Method and system for jointly identifying suspected illegal water taking facilities by remote sensing and unmanned aerial vehicle

Abstract

This invention discloses a method and system for jointly identifying suspected illegal water extraction facilities using remote sensing and unmanned aerial vehicles (UAVs). The method includes: acquiring a multi-source dataset of the target area; analyzing the spectral, texture, and shape features of the images based on known water extraction facility locations and water extraction permit data to establish interpretation markers for water extraction facilities and generate an image sample library; and classifying areas with high incidence of suspected illegal water extraction based on land use status data and distribution data of high water-consuming industries. After acquiring the images to be identified, the suspected water extraction facility locations are extracted from the high-incidence areas using the image sample library, and legal facilities are eliminated by comparing them with the water extraction permit data. Key locations are selected for UAV field verification to collect on-site evidence to confirm their legality, and a list of suspected illegal water extraction facilities is generated. This invention effectively improves the efficiency and accuracy of identifying suspected illegal water extraction facilities by combining remote sensing and UAVs, integrating multi-source data fusion, and conducting joint field and office verification.

Description

Technical Field

[0001] This invention belongs to the field of water resource monitoring, specifically relating to a method and system for jointly identifying suspected illegal water extraction facilities using remote sensing and unmanned aerial vehicles. Background Technology

[0002] Currently, the identification and investigation of suspected illegal water extraction facilities mainly relies on manual patrols, but these methods have many limitations. When investigating large areas, it not only requires a significant amount of manpower, resources, and time, but is also limited by geographical conditions, making it difficult to achieve comprehensive coverage in remote areas and areas with complex terrain, resulting in low efficiency and poor effectiveness. Furthermore, illegal water extraction activities are random, accidental, and seasonal, and manual patrols cannot grasp their patterns, making it difficult to detect and prosecute some illegal water extraction activities in a timely manner.

[0003] Therefore, there is an urgent need for a method to identify suspected illegal water extraction facilities that can combine multi-source data, balance coverage and identification accuracy, and adapt to complex geographical environments, in order to solve the problems of low investigation efficiency, incomplete coverage, and difficulty in detecting regular illegal water extraction behavior in existing technologies. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies, such as low efficiency, limited coverage, and difficulty in detecting regular illegal water extraction activities, through manual inspections. This invention provides a method and system for jointly identifying suspected illegal water extraction facilities using remote sensing and unmanned aerial vehicles (UAVs). By synergistically applying remote sensing and UAV technologies, combined with multi-source data fusion analysis and joint verification by field and office operations, this invention effectively avoids limitations caused by factors such as terrain, weather, and seasonal changes, significantly improving the efficiency and accuracy of identifying suspected illegal water extraction facilities, and providing reliable technical support for water resource supervision.

[0005] The technical solution of this invention is: In a first aspect, the present invention provides a method for jointly identifying suspected illegal water extraction facilities using remote sensing and unmanned aerial vehicles, comprising: S1. Obtain a multi-source data set of the target area, including historical remote sensing image data, historical UAV aerial photography data, water abstraction permit data, basic geographic information data, and land use status data; S2. Based on the known locations of water intake facilities and the water intake permit data in the basic geographic information data, analyze the spectral, texture and shape characteristics of the historical remote sensing image data and the historical UAV aerial photography data, establish common water intake facility interpretation marks and generate a corresponding image sample library. S3. Based on the land use status data and the preset distribution data of high water-consuming industries, identify areas with high incidence of suspected illegal water extraction. S4. Obtain remote sensing image data and UAV aerial photography data to be identified, and with reference to the image sample library, identify and extract suspected water intake facility locations in high-incidence areas through human-computer interaction. S5. Compare the suspected water intake facility locations with the water intake permit data, eliminate facilities with legal permits, and determine the water intake type of the remaining suspected facilities based on the land use status data. S6. For the remaining suspected facility locations after comparison, select key locations for on-site verification using drones, collect on-site evidence to confirm the legality of their water extraction activities; integrate the identification results and verification results to generate a structured list of suspected illegal water extraction facilities.

[0006] Furthermore, S1 includes: S11. Acquire historical remote sensing image data with different spatial resolutions; S12. Acquire historical UAV aerial photography data, wherein the spatial resolution of the historical UAV aerial photography data is higher than that of historical remote sensing image data. S13. Obtain the latest water abstraction permit information and water abstraction project registration information to form the water abstraction permit data; S14. Obtain basic surveying data including the locations of sluice gates and pumping stations to form the basic geographic information data; S15. Obtain annual land use change survey data containing land type patches to form the land use status data.

