Water quality inspection system based on full-spectrum unmanned aerial vehicle

Abstract

The invention discloses a water quality inspection system based on a full-spectrum unmanned aerial vehicle, and relates to the technical field of environmental monitoring, the system comprises a spectrum acquisition module for acquiring an original spectrum data set containing related information of pollutant types and concentrations to obtain preliminary spectral feature distribution, a data preprocessing module for preprocessing the initial spectral feature distribution according to the preliminary spectral feature distribution, and a water quality inspection module for inspecting water quality according to the data preprocessing module. The system comprises an original spectrum data set, a feature extraction module for denoising the original spectrum data set to obtain a pure spectrum signal set, and a feature extraction module for extracting specific signal features of pollutants for the pure spectrum signal set and determining a spectrum band range corresponding to each feature to obtain a characterized spectrum signal subset; the water quality inspection system based on the full-spectrum unmanned aerial vehicle realizes real-time monitoring and analysis of water body pollution, can quickly identify the types, concentrations and diffusion conditions of pollutants, provides scientific basis and decision support for water body pollution treatment, and has important environmental protection significance.

Description

Technical Field

[0001] This invention relates to the field of environmental monitoring technology, specifically to a water quality inspection system based on a full-spectrum unmanned aerial vehicle (UAV). Background Technology

[0002] Water quality monitoring, as a crucial component of environmental protection and public health, bears the critical mission of ensuring water resource security and preventing the spread of pollution. With accelerated industrialization and urbanization, water pollution problems are becoming increasingly severe, making the rapid and accurate identification of pollutant types and concentrations a pressing issue. Effective water quality monitoring not only concerns ecological balance but also directly impacts human quality of life and economic development.

[0003] However, current water quality testing methods have significant limitations in practical applications. These methods often rely on complex laboratory analyses or single chemical reagent detection, and generally suffer from cumbersome operation, slow response speed, and difficulty adapting to complex environments. Especially in the dynamic monitoring of large-scale water bodies, existing technologies struggle to achieve real-time and comprehensive coverage, leading to delays in the tracking and treatment of pollution sources and missed opportunities for optimal intervention. Against this backdrop, the core challenges facing the research field are becoming increasingly apparent. The primary issue lies in how to capture the unique signal characteristics of pollutants in water bodies under specific conditions. These characteristics are often hidden in complex environmental backgrounds and are difficult to effectively separate and identify using traditional methods. Furthermore, the accurate acquisition of these signal characteristics directly affects the rapid determination of pollutant types and concentrations. If multidimensional data collection and analysis cannot be completed in a short time, it is difficult to meet the real-time requirements of large-scale water body inspections. These two technical factors are closely related and together constitute the bottleneck for breakthroughs in water quality monitoring technology. Summary of the Invention

[0004] The purpose of this invention is to provide a water quality inspection system based on a full-spectrum unmanned aerial vehicle (UAV), which, based on the unique resonance characteristics of spectral signals, constructs a technical system capable of rapidly identifying the types and concentrations of pollutants in complex environments.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a water quality inspection system based on a full-spectrum unmanned aerial vehicle (UAV), the system comprising: The spectral acquisition module acquires raw spectral datasets containing information on pollutant types and concentrations, and obtains preliminary spectral feature distributions. The data preprocessing module performs noise reduction on the original spectral dataset based on the preliminary spectral feature distribution to obtain a clean spectral signal set. The feature extraction module extracts the unique signal features of pollutants from the pure spectral signal set, determines the spectral band range corresponding to each feature, and obtains a characteristic subset of spectral signals. The classification and judgment module performs preliminary classification and judgment of pollutant types through a featured subset of spectral signals. If the classification confidence is higher than a preset threshold, the corresponding pollutant type identifier is output. The concentration analysis module performs multidimensional analysis on a characteristic subset of spectral signals based on pollutant type identification to obtain the predicted range of pollutant concentrations. The anomaly detection module dynamically updates a subset of the spectral signal for the predicted value range. If the predicted value range exceeds the safe range, an anomaly signal is triggered, and the marked anomaly data group is obtained. The distribution map generation module generates a dynamic distribution map of water quality monitoring based on the marked abnormal data groups, thereby determining the potential location and range of water pollution sources. The diffusion trend analysis module obtains preliminary judgment results on the diffusion trend of pollution sources based on the potential location range and outputs the distribution characteristics of the diffusion trend; The governance decision-making module acquires real-time data on the dynamic changes of pollution sources based on the distribution characteristics of the diffusion trend, and determines the priority areas for pollution control.

[0006] Preferably, the spectral acquisition module acquires a raw spectral dataset containing information related to pollutant types and concentrations to obtain a preliminary spectral feature distribution. This includes acquiring initial spectral data from a pre-established spectral signal database, scanning the water sample with a spectrometer to obtain a raw spectral dataset containing information related to pollutant types and concentrations, and obtaining a preliminary spectral feature distribution map. Based on the preliminary spectral feature distribution map, a data preprocessing tool is used to denoise and perform baseline correction on the raw spectral dataset. Key band features are extracted from the processed spectral data to determine the positions of spectral absorption peaks related to pollutant types. If the key band features do not match a preset threshold range, a spectral comparison tool is used to match the processed spectral data with standard spectral data in the database one by one to determine the specific classification and possible concentration range of the pollutant. For the matched pollutant types and concentration ranges, a regression analysis tool is used to quantify the intensity data of the spectral absorption peaks to obtain the precise value of the pollutant concentration, thus obtaining the final pollutant distribution result of the water sample.

