Soil pollution tracing method and system based on remote sensing image

Abstract

The invention discloses a soil pollution traceability method and system based on a remote sensing image, and the method comprises the steps: inputting each target remote sensing image data into a preset pollution recognition model, and enabling the pollution recognition model to output and obtain a regional pollution distribution diagram corresponding to each target remote sensing image data; according to a preset key point extraction rule, initial pollution key points are extracted from the initial pollution boundary line, other pollution key points are extracted from other pollution boundary lines, the initial pollution points and the other pollution key points are connected with an initial center point of the initial area, and at least one soil pollution doubt path is obtained; and based on the at least one soil pollution doubt path, positioning the pollution source candidate area by adopting a preset area positioning strategy, and determining final position information of a pollution emission source in the pollution source candidate area. Automatic traceability from macroscopic monitoring to precise positioning is realized, and traceability efficiency and precision are effectively improved.

Description

Technical Field

[0001] This invention belongs to the field of soil testing technology, and in particular relates to a method and system for tracing the source of soil pollution based on remote sensing images. Background Technology

[0002] Soil pollution source tracing is a core component of environmental governance and law enforcement, and its key lies in accurately locating pollution emission sources. Traditional source tracing methods heavily rely on manual ground investigation and intensive sampling analysis, which have inherent drawbacks such as high cost, low efficiency, long cycle time, and inability to cope with large areas.

[0003] In recent years, the development of remote sensing technology has provided new avenues for large-scale and rapid monitoring of soil pollution. By retrieving the spectral characteristics of soil pollutants from remote sensing images, it is possible to identify and delineate polluted areas, creating pollution distribution maps. However, most existing remote sensing-based analysis methods remain at the level of static monitoring and extent assessment.

[0004] Specifically, existing technologies are unable to automatically and objectively extract key spatial features that indicate the main migration direction of pollutants from the complex and ever-changing morphology of polluted areas during the diffusion process (instead of relying on subjective experience judgment) from these complex boundary morphologies. As a result, the efficiency of soil pollution source tracing is not high while ensuring the accuracy of source tracing. Summary of the Invention

[0005] This invention provides a method and system for tracing soil pollution sources based on remote sensing images, which addresses the technical problem of low efficiency in tracing soil pollution sources while ensuring accuracy.

[0006] In a first aspect, the present invention provides a method for tracing the source of soil pollution based on remote sensing images, comprising: Acquire remote sensing image data of the soil area to be processed at different time points; The remote sensing image data of each target is input into a preset pollution identification model, and the pollution identification model outputs a regional pollution distribution map corresponding to the remote sensing image data of each target, wherein the regional pollution distribution map includes pollution boundary lines; According to the preset key point extraction rules, initial pollution key points are extracted on the initial pollution boundary line, and other pollution key points are extracted on other pollution boundary lines. The initial pollution points and other pollution key points are connected to the initial center point of the initial area to obtain at least one soil pollution suspected path. The initial pollution boundary line is the pollution boundary line in the pollution distribution map of the initial area, and the other pollution boundary lines are the pollution boundary lines in the pollution distribution maps of other areas. Based on the at least one suspected soil pollution path, a preset regional positioning strategy is used to locate the candidate pollution source area, and the final location information of the pollution emission source is determined within the candidate pollution source area.

[0007] Secondly, the present invention provides a soil pollution source tracing system based on remote sensing images, comprising: The acquisition module is configured to acquire remote sensing image data of the soil area to be processed at different time points. The output module is configured to input remote sensing image data of each target into a preset pollution identification model, wherein the pollution identification model outputs a regional pollution distribution map corresponding to the remote sensing image data of each target, wherein the regional pollution distribution map includes pollution boundary lines; The extraction module is configured to extract initial pollution key points on the initial pollution boundary line and other pollution key points on other pollution boundary lines according to preset key point extraction rules, and connect the initial pollution point and the other pollution key points to the initial center point of the initial area to obtain at least one soil pollution suspected path, wherein the initial pollution boundary line is the pollution boundary line in the pollution distribution map of the initial area, and the other pollution boundary lines are the pollution boundary lines in the pollution distribution maps of other areas; The determination module is configured to locate the candidate pollution source area based on the at least one suspected soil pollution path, using a preset regional positioning strategy, and determine the final location information of the pollution emission source within the candidate pollution source area.

[0008] Thirdly, an electronic device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the steps of the soil pollution tracing method based on remote sensing images according to any embodiment of the present invention.

[0009] Fourthly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the program instructions are executed by a processor, the processor performs the steps of the soil pollution source tracing method based on remote sensing images according to any embodiment of the present invention.

