Pollution source discovery method and device based on remote sensing image
By combining multi-source remote sensing images and water quality parameter inversion models with image data from periods without human activity, changes in river and lake water quality can be accurately monitored. This solves the problem of inaccurate identification of pollution sources in river and lake water quality monitoring using remote sensing images, and achieves efficient and accurate discovery of pollution sources and water quality monitoring.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-09
- Publication Date
- 2026-03-24
AI Technical Summary
In existing technologies, remote sensing images are difficult to accurately identify sources of pollution in river and lake water quality monitoring, and traditional methods are time-consuming and costly, making it difficult to achieve large-scale, high-frequency real-time monitoring.
Using multi-source remote sensing image data, including multi-temporal, multi-spectral, multi-sensor, multi-platform, and multi-resolution images, combined with water quality parameter remote sensing inversion models and machine learning algorithms, and using remote sensing images from periods without human activity as a comparison benchmark, the influence of the water body itself is eliminated, and areas with abnormal water quality changes are accurately monitored and the source of pollution is traced.
It has enabled efficient, accurate, and real-time monitoring of river and lake water quality, reduced costs, improved the accuracy and monitoring effectiveness of pollution source lists, and provided a scientific basis for water body management.
Smart Images
Figure CN120673254B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of satellite remote sensing technology, and in particular to a method and apparatus for detecting pollution sources based on remote sensing images. Background Technology
[0002] Rivers and lakes, as vital components of the Earth's water cycle, directly impact human production, daily life, and the health and stability of the natural environment. With global climate change and intensifying human activities, river and lake regions face a series of severe problems, including water pollution and ecological degradation. Monitoring the water quality, ecological status, and trends of these regions has become a crucial task for environmental protection and water resource management. While traditional methods for monitoring river and lake water quality, such as on-site sampling and laboratory analysis, can provide relatively accurate water quality and ecological data, they are time-consuming, costly, and limited by time, human, and material resources. This makes large-scale, high-frequency real-time monitoring difficult, and long-term retrospective investigations and evidence collection challenging.
[0003] In recent years, the rapid development of remote sensing technology has provided new solutions for river and lake monitoring. Remote sensing imagery, with its advantages of wide coverage, fast acquisition speed, and large information volume, has been widely used in the field of environmental monitoring.
[0004] In existing technologies, water quality parameters are typically retrieved from remote sensing images. Anomaly areas are then identified based on these retrieved parameters, and potential sources of pollution are analyzed based on these anomaly areas. However, this method does not accurately determine the distribution of pollution sources. One reason is that the retrieved water quality parameters are affected not only by the pollution sources but also by the water quality itself, leading to inaccurate extraction of anomaly areas and consequently affecting the accuracy of pollution source identification.
[0005] Moreover, in practical applications, relying on remote sensing imagery data for river and lake monitoring still presents some challenges. The water quality and ecological conditions of river and lake areas are affected by a variety of factors, such as human activities, and changes in these factors can lead to significant changes in the water quality and ecological conditions of river and lake areas.
[0006] Therefore, in practical applications, how to effectively utilize remote sensing image data to achieve precise monitoring of high-value areas of river and lake water quality and efficiently, accurately, and in real time identify sources of pollution remains an urgent problem to be solved. Summary of the Invention
[0007] To address the aforementioned technical problems, this invention provides a method and apparatus for detecting sewage discharge sources based on remote sensing imagery, which enables precise monitoring of areas with abnormal water quality changes and can efficiently, accurately, and in real time detect sewage discharge sources.
[0008] The technical solution provided by this invention is as follows:
[0009] A method for identifying pollution sources based on remote sensing imagery, the method comprising:
[0010] S1: Acquire the remote sensing images to be processed and the background remote sensing images covering the area to be studied;
[0011] The background remote sensing image refers to the remote sensing image of the area under study during a period of no human activity or little human activity.
[0012] S2: Preprocess the remote sensing image to be processed and the background remote sensing image;
[0013] S3: Using the constructed water quality parameter remote sensing inversion model, water quality parameters are inverted on the remote sensing image to be processed and the background remote sensing image to obtain the water quality parameter inversion results of the remote sensing image to be processed and the background remote sensing image.
