system includes a data visualization module and a 3D modeling module. The 3D modeling module is used to provide 3D point cloud processing technology and to acquire corresponding water body graphic models through lidar to display the water body conditions.

[0051] The method is based on an unmanned aerial vehicle (UAV) system and a detection platform system, and includes the following steps:

[0052] S1. Preparation for UAV flight positioning and data acquisition;

[0053] In this embodiment, it should also be noted that the implementation steps of step S1 are as follows: high-precision positioning of the UAV is achieved by combining RTK and PPK with GPS, Beidou and IMU navigation technologies, and a three-axis gimbal stabilized multispectral or hyperspectral detector is used to detect visible light, near infrared and short-wave infrared bands to ensure data acquisition stability.

[0054] S2. Perform multi-source spectral data acquisition;

[0055] In this embodiment, it should also be noted that the implementation steps of step S2 are as follows: the UAV flies along a preset route and simultaneously collects multi-band spectral data of spectral reflectance information of dissolved oxygen, nitrogen and phosphorus indices in the target water area, and obtains 3D point cloud data of the water body through lidar to assist in the construction of an underwater terrain model.

[0056] S3. Perform real-time data transmission and edge computing;

[0057] In this embodiment, it should also be noted that the implementation steps of step S3 are as follows: the collected spectral data and point cloud data are transmitted to the edge computing node in real time via 5G or wireless network for data preprocessing such as noise reduction, filtering and removal of interference information.

[0058] S4. Perform multi-source data fusion and feature extraction.

[0059] In this embodiment, it should also be noted that the implementation steps of step S4 are as follows:

[0060] S41. Employs the TCNet fusion model, i.e., a CNN fusion Transformer architecture:

[0061] S411. Local features are extracted using CNN and global dependencies are constructed using a multi-head self-attention mechanism;

[0062] S412. Integrating multispectral data with spatial distribution information of 3D point clouds to enhance pollutant identification capabilities;

[0063] S42. Combining the YOLO-V11 algorithm, feature extraction is optimized through the C3k2 module, and target association is enhanced under complex backgrounds through the C2PSA module, identifying pollutants, algae, and suspended matter targets and labeling rotated bounding boxes;

[0064] S5. Conduct water quality parameter modeling and dynamic analysis;

[0065] In this embodiment, it should also be noted that the implementation steps of step S5 are as follows: establish a water quality parameter inversion model including pollutant concentration and turbidity based on the fused data, combine historical data and real-time data, and predict the water quality change trend through AI algorithm;

[0066] S6. Provide real-time early warnings and visual output;

[0067] In this embodiment, it should also be noted that the implementation steps of step S6 are as follows: constructing an abnormal data triggering early warning mechanism including pollution events, automatically generating reports, integrating data through a GIS platform, and generating a multi-dimensional monitoring map;

[0068] S7. Building cloud-based collaboration and system management;

[0069] In this embodiment, it should also be noted that the implementation steps of step S7 are as follows: using the K8S cluster management platform to elastically schedule computing resources, perform real-time processing at the edge and in-depth analysis in the cloud, and integrate Prometheus and Grafana to achieve full-link monitoring.

[0070] S8. Establish adaptive optimization and feedback;

[0071] In this embodiment, it should also be noted that the implementation steps of step S8 are as follows: adjust the frequency and route of the drone inspection according to the monitoring results, update the AI ​​model threshold, and adapt to the needs of different aquatic environments.

[0072] Example 2: In practical applications, the present invention provides an intelligent water quality monitoring method based on the fusion of multi-source spectral data from unmanned aerial vehicles. Specifically, the implementation steps are as follows:

[0073] Please see Figure 1 This invention constructs a drone water quality detection system that integrates efficient data acquisition, precise analysis, and intelligent decision support.

[0074] It utilizes RPK and PPK technologies to achieve accurate flight positioning, multispectral and hyperspectral technologies to achieve efficient water quality data acquisition, YOLO-V11 algorithm and CTNet fusion model to achieve stable pollutant identification, and finally uses 3D point cloud processing and data visualization to build a big data analysis system for water quality testing, providing a scientific basis for water resource protection.