[0007] Furthermore, S2 includes: S21. Select the sluice gates and pumping stations in the basic geographic information data, along with the drinking water intake points, as reference samples, and extract the shape feature parameters of the reference samples, including the compactness of the boundary contour and the aspect ratio. S22. Obtain the spectral curve corresponding to the reference sample from the historical remote sensing image data and extract the spectral feature parameters; obtain the reference sample image from the historical UAV aerial photography data and extract the texture feature parameters. S23. Using spectral feature parameters, texture feature parameters, and shape feature indicators as typical feature data, establish a quantitative threshold range for interpreting common water intake facility indicators. S24. Referring to the permit type in the water intake permit data, classify and label the quantitative threshold range to form a subset of interpretation markers for different types of common water intake facilities; S25. Using the aforementioned subset of interpretation markers, extract common water intake facility examples from historical remote sensing image data and historical UAV aerial photography data, perform manual verification, remove erroneously extracted samples, and generate an image example library.

[0008] Furthermore, S3 includes: Analyze the land use status data of the land type map, filter out specific land types, including industrial land and aquaculture water surface, overlay the preset high water consumption industry distribution data, identify the high water consumption enterprise concentration area in the specific land type, and classify it as a suspected high-incidence area of ​​illegal water extraction. The area where high water-consuming enterprises are concentrated is extended by a preset distance and designated as a preliminary high-incidence area. If the preliminary high-incidence area intersects with water body elements, it is marked as a suspected high-incidence area for illegal water extraction and potential water extraction paths are identified.

[0009] Furthermore, S4 includes: S41. Acquire the remote sensing image data and UAV aerial photography data to be identified, and load the image sample library into the human-computer interaction interface. S42. In areas with high incidence of remote sensing image data and UAV aerial photography data to be identified, suspected locations are initially marked by visually comparing the image sample library; S43. For the initially marked suspected locations, filter the remote sensing image data and UAV aerial photography data to be identified by the quantitative threshold range of common water intake facility interpretation marks to obtain the suspected water intake facility locations and record the geographical coordinates and feature descriptions.

[0010] Furthermore, S5 includes: S51. Spatial matching of the geographic coordinates of suspected water intake facility locations with the registered coordinates in the water intake permit data; If a match is found, legal facilities are removed based on the permit status in the water withdrawal permit data; for the remaining suspected facility locations, the land use category attribute of their corresponding land use status data is extracted. S52. Determine the water intake type based on land use attributes; determine the water intake type of the remaining suspected facilities, and mark them as priority verification objects if the water intake type is industrial water or building water.

[0011] Furthermore, S6 confirms the legality of its water extraction activities by including: S61. Select the suspected or typical locations among the remaining suspected facility locations as key locations, and plan the drone field flight path to cover the key locations. S62. Collect on-site photos and video data of the key locations using drones, analyze the signs of facility operation in the on-site photos and video data, and determine the existence of water extraction behavior; S63. If it is determined that water extraction has occurred, supplementary evidence shall be obtained by collecting interview records through additional ground surveys. The legality of the water extraction shall be confirmed by comparing the supplementary evidence with the water extraction permit data.

[0012] Furthermore, the structured list of suspected illegal water extraction facilities generated in S6 includes: S64. Summarize the identification results of suspected water intake facility locations and the evidence data from the on-site verification by drones, and standardize the attribute fields of the summarized data to form a unified data table structure; S65. Based on the legality status of suspected water intake facility locations, classify and organize the location information in the data table structure, generate a list of suspected illegal water intake facilities including location coordinates, water intake type, and evidence links, and push it to the relevant law enforcement system through a preset interface.

[0013] Secondly, the present invention provides a system for jointly identifying suspected illegal water extraction facilities using remote sensing and unmanned aerial vehicles, comprising: The data acquisition module is configured to acquire a multi-source data set of the target area, which includes historical remote sensing image data, historical UAV aerial photography data, water abstraction permit data, basic geographic information data, and land use status data. The interpretation marker establishment module is configured to analyze the spectral, texture and shape features of the historical remote sensing image data and the historical UAV aerial photography data based on the known water intake facility locations and the water intake permit data in the basic geographic information data, establish common water intake facility interpretation markers and generate a corresponding image sample library. The high-incidence area division module is configured to divide suspected high-incidence areas of illegal water extraction based on the land use status data and the preset distribution data of high water-consuming industries. The suspected location identification module is configured to acquire remote sensing image data and UAV aerial photography data to be identified, and, with reference to the image sample library, identify and extract suspected water intake facility locations in the high-incidence area through human-computer interaction. The legality comparison module is configured to compare the suspected water intake facility locations with the water intake permit data, eliminate facilities with legal permits, and determine the water intake type of the remaining suspected facilities based on the land use status data. The field verification and list generation module is configured to select key locations from the remaining suspected facility locations after comparison for on-site verification using drones, collect on-site evidence to confirm the legality of their water extraction activities, and integrate the identification results and verification results to generate a structured list of suspected illegal water extraction facilities.

[0014] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed, implements the method described thereon.