[0007] Preferably, the data preprocessing module performs denoising processing on the original spectral dataset based on the preliminary spectral feature distribution to obtain a clean spectral signal set. This includes using a data preprocessing tool to perform preliminary noise filtering on the original spectral dataset based on the initially obtained spectral feature distribution data, and performing basic corrections for environmental noise and background interference to obtain a set of spectral signals that have undergone preliminary denoising. For the set of spectral signals that have undergone preliminary denoising, a band separation tool is used to perform feature decomposition on the data, extracting the main spectral features from the data while removing irrelevant interference to determine the separated feature band dataset. If residual noise is still detected in the separated feature band dataset, a signal enhancement tool is used to perform secondary correction on the feature band dataset, adjusting the waveform in conjunction with a preset threshold range to obtain a spectral signal group with higher purity. Finally, an interference removal tool is used to perform a final comparison of the spectral signal group with higher purity, and a comparative analysis method is used to determine whether the signal group meets the preset purity standard to obtain the final clean spectral signal set.

[0008] Preferably, the feature extraction module, for a pure spectral signal set, extracts signal features unique to pollutants, determines the spectral band range corresponding to each feature, and obtains a characteristic spectral signal subset. This includes processing the pure spectral signal set, employing the following process for feature extraction: Based on the pure spectral signal set, a signal decomposition tool is used to perform layered processing on the data, separating waveform segments related to pollutant features to obtain a preliminary decomposed signal dataset; for the preliminary decomposed signal dataset, a band segmentation tool is used to define the band range; if the signal purity within the band range is lower than a preset threshold, a waveform adjustment tool is used for optimization to obtain an adjusted band signal group; using the adjusted band signal group, a feature matching tool is used to compare and analyze the band signals, extracting parts that match the resonance characteristics from the data to determine a characteristic signal segment set; based on the characteristic signal segment set, a data integration tool is used to classify and organize the segments, ranking them according to their correlation with pollutant features to obtain the final characteristic spectral signal subset.

[0009] Preferably, the classification and judgment module performs preliminary classification and judgment of pollutant types using a featured subset of spectral signals. If the classification confidence level is higher than a preset threshold, it outputs the corresponding pollutant type identifier. This includes comparing the spectral features in the featured subset of signals one by one using a data comparison tool to extract feature matching information related to the pollutant class, thus obtaining a preliminary feature comparison dataset. For the preliminary feature comparison dataset, a signal processing tool is used to standardize the signal intensity of the features. If the standardized signal intensity is higher than a preset threshold, it is classified as a valid signal segment, determining a valid feature signal combination. Using the valid feature signal combination, a classification tool is used to classify the signal segment, and a classification judgment is performed based on the confidence level. If the confidence level is higher than a preset threshold, the corresponding pollutant class judgment basis is obtained. Based on the pollutant class judgment basis, a category output tool is used to associate and map the classification result with the type identifier, generating the final category output information and determining the type identifier data corresponding to the pollutant class.

[0010] Preferably, the concentration analysis module performs multidimensional analysis on a characteristic subset of spectral signals based on the pollutant type identifier to obtain the predicted value range of pollutant concentration. This includes matching features within the spectral signal subset one by one using a data comparison tool to obtain feature information related to the pollutant type, resulting in a preliminary feature comparison set; standardizing the feature information using a signal processing tool; classifying the standardized signal value as a valid signal segment if it exceeds a preset threshold, thus determining the valid feature combination for concentration prediction; using the valid feature combination, calling a pre-established concentration prediction tool and combining it with multidimensional analysis methods to perform in-depth processing on the valid signal segment to obtain prediction range data related to pollutant concentration; and using a result mapping tool to associate and bind the prediction range with the pollutant type identifier to generate the final concentration prediction output information, thus determining the concentration range corresponding to the pollutant type.

[0011] Preferably, the anomaly detection module dynamically updates the spectral signal subset for the predicted value range. If the predicted value range exceeds the safe range, anomaly signal marking is triggered, resulting in a marked anomalous data group. This includes comparing the predicted range with a preset safe range using a range judgment tool; if the predicted range exceeds the safe range, an anomaly marking process is triggered, resulting in a marked preliminary anomalous data fragment. For the preliminary anomalous data fragment, the spectral signal subset is dynamically updated using a signal adjustment tool, and the updated signal subset content is determined by combining the latest signal features obtained from the real-time demand module. Based on the updated signal subset, the signal features are classified and organized using a data grouping tool. If the classified signal features meet the anomalous data standard, they are included in the anomalous data group, resulting in a classified anomalous data set. The classified anomalous data set is then used to perform a secondary verification of the anomalous data group using a marking processing tool, and the accuracy of the final marked anomalous data group is determined by combining the real-time feedback from the demand module, resulting in the final processing result.