[0014] S4: Based on the water quality parameter inversion results of the remote sensing image to be processed and the background remote sensing image, and the preset water quality parameter threshold, the high value region of the water quality parameters of the remote sensing image to be processed and the background remote sensing image is obtained.
[0015] S5: Subtract the high-value areas of water quality parameters in the background remote sensing image from the high-value areas of water quality parameters in the remote sensing image to be processed to obtain the areas of abnormal water quality changes.
[0016] S6: Conduct pollution source tracing analysis based on the areas of abnormal water quality changes to obtain a list of pollution discharge cases.
[0017] Furthermore, the remote sensing image to be processed and the background remote sensing image are multi-source remote sensing images, which include remote sensing image data of multiple time phases, multiple spectra, multiple sensors, multiple platforms and / or multiple resolutions.
[0018] Furthermore, the preprocessing includes atmospheric correction, geometric correction, and radiometric correction.
[0019] Furthermore, a remote sensing inversion model for water quality parameters is constructed using the following method:
[0020] Based on measured water quality parameters and remote sensing image reflectance data, a remote sensing inversion model for the water quality parameters is constructed using a semi-empirical model or machine learning algorithm.
[0021] Furthermore, the water quality parameter is the total phosphorus concentration, and a remote sensing inversion model for the water quality parameter is constructed using the following method:
[0022] Using the ratio of total phosphorus concentration to remote sensing image bands, the XGBoost model is used to retrieve total phosphorus, or multiple linear regression is performed to obtain the remote sensing retrieval model of water quality parameters corresponding to the total phosphorus concentration.
[0023] Furthermore, the water quality parameter is the suspended solids concentration, and the remote sensing inversion model for the water quality parameter retrieves the suspended solids concentration using the following formula:
[0024]
[0025] Where T is the suspended solids concentration, R red R Green and R Nir These are the reflectance data for the red, green, and near-infrared bands of the remote sensing image to be processed or the background remote sensing image, respectively.
[0026] Furthermore, S6 includes:
[0027] S61: Extract the distribution and area of the abnormal water quality areas;
[0028] S62: Obtain land use data for the area with abnormal water quality changes and its surrounding areas;
[0029] S63: Using the distribution and area of areas with abnormal water quality changes, as well as the land use data, conduct pollution source tracing analysis to obtain a list of pollution discharge cases.
[0030] A device for detecting pollution sources based on remote sensing imagery, the device comprising:
[0031] The data acquisition module is used to acquire remote sensing images to be processed and background remote sensing images covering the area to be studied;
[0032] The background remote sensing image refers to the remote sensing image of the area under study during a period of no human activity or little human activity.
[0033] The preprocessing module is used to preprocess the remote sensing image to be processed and the background remote sensing image;
[0034] The water quality inversion module is used to perform water quality parameter inversion on the remote sensing image to be processed and the background remote sensing image using the constructed water quality parameter remote sensing inversion model, and obtain the water quality parameter inversion results of the remote sensing image to be processed and the background remote sensing image.
[0035] The high-value area determination module is used to obtain the high-value areas of water quality parameters in the remote sensing image to be processed and the background remote sensing image based on the water quality parameter inversion results of the remote sensing image to be processed and the background remote sensing image and the preset water quality parameter threshold.
[0036] The abnormal region determination module is used to subtract the high-value regions of water quality parameters in the background remote sensing image from the high-value regions of water quality parameters in the remote sensing image to be processed, so as to obtain the abnormal regions of water quality change.
[0037] The pollution source tracing module is used to perform pollution source tracing analysis based on the areas of abnormal water quality changes, and to obtain a list of pollution discharge cases.
[0038] Furthermore, the remote sensing image to be processed and the background remote sensing image are multi-source remote sensing images, which include remote sensing image data of multiple time phases, multiple spectra, multiple sensors, multiple platforms and / or multiple resolutions.
[0039] Furthermore, the preprocessing includes atmospheric correction, geometric correction, and radiometric correction.
[0040] Furthermore, a remote sensing inversion model for water quality parameters is constructed through the following process:
[0041] Based on measured water quality parameters and remote sensing image reflectance data, a remote sensing inversion model for the water quality parameters is constructed using a semi-empirical model or machine learning algorithm.