[0015] The beneficial effects of this invention are: 1. Wide coverage and high screening efficiency: Remote sensing imagery can cover a large area in a short time, enabling macro-level screening; UAV aerial photography has higher resolution, allowing for low-altitude, detailed inspections of key areas of interest. The combination of the two forms a highly efficient identification model of "macro-level screening + micro-level verification," significantly shortening the screening cycle and expanding the screening coverage.

[0016] 2. High recognition accuracy and low false positive rate: By establishing a library of interpretation marks and image samples of common water intake facilities, and combining multi-source data for human-computer interactive recognition, and conducting on-site verification of suspected locations by drones, the system achieves mutual verification between remote sensing images, drone aerial photography, and on-site verification data, avoiding the problem of false positives from a single technical means and effectively improving the accuracy of the investigation results.

[0017] 3. Possesses dynamic monitoring capabilities, facilitating long-term tracking: Remote sensing technology acquires information quickly, enabling dynamic monitoring and regular updates, and promptly detecting newly added suspected illegal water extraction facilities; drones can continuously film key areas of concern to monitor changes in facilities. The combination of these two technologies provides dynamic and continuous data support for water resources regulatory departments.

[0018] 4. Reduce labor costs and form a traceable chain of evidence: The combination of remote sensing and drone technology can significantly reduce the investment in manual inspections and lower the cost of manual investigation; remote sensing data can be reused, making it easy to compare and discover changes in facilities over a long period of time, forming a visualized and traceable chain of law enforcement evidence, and improving the standardization and authority of law enforcement supervision.

[0019] 5. Support regulatory decision-making and improve management level: By generating a structured list of suspected illegal water extraction facilities and establishing a ledger of suspected illegal water extraction facilities, the distribution patterns of suspected facilities, concentrated high-incidence areas and their impact on water resource management can be analyzed, providing a data foundation and technical support for water conservancy-related departments to formulate water resource supervision policies and optimize the allocation of water resources.

[0020] Other features and advantages of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0021] The above and other objects, features and advantages of the present invention will become more apparent from the more detailed description of exemplary embodiments of the invention in conjunction with the accompanying drawings, wherein the same reference numerals generally represent the same components in the exemplary embodiments of the invention.

[0022] Figure 1 A flowchart of the method for jointly identifying suspected illegal water extraction facilities using remote sensing and unmanned aerial vehicles according to the present invention is shown. Detailed Implementation

[0023] Preferred embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While preferred embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0024] Example 1 This invention provides a method for jointly identifying suspected illegal water extraction facilities using remote sensing and unmanned aerial vehicles (UAVs). Figure 1 This is a flowchart illustrating the method for jointly identifying suspected illegal water extraction facilities using remote sensing and unmanned aerial vehicles (UAVs) according to an embodiment of the present invention. Figure 1 As shown, the method includes: S1. Obtain a multi-source data set of the target area, including historical remote sensing image data, historical UAV aerial photography data, water abstraction permit data, basic geographic information data, and land use status data; Specifically, when implementing this method, it is first necessary to collect data from multiple sources within the target area to ensure the comprehensiveness and accuracy of subsequent analysis. The target area can be a river basin, the vicinity of a lake, or a densely populated industrial area—areas where water extraction activities may occur. Acquiring these data sets is to build a complete data foundation to support the identification and verification of suspected illegal water extraction facilities. By integrating multi-source data, the characteristics of water extraction facilities can be captured from different dimensions, avoiding the limitations of a single data source. For example, in river management scenarios, multi-source data can help identify pumping stations or sluice gates hidden in vegetation, thereby improving monitoring efficiency.

[0025] In one implementation, S1 includes: S11. Acquire historical remote sensing imagery data at different spatial resolutions. Historical remote sensing imagery data refers to imagery information acquired over a past period using satellite or airborne remote sensing platforms. This data typically covers a wide area and has different spatial resolutions. Spatial resolution indicates the actual ground distance represented by each pixel in the image. For example, low-resolution data might be 2 meters per pixel, suitable for large-scale monitoring, while high-resolution data such as 0.5 meters or 0.2 meters per pixel are more suitable for detailed observation. This data can be downloaded from national remote sensing centers or commercial satellite platforms, such as medium-resolution imagery provided by the Landsat series satellites or high-resolution imagery from the Sentinel satellite. The time span of this historical data can be from the past 5 to 10 years to analyze the evolution trends of water intake facilities.

[0026] In the preprocessing stage, the remote sensing image data is preprocessed using image processing software, including orthorectification, image color balancing, image mosaicking, and frame cropping.