[0012] Preferably, the distribution map generation module generates a dynamic distribution map of water quality monitoring based on the marked abnormal data groups. Determining the potential location range of water pollution sources includes: classifying the marked abnormal data groups using a data integration tool; extracting water quality monitoring-related feature information from the data groups to obtain a classified pollution feature set; identifying key data points in the set corresponding to the water body range; generating a dynamic distribution map using a graphics drawing tool based on the pollution feature set; mapping and comparing the key data points in the map with the monitoring area; marking data points exceeding a preset threshold range as high-risk areas to obtain an annotated distribution map; continuously updating the abnormal data groups of the high-risk areas using a data acquisition interface based on the annotated distribution map; obtaining updated pollution feature data by combining real-time information within the water body range; determining whether the data meets the conditions for calculating potential locations; and using a location estimation tool to perform multi-dimensional comparisons of the high-risk areas based on the updated pollution feature data, extracting spatial information related to pollution source location from the data, and determining the potential location range of water pollution sources.

[0013] Preferably, the diffusion trend analysis module, based on the potential location range, obtains a preliminary judgment result of the pollution source diffusion trend and outputs the distribution characteristics of the diffusion trend. This includes: using a data integration tool to initially screen the spatial information related to the pollution source location based on the abnormal data; extracting key indicators related to the diffusion trend from the potential range to obtain a pre-organized data set; using a data comparison tool to perform multi-dimensional comparison on the pre-organized data set; combining the region mapping and preset threshold rules; if the indicators of a certain region exceed the threshold range, it is marked as a high-risk diffusion region, and the distribution characteristics of the high-risk region are determined; using the distribution characteristics of the high-risk region, a data update interface is used to obtain new information related to the data depth; spatial data matching the diffusion trend is extracted from the new information to determine the preliminary direction of the diffusion trend; based on the preliminary direction, a graphics drawing tool is used to generate a visualization map corresponding to the distribution characteristics; and a secondary comparison of the spatial data of the high-risk region in the map is performed to obtain the final distribution characteristics of the pollution source diffusion trend.

[0014] Preferably, the governance decision-making module, based on the distribution characteristics of the diffusion trend, acquires real-time data on the dynamic changes of pollution sources and determines the final priority areas for pollution control. This includes, based on the distribution characteristics of the diffusion trend, using a spectral signal acquisition tool to extract signal data related to the pollution source from a pre-established database; performing preliminary screening on the signal data to obtain a set of spectral signals matching the dynamic changes, thus obtaining a basic information set for subsequent comparison; and using a data comparison tool to continuously track and analyze the basic information set, extracting real-time data related to the distribution characteristics from the spectral signal set. If the fluctuation of the real-time data exceeds a preset threshold... The range is then marked as an abnormal signal area, and the preliminary distribution range of the abnormal signal area is determined. Based on the preliminary distribution range of the abnormal signal area, the latest dynamic change information related to the pollution source is obtained using a data update interface. The latest information is matched and compared with the regional division to determine whether it is consistent with the priority goal of pollution control, and the distribution information of high-priority pollution control areas is obtained. Based on the distribution information of the high-priority pollution control areas, a visual distribution map corresponding to the diffusion trend is generated using a graphics drawing tool. The abnormal areas in the distribution map are subjected to signal matching verification to obtain the final regional division data related to the priority of pollution control, and the priority ranking basis for pollution control is determined.

[0015] As can be seen from the above technical solution, the present invention has the following beneficial effects: This water quality monitoring system based on a full-spectrum UAV collects raw spectral data of water samples from a pre-established spectral signal database and performs noise reduction processing in complex environments to obtain clean spectral signals. Resonance characteristics are analyzed to extract pollutant features, and a support vector machine classification model is used to determine pollutant types and predict concentrations. Abnormal data is marked and dynamically updated to generate a water quality monitoring distribution map, identifying pollution source locations and diffusion trends. This invention enables real-time monitoring and analysis of water pollution, rapidly identifying pollutant types, concentrations, and diffusion patterns, providing a scientific basis and decision support for water pollution control, and has significant environmental protection implications. Attached Figure Description

[0016] Figure 1 This is a system connection diagram of the present invention. Detailed Implementation

[0017] 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, and 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.

[0018] like Figure 1 As shown, the present invention provides a technical solution: a water quality inspection system based on a full-spectrum unmanned aerial vehicle (UAV), the system comprising: The spectral acquisition module acquires raw spectral datasets containing information on pollutant types and concentrations, and obtains preliminary spectral feature distributions. The data preprocessing module performs noise reduction on the original spectral dataset based on the preliminary spectral feature distribution to obtain a clean spectral signal set. The feature extraction module extracts the unique signal features of pollutants from the pure spectral signal set, determines the spectral band range corresponding to each feature, and obtains a characteristic subset of spectral signals. The classification and judgment module performs preliminary classification and judgment of pollutant types through a featured subset of spectral signals. If the classification confidence is higher than a preset threshold, the corresponding pollutant type identifier is output. The concentration analysis module performs multidimensional analysis on a characteristic subset of spectral signals based on pollutant type identification to obtain the predicted range of pollutant concentrations. The anomaly detection module dynamically updates a subset of the spectral signal for the predicted value range. If the predicted value range exceeds the safe range, an anomaly signal is triggered, and the marked anomaly data group is obtained. The distribution map generation module generates a dynamic distribution map of water quality monitoring based on the marked abnormal data groups, thereby determining the potential location and range of water pollution sources. The diffusion trend analysis module obtains preliminary judgment results on the diffusion trend of pollution sources based on the potential location range and outputs the distribution characteristics of the diffusion trend; The governance decision-making module acquires real-time data on the dynamic changes of pollution sources based on the distribution characteristics of the diffusion trend, and determines the priority areas for pollution control.