[0042] Furthermore, the water quality parameter is the total phosphorus concentration, and a remote sensing inversion model for the water quality parameter is constructed through the following process:
[0043] Using the ratio of total phosphorus concentration to remote sensing image bands, the XGBoost model is used to retrieve total phosphorus, or multiple linear regression is performed to obtain the remote sensing retrieval model of water quality parameters corresponding to the total phosphorus concentration.
[0044] Furthermore, the water quality parameter is the suspended solids concentration, and the remote sensing inversion model for the water quality parameter retrieves the suspended solids concentration using the following formula:
[0045]
[0046] Where T is the suspended solids concentration, R red R Green and R Nir These are the reflectance data for the red, green, and near-infrared bands of the remote sensing image to be processed or the background remote sensing image, respectively.
[0047] Furthermore, the pollution source tracing module includes:
[0048] An abnormal region extraction unit is used to extract the distribution and area of the abnormal water quality change areas;
[0049] The land use data acquisition unit is used to acquire land use data in and around the area of abnormal water quality changes.
[0050] The pollution source tracing unit is used to conduct pollution source tracing analysis by utilizing the distribution and area of areas with abnormal water quality changes, as well as the land use data, to obtain a list of pollution discharge cases.
[0051] The present invention has the following beneficial effects:
[0052] This invention utilizes remote sensing technology to achieve real-time monitoring of pollution sources in water bodies, enabling timely acquisition of water quality and ecological status information, thus improving monitoring efficiency. Compared to traditional monitoring methods, it reduces monitoring costs and enhances monitoring effectiveness. This invention introduces remote sensing image data from periods with little or no human activity as a comparative benchmark. During these periods, water quality is less affected by human activities and other factors, better reflecting changes in the water quality itself. By comparing and analyzing this data with current remote sensing image data, it can more accurately reveal the changing trends of water quality under the influence of human activities and other factors, especially changes in high-value areas. This comparative analysis method minimizes the influence of the river and lake's own water quality, enabling precise source tracing analysis, providing support for the discovery of pollution sources, improving the accuracy of pollution source lists, and thus enhancing the precision and reliability of water quality monitoring in river and lake areas. This invention provides a scientific basis and decision support for water body management and protection, helps identify potential environmental problems, and facilitates the development of relevant measures targeting the water quality of rivers and lakes and the impact of human activities, thereby improving water quality and the ecological environment. Attached Figure Description
[0053] Figure 1 This is a flowchart of the pollution source detection method based on remote sensing imagery of the present invention;
[0054] Figure 2 This is a schematic diagram of the pollution source detection device based on remote sensing images according to the present invention. Detailed Implementation
[0055] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.
[0056] Example 1:
[0057] This invention provides a method for discovering pollution sources based on remote sensing imagery, such as... Figure 1 As shown, the method includes:
[0058] S1: Acquire the remote sensing images to be processed and the background remote sensing images covering the area to be studied.
[0059] Among them, the background remote sensing image is a remote sensing image of the area under study during a period of no human activity or little human activity.
[0060] The study area in this invention generally refers to an area containing water bodies such as lakes, rivers, and reservoirs. The remote sensing images of periods with little or no human activity described in the embodiments of this invention refer to remote sensing images of the study area during periods when it is unaffected or minimally affected by human activities, serving as background remote sensing images. Background remote sensing images can eliminate the interference of human activities on water quality and reflect the water quality of the water body itself.
[0061] In practical applications, the acquisition and processing of remote sensing image data presents certain technical challenges and cost limitations. To address these issues, the remote sensing images to be processed and the background remote sensing images of this invention are multi-source remote sensing images. Multi-source remote sensing images include remote sensing image data from multiple time periods, multiple spectra, multiple sensors, multiple platforms (such as satellites, UAVs, and other remote sensing platforms), and / or multiple resolutions.