[0075] Please see Figure 2 This invention consists of two main parts: an unmanned aerial vehicle (UAV) system and a detection platform system (K8S).

[0076] The unmanned aerial vehicle (UAV) system includes a flight positioning module (using RTK and PPK technologies), a data acquisition module (multispectral and hyperspectral), and a pollutant identification module (using the YOLO-V11 algorithm and CTNet fusion model).

[0077] The detection platform system (K8S) includes a data visualization module and a 3D modeling module. The 3D modeling module uses 3D point cloud processing technology, and the lidar it uses can display the water condition by acquiring the corresponding water body graphic model.

[0078] Please see Figure 3 The hardware components of this invention include:

[0079] (1) The photoelectric pod with high-precision motor control has stable attitude performance.

[0080] (2) High-precision RPK and PPK fusion navigation hardware, which integrates GPS, Beidou and other navigation systems and IMU (including geomagnetism, accelerometer and gyroscope) to provide high-precision positioning information;

[0081] (3) The three-axis gimbal provides a highly stable platform for imaging or sensing equipment carried by the UAV. It can rotate along three axes and be controlled independently to achieve pitch, yaw and roll movements, which is a key guarantee for obtaining high-quality and high-precision data.

[0082] (4) Multispectral detectors can efficiently monitor water quality parameters, accurately track pollution sources and analyze water quality change trends by forming specific spectral curves based on the unique reflection, absorption and transmission characteristics of different substances to different wavelengths of light.

[0083] To achieve more accurate water quality safety monitoring and non-contact safe operation;

[0084] Please see Figure 4 The present invention also includes a drone monitoring platform:

[0085] The platform uses the K8S cluster framework to ensure its stability and features high concurrency, strong stress resistance, and ease of operation.

[0086] K8S clusters provide a flexible, highly available, and easily scalable underlying architecture for drone monitoring platforms. By integrating toolchains such as Prometheus and Grafana, they enable end-to-end monitoring of drone status and cluster resources.

[0087] Its core advantages lie in automated operation and maintenance, multi-environment adaptability (cloud / edge) and flexible scalability, making it suitable for large-scale drone swarm management and complex mission scenarios;

[0088] Please see Figure 5 and Figure 6 This invention also innovatively optimizes the TCNet fusion model, YOLO-V11 algorithm, and 3D point cloud processing technology:

[0089] The TCNet fusion model is designed for efficient detection of water pollutants by UAVs, significantly improving the detection capability of targets at different scales (from tiny particles to large-area variations) and in occluded environments;

[0090] By combining the local feature extraction advantages of convolutional neural networks (CNN) with the global feature modeling capabilities of Transformers, complex features in water bodies can be effectively captured.

[0091] Its multi-head self-attention mechanism particularly enhances the recognition accuracy of small targets and occluded targets;

[0092] The architecture's newly added Rotated Bounding Box (OBB) and efficient attitude estimation model support rotation angle detection and attitude analysis of dynamic targets (such as floating objects and aquatic organisms), and optimize multi-task collaboration, making it suitable for UAVs to conduct comprehensive analysis of pollutant location, shape, and direction of movement in all dimensions.

[0093] Please see Figure 7 The YOLO-V11 algorithm significantly improves the ability of UAVs to detect water quality in complex environments through architectural innovation;

[0094] The algorithm first introduces the C3k2 module to replace the C2f module of YOLOv8. With its unique structure, it can achieve more efficient feature extraction and fusion, and capture richer detailed features when processing complex images, laying a better foundation for subsequent detection.

[0095] Meanwhile, the algorithm adds a C2PSA module with integrated multi-head self-attention mechanism after the SPPF layer, which effectively captures long-distance spatial dependencies within the feature map. Even under complex background interference, it can fully associate the target object with its surrounding environment information, thereby greatly improving the comprehensiveness and accuracy of feature extraction.

[0096] These two core improvements work together to make the YOLO-V11 particularly adept at handling the challenges of water quality testing by drones in complex scenarios;

[0097] Please see Figure 8 3D point cloud processing technology generates high-precision water terrain models and fuses multi-source data, allowing users to understand the underwater environment more clearly.