[0019] This system is based on a multi-module integrated design. It uses sensors mounted on a full-spectrum UAV to collect high-resolution spectral data covering the visible to near-infrared bands, forming a raw spectral dataset. The spectral acquisition module automatically acquires data along a preset flight path and uploads it to the data processing terminal in real time. The data preprocessing module uses wavelet transform and high-pass filtering algorithms to remove system noise and background interference from the raw data, retaining effective spectral response information. The feature extraction module uses principal component analysis (PCA) and an improved spectral reflectance algorithm to identify key pollutant characteristic intervals and extract stable feature vectors. The classification module classifies the feature vectors based on a support vector machine (SVM) model and uses a confidence calculation mechanism to filter valid classification results. The concentration analysis module combines multivariate linear regression and spectral intensity mapping to output dynamic predicted values ​​of pollutant concentrations. The anomaly detection module sets safety thresholds according to national water quality safety standards and, when anomalies in predicted values ​​are detected, marks data groups using a dynamic masking algorithm for subsequent analysis. The distribution map generation module uses spatial interpolation and heatmap generation methods, combined with GPS positioning information, to create a spatial distribution map of pollutants. The diffusion trend analysis module further analyzes the diffusion trajectory of pollution sources in water bodies and uses a particle diffusion model for trend prediction. The governance decision-making module integrates the output results of each module, combines historical data and expert decision-making rules, and determines the priority response areas for pollution control.

[0020] Compared to traditional water quality monitoring methods, this system has the following advantages: First, it achieves a high degree of automation and real-time performance in pollutant identification and concentration estimation, effectively improving water quality monitoring efficiency; second, through modular design and multi-dimensional data fusion, it significantly enhances the system's adaptability and stability; third, combined with spatial positioning and graphical visualization functions, it facilitates environmental protection departments to quickly grasp pollution dynamics and make precise interventions; in addition, the system supports flexible deployment in different water environments, possessing strong promotional value and environmental adaptability.

[0021] The spectral acquisition module acquires a raw spectral dataset containing information on pollutant types and concentrations to obtain a preliminary spectral feature distribution. This includes initial spectral data acquisition of water samples using a pre-established spectral signal database, scanning the water samples with a spectrometer to obtain a raw spectral dataset containing information on pollutant types and concentrations, resulting in a preliminary spectral feature distribution map. Based on the preliminary spectral feature distribution map, data preprocessing tools are used to denoise and perform baseline correction on the raw spectral dataset. Key band features are extracted from the processed spectral data to determine the positions of spectral absorption peaks related to pollutant types. If the key band features do not match a preset threshold range, a spectral comparison tool is used to match the processed spectral data with standard spectral data in the database one by one to determine the specific classification and possible concentration range of the pollutant. For the matched pollutant types and concentration ranges, regression analysis tools are used to quantify the intensity data of the spectral absorption peaks to obtain precise values ​​of the pollutant concentrations, resulting in the final pollutant distribution results for the water samples.

[0022] In this implementation, the spectral acquisition module uses a drone as a carrier, equipped with a portable full-spectrum spectrometer, to perform high-frequency scanning of different water areas according to a preset flight path. The acquired raw spectral data is transmitted back in real time via a wireless link and compared with preset standard spectra in a spectral signal database. This database contains standard spectral response curves of various typical pollutants under different environmental conditions, supporting automatic data matching. The data preprocessing tool uses the Savitzky-Golay smoothing filter algorithm and baseline drift correction model to process the raw data, thereby eliminating non-target interference signals and highlighting the true spectral characteristics of the pollutants. After extracting key band features, the system initially screens for possible pollutant types by setting the wavelength and intensity range of the pollutant's characteristic absorption peaks. If the key feature values ​​deviate from the standard threshold, the spectral comparison tool is called to compare all database records, using a cosine similarity algorithm and cross-correlation function to improve the comparison accuracy, and finally outputs the pollutant type identification and concentration range. Furthermore, the system uses multiple linear regression and ridge regression models to quantitatively analyze the absorption peak intensity values, calculate the pollutant concentration values, and generate a final pollutant spatial distribution map for water pollution level assessment and management response basis.