[0027] S12. Acquire historical UAV aerial imagery data. This historical UAV aerial imagery data has a higher spatial resolution than historical remote sensing imagery data. Historical UAV aerial imagery data consists of high-resolution images collected by cameras or sensors mounted on the UAV during past flight missions. The resolution is typically at the centimeter level, such as 0.1 meters per pixel, which is far higher than the resolution of satellite remote sensing data. This high resolution allows the images to capture the fine details of water intake facilities, such as the diameter of pump pipes or the details of sluice gates. This data can be extracted from the archives of local water authorities or UAV service providers, with a time span covering the past few years to match historical remote sensing data. The UAV imagery acquisition process involves flight path planning, such as using a grid-like flight pattern to cover the entire target area. It should be noted that this data may include orthophotos and oblique views. Orthophotos are suitable for measurement, while oblique views are helpful for 3D reconstruction.

[0028] In the preprocessing stage, aerial photography processing software is used to complete the steps of importing photos, selecting coordinate systems, selecting scenes, and stitching orthophotos of the UAV aerial photography data.

[0029] S13. Obtain the latest water abstraction permit information and water abstraction project registration information to constitute the aforementioned water abstraction permit data; the water abstraction permit data is officially recorded information on legal water abstraction activities, including details of water abstraction permit issuance and registration files of water abstraction projects. This data is usually obtained from the database of the water administration department, such as the provincial water resources management system.

[0030] S14. Obtain basic surveying and mapping data containing the locations of sluice gates and pumping stations to constitute the basic geographic information data. Basic geographic information data refers to surveying and mapping data containing topography, rivers, and infrastructure, particularly the location information of sluice gates and pumping stations. This data is obtained from the National Bureau of Surveying and Mapping or local geographic information centers. Location data includes coordinates, type, and attributes, such as the size of the sluice gate or the power of the pumping station.

[0031] S15. Obtain annual land use change survey data containing land use category patches to constitute the aforementioned land use status data. The land use status data is the result of the annual land survey and includes land use category patches such as cultivated land, industrial land, and water areas. This data is obtained from the natural resources department; the patches are represented by polygons and include attributes such as land use code and area.

[0032] S2. Based on the known locations of water intake facilities and the water intake permit data in the basic geographic information data, analyze the spectral, texture and shape characteristics of the historical remote sensing image data and the historical UAV aerial photography data, establish common water intake facility interpretation marks and generate corresponding image sample libraries.

[0033] After acquiring data in step S1, the feature analysis stage begins. This step aims to extract typical features of water intake facilities from historical data, forming a reusable library of interpretable symbols and examples. These symbols and libraries serve as core references for subsequent identification, significantly improving the efficiency of human-computer interaction. By analyzing spectrum, texture, and shape, the method captures the multidimensional characteristics of the facilities, avoiding the limitations of single features. For example, in a river environment, the spectral characteristics of a sluice gate might show low reflectivity, while its texture might exhibit a linear structure.

[0034] Specifically, S2 includes: S21. Select sluice gates and pumping stations from the basic geographic information data, along with drinking water intake points, as reference samples. Extract shape feature parameters from the reference samples, including the compactness and aspect ratio of the boundary contour. The selection of reference samples is based on known points in the basic geographic information data, which are reliable starting points. For example, sluice gates are usually rectangular, while pumping stations may be circular or square. When extracting shape features, first locate the sample's position in historical imagery, then use image processing tools to calculate the compactness of the boundary contour. Compactness is the ratio of the square of the perimeter to the area, used to measure the compactness of the shape. The aspect ratio is the ratio of length to width, reflecting the extension length of the facility.

[0035] S22. Obtain the spectral curve corresponding to the reference sample from the historical remote sensing image data and extract spectral feature parameters; obtain the reference sample image from the historical UAV aerial image data and extract texture feature parameters; spectral feature extraction involves obtaining the spectral curve of the sample pixel from the remote sensing image, where the spectral curve is a sequence of reflectance in different bands. For example, in the visible and near-infrared bands, water features show low reflectance, while surrounding vegetation shows high reflectance. Extracted parameters such as average reflectance or band ratios, such as the NDWI index, are used to highlight water features. For UAV imagery, texture feature extraction uses a gray-level co-occurrence matrix to calculate parameters such as contrast, correlation, and entropy. Contrast reflects texture coarseness, and correlation represents directional consistency. For example, the texture of a pumping station may show high contrast due to the metal surface.

[0036] S23. Using spectral feature parameters, texture feature parameters, and shape feature indicators as typical feature data, establish quantitative threshold ranges for interpreting common water intake facility indicators; for example, the tightness threshold for a sluice gate is 1.4-2.2, the spectral ratio is 0.7-1.0, and the texture contrast is >1.5. These thresholds are derived from statistical reference samples. Specifically, the threshold ranges can be established by type, such as separating them for pump stations and sluice gates.

[0037] S24. Referring to the permit types in the water intake permit data, classify and label the quantified threshold range to form interpretation label subsets for different types of common water intake facilities; classify and label the thresholds according to the permit type, such as industrial or agricultural. For example, industrial pumping stations are labeled with a high-texture-contrast subset. For example, agricultural water intakes have a more lenient threshold subset, taking into account the field environment. This classification makes the labels more targeted.