[0098] This technology processes point cloud data acquired by devices such as lidar, and removes irrelevant information such as impurities on the water surface (e.g., birds) through preprocessing such as noise reduction and filtering, while accurately preserving point cloud data of topographic features such as water bodies, riverbeds, and lake bottoms.

[0099] Accurate digital elevation models (DEM) and digital surface models (DSM) are generated using interpolation algorithms, clearly presenting the topographic undulations and depth variations of water bodies, providing key topographic references for water quality testing;

[0100] More importantly, this technology achieves comprehensive monitoring by fusing multi-source data: combining point clouds with multispectral images to integrate water spectral information with topographic / target spatial distribution, accurately analyzing regional differences in water quality parameters; and fusing point clouds with real-time water quality sensor data to correlate water quality indicators with three-dimensional space.

[0101] Information enables multi-dimensional monitoring of water quality.

[0102] This invention also provides a spectral multi-band fusion technology based on UAV remote sensing:

[0103] Multi-mode fusion technology enables more accurate water quality safety monitoring and more comprehensive monitoring of water quality indicators such as dissolved oxygen, nitrogen, and phosphorus. When combined with relevant hardware, it can display the water quality status to the analysis platform in real time and continuously.

[0104] It comprehensively utilizes the spectral reflectance information of water bodies on different indicators in multiple bands such as visible light, near infrared, and infrared;

[0105] Hyperspectral imagers can provide high spatial resolution images, clearly distinguishing different areas in water bodies and subtle differences in water quality.

[0106] This is of great significance for identifying the location of pollution sources, monitoring the spread of water pollution, and assessing the water quality of local water bodies;

[0107] By combining with geographic information systems, water quality monitoring results can be displayed intuitively on maps, providing more detailed and accurate information for water resource management and planning.

[0108] Through the above steps, the present invention achieves the following technical effects:

[0109] 1. Flexible and efficient water quality monitoring capabilities:

[0110] The frequency of daily monitoring can be flexibly determined according to needs. Each time the drone takes off, it can operate continuously for about 90 minutes, effectively covering the target area and completing water quality monitoring tasks, thus improving the adaptability of emergency response and routine monitoring.

[0111] 2. Improved monitoring accuracy and efficiency through artificial intelligence:

[0112] AI application systems have greatly improved the accuracy and processing efficiency of monitoring data. Compared with traditional manual methods, the efficiency of drone-based intelligent monitoring has increased from 70.5% to 98.5%, enabling rapid and accurate acquisition and analysis of water quality status.

[0113] 3. Intelligent inspection capabilities adaptable to complex environments:

[0114] This effectively solves the problems of high labor intensity and difficulty in reaching certain areas during traditional inspections. Drones possess fully automated inspection capabilities, and can still complete 98% of tasks even under extreme operating conditions, demonstrating strong environmental adaptability.

[0115] 4. Real-time transmission and intelligent early warning mechanism for water quality data:

[0116] Monitoring data can be transmitted back to the big data monitoring platform in real time and automatically generate analysis reports; the system supports remote real-time monitoring and can switch to manual intervention in case of abnormalities to quickly identify and preserve evidence of abnormal events.

[0117] 5. Cost control and operation and maintenance optimization during long-term operation:

[0118] The invention significantly reduces long-term monitoring costs by minimizing human intervention, saving approximately 3,000 yuan in labor expenses compared to traditional methods, and providing an economical and efficient solution for continuous environmental supervision.

[0119] In the description of this invention, it should be understood that the terms "coaxial," "bottom," "one end," "top," "middle," "other end," "upper," "side," "top," "inner," "front," "center," "both ends," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention.

[0120] Furthermore, the terms "first," "second," "third," and "fourth" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first," "second," "third," or "fourth" may explicitly or implicitly include at least one of those features.