[0062] Multi-source remote sensing imagery encompasses rich information from multiple time periods, spectra, sensors, platforms, and resolutions (e.g., meter-level, sub-meter-level). It can acquire remote sensing imagery data from multiple sources, thus addressing the technical difficulties and cost limitations in acquiring and processing remote sensing imagery. Furthermore, multi-source remote sensing imagery can fuse imagery information from different time, space, and spectral resolutions, providing a richer and more comprehensive data source for river and lake monitoring. Effectively integrating and utilizing this multi-source data is crucial for a comprehensive and in-depth understanding of the water quality, ecological conditions, and changing trends of river and lake areas. It facilitates the accurate identification and real-time monitoring of pollution discharge, enabling source tracing and precise investigation of pollution sources.
[0063] S2: Preprocess the remote sensing image to be processed and the background remote sensing image.
[0064] Preprocessing includes atmospheric correction, geometric correction, radiometric correction, and image fusion (for multi-source remote sensing images). The purpose of preprocessing is to improve the quality and usability of remote sensing data. Both the remote sensing image to be processed and the background remote sensing image undergo the same preprocessing to ensure comparability between the image data.
[0065] S3: Using the constructed water quality parameter remote sensing inversion model, water quality parameters are inverted from the remote sensing image to be processed and the background remote sensing image to obtain the water quality parameter inversion results of the remote sensing image to be processed and the background remote sensing image.
[0066] In this invention, a remote sensing inversion model for water quality parameters can be constructed using the following method:
[0067] Based on measured water quality parameters (such as suspended solids concentration and total phosphorus) and remote sensing image reflectance data, a remote sensing inversion model for water quality parameters is constructed using a semi-empirical model or machine learning algorithm.
[0068] For example, when the water quality parameter is total phosphorus concentration, a remote sensing inversion model for water quality parameters can be constructed using the following method:
[0069] Using the ratio of total phosphorus concentration to remote sensing image bands, the XGBoost model is used to retrieve total phosphorus, or multiple linear regression is performed to obtain the remote sensing retrieval model of water quality parameters corresponding to the total phosphorus concentration.
[0070] After obtaining the remote sensing inversion model for water quality parameters, water quality parameter inversion is performed to obtain the inversion results for the remote sensing image to be processed and the background remote sensing image. If there is only one water quality parameter, the remote sensing image to be processed and the background remote sensing image will each yield the inversion result for that single water quality parameter. If there are multiple water quality parameters, the remote sensing image to be processed and the background remote sensing image will each yield the inversion results for multiple water quality parameters.
[0071] In one example, when the water quality parameter is suspended solids concentration, there is a certain correlation between the suspended solids concentration and the differences between the green and red light bands, and between the green and near-infrared light bands. This invention applies a multi-band combination method to retrieve the suspended solids concentration. Specifically, the water quality parameter remote sensing inversion model retrieves the suspended solids concentration using the following formula:
[0072]
[0073] Where T is the suspended solids concentration, R red R Green and R Nir These are the reflectance data for the red, green, and near-infrared bands of the remote sensing image to be processed or the background remote sensing image, respectively.
[0074] The water quality parameter inversion results from the remote sensing image to be processed represent the actual water quality parameters at the time of monitoring, while the water quality parameter inversion results from the background remote sensing image represent the water quality parameter values of the water body itself. By comparing and analyzing the water quality parameter inversion results from these two types of remote sensing images, source tracing analysis of pollution sources can be achieved.
[0075] S4: Based on the water quality parameter inversion results of the remote sensing image to be processed and the background remote sensing image, as well as the preset water quality parameter thresholds, the high-value areas of water quality parameters in the remote sensing image to be processed and the background remote sensing image are obtained.
[0076] This step is used to screen out areas with high water quality parameters in both the remote sensing image to be processed and the background remote sensing image. The screening method is the threshold comparison method. The relevant thresholds can be set according to relevant standards. For example, according to the "Surface Water Environmental Quality Standard GB 3838-2002", areas with total phosphorus concentration thresholds worse than Class III water quality are screened out; and high-value areas are screened out according to the Class I standard for suspended solids concentration in the "Integrated Wastewater Discharge Standard GB8978-1996".
[0077] S5: Subtract the high-value areas of water quality parameters in the background remote sensing image from the high-value areas of water quality parameters in the remote sensing image to be processed to obtain the areas of abnormal water quality changes.