[0023] The data preprocessing module performs denoising on the original spectral dataset based on the preliminary spectral feature distribution to obtain a clean spectral signal set. This includes using a data preprocessing tool to perform preliminary noise filtering on the original spectral dataset based on the initially obtained spectral feature distribution data, and performing basic corrections for environmental noise and background interference to obtain a set of spectral signals that have undergone preliminary denoising. For the set of spectral signals that have undergone preliminary denoising, a band separation tool is used to perform feature decomposition on the data, extracting the main spectral features from the data while removing irrelevant interference to determine the separated feature band dataset. If residual noise is still detected in the separated feature band dataset, a signal enhancement tool is used to perform secondary correction on the feature band dataset, adjusting the waveform in conjunction with a preset threshold range to obtain a spectral signal group with higher purity. Finally, an interference removal tool is used to perform a final comparison of the spectral signal group with higher purity, and a comparative analysis method is used to determine whether the signal group meets the preset purity standard to obtain the final clean spectral signal set.

[0024] In this implementation, the data preprocessing module, serving as the system's data quality assurance link, first determines the type of noise source in the data based on the preliminary feature distribution map transmitted by the spectral acquisition module, including sensor thermal noise, atmospheric scattering effects, and water reflection interference. A wavelet threshold filtering algorithm is used to perform preliminary noise filtering, combined with a linear baseline correction method to compensate for background drift, thereby obtaining a pre-denoised spectral signal set. Subsequently, band separation tools (such as bandpass filters or principal component dimension degradation algorithms) are used to decompose the data into spectral bands, extracting key feature spectral bands related to pollutants and removing other redundant or repetitive signals. If system interference or external residual noise still exists in the extracted bands, the system further performs secondary correction using signal enhancement tools (such as a multi-scale normalization and Gaussian smoothing joint algorithm), adjusting the thresholds of waveform regions that do not conform to morphological standards to enhance the stability and recognizability of the target signal. Finally, an interference removal tool based on a hybrid model of Mahalanobis distance and Euclidean distance is used to match and analyze the processed signal with a standard clean signal template to determine whether its purity meets the set quality standards. After confirmation, a set of clean spectral signals that can be used for subsequent analysis is output.

[0025] The feature extraction module extracts unique signal features of pollutants from a pure spectral signal set, determines the spectral band range corresponding to each feature, and obtains a characteristic spectral signal subset. This includes processing the pure spectral signal set using the following process: Based on the pure spectral signal set, a signal decomposition tool is used to perform hierarchical processing of the data, separating waveform segments related to pollutant features to obtain a preliminary decomposed signal dataset; for the preliminary decomposed signal dataset, a band segmentation tool is used to define the band range. If the signal purity within the band range is lower than a preset threshold, it is optimized using a waveform adjustment tool to obtain an adjusted band signal group; using the adjusted band signal group, a feature matching tool is used to compare and analyze the band signals, extracting the parts that match the resonance characteristics from the data to determine a characteristic signal segment set; based on the characteristic signal segment set, a data integration tool is used to classify and organize the segments, ranking them according to their correlation with pollutant features to obtain the final characteristic spectral signal subset.

[0026] In this embodiment, the feature extraction module, based on a multi-stage signal processing flow, effectively identifies and filters key bands in the spectrum that possess pollutant characteristics. First, signal decomposition tools such as wavelet decomposition or empirical mode decomposition (EMD) are used to perform multi-scale hierarchical processing on the pure spectral signal, separating waveform segments that correspond to the spectral characteristics of pollutants, removing high-frequency noise and low-frequency baseline drift, and generating a preliminary signal dataset. Subsequently, band segmentation tools (such as bandpass filter banks or segmentation algorithms based on spectral density estimation) are used to define the range of each candidate band. If the signal purity of a certain band does not meet the set threshold standard, a waveform adjustment tool combining polynomial fitting and curve smoothing is used to optimize it, enhancing the stability and discriminative power of its spectral peak shape.

[0027] Subsequently, for the adjusted and optimized band signal group, a feature matching tool was used to compare it with standard pollutant spectral templates. Resonance frequency matching and cross-spectral analysis techniques were employed to screen signal segments that highly matched the known characteristics of pollutants, forming a set of characteristic signal segments. Finally, data integration tools (such as K-means clustering, correlation scoring algorithms, etc.) were used to classify and organize the characteristic segments, and they were scored and ranked according to their consistency with the spectral response of pollutants. The most representative and identifiable set of characteristic bands was output first, constituting the final subset of characteristic spectral signals.

[0028] The classification and judgment module performs preliminary classification and judgment of pollutant types using a featured subset of spectral signals. If the classification confidence level is higher than a preset threshold, it outputs the corresponding pollutant type identifier. This includes comparing the spectral features in the featured subset of signals one by one using a data comparison tool to extract feature matching information related to the pollutant class, thus obtaining a preliminary feature comparison dataset. For the preliminary feature comparison dataset, a signal processing tool is used to standardize the signal intensity of the features. If the standardized signal intensity is higher than a preset threshold, it is classified as a valid signal segment, determining a valid feature signal combination. Using the valid feature signal combination, a classification tool is used to classify the signal segment, and a classification judgment is made based on the confidence level. If the confidence level is higher than a preset threshold, the corresponding pollutant class judgment basis is obtained. Based on the pollutant class judgment basis, a category output tool is used to associate and map the classification result with the type identifier, generating the final category output information and determining the type identifier data corresponding to the pollutant class.