[0038] S25. Using the aforementioned subset of interpretation markers, common water intake facility examples are extracted from historical remote sensing image data and historical UAV aerial photography data. These are then manually verified, erroneously extracted samples are removed, and an image example library is generated. Examples are generated in remote sensing images and UAV aerial photography at different spatial resolutions, but the clarity of the examples varies across different images. Naturally, higher image resolution results in higher example clarity. The corresponding conclusion is that large and medium-sized water intake facilities can be identified in remote sensing images with a resolution better than 0.5 meters, while small facilities can be identified in remote sensing images with a resolution better than 0.2 meters. In specific identification processes, factors such as cost, demand, and availability should be comprehensively considered to select images with appropriate resolutions for identification.

[0039] S3. Based on the land use status data and the preset distribution data of high water consumption industries, identify areas with high incidence of suspected illegal water extraction.

[0040] Based on the output of step S2, step S3 focuses on screening risk areas. High-risk areas, such as industrial land near water bodies, are identified using land use and industry data. This narrows the search scope and improves efficiency. Land use patches in the current land use data are analyzed to filter specific land types, including industrial land and aquaculture water surfaces. Preset high water-consuming industry distribution data is overlaid to identify areas with concentrated high water-consuming enterprises within these specific land types, classifying them as areas with suspected high incidence of illegal water extraction. Land use patches are the core of land use data; specific land types, such as industrial land, are often associated with high water consumption. Preset high water-consuming industry data can be enterprise distributions provided by the statistics bureau, such as the chemical or metallurgical industries.

[0041] Specifically, S3 includes: analyzing land use patches in the current land use data, filtering out specific land types, including industrial land and aquaculture water surfaces, overlaying preset high water-consuming industry distribution data, identifying areas where high water-consuming enterprises are concentrated in the specific land types, and classifying them as areas with high incidence of suspected illegal water extraction; extending the areas where high water-consuming enterprises are concentrated by a preset distance and using them as preliminary high-incidence areas; if the preliminary high-incidence areas intersect with water body elements, they are marked as areas with high incidence of suspected illegal water extraction and potential water extraction paths are determined.

[0042] For example, in an aquaculture area, the land parcel map shows a large area of ​​fishponds. Overlaying data from high water-consuming industries identifies 10 large-scale aquaculture enterprises with high density. A 200-meter buffer zone is extended to intersect with the surrounding water channels, and the intersection is marked as a high-risk area. The potential path is the pumping pipeline at the edge of the fishponds. This step establishes a logical connection between land attributes and industries, reduces interference from irrelevant areas, and focuses the identification on high-risk points, thereby improving the overall accuracy of monitoring.

[0043] S4. Acquire remote sensing image data and UAV aerial photography data to be identified. Referring to the image sample library, identify and extract suspected water intake facility locations in high-incidence areas through human-computer interaction.

[0044] After identifying high-incidence areas in step S3, step S4 proceeds to the specific facility identification stage. The high-incidence areas, acting as spatial constraints, significantly narrow the image range requiring analysis, making the identification work more targeted. The remote sensing image data and UAV aerial photography data to be identified are currently or recently acquired, reflecting the latest status of the target area. These data correspond in type to the historical data obtained in step S1, but are updated in time, capturing newly constructed facilities or changes in facility status. Referring to the image sample library generated in step S2 provides identification personnel with intuitive and standard visual references, reducing the subjectivity and difficulty of visual interpretation. The human-computer interaction method combines the computer's rapid filtering capabilities with the experienced judgment of human experts, enabling accurate identification of suspected locations against complex backgrounds.

[0045] Specifically, S4 includes: S41, acquiring remote sensing image data and UAV aerial photography data to be identified, and loading the image sample library into the human-computer interaction interface; S42, in the high-incidence areas of the remote sensing image data and UAV aerial photography data to be identified, visually comparing the image sample library to initially mark suspected locations; S43, filtering the initially marked suspected locations from the remote sensing image data and UAV aerial photography data to be identified using the quantization threshold range of common water intake facility interpretation marks, obtaining suspected water intake facility locations, and recording their geographical coordinates and feature descriptions.

[0046] On the software interface, interpreters used the boundaries of high-incidence areas as their workspace, browsing remote sensing and drone images within that area one by one. Through visual observation, they compared the shapes, textures, colors, and spatial layouts of features in the images with typical water intake facility images in the sample library. When a feature was found to be highly similar in visual features to a sample and located near a body of water or on a potential water intake path, the interpreter used the software's point annotation tool to make an initial annotation at the center of the feature and recorded the preliminary assessment of the facility type.