[0078] This step involves comparing and analyzing high-value areas of water quality parameters in the baseline remote sensing image during periods of no or little human activity with high-value areas of water quality parameters in the current remote sensing image to be processed. This preliminary identification identifies areas of abnormal changes in river and lake water quality during periods of no or little human activity, as well as areas of abnormal river and lake water quality during periods of human activity, i.e., areas of severe water pollution.
[0079] Specifically, the areas with high water quality parameters in the remote sensing image to be processed can be subtracted from the areas with high water quality parameters in the background remote sensing image to obtain the areas with abnormal water quality changes.
[0080] S6: Conduct pollution source tracing analysis based on areas with abnormal water quality changes to obtain a list of pollution discharge cases.
[0081] In this invention, pollution source tracing analysis can be performed based on any pollution source tracing analysis method or model disclosed in the prior art. Specific source tracing analysis methods are not detailed herein. The resulting list of pollution discharge cases includes the case number, location, etc.
[0082] As an improvement to an embodiment of the present invention, the aforementioned S6 includes:
[0083] S61: Extract the distribution and area of areas with abnormal water quality changes.
[0084] S62: Obtain land use data for areas with abnormal water quality changes and their surrounding areas.
[0085] Land use data is used to determine the impact of surrounding land use patterns on areas with abnormal water quality changes. Land use data can be obtained from platforms such as the National Geomatics Center of China, or it can be interpreted using sub-meter resolution remote sensing imagery.
[0086] S63: Utilize the distribution and area of areas with abnormal water quality changes, as well as land use data, to conduct pollution source tracing analysis and obtain a list of pollution discharge cases.
[0087] This invention utilizes the distribution and area of areas with abnormal water quality changes, as well as land use data, to conduct pollution source analysis on the changes in high-value areas of water quality parameters. Combining factors such as the surrounding environment of the water body and human activities, it analyzes the possible natural and anthropogenic factors that cause water environment damage, focuses on investigating the potential link between human activities and water environment damage, and generates a list of pollution sources during periods of human activity.
[0088] Furthermore, this invention can establish a continuous monitoring mechanism to conduct long-term and stable monitoring of the duration and development trend of high water quality areas, in order to track their changing trends and influencing factors. Finally, based on the monitoring results, a remote sensing monitoring report on sewage discharge is generated, providing scientific basis and decision support for river and lake management and protection.
[0089] This invention utilizes remote sensing technology to achieve real-time monitoring of pollution sources in water bodies, enabling timely acquisition of water quality and ecological status information, thus improving monitoring efficiency. Compared to traditional monitoring methods, it reduces monitoring costs and enhances monitoring effectiveness. This invention introduces remote sensing image data from periods with little or no human activity as a comparative benchmark. During these periods, water quality is less affected by human activities and other factors, better reflecting changes in the water quality itself. By comparing and analyzing this data with current remote sensing image data, it can more accurately reveal the changing trends of water quality under the influence of human activities and other factors, especially changes in high-value areas. This comparative analysis method minimizes the influence of the river and lake's own water quality, enabling precise source tracing analysis, providing support for the discovery of pollution sources, improving the accuracy of pollution source lists, and thus enhancing the precision and reliability of water quality monitoring in river and lake areas. This invention provides a scientific basis and decision support for water body management and protection, helps identify potential environmental problems, and facilitates the development of relevant measures targeting the water quality of rivers and lakes and the impact of human activities, thereby improving water quality and the ecological environment.
[0090] Example 2:
[0091] This invention provides a device for detecting pollution sources based on remote sensing imagery, such as... Figure 2 As shown, the device includes:
[0092] Data acquisition module 1 is used to acquire remote sensing images to be processed and background remote sensing images covering the area to be studied.
[0093] Among them, the background remote sensing image is a remote sensing image of the area under study during a period of no human activity or little human activity.
[0094] Preprocessing module 2 is used to preprocess the remote sensing image to be processed and the background remote sensing image.
[0095] The water quality inversion module 3 is used to invert water quality parameters of the remote sensing image to be processed and the background remote sensing image using the constructed water quality parameter remote sensing inversion model, and obtain the water quality parameter inversion results of the remote sensing image to be processed and the background remote sensing image.