[0029] In this embodiment, the classification and judgment module operates collaboratively through multiple stages, including hierarchical signal matching, feature intensity evaluation, classification modeling, and confidence analysis, to improve the accuracy and stability of pollutant type identification. First, based on the characteristic spectral signal subset output by the preprocessing module, a data comparison tool is used to accurately compare each band signal with the pollutant feature database. Algorithms such as cosine similarity and multi-scale correlation analysis are employed to extract key feature matching indicators, forming a preliminary feature comparison dataset. Subsequently, a signal processing tool normalizes, removes bias, and standardizes the signal intensity of the spectral features extracted from the comparison dataset. If the standardized intensity exceeds a system-set threshold, it is marked as a valid signal segment and combined to form a high-confidence feature signal set.

[0030] Next, the system uses classification tools (such as Support Vector Machine (SVM), Random Forest (RF), or Convolutional Neural Network (CNN)) to classify the effective signal set and judges the classification validity based on the confidence score mechanism of the model output. If the confidence score of a certain category exceeds a set threshold, the system outputs the pollutant identification criteria for that category. Finally, the system uses a category output tool to map the classification results to pollutant type numbers or names, generating formatted category output data for subsequent concentration analysis, anomaly detection, and visualization.

[0031] The concentration analysis module performs multidimensional analysis on a characteristic subset of spectral signals based on pollutant type identification to obtain the predicted concentration range of pollutants. This includes matching features within the spectral signal subset one by one using a data comparison tool to obtain feature information related to the pollutant type, resulting in a preliminary feature comparison set. For this preliminary feature comparison set, a signal processing tool is used to standardize the feature information. If the standardized signal value is higher than a preset threshold, it is classified as a valid signal segment, determining the valid feature combination for concentration prediction. Using the valid feature combination, a pre-established concentration prediction tool is invoked, and multidimensional analysis methods are combined to perform in-depth processing on the valid signal segment to obtain prediction range data related to pollutant concentration. Based on the prediction range data, a result mapping tool is used to associate and bind the prediction range with the pollutant type identification, generating the final concentration prediction output information and determining the concentration range corresponding to the pollutant type.

[0032] In this implementation, the concentration analysis module plays a crucial role in the further processing after pollutant identification, ensuring a smooth transition from classification results to quantitative analysis. First, based on a subset of characteristic spectral signals and pollutant type identifiers, the system retrieves standard absorption feature templates for the corresponding pollutants from the database. A data comparison tool performs line-by-line matching, filtering for matching features based on peak position, intensity, and morphology, forming a preliminary feature comparison set. Subsequently, a signal processing tool normalizes and filters outliers from the feature signals in the set. If the intensity value of the standardized signal exceeds a set threshold, it is considered a valid signal segment with concentration prediction value, constructing a feature combination for analysis. The system then calls a trained concentration prediction tool, which can be a multiple regression model, a random forest regressor, or a deep neural network. By modeling the valid feature combinations and performing multidimensional feature cross-analysis, it outputs the predicted range for pollutant concentration. The multidimensional analysis methods may include spectral peak area integration, multi-channel feature coupling analysis, and joint modeling with environmental parameters (such as water temperature and pH), improving the accuracy and robustness of concentration prediction. Finally, the results mapping tool binds the concentration prediction range with the corresponding pollutant type identifier, outputting structured prediction results for subsequent anomaly detection, visualization, and decision support.

[0033] The anomaly detection module dynamically updates a subset of spectral signals for a predicted value range. If the predicted value range exceeds the safe range, anomaly signal marking is triggered, resulting in a marked anomalous data group. This includes comparing the predicted range with a preset safe range using a range judgment tool; if the predicted range exceeds the safe range, an anomaly marking process is triggered, resulting in a marked preliminary anomalous data fragment. For this preliminary anomalous data fragment, a signal adjustment tool dynamically updates the subset of spectral signals, combining the latest signal features obtained from the real-time demand module to determine the updated signal subset content. Based on the updated signal subset, a data grouping tool classifies and organizes the signal features. If the classified signal features meet the anomalous data criteria, they are assigned to the anomalous data group, resulting in a classified anomalous data set. Using the classified anomalous data set, a marking processing tool performs a secondary verification on the anomalous data group, combining real-time feedback from the demand module to determine the accuracy of the final marked anomalous data group, thus obtaining the final processing result.

[0034] In this embodiment, the anomaly detection module is used to identify abnormal trends in the concentration of pollutants in the water, ensuring that changes in water quality can be promptly reported and traced back to their source. The system first calls the interval judgment tool to compare the predicted pollutant concentration interval output by the concentration analysis module with national or user-defined safety concentration standards. If the predicted value exceeds this range, it is considered a potential risk point, and the anomaly signal marking process is initiated to mark preliminary abnormal data segments. Next, based on the preliminary anomaly results, the system applies a signal adjustment tool to dynamically process the corresponding spectral signal subset, including updating the filtering strategy, resampling the signal, and implementing real-time compensation algorithms to ensure that the signal set reflects the latest state. Simultaneously, referencing the latest task instructions, dynamic safety thresholds, and environmental information in the real-time requirements module, the system optimizes the signal update strategy to obtain an updated signal subset.