[0047] For example, a location initially labeled as a "pump station" has a calculated aspect ratio of 1.1, falling within the threshold range (1.0-1.5) for pump stations; its texture contrast is 1.8, higher than the lower threshold of 1.5; however, its NDWI value is 0.3, lower than the common lower threshold of 0.5 for water-related facilities. Based on this, the location may be a pump station that is not currently drawing water, or it may not be a water intake facility at all. A filtering rule can be set: if two or more of the three main features meet the threshold, the location is retained; otherwise, it is discarded. According to this rule, the location may be filtered out due to inconsistent spectral characteristics.

[0048] Filtering can be performed in batches via scripts or plugins in human-computer interaction software. Interpreters can review the filtering results and manually adjudicate boundary cases. The points retained after filtering are the final suspected water intake facility locations. For these points, their geographical coordinates (latitude and longitude), the facility type confirmed by filtering, the extracted feature values ​​(such as compactness 1.6, NDWI 0.6, etc.), and the temporal information of the image they are located in need of being recorded in their attribute table.

[0049] For example, in a high-risk area of ​​an industrial zone, 50 suspected locations were initially identified through labeling. After quantization threshold filtering, 25 locations with high feature matching were retained. The coordinates and feature descriptions of these locations were structured and saved as input for step S5.

[0050] S5. Compare the suspected water intake facility locations with the water intake permit data, eliminate facilities with legal permits, and determine the water intake type of the remaining suspected facilities based on the land use status data.

[0051] After identifying suspected water intake facility locations, this step involves preliminary legality screening and usage determination to further focus on truly suspicious entities. By comparing with official permit data, a large number of legally registered water intake facilities can be directly ruled out, avoiding a waste of enforcement resources. Simultaneously, combining land use data to determine the type of water intake helps assess the potential impact of the activity; for example, illegal industrial water intake may consume more water resources and pose a higher pollution risk than illegal agricultural water intake.

[0052] Specifically, S5 includes S51, spatially matching the geographic coordinates of suspected water intake facility locations with the registered coordinates in the water intake permit data; if the match is successful, legal facilities are removed based on the permit status in the water intake permit data; for the remaining suspected facility locations, the land use category attribute of their corresponding land use status data is extracted; S52, the water intake type is determined based on the land use category attribute; the water intake type of the remaining suspected facilities is determined, and if the water intake type is industrial water or building water, it is marked as a priority verification object.

[0053] For example, water use types include agricultural water (irrigation of farmland, water use in aquaculture ponds, etc.), industrial water (water use in textile factories, cement plants, etc.), urban water (water use in car washes, bathhouses, restaurants, etc.), construction water (water use for construction sites, etc.), and other water use (temporary water use).

[0054] S6. For the remaining suspected facility locations after comparison, select key locations for on-site verification using drones, collect on-site evidence to confirm the legality of their water extraction activities; integrate the identification results and verification results to generate a structured list of suspected illegal water extraction facilities.

[0055] Step S5 outputs a list of remaining suspected facility locations with priority markers. However, this list may still contain misclassifications (such as mistaking other structures for water intake facilities) or cases where the legality of a water intake facility is unclear. Step S6 introduces a crucial step of on-site verification using drones to obtain firsthand evidence and ultimately confirm the authenticity of the locations and the legality of the water intake activities. This demonstrates the close integration of remote sensing monitoring and ground verification, ensuring the solid reliability of the enforcement evidence. Finally, all information is integrated into a structured list, forming a result that can be directly used for management decisions or enforcement actions.

[0056] Specifically, in S6, the status of confirming the legality of its water-taking behavior includes: S61. Select the suspected or typical locations among the remaining suspected facility locations as key locations, and plan the drone field flight path to cover the key locations; S62. Collect on-site photos and video data of the key locations using drones, analyze the signs of facility operation in the on-site photos and video data, and determine the existence of water extraction behavior; S63. If it is determined that water extraction behavior exists, collect additional ground survey and interview records as supplementary evidence, and compare the supplementary evidence with the water extraction permit data again to confirm the legality of the water extraction behavior.

[0057] In S6, generating a structured list of suspected illegal water extraction facilities includes: S64, summarizing the identification results of suspected water extraction facility locations and the evidence data from on-site verification by drones, and standardizing the attribute fields of the summarized data to form a unified data table structure; S65, classifying and organizing the location information in the data table structure according to the legality status of the suspected water extraction facility locations, generating a list of suspected illegal water extraction facilities containing location coordinates, water extraction type, and evidence links, and pushing it to the relevant law enforcement system through a preset interface.

[0058] Example 2 This invention also provides a system for jointly identifying suspected illegal water extraction facilities using remote sensing and unmanned aerial vehicles. The system includes modules that correspond one-to-one with the steps of the above method and are used to implement all or part of the process of the method. Each module can be implemented by hardware, software or a combination of both.