[0096] The high-value area determination module 4 is used to obtain the high-value areas of water quality parameters in the remote sensing image to be processed and the background remote sensing image based on the water quality parameter inversion results of the remote sensing image to be processed and the background remote sensing image and the preset water quality parameter threshold.
[0097] The abnormal region determination module 5 is used to subtract the high-value areas of water quality parameters in the background remote sensing image from the high-value areas of water quality parameters in the remote sensing image to be processed, so as to obtain the abnormal water quality change areas.
[0098] Pollution source tracing module 6 is used to conduct pollution source tracing analysis based on areas with abnormal water quality changes, and to obtain a list of pollution discharge cases.
[0099] In one example, the aforementioned remote sensing image to be processed and the background remote sensing image are multi-source remote sensing images, which include remote sensing image data of multiple time phases, multiple spectra, multiple sensors, multiple platforms and / or multiple resolutions.
[0100] In this invention, the preprocessing includes operations such as atmospheric correction, geometric correction, and radiometric correction.
[0101] As part of this invention, a remote sensing inversion model for water quality parameters can be constructed through the following process:
[0102] Based on measured water quality parameters and remote sensing image reflectance data, a remote sensing inversion model for water quality parameters is constructed using a semi-empirical model or machine learning algorithm.
[0103] Specifically, when the water quality parameter is total phosphorus concentration, a remote sensing inversion model for water quality parameters can be constructed through the following process:
[0104] Using the ratio of total phosphorus concentration to remote sensing image bands, the XGBoost model is used to retrieve total phosphorus, or multiple linear regression is performed to obtain the remote sensing retrieval model of water quality parameters corresponding to the total phosphorus concentration.
[0105] When the water quality parameter is the suspended solids concentration, the remote sensing inversion model for water quality parameters retrieves the suspended solids concentration using the following formula:
[0106]
[0107] Where T is the suspended solids concentration, R red R Green and R Nir These are the reflectance data for the red, green, and near-infrared bands of the remote sensing image to be processed or the background remote sensing image, respectively.
[0108] As an improvement to this embodiment of the invention, the aforementioned pollution source tracing module includes:
[0109] The abnormal area extraction unit is used to extract the distribution and area of abnormal water quality changes.
[0110] The land use data acquisition unit is used to acquire land use data in and around areas with abnormal water quality changes.
[0111] The pollution source tracing unit is used to conduct pollution source tracing analysis by utilizing the distribution and area of areas with abnormal water quality changes, as well as land use data, to obtain a list of pollution discharge cases.
[0112] The apparatus provided in this embodiment of the invention has the same implementation principle and technical effects as the aforementioned method embodiment. For the sake of brevity, any parts not mentioned in the apparatus embodiment can be referred to the corresponding content in the aforementioned method embodiment 1. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the apparatus and units described above can all be referred to the corresponding processes in the above method embodiments, and will not be repeated here.
[0113] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit them. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the scope of the technology disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention.
Claims
1. A method for discovering pollution sources based on remote sensing imagery, characterized in that, The method includes: S1: Acquire the remote sensing images to be processed and the background remote sensing images covering the area to be studied; The background remote sensing image refers to the remote sensing image of the area under study during a period of no human activity or little human activity. S2: Preprocess the remote sensing image to be processed and the background remote sensing image; S3: Using the constructed water quality parameter remote sensing inversion model, water quality parameters are inverted on the remote sensing image to be processed and the background remote sensing image to obtain the water quality parameter inversion results of the remote sensing image to be processed and the background remote sensing image. S4: Based on the water quality parameter inversion results of the remote sensing image to be processed and the background remote sensing image, and the preset water quality parameter threshold, the high value region of the water quality parameters of the remote sensing image to be processed and the background remote sensing image is obtained. S5: Subtract the high-value areas of water quality parameters in the background remote sensing image from the high-value areas of water quality parameters in the remote sensing image to be processed to obtain the areas of abnormal water quality changes. S6: Conduct pollution source tracing analysis based on the areas of abnormal water quality changes to obtain a list of pollution discharge sources; The water quality parameter is the total phosphorus concentration. A remote sensing inversion model for the water quality parameter is constructed using the following method: Using the ratio of total phosphorus concentration to remote sensing image bands, the XGBoost model is used to retrieve total phosphorus, or multiple linear regression is performed to obtain the remote sensing retrieval model of water quality parameters corresponding to total phosphorus concentration. Alternatively, the water quality parameter is the suspended solids concentration, and the remote sensing inversion model for the water quality parameter inverts the suspended solids concentration using the following formula: T=5556.2×[R Red -0.75×R Green +0.25×R Nir ]+7.34 Where T is the suspended solids concentration, R red R Green and R Nir These are the reflectance data for the red, green, and near-infrared bands of the remote sensing image to be processed or the background remote sensing image, respectively.