[0010] This application presents a method and system for tracing soil pollution sources based on remote sensing images. Through a progressive spatial analysis method of "key points-path-candidate regions," it transforms static pollution distribution maps into dynamic source tracing paths, intelligently locking high-probability candidate pollution source regions, and effectively converging the source tracing scope from area to line. Finally, within the candidate regions, it combines the spatiotemporal accumulation index of pollution concentration with an environmental diffusion model for refined positioning. This not only improves the source tracing accuracy to the pixel level but also ensures that the positioning results conform to the physical laws of pollutant migration, significantly improving the reliability, scientific validity, and practicality of the source tracing results, and providing efficient and reliable technical support for environmental supervision and governance. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 A flowchart illustrating a method for tracing soil pollution sources based on remote sensing images, as provided in an embodiment of the present invention; Figure 2 This is a structural block diagram of a soil pollution source tracing system based on remote sensing images, provided in an embodiment of the present invention. Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0014] Please see Figure 1 The diagram shows a flowchart of a method for tracing soil pollution sources based on remote sensing images, as proposed in this application.

[0015] like Figure 1 As shown, the soil pollution source tracing method based on remote sensing images specifically includes the following steps: Step S101: Obtain remote sensing image data of the soil area to be processed at different time points.

[0016] In this step, the original remote sensing image data of the soil area to be processed at different time points are obtained, wherein the original remote sensing image data have different sizes and spatial resolutions; Based on the georeferenced information of the original remote sensing image data, the geographic extent of all the original remote sensing image data is extracted, and a common geographic bounding box is calculated, wherein the common geographic bounding box covers the soil area to be processed in all the original remote sensing image data. The common geographic bounding box is divided into a uniform grid, and the target pixel size is determined, where the target pixel size is set to the highest spatial resolution among all the original remote sensing image data; The nearest neighbor interpolation method is used to resample each original remote sensing image data to a common geographic bounding box and target pixel size, resulting in remote sensing image data of each target with consistent size.

[0017] In this embodiment, firstly, data is collected from a satellite data platform (such as...). Remote sensing systems (such as high-resolution series) or aerial / unmanned aerial vehicles (UAVs) acquire raw remote sensing image data covering the soil area to be treated, collected at different time points (e.g., monthly or quarterly). This raw data is typically... Data is stored in formats that include georeferenced information. When reading the data, it is necessary to simultaneously parse and extract the embedded georeferenced information, including but not limited to coordinate systems (such as coordinate systems). The original images contain projection parameters, geographic coordinates of the top left corner of the image, and spatial resolution of the pixels (e.g., 10 meters, 30 meters, etc.). Due to differences in data sources and time periods, these original images inevitably exhibit inconsistencies in size and spatial resolution.

[0018] The geographic extent of all images is extracted. By comparing geographic coordinates, the smallest bounding rectangle that can completely cover the soil area to be processed in all images is calculated, i.e., the common geographic bounding box. This bounding box is defined by the maximum and minimum longitude and latitude (or projected coordinates), ensuring that subsequent processing can focus on the same geographic extent and avoid missing or misaligned areas.

[0019] The shared geographic bounding boxes are divided into a uniform grid on the plane. The target pixel size of the grid is a key parameter, set to the highest spatial resolution (i.e., the smallest numerical resolution, such as 10 meters) among all the original remote sensing image data. For example, if the original data contains images with 10-meter and 30-meter resolutions, the target pixel size should be set to 10 meters. The aim is to preserve the finest spatial details, prevent the loss of high-resolution information during data processing, and thus maximize the accuracy and clarity of the contaminated boundary lines.

[0020] The nearest neighbor interpolation method is used to resample each original remote sensing image data onto a defined common geographic bounding box and target pixel size grid. Nearest neighbor interpolation is preferred because it resamples by directly assigning the values ​​of the nearest neighbor pixels, minimizing the smoothing or distortion of pixel values ​​and effectively preserving the original spectral features. This is crucial for accurately extracting pollution boundaries in spectral analysis-based pollution identification models. After this step, images from all time points are standardized to target remote sensing image data with the exact same geographic extent, pixel size, and grid alignment.

[0021] By calculating a common geographic bounding box and setting the target pixel size based on the highest resolution, and using the nearest neighbor interpolation method for resampling, all raw data were successfully standardized to a unified spatial reference. This process ensured that images at different time points were strictly registered in space, with each pixel representing the exact same location and area on the ground, laying a solid foundation for accurately capturing the spatiotemporal evolution of contaminated areas.

[0022] Step S102: Input the remote sensing image data of each target into a preset pollution identification model. The pollution identification model outputs a regional pollution distribution map corresponding to the remote sensing image data of each target, wherein the regional pollution distribution map includes pollution boundary lines.

[0023] In this step, the input to the pollution identification model is the multispectral bands (e.g., blue, green, red, near-infrared, short-wave infrared, etc.) of the target remote sensing image data after preprocessing in step S101.

[0024] The output of the pollution identification model is a single-channel probability map or classification map with the same spatial size as the input image, where each pixel value represents the probability or category label that the location belongs to "contaminated soil".