[0035] Subsequently, the data grouping tool reclassifies the signal features based on band characteristics, pollutant correspondences, and timestamp information, forming preliminary signal groups. If the signal features in a group meet the set anomaly identification rules (such as abnormal peak shifts, abrupt intensity changes, nonlinear growth, etc.), the group is identified as anomalous data and added to the anomalous data group, generating a classified set of anomalous data. Finally, the system calls the labeling processing tool to perform secondary verification on this set, verifying the accuracy of the final labeling through comparison with historical data, anomaly fluctuation backtracking analysis, and feedback from the real-time monitoring module, outputting a reliable set of anomalous labeled data for subsequent pollution diffusion analysis and governance decisions.

[0036] The distribution map generation module generates a dynamic distribution map of water quality monitoring based on marked abnormal data groups. This determines the potential location range of water pollution sources. The process includes: classifying the marked abnormal data groups using a data integration tool; extracting water quality monitoring-related feature information from the data groups to obtain a classified pollution feature set; identifying key data points in the set corresponding to the water body range; generating a dynamic distribution map using a graphics drawing tool based on the pollution feature set; mapping and comparing the key data points in the map with the monitoring area; marking data points exceeding a preset threshold range as high-risk areas to obtain an annotated distribution map; continuously updating the abnormal data groups in the high-risk areas using a data acquisition interface based on the annotated distribution map; obtaining updated pollution feature data by combining real-time information within the water body range; determining whether the data meets the conditions for calculating potential locations; and using a location estimation tool to perform multi-dimensional comparisons of the high-risk areas based on the updated pollution feature data, extracting spatial information related to pollution source location from the data, and determining the potential location range of water pollution sources.

[0037] In this embodiment, the distribution map generation module, based on spectral recognition and spatial visualization technologies, identifies, locates, and displays the distribution of areas with abnormal water quality. First, the system receives labeled data sets from the anomaly detection module and uses data integration tools (such as attribute-based clustering or pollutant type indexing) to automatically classify the abnormal data, identifying signal indicators highly correlated with water pollution characteristics, such as specific absorption peak intensities and spectral band variation amplitudes. Combining UAV-borne GPS positioning information and flight path planning data, spatial coordinate points are extracted from the data to determine key pollution data points within the corresponding water body area, forming a pollution feature set.

[0038] The system then uses graphics drawing tools (such as GIS rendering engines and visualization frameworks) to generate heat maps or isosurface maps from the pollution feature set. It performs spatial matching and concentration threshold comparison between key points and preset monitoring areas. If pollution features in certain areas exceed critical values, they are marked as "high-risk areas" on the map, forming a labeled pollution distribution map. This map supports dynamic layer updates and can be refreshed in real time with new data input.

[0039] Furthermore, the system continuously samples data from high-risk areas through data acquisition interfaces (such as interacting with edge devices on drones or accessing ground sensor networks), collecting trends in the temporal and spatial dimensions of pollution characteristics. The system determines whether these changes exhibit pollution source aggregation characteristics (such as concentration gradient focusing, centripetal diffusion of peaks, etc.). If so, location estimation tools (such as triangulation and trajectory fitting analysis) jointly compare the pollution intensity gradient, geographical boundary information, and water flow vectors of key areas to estimate the possible spatial range of the pollution source, representing it as a polygonal region, coordinate set, etc., for subsequent diffusion analysis and governance decisions.

[0040] The diffusion trend analysis module, based on the potential location range, obtains a preliminary judgment result of the pollution source diffusion trend and outputs the distribution characteristics of the diffusion trend. This includes: using data integration tools to initially screen the spatial information related to the pollution source location based on the abnormal data; extracting key indicators related to the diffusion trend from the potential range to obtain a pre-organized data set; using data comparison tools to perform multi-dimensional comparisons on the pre-organized data set; combining the region mapping and preset threshold rules; if the indicators of a certain region exceed the threshold range, it is marked as a high-risk diffusion region, and the distribution characteristics of the high-risk region are determined; using the distribution characteristics of the high-risk region, a data update interface is used to obtain new information related to data depth; from the new information, spatial data matching the diffusion trend is extracted to determine the preliminary direction of the diffusion trend; based on the preliminary direction, a graphics drawing tool is used to generate a visualization map corresponding to the distribution characteristics; and a secondary comparison of the spatial data of the high-risk region in the map is performed to obtain the final distribution characteristics of the pollution source diffusion trend.

[0041] In this embodiment, the diffusion trend analysis module constructs a preliminary prediction model of the pollution source diffusion trend based on high-precision spatial information and pollution dynamic data. First, the system receives the potential pollution source locations and related anomaly data provided by the distribution map generation module. It then calls the data integration tool to perform preliminary screening and organization of information related to the spatial location of pollution sources, such as water flow velocity, direction, terrain slope, and historical pollution trajectories. This extracts indicators highly correlated with the diffusion trend analysis, such as the pollution concentration gradient change rate and spectral diffusion slope, forming a structured preliminary data set.