2. The method for discovering pollution sources based on remote sensing imagery according to claim 1, characterized in that, The remote sensing images to be processed and the background remote sensing images are multi-source remote sensing images, which include remote sensing image data of multiple time phases, multiple spectra, multiple sensors, multiple platforms and / or multiple resolutions.
3. The method for discovering pollution sources based on remote sensing imagery according to claim 1, characterized in that, The preprocessing includes atmospheric correction, geometric correction, and radiometric correction.
4. The method for discovering pollution sources based on remote sensing imagery according to claim 1, characterized in that, A remote sensing inversion model for water quality parameters was constructed using the following method: Based on measured water quality parameters and remote sensing image reflectance data, a remote sensing inversion model for the water quality parameters is constructed using a semi-empirical model or machine learning algorithm.
5. The method for discovering pollution sources based on remote sensing imagery according to any one of claims 1-4, characterized in that, S6 includes: S61: Extract the distribution and area of the abnormal water quality areas; S62: Obtain land use data for the area with abnormal water quality changes and its surrounding areas; S63: Using the distribution and area of areas with abnormal water quality changes, as well as the land use data, conduct pollution source tracing analysis to obtain a list of pollution discharge cases.
6. A device for detecting pollution sources based on remote sensing imagery, characterized in that, The device includes: The data acquisition module is used to acquire remote sensing images to be processed and background remote sensing images covering the area to be studied; The background remote sensing image refers to the remote sensing image of the area under study during a period of no human activity or little human activity. The preprocessing module is used to preprocess the remote sensing image to be processed and the background remote sensing image; The water quality inversion module is used to perform water quality parameter inversion on the remote sensing image to be processed and the background remote sensing image using the constructed water quality parameter remote sensing inversion model, and obtain the water quality parameter inversion results of the remote sensing image to be processed and the background remote sensing image. The high-value area determination module is used to obtain the high-value areas of water quality parameters in the remote sensing image to be processed and the background remote sensing image based on the water quality parameter inversion results of the remote sensing image to be processed and the background remote sensing image and the preset water quality parameter threshold. The abnormal region determination module is used to subtract the high-value regions of water quality parameters in the background remote sensing image from the high-value regions of water quality parameters in the remote sensing image to be processed, so as to obtain the abnormal regions of water quality change. The pollution source tracing module is used to perform pollution source tracing analysis based on the areas of abnormal water quality changes and obtain a list of pollution discharge cases. The water quality parameter is the total phosphorus concentration. A remote sensing inversion model for the water quality parameter is constructed using the following method: Using the ratio of total phosphorus concentration to remote sensing image bands, the XGBoost model is used to retrieve total phosphorus, or multiple linear regression is performed to obtain the remote sensing retrieval model of water quality parameters corresponding to total phosphorus concentration. Alternatively, the water quality parameter is the suspended solids concentration, and the remote sensing inversion model for the water quality parameter inverts the suspended solids concentration using the following formula: T=5556.2×[R Red -0.75×R Green +0.25×R Nir ]+7.34 Where T is the suspended solids concentration, R red R Green and R Nir These are the reflectance data for the red, green, and near-infrared bands of the remote sensing image to be processed or the background remote sensing image, respectively.
7. The pollution source detection device based on remote sensing imagery according to claim 6, characterized in that, The pollution source tracing module includes: An abnormal region extraction unit is used to extract the distribution and area of the abnormal water quality change areas; The land use data acquisition unit is used to acquire land use data in and around the area of abnormal water quality changes. The pollution source tracing unit is used to conduct pollution source tracing analysis by utilizing the distribution and area of areas with abnormal water quality changes, as well as the land use data, to obtain a list of pollution discharge cases.