[0121] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "setting," "connection," "fixing," "screw connection," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal connection of two components or the interaction between two components. Unless otherwise explicitly limited, those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0122] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A water quality intelligent monitoring method based on UAV multi-source spectral data fusion, the method being implemented based on a UAV system and a detection platform system, characterized in that, Includes the following steps: S1. Preparation for UAV flight positioning and data acquisition; S2. Perform multi-source spectral data acquisition; S3. Perform real-time data transmission and edge computing; S4. Perform multi-source data fusion and feature extraction. S5. Conduct water quality parameter modeling and dynamic analysis; S6. Provide real-time early warnings and visual output; S7. Building cloud-based collaboration and system management; S8. Establish adaptive optimization and feedback.

2. The intelligent water quality monitoring method based on UAV multi-source spectral data fusion according to claim 1, characterized in that, The implementation steps of step S1 are as follows: high-precision positioning of the UAV is achieved by combining RTK and PPK with GPS, Beidou and IMU navigation technologies, and a three-axis gimbal stabilized multispectral or hyperspectral detector is used to detect visible light, near infrared and short-wave infrared bands to ensure data acquisition stability.

3. The intelligent water quality monitoring method based on UAV multi-source spectral data fusion according to claim 2, characterized in that, The implementation steps of step S2 are as follows: the UAV flies along a preset route and simultaneously collects multi-band spectral data of spectral reflectance information of dissolved oxygen, nitrogen and phosphorus in the target water area. It also acquires 3D point cloud data of the water body through lidar to assist in the construction of an underwater terrain model.

4. The intelligent water quality monitoring method based on UAV multi-source spectral data fusion according to claim 3, characterized in that, The implementation steps of step S3 are as follows: the collected spectral data and point cloud data are transmitted to the edge computing node in real time via 5G or wireless network for data preprocessing such as noise reduction, filtering and removal of interference information.

5. The intelligent water quality monitoring method based on UAV multi-source spectral data fusion according to claim 4, characterized in that, The implementation steps of step S4 are as follows: S41. Employs the TCNet fusion model, i.e., a CNN fusion Transformer architecture: S411. Local features are extracted using CNN and global dependencies are constructed using a multi-head self-attention mechanism; S412. Integrating multispectral data with spatial distribution information of 3D point clouds to enhance pollutant identification capabilities; S42. Combining the YOLO-V11 algorithm, feature extraction is optimized through the C3k2 module, and target association is enhanced under complex backgrounds through the C2PSA module, identifying pollutants, algae, and suspended matter targets and labeling rotated bounding boxes.

6. The intelligent water quality monitoring method based on UAV multi-source spectral data fusion according to claim 5, characterized in that, The implementation steps of step S5 are as follows: Based on the fused data, a water quality parameter inversion model including pollutant concentration and turbidity is established, and by combining historical data and real-time data, the water quality change trend is predicted through AI algorithms.

7. The intelligent water quality monitoring method based on UAV multi-source spectral data fusion according to claim 6, characterized in that, The implementation steps of step S6 are as follows: construct an abnormal data triggering early warning mechanism including pollution events, automatically generate reports, integrate data through a GIS platform, and generate a multi-dimensional monitoring map.

8. The intelligent water quality monitoring method based on UAV multi-source spectral data fusion according to claim 7, characterized in that, The implementation steps of step S7 are as follows: use the K8S cluster management platform to elastically schedule computing resources, perform real-time processing at the edge and in-depth analysis in the cloud, and integrate Prometheus and Grafana to achieve full-link monitoring.

9. The intelligent water quality monitoring method based on UAV multi-source spectral data fusion according to claim 8, characterized in that, The implementation steps of step S8 are as follows: adjust the frequency and route of drone inspections based on the monitoring results, update the AI ​​model thresholds, and adapt to the needs of different aquatic environments.

10. The intelligent water quality monitoring method based on UAV multi-source spectral data fusion according to claim 9, characterized in that, The unmanned aerial vehicle (UAV) system includes a flight positioning module, a data acquisition module, and a pollutant identification module. The flight positioning module is used to provide RTK and PPK technologies, the data acquisition module is used to acquire multispectral and hyperspectral data, and the pollutant identification module is used to provide a YOLO-V11 algorithm and a CTNet fusion model. The detection platform system includes a data visualization module and a 3D modeling module. The 3D modeling module is used to provide 3D point cloud processing technology and to acquire corresponding water body graphic models through lidar to display the water body conditions.

Images