[0025] Structure of the contamination identification model: The contamination detection model employs an encoder-decoder structure. The encoder (downsampling path) uses pre-trained... The network serves as the backbone, responsible for extracting multi-level features from multispectral images, ranging from shallow texture and edge features to deep abstract semantic features.

[0026] The decoder (upsampling path) gradually restores the spatial size of the feature map through transposed convolution or upsampling interpolation operations, and makes skip connections with the high-resolution feature map of the corresponding level of the encoder, fusing shallow details and deep semantic information to achieve pixel-level accurate classification.

[0027] The pollution identification model is connected to the last one Activation function or The activation function outputs the contamination probability for each pixel.

[0028] Use the training dataset to minimize the loss function (such as cross-entropy loss function) between the predicted results and the ground truth. The model parameters are iteratively optimized using the backpropagation algorithm (with the loss function as the objective) until the pollution identification model converges.

[0029] After the model training is completed, the target remote sensing image data at each time point obtained in step S101 are input into the trained contamination identification model one by one. The contamination identification model calculates for each pixel in the image and outputs a continuous contamination probability value (range 0-1).

[0030] Set a probability threshold (e.g., 0.5). Pixels with a probability value greater than this threshold are classified as "contaminated pixels," and the rest are classified as "non-contaminated pixels." All spatially connected "contaminated pixels" together constitute a contaminated patch.

[0031] Employing edge detection algorithms (such as...) By using operators or by directly extracting the outline of each contaminated patch, a clear and continuous contaminated boundary line can be obtained.

[0032] Finally, all pollution patches and their boundaries are overlaid on the base map to generate the regional pollution distribution map corresponding to that time point. This map visually displays the extent, shape, and spatial location of the pollution.

[0033] Step S103: Extract initial pollution key points on the initial pollution boundary line and other pollution key points on other pollution boundary lines according to the preset key point extraction rules, and connect the initial pollution points and other pollution key points to the initial center point of the initial area to obtain at least one soil pollution suspected path, wherein the initial pollution boundary line is the pollution boundary line in the pollution distribution map of the initial area, and the other pollution boundary lines are the pollution boundary lines in the pollution distribution maps of other areas.

[0034] In this step, the initial regional pollution distribution map is the regional pollution distribution map corresponding to the earliest time point among different time points, and the other regional pollution distribution maps are the regional pollution distribution maps corresponding to other time points among different time points.

[0035] It should be noted that when the initial regional pollution distribution map is input into the preset two-dimensional coordinate system, the boundary line of one map of the initial regional pollution distribution map coincides with the horizontal axis of the two-dimensional coordinate system, the boundary line of the other map of the initial regional pollution distribution map coincides with the vertical axis of the two-dimensional coordinate system, and the boundary lines of one map are adjacent to the boundary lines of the other map. Determine the center coordinates of the initial center point of the initial region, calculate the initial distance from the initial center point to each boundary point on the initial pollution boundary line based on the center coordinates, and take the boundary point corresponding to the farthest initial distance as the initial pollution key point; Input a pollution distribution map of a certain area into a preset two-dimensional coordinate system, calculate a certain distance from the initial center point to each boundary point on a certain pollution boundary line, and take the boundary point corresponding to the farthest distance as a certain pollution key point. Here, a certain pollution boundary line is the pollution boundary line in the pollution distribution map of a certain area.

[0036] In this embodiment, firstly, the regional pollution distribution maps of all time nodes obtained in step S102 are sorted in chronological order, and the distribution map corresponding to the earliest time node is defined as the initial regional pollution distribution map. The pollution area represented by this map is regarded as the "first identified pollution area", and its pollution boundary line is the initial pollution boundary line. The geometric center of this area is calculated and defined as the initial center point. The regional pollution distribution maps of the remaining time nodes are collectively referred to as other regional pollution distribution maps, and their boundary lines are other pollution boundary lines.

[0037] Establish a unified two-dimensional coordinate system: Create a preset two-dimensional coordinate system, input the initial regional pollution distribution map, and register it into this coordinate system. The specific registration method is: align the image boundaries of the distribution map with the coordinate axes, so that the boundary line of one image (e.g., the lower boundary of the image) aligns with the horizontal axis of the coordinate system. The axis coincides with the lower boundary, while aligning another graph boundary line (e.g., the left boundary of the image) adjacent to the lower boundary with the vertical axis of the coordinate system. The axes coincide. This ensures that all subsequent spatial calculations are performed under a unified and accurate coordinate reference.

[0038] In the registered coordinate system, obtain the precise center coordinates of the initial center point. , ).

[0039] Iterate through each boundary point on the initial contaminated boundary line (these points are a set of discrete coordinate pairs) and use the Euclidean distance formula to calculate the distance from the initial center point to each boundary point.

[0040] Compare all these distance values ​​and find the farthest initial distance. , will be the farthest initial distance The corresponding unique boundary point is marked as the initial pollution critical point. This point represents the spatial extent of the pollution that extends furthest from the center point when the pollution was first discovered.