[0059] The data acquisition module is configured to acquire a multi-source data set of the target area, which includes historical remote sensing image data, historical UAV aerial photography data, water withdrawal permit data, basic geographic information data, and land use status data. This module can integrate functions such as data download, API access, and reading local databases, and is responsible for data collection and preliminary formatting.

[0060] The interpretation marker establishment module is configured to analyze the spectral, texture, and shape features of the historical remote sensing image data and the historical UAV aerial photography data based on the known water intake facility locations and the water intake permit data in the basic geographic information data, establish interpretation markers for common water intake facilities, and generate a corresponding image sample library. This module further includes a feature extraction submodule, a statistical analysis submodule, and a sample library management submodule to realize automatic feature calculation, threshold analysis, and sample storage and retrieval.

[0061] The high-incidence area division module is configured to divide suspected high-incidence areas of illegal water extraction based on the land use status data and the preset distribution data of high water-consuming industries. This module mainly realizes spatial overlay analysis, extended buffer analysis and spatial query functions, and can automatically perform land type screening, enterprise density calculation and area marking.

[0062] The suspected location identification module is configured to acquire remote sensing image data and UAV aerial photography data to be identified, and, with reference to the image sample library, identify and extract suspected water intake facility locations in the high-incidence area through human-computer interaction. This module usually exists in the form of a software plug-in or a standalone application, providing a graphical human-computer interaction interface, integrating image display, sample library access, manual annotation, and threshold-based filtering tools.

[0063] The legality comparison module is configured to compare the suspected water intake facility locations with the water intake permit data, eliminate facilities with legal permits, and determine the water intake type of the remaining suspected facilities based on the land use status data; this module realizes spatial matching, attribute query and rule-based water intake type inference.

[0064] The field verification and list generation module is configured to select key locations from the remaining suspected facility sites after comparison for on-site verification using drones, collecting on-site evidence to confirm the legality of their water extraction activities; and integrating the identification and verification results to generate a structured list of suspected illegal water extraction facilities. This module may include a flight mission planning submodule, a verification data management submodule, and a data integration and export submodule, supporting the entire chain of operations from planning to result generation.

[0065] The aforementioned system modules can be deployed on the same server or distributed computing nodes and communicate via an internal data bus. Users can access the system's various functions through a unified web portal or desktop client. The system's modular design enables the entire identification process to be executed automatically or semi-automatically, significantly improving the efficiency and technological level of water resource monitoring.

[0066] Example 3 The present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed, implements the method described thereon.

[0067] The various embodiments of the present invention have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments.

Claims

1. A method for jointly identifying suspected illegal water-taking facilities by remote sensing and unmanned aerial vehicles, characterized in that, include: S1. Obtain a multi-source data set for the target area, including historical remote sensing image data, historical UAV aerial photography data, water abstraction permit data, basic geographic information data, and land use status data; S2. Based on the known locations of water intake facilities and the water intake permit data in the basic geographic information data, analyze the spectral, texture and shape characteristics of the historical remote sensing image data and the historical UAV aerial photography data, establish common water intake facility interpretation marks and generate a corresponding image sample library. S3. Based on the land use status data and the preset distribution data of high water-consuming industries, identify areas with high incidence of suspected illegal water extraction. S4. Obtain remote sensing image data and UAV aerial photography data to be identified, and with reference to the image sample library, identify and extract suspected water intake facility locations in high-incidence areas through human-computer interaction. S5. Compare the suspected water intake facility locations with the water intake permit data, eliminate facilities with legal permits, and determine the water intake type of the remaining suspected facilities based on the land use status data. S6. For the remaining suspected facility locations after comparison, select key locations for on-site verification using drones, collect on-site evidence to confirm the legality of their water extraction activities; integrate the identification results and verification results to generate a structured list of suspected illegal water extraction facilities.

2. The method of claim 1, wherein the remote sensing and drone combined method of identifying suspected illegal water diversion facilities is characterized by S1 includes: S11. Acquire historical remote sensing image data with different spatial resolutions; S12. Acquire historical UAV aerial photography data, wherein the spatial resolution of the historical UAV aerial photography data is higher than that of historical remote sensing image data. S13. Obtain the latest water abstraction permit information and water abstraction project registration information to form the water abstraction permit data; S14. Obtain basic surveying data including the locations of sluice gates and pumping stations to form the basic geographic information data; S15. Obtain annual land use change survey data containing land type patches to form the land use status data.