[0041] For each of the other regional pollution distribution maps (denoted as a regional pollution distribution map), align it in the same way (i.e., align the lower boundary of the image). Axis, left boundary aligned The axes are registered to the unified two-dimensional coordinate system mentioned above.

[0042] For each image, the initial center point is used in the same way ( , The coordinates of the graph are used to calculate the distance from the graph to each boundary point on a certain pollution boundary line. Similarly, the farthest distance is identified, and the boundary point corresponding to this farthest distance is marked as a pollution key point at that time node. This point represents the main expansion direction of the pollution area relative to the initial center point at that time node.

[0043] In a unified two-dimensional coordinate system, the extracted initial pollution key points are... Connecting the initial center point with a straight line forms the first path; subsequently, each other key pollution point is connected to the initial center point with a straight line. These rays from the initial center point to each key point together constitute at least one suspected soil pollution path. These paths geometrically show the direction from the "regional center where pollution was first discovered" to the "farthest expansion point of pollution at each period," providing clear geometric clues for inferring the source path of pollutants.

[0044] In summary, by registering all distribution maps to a unified coordinate system and extracting key points based on the stable and physically meaningful rule of "from the initial center point to the farthest point of the boundary," this method condenses complex pollution morphological changes into feature vectors representing their main expansion direction. This effectively overcomes the subjectivity and inaccuracy brought about by relying solely on visual observation of pollution pattern changes, and provides objective and quantifiable data support for source tracing analysis.

[0045] By constructing suspected pollution pathways, a direct and robust chain of spatial evidence is provided for the source tracing hypothesis. These pathways are not randomly connected, but rather point to the furthest point in the pollution range that continuously extends outward from the initial center point over time. Physically, this typically aligns with the pattern of pollutants spreading or migrating outward from a fixed source. Therefore, the intersection or reverse extension of these pathways highly likely points to the source of pollution emissions, significantly narrowing the subsequent search area.

[0046] Step S104: Based on the at least one suspected soil pollution path, a preset regional positioning strategy is used to locate the candidate pollution source area, and the final location information of the pollution emission source is determined within the candidate pollution source area.

[0047] In this step, a first target soil pollution suspected path and a second target soil pollution suspected path are extracted from at least one soil pollution suspected path, wherein the angle formed between the first target soil pollution suspected path and the second target soil pollution suspected path is the largest. Determine whether the angle formed between the suspected soil pollution path of the first target and the suspected soil pollution path of the second target is greater than a preset angle threshold; If the angle exceeds the preset threshold, the area enclosed by the first target soil pollution suspected path, the second target soil pollution suspected path, and the initial pollution boundary line is directly defined as the pollution source candidate area. If the angle is not greater than the preset threshold, the intersection of the first target soil pollution suspected path and the initial pollution boundary line is obtained and defined as the first intersection point, or the intersection of the second target soil pollution suspected path and the initial pollution boundary line is obtained and defined as the second intersection point. A first region is obtained by drawing a circle with the first distance between the first intersection point and the initial center point as the diameter and the first center point between the first intersection point and the initial center point as the center; or a second region is obtained by drawing a circle with the second distance between the second intersection point and the initial center point as the diameter and the second center point between the second intersection point and the initial center point as the center. The first or second region is defined as a candidate region for pollution sources.

[0048] In this embodiment, the angle between every two soil pollution suspected paths obtained in step S103 is calculated. By traversing and comparing, the two paths forming the largest angle are identified and named the first target soil pollution suspected path. These paths are then connected to the initial center point with a straight line to form the first path. Subsequently, each other pollution key point is connected to the initial center point with a straight line. These rays from the initial center point to each key point collectively constitute at least one soil pollution suspected path. These paths geometrically and intuitively show the direction from the "center of the area where pollution was first discovered" to the "farthest expansion point of the pollution range at each period," providing clear geometric clues for inferring the source path of pollutants.

[0049] In summary, by registering all distribution maps to a unified coordinate system and extracting key points based on the stable and physically meaningful rule of "from the initial center point to the farthest point of the boundary," this method condenses complex pollution morphological changes into feature vectors representing their main expansion direction. This effectively overcomes the subjectivity and inaccuracy brought about by relying solely on visual observation of pollution pattern changes, and provides objective and quantifiable data support for source tracing analysis.

[0050] By constructing suspected pollution pathways, a direct and robust chain of spatial evidence is provided for the source tracing hypothesis. These pathways are not randomly connected, but rather point to the furthest point in the pollution range that continuously extends outward from the initial center point over time. Physically, this typically aligns with the pattern of pollutants spreading or migrating outward from a fixed source. Therefore, the intersection or reverse extension of these pathways highly likely points to the source of pollution emissions, significantly narrowing the subsequent search area.

[0051] Step S104: Based on the at least one suspected soil pollution path, a preset regional positioning strategy is used to locate the candidate pollution source area, and the final location information of the pollution emission source is determined within the candidate pollution source area.

[0052] In this step, a first target soil pollution suspected path and a second target soil pollution suspected path are extracted from at least one soil pollution suspected path, wherein the angle formed between the first target soil pollution suspected path and the second target soil pollution suspected path is the largest. Determine whether the angle formed between the suspected soil pollution path of the first target and the suspected soil pollution path of the second target is greater than a preset angle threshold; If the angle exceeds the preset threshold, the area enclosed by the first target soil pollution suspected path, the second target soil pollution suspected path, and the initial pollution boundary line is directly defined as the pollution source candidate area. If the angle is not greater than the preset threshold, the intersection of the first target soil pollution suspected path and the initial pollution boundary line is obtained and defined as the first intersection point, or the intersection of the second target soil pollution suspected path and the initial pollution boundary line is obtained and defined as the second intersection point. A first region is obtained by drawing a circle with the first distance between the first intersection point and the initial center point as the diameter and the first center point between the first intersection point and the initial center point as the center; or a second region is obtained by drawing a circle with the second distance between the second intersection point and the initial center point as the diameter and the second center point between the second intersection point and the initial center point as the center. The first or second region is defined as a candidate region for pollution sources.

[0053] In this embodiment, the angle between any two paths obtained in step S103 is calculated from all suspected soil pollution paths. By traversing and comparing these paths, the two paths with the largest angle are identified and named the first target suspected soil pollution paths. Second target soil pollution questionable pathway The purpose of choosing the largest possible angle is to expand the search sector as much as possible and increase the probability of covering the actual pollution source.

[0054] The specific method for calculating the included angle can be as follows: In a unified two-dimensional coordinate system, [the following is a possible interpretation of the given information] and Consider them as two vectors originating from the initial center point. Calculate the angle between these two vectors using the cross product or dot product formulas. .

[0055] When the maximum included angle is greater than the preset threshold, it indicates that the difference in direction indicated by the two paths is large enough, and the fan-shaped area formed by them has high credibility and can better limit the potential range of the pollution source.

[0056] At this point, the primary target of questionable soil pollution pathways should be directly identified. The second target is the questionable pathway of soil pollution. The approximately fan-shaped region enclosed by the initial pollution boundary line, the initial pollution boundary line, and the surrounding area is defined as the pollution source candidate region. This region is a closed polygon whose boundary consists of two straight-line paths and an arc-shaped boundary of the initial pollution.

[0057] When the maximum included angle is not greater than a preset threshold, it indicates that all path directions are relatively concentrated, and the directly enclosed area may be too narrow to effectively cover the pollution source, or it indicates that the pollution diffusion pattern is relatively complex. In this case, a conservative but more comprehensive circular area positioning method is adopted.

[0058] The method in this embodiment, by introducing the core logic of "maximum angle judgment," can adapt to different pollution diffusion patterns. When the path direction is distinct, direct geometric enclosure is used for accurate positioning; when the path direction is concentrated, a circular area is constructed to expand the search range, effectively avoiding the risk of missing pollution sources due to insufficient directional information. This flexibility makes the method well adaptable to both point source pollution and area source pollution.

[0059] Furthermore, regional remote sensing data at different time points within the candidate pollution source area are acquired, and the pollution concentration values ​​of each pixel in each regional remote sensing data are extracted to obtain a pollution concentration value sequence corresponding to each pixel. Based on each pollution concentration value sequence, a preset spatiotemporal analysis model is used to calculate the pollution accumulation index of each pixel. The local maxima of the pollution accumulation index are identified based on the peak detection algorithm, and the point with the highest pollution accumulation index among all local maxima is determined as the preliminary location of the pollution emission source. Based on the topographic and hydrological data of the candidate pollution source area, a preset diffusion correction model is used to optimize the preliminary location, and the optimized location is used as the final location information of the pollution emission source.

[0060] The specific implementation steps for determining the final location information of pollution emission sources within the candidate pollution source area are as follows: 1. Construct a pixel-level pollution concentration value sequence Data Acquisition: Acquire high spatial resolution regional remote sensing data of candidate pollution source areas at different time points (which should be consistent with or more densely packed with the main process time series). This data can be surface reflectance data that has undergone rigorous atmospheric correction.

[0061] Concentration Inversion and Sequence Construction: For each pixel within the candidate region, its multispectral data is used to calculate its pollution concentration value at each time node using a pre-established pollutant concentration inversion model (e.g., a statistical regression model or physical model based on the characteristic bands of specific pollutants). Thus, each pixel obtains an array of concentration values ​​arranged in chronological order, i.e., a pollution concentration value sequence.

[0062] 2. Calculate the pollution accumulation index Spatiotemporal analysis model: A pre-defined spatiotemporal analysis model is used to analyze the concentration value sequence of each pixel. This model aims to quantify the accumulation pattern of pollutants in time and space. An effective model can be a weighted accumulation model, the calculation formula of which is as follows: , In the formula, For all time points; Let be the pollution concentration value of pixel i at time t; The background pollution concentration value of the region can be determined by historical clean-up period data or regional background values. The time-weighted factor can be designed to be positively correlated with time (higher weight for more recent events, emphasizing recent emissions) or related to the duration of the pollution event. The pollution accumulation index comprehensively reflects the cumulative intensity and duration of pollutants exceeding the background value for each pixel during the observation period.