3. The method of claim 1, wherein S2 include: S21. Select the sluice gates and pumping stations in the basic geographic information data, along with the drinking water intake points, as reference samples, and extract the shape feature parameters of the reference samples, including the compactness of the boundary contour and the aspect ratio. S22. Obtain the spectral curve corresponding to the reference sample from the historical remote sensing image data, and extract the spectral feature parameters; Obtain reference sample images from the historical UAV aerial photography data and extract texture feature parameters; S23. Using spectral feature parameters, texture feature parameters, and shape feature indicators as typical feature data, establish a quantitative threshold range for interpreting common water intake facility indicators. S24. Referring to the permit type in the water intake permit data, classify and label the quantitative threshold range to form a subset of interpretation markers for different types of common water intake facilities; S25. Using the aforementioned subset of interpretation markers, extract common water intake facility examples from historical remote sensing image data and historical UAV aerial photography data, perform manual verification, remove erroneously extracted samples, and generate an image example library.

4. The method of claim 1, wherein S3 include: Analyze the land use status data of the land type map, filter out specific land types, including industrial land and aquaculture water surface, overlay the preset high water consumption industry distribution data, identify the high water consumption enterprise concentration area in the specific land type, and classify it as a suspected high-incidence area of ​​illegal water extraction behavior. The area where high water-consuming enterprises are concentrated is extended by a preset distance and designated as a preliminary high-incidence area. If the preliminary high-incidence area intersects with water body elements, it is marked as a suspected high-incidence area for illegal water extraction and potential water extraction paths are identified.

5. The method of claim 1, wherein S4 include: S41. Acquire the remote sensing image data and UAV aerial photography data to be identified, and load the image sample library into the human-computer interaction interface. S42. In areas with high incidence of remote sensing image data and UAV aerial photography data to be identified, suspected locations are initially marked by visually comparing the image sample library; S43. For the initially marked suspected locations, filter the remote sensing image data and UAV aerial photography data to be identified by the quantitative threshold range of common water intake facility interpretation marks to obtain the suspected water intake facility locations and record the geographical coordinates and feature descriptions.

6. The method of claim 1, wherein S5 include: S51. Spatial matching of the geographic coordinates of suspected water intake facility locations with the registered coordinates in the water intake permit data; If a match is found, legal facilities are removed based on the permit status in the water withdrawal permit data; for the remaining suspected facility locations, the land use category attribute of their corresponding land use status data is extracted. S52. Determine the water intake type based on land use attributes; determine the water intake type of the remaining suspected facilities, and mark them as priority verification objects if the water intake type is industrial water or building water.

7. The method of claim 1, wherein In S6, the legality of the water-taking behavior is confirmed by the following conditions: S61. Select the suspected or typical locations among the remaining suspected facility locations as key locations, and plan the drone field flight path to cover the key locations. S62. Collect on-site photos and video data of the key locations using drones, analyze the signs of facility operation in the on-site photos and video data, and determine the existence of water extraction behavior; S63. If it is determined that water extraction has occurred, supplementary evidence shall be obtained by collecting interview records from additional ground surveys. The legality of the water extraction shall be confirmed by comparing the supplementary evidence with the water extraction permit data.

8. The method of claim 7, wherein the remote sensing and drone combined method of identifying suspected illegal water diversion facilities is characterized by In S6, the generated structured list of suspected illegal water extraction facilities includes: S64. Summarize the identification results of suspected water intake facility locations and the evidence data from the on-site verification by drones, and standardize the attribute fields of the summarized data to form a unified data table structure; S65. Based on the legality status of suspected water intake facility locations, classify and organize the location information in the data table structure, generate a list of suspected illegal water intake facilities including location coordinates, water intake type, and evidence links, and push it to the relevant law enforcement system through a preset interface.

9. A system for use in a method of remotely sensing and identifying, by a drone, suspected illegal water taking facilities according to any one of claims 1-8, characterised in that, include: The data acquisition module is configured to acquire a multi-source data set of the target area, which includes historical remote sensing image data, historical UAV aerial photography data, water abstraction permit data, basic geographic information data, and land use status data. The interpretation marker establishment module is configured to analyze the spectral, texture and shape features of the historical remote sensing image data and the historical UAV aerial photography data based on the known water intake facility locations and the water intake permit data in the basic geographic information data, establish common water intake facility interpretation markers and generate a corresponding image sample library. The high-incidence area division module is configured to divide suspected high-incidence areas of illegal water extraction based on the land use status data and the preset distribution data of high water-consuming industries. The suspected location identification module is configured to acquire remote sensing image data and UAV aerial photography data to be identified, and, with reference to the image sample library, identify and extract suspected water intake facility locations in the high-incidence area through human-computer interaction. The legality comparison module is configured to compare the suspected water intake facility locations with the water intake permit data, eliminate facilities with legal permits, and determine the water intake type of the remaining suspected facilities based on the land use status data. The field verification and list generation module is configured to select key locations from the remaining suspected facility locations after comparison for on-site verification using drones, collect on-site evidence to confirm the legality of their water extraction activities, and integrate the identification results and verification results to generate a structured list of suspected illegal water extraction facilities.

10. A computer readable storage medium having stored thereon a computer program, characterized in that When the program is executed, it implements the method as described in any one of claims 1-8.

Images