[0063] 3. Identify the initial location Peak detection: The calculated pollution accumulation index of the entire candidate region is treated as a two-dimensional grayscale image. Peak detection algorithms (e.g., image morphology-based reconstruction opening operation combined with local maximum search, or finding points larger than all neighbors within a specified neighborhood window) are used to identify all local maxima.

[0064] Determining the initial location: Among all identified local maxima, compare their pollution accumulation index values. The point with the highest value is identified as the initial location of the pollution source. This point represents the core point where the pollutant accumulation signal is strongest in both space and time.

[0065] 4. Perform diffusion correction optimization Data preparation: Collect high-precision digital elevation models of candidate regions ( Data and hydrogeological data (such as water flow direction, groundwater depth, soil permeability, etc.).

[0066] Diffusion Correction Model: The initial position is optimized using a pre-defined diffusion correction model. This model can be a physics-based simulation of the particle's reverse trajectory. Forward simulation: The initial location is assumed to be the leak point, combined with the terrain slope and water flow direction (from... Information such as surface roughness is extracted to simulate the transport path and diffusion range of pollutants under the action of gravity and hydraulic forces.

[0067] Reverse source tracing: Using the spatial distribution of pollution identified in the previous steps (especially the early pollution distribution) as the "target," the simulated leak point location is adjusted to achieve an optimal fit between the results of the forward simulation and the observed pollution distribution. The leak point location at the optimal fit is the optimized location.

[0068] Alternatively, a simplified empirical model can be used, for example, shifting the initial location upstream (against the flow direction and slope) by a distance determined by historical data or empirical formulas, based on the dominant flow direction and slope.

[0069] Final output location: The optimized location output by the diffusion correction model is the final location information of the pollution emission source. This location, after taking into account the physical laws of pollutant migration in the environment, is more scientific and reliable than simply the peak point of the remote sensing signal.

[0070] In summary, the method presented in this application, through a progressive spatial analysis approach of "key points-path-candidate regions," transforms static pollution distribution maps into dynamic source tracing paths, intelligently identifying high-probability candidate pollution sources and effectively converging the source tracing scope from surface to line. Finally, within the candidate regions, it combines the spatiotemporal accumulation index of pollution concentration with an environmental diffusion model for refined positioning. This not only improves the source tracing accuracy to the pixel level but also ensures that the positioning results conform to the physical laws of pollutant migration, significantly enhancing the reliability, scientific validity, and practicality of the source tracing results. This provides efficient and reliable technical support for environmental supervision and governance.

[0071] Please see Figure 2 The diagram shows a structural block diagram of a soil pollution tracing system based on remote sensing images according to this application.

[0072] like Figure 2 As shown, the soil pollution tracing system 200 based on remote sensing images includes an acquisition module 210, an output module 220, an extraction module 230, and a determination module 240.

[0073] The system includes: an acquisition module 210 configured to acquire remote sensing image data of the soil region to be processed at different time points; an output module 220 configured to input each target remote sensing image data into a preset pollution identification model, wherein the pollution identification model outputs a regional pollution distribution map corresponding to each target remote sensing image data, wherein the regional pollution distribution map includes a pollution boundary line; an extraction module 230 configured to extract initial pollution key points on the initial pollution boundary line and other pollution key points on other pollution boundary lines according to preset key point extraction rules, and connect the initial pollution point and the other pollution key points to the initial center point of the initial region to obtain at least one soil pollution suspected path, wherein the initial pollution boundary line is the pollution boundary line in the initial regional pollution distribution map, and the other pollution boundary lines are the pollution boundary lines in other regional pollution distribution maps; and a determination module 240 configured to locate the pollution source candidate region based on the at least one soil pollution suspected path using a preset regional positioning strategy, and determine the final location information of the pollution emission source within the pollution source candidate region.

[0074] It should be understood that Figure 2 The modules and references described in the document Figure 1 The steps described in the text correspond to those in the method described above. Therefore, the operations, features, and corresponding technical effects described above also apply to the method described in the text. Figure 2 The various modules in the document will not be described in detail here.

[0075] In other embodiments, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the program instructions are executed by a processor, the processor performs the soil pollution source tracing method based on remote sensing images in any of the above method embodiments. In one embodiment, the computer-readable storage medium of the present invention stores computer-executable instructions, which are configured as follows: Acquire remote sensing image data of the soil area to be processed at different time points; The remote sensing image data of each target is input into a preset pollution identification model, and the pollution identification model outputs a regional pollution distribution map corresponding to the remote sensing image data of each target, wherein the regional pollution distribution map includes pollution boundary lines; According to the preset key point extraction rules, initial pollution key points are extracted on the initial pollution boundary line, and other pollution key points are extracted on other pollution boundary lines. The initial pollution points and other pollution key points are connected to the initial center point of the initial area to obtain at least one soil pollution suspected path. The initial pollution boundary line is the pollution boundary line in the pollution distribution map of the initial area, and the other pollution boundary lines are the pollution boundary lines in the pollution distribution maps of other areas. Based on the at least one suspected soil pollution path, a preset regional positioning strategy is used to locate the candidate pollution source area, and the final location information of the pollution emission source is determined within the candidate pollution source area.

[0076] Computer-readable storage media may include a stored program area and a stored data area, wherein the stored program area may store an operating system and an application program required for at least one function; the stored data area may store data created based on the use of the remote sensing image-based soil pollution tracing system, etc. Furthermore, the computer-readable storage medium may include high-speed random access memory, and may also include memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, the computer-readable storage medium may optionally include memory remotely configured relative to a processor, which can be connected to the remote sensing image-based soil pollution tracing system via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0077] Figure 3 This is a schematic diagram of the structure of the electronic device provided in the embodiment of the present invention, such as... Figure 3 As shown, the device includes a processor 310 and a memory 320. The electronic device may also include an input device 330 and an output device 340. The processor 310, memory 320, input device 330, and output device 340 can be connected via a bus or other means. Figure 3 Taking a bus connection as an example, the memory 320 is the computer-readable storage medium described above. The processor 310 executes various server functions and data processing by running non-volatile software programs, instructions, and modules stored in the memory 320, thereby implementing the soil pollution tracing method based on remote sensing images described in the above embodiment. The input device 330 can receive input digital or character information and generate key signal inputs related to user settings and function control of the soil pollution tracing system based on remote sensing images. The output device 340 may include a display screen or other display device.

[0078] The aforementioned electronic device can execute the method provided in the embodiments of the present invention, and has the corresponding functional modules and beneficial effects for executing the method. Technical details not described in detail in this embodiment can be found in the method provided in the embodiments of the present invention.

[0079] In one implementation, the above-described electronic device is applied to a soil pollution tracing system based on remote sensing images, and is used as a client. It includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enable the at least one processor to: Acquire remote sensing image data of the soil area to be processed at different time points; The remote sensing image data of each target is input into a preset pollution identification model, and the pollution identification model outputs a regional pollution distribution map corresponding to the remote sensing image data of each target, wherein the regional pollution distribution map includes pollution boundary lines; According to the preset key point extraction rules, initial pollution key points are extracted on the initial pollution boundary line, and other pollution key points are extracted on other pollution boundary lines. The initial pollution points and other pollution key points are connected to the initial center point of the initial area to obtain at least one soil pollution suspected path. The initial pollution boundary line is the pollution boundary line in the pollution distribution map of the initial area, and the other pollution boundary lines are the pollution boundary lines in the pollution distribution maps of other areas. Based on the at least one suspected soil pollution path, a preset regional positioning strategy is used to locate the candidate pollution source area, and the final location information of the pollution emission source is determined within the candidate pollution source area.

[0080] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus necessary general-purpose hardware platforms, and of course, it can also be implemented using hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as... Disks, optical disks, etc., including a number of instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform the methods of various embodiments or parts thereof.

[0081] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications 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.

Claims

1. A method for tracing the source of soil pollution based on remote sensing images, characterized in that, include: Acquire remote sensing image data of the soil area to be processed at different time points; The remote sensing image data of each target is input into a preset pollution identification model, and the pollution identification model outputs a regional pollution distribution map corresponding to the remote sensing image data of each target, wherein the regional pollution distribution map includes pollution boundary lines; According to the preset key point extraction rules, initial pollution key points are extracted on the initial pollution boundary line, and other pollution key points are extracted on other pollution boundary lines. The initial pollution points and other pollution key points are connected to the initial center point of the initial area to obtain at least one soil pollution suspected path. The initial pollution boundary line is the pollution boundary line in the pollution distribution map of the initial area, and the other pollution boundary lines are the pollution boundary lines in the pollution distribution maps of other areas. Based on the at least one suspected soil pollution path, a preset regional positioning strategy is used to locate the candidate pollution source area, and the final location information of the pollution emission source is determined within the candidate pollution source area.

2. The method for tracing soil pollution sources based on remote sensing images according to claim 1, characterized in that, The acquisition of remote sensing image data of the soil area to be processed at different time points includes: Acquire raw remote sensing image data of the soil area to be processed at different time points, wherein the raw remote sensing image data have different sizes and spatial resolutions; Based on the georeferenced information of the original remote sensing image data, the geographic extent of all original remote sensing image data is extracted, and a common geographic bounding box is calculated, wherein the common geographic bounding box covers the soil area to be processed in all original remote sensing image data. The common geographic bounding box is divided into a uniform grid, and the target pixel size is determined, wherein the target pixel size is set to the highest spatial resolution among all the original remote sensing image data; Each original remote sensing image data is resampled to the common geographic bounding box and the target pixel size using the nearest neighbor interpolation method, resulting in remote sensing image data of each target with consistent size.

3. The method for tracing soil pollution sources based on remote sensing images according to claim 1, characterized in that, in, The initial regional pollution distribution map is the regional pollution distribution map corresponding to the earliest time point among the different time points, and the other regional pollution distribution maps are the regional pollution distribution maps corresponding to other time points among the different time points; The step of extracting at least one initial pollution key point on the initial pollution boundary line according to a preset key point extraction rule, and extracting other pollution key points on other pollution boundary lines, includes: The initial regional pollution distribution map is input into a preset two-dimensional coordinate system, wherein the boundary line of one map of the initial regional pollution distribution map coincides with the horizontal axis of the two-dimensional coordinate system, the boundary line of the other map of the initial regional pollution distribution map coincides with the vertical axis of the two-dimensional coordinate system, and the boundary line of the first map is adjacent to the boundary line of the other map. Determine the center coordinates of the initial center point of the initial region, calculate the initial distance from the initial center point to each boundary point on the initial pollution boundary line based on the center coordinates, and take the boundary point corresponding to the farthest initial distance as the initial pollution key point; Input a pollution distribution map of a certain area into a preset two-dimensional coordinate system, calculate a certain distance from the initial center point to each boundary point on a certain pollution boundary line, and take the boundary point corresponding to the farthest distance as a certain pollution key point, wherein the certain pollution boundary line is the pollution boundary line in the pollution distribution map of the certain area.

4. The method for tracing soil pollution sources based on remote sensing images according to claim 1, characterized in that, The step of locating candidate pollution source areas based on the at least one suspected soil pollution path using a preset regional positioning strategy includes: From the at least one suspected soil pollution path, a first suspected soil pollution path and a second suspected soil pollution path are extracted, wherein the angle formed between the first suspected soil pollution path and the second suspected soil pollution path is the largest. Determine whether the angle formed between the first target soil pollution suspected path and the second target soil pollution suspected path is greater than a preset angle threshold; If the angle is greater than the preset threshold, the area enclosed by the first target soil pollution suspected path, the second target soil pollution suspected path, and the initial pollution boundary line is directly defined as the pollution source candidate area. If the angle is not greater than a preset threshold, the intersection of the first target soil pollution suspected path and the initial pollution boundary line is obtained and defined as the first intersection point, or the intersection of the second target soil pollution suspected path and the initial pollution boundary line is obtained and defined as the second intersection point. A first region is obtained by drawing a circle with the first distance between the first intersection point and the initial center point as the diameter and the first center point between the first intersection point and the initial center point as the center; or a second region is obtained by drawing a circle with the second distance between the second intersection point and the initial center point as the diameter and the second center point between the second intersection point and the initial center point as the center. The first region or the second region is defined as a candidate region for pollution sources.

5. The method for tracing soil pollution sources based on remote sensing images according to claim 1, characterized in that, The final location information for determining the pollution emission source within the candidate pollution source area includes: The region remote sensing data of the candidate pollution source area at different time points is obtained, and the pollution concentration value of each pixel in each region remote sensing data is extracted to obtain the pollution concentration value sequence corresponding to each pixel. Based on the various pollution concentration value sequences, the pollution accumulation index of each pixel is calculated using a preset spatiotemporal analysis model; The local maxima of the pollution accumulation index are identified based on the peak detection algorithm, and the point with the highest pollution accumulation index among all local maxima is determined as the preliminary location of the pollution emission source. Based on the topographic and hydrological data of the candidate pollution source area, the preliminary location is optimized using a preset diffusion correction model, and the optimized location is used as the final location information of the pollution emission source.

6. A soil pollution source tracing system based on remote sensing images, characterized in that, include: The acquisition module is configured to acquire remote sensing image data of the soil area to be processed at different time points. The output module is configured to input remote sensing image data of each target into a preset pollution identification model, wherein the pollution identification model outputs a regional pollution distribution map corresponding to the remote sensing image data of each target, wherein the regional pollution distribution map includes pollution boundary lines; The extraction module is configured to extract initial pollution key points on the initial pollution boundary line and other pollution key points on other pollution boundary lines according to preset key point extraction rules, and connect the initial pollution point and the other pollution key points to the initial center point of the initial area to obtain at least one soil pollution suspected path, wherein the initial pollution boundary line is the pollution boundary line in the pollution distribution map of the initial area, and the other pollution boundary lines are the pollution boundary lines in the pollution distribution maps of other areas; The determination module is configured to locate the candidate pollution source area based on the at least one suspected soil pollution path, using a preset regional positioning strategy, and determine the final location information of the pollution emission source within the candidate pollution source area.

7. An electronic device, characterized in that, include: At least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method described in any one of claims 1 to 5.

Images