[0042] Subsequently, the data comparison tool performs multi-dimensional spatial and temporal comparative analysis based on the correlation between historical diffusion templates and current data. Combined with a regional mapping model and set thresholds (such as concentration increase rate and abrupt changes in water flow vector direction), it identifies and judges potentially polluted areas. If an indicator in a certain area exceeds the limit, it is marked as a high-risk diffusion area. The system further identifies the distribution characteristics of these areas based on their spatial distribution density, expansion path, and dynamic change trends, constructing a distribution pattern of high-risk areas.

[0043] Based on the identified high-risk areas, the system continuously calls the latest monitoring data (such as continuous drone flight data, data transmitted from shore-based sensors, etc.) through a data update interface to supplement the spatial and temporal data dimensions, and extracts spatial vector features that conform to the diffusion trend and the dynamic evolution behavior of pollution sources. Based on this in-depth information, the system determines the initial direction of pollutant diffusion and uses graphics drawing tools to create visualization maps such as trend maps and contour maps, intuitively displaying the propagation path and affected area of ​​pollutants.

[0044] Finally, the system performs a secondary spatial comparison operation on high-risk areas in the map, including multi-source data cross-validation, trend curve fitting, and diffusion rate prediction, to obtain the final distribution characteristics of pollution source diffusion trends, providing spatial judgment basis and intervention path suggestions for subsequent decision-making in the governance module.

[0045] The governance decision-making module, based on the distribution characteristics of the diffusion trend, acquires real-time data on the dynamic changes of pollution sources and determines the final priority areas for pollution control. This includes extracting signal data related to the pollution sources from a pre-established database using spectral signal acquisition tools, performing preliminary screening of the signal data to obtain a set of spectral signals matching the dynamic changes, and obtaining a basic information set for subsequent comparison. For this basic information set, a data comparison tool is used for continuous tracking and analysis, extracting real-time data related to the distribution characteristics from the spectral signal set. If the fluctuation of the real-time data exceeds a preset threshold range, it is marked as an abnormal signal area, and the preliminary distribution range of the abnormal signal area is determined. Based on the preliminary distribution range of the abnormal signal area, a data update interface is used to acquire the latest dynamic change information related to the pollution sources. This latest information is matched and compared with the regional division to determine whether it is consistent with the governance priority target, obtaining the distribution information of high-priority governance areas. Based on the distribution information of the high-priority governance areas, a graphical drawing tool is used to generate a visual distribution map corresponding to the diffusion trend. Signal matching verification is performed on the abnormal areas in the distribution map to obtain the final regional division data related to governance priority and determine the priority ranking criteria for pollution control.

[0046] In this embodiment, after acquiring the diffusion trend distribution characteristics, the governance decision-making module dynamically identifies the evolution path of pollution sources and key areas for governance response through cross-analysis of signal data and spatial regions. The system first uses a spectral signal acquisition tool to retrieve historical pollutant spectral templates matching the current diffusion trend from the pollutant database. Combined with the current abnormal distribution areas, the data is initially screened to extract a basic spectral signal set that can be used for dynamic comparison. Next, real-time tracking is performed using a data comparison tool, focusing on monitoring dynamic indicators such as spectral intensity and morphological changes in characteristic bands. If signal fluctuations exceed a set threshold (such as daily concentration change rate or spectral shift), it is automatically marked as an abnormal signal area, generating a preliminary distribution area of ​​pollution evolution.

[0047] The system further calls the data update interface to extract the latest pollution data collected in real time from edge devices, drone platforms, and environmental monitoring systems, and matches it with the preliminary anomaly areas. If the distribution location, pollution type, and trend are consistent with the priority indicators of the governance strategy (such as proximity to drinking water sources, rapid diffusion rate, etc.), it is designated as a high-priority governance area. Finally, the governance decision module visualizes the high-priority area information spatially using a graphics drawing tool, generates a governance priority map, and performs signal matching verification on the anomaly points in the map to verify the consistency of their band intensity with the characteristics of the pollution source. The final governance ranking layer is output to guide the regional allocation and intervention order of water pollution governance work.

[0048] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A water quality inspection system based on a full-spectrum unmanned aerial vehicle (UAV), characterized in that, The system includes: The spectral acquisition module acquires raw spectral datasets containing information on pollutant types and concentrations, and obtains preliminary spectral feature distributions. The data preprocessing module performs noise reduction on the original spectral dataset based on the preliminary spectral feature distribution to obtain a clean spectral signal set. The feature extraction module extracts the unique signal features of pollutants from the pure spectral signal set, determines the spectral band range corresponding to each feature, and obtains a characteristic subset of spectral signals. The classification and judgment module performs preliminary classification and judgment of pollutant types through a featured subset of spectral signals. If the classification confidence is higher than a preset threshold, the corresponding pollutant type identifier is output. The concentration analysis module performs multidimensional analysis on a characteristic subset of spectral signals based on pollutant type identification to obtain the predicted range of pollutant concentrations. The anomaly detection module dynamically updates a subset of the spectral signal for the predicted value range. If the predicted value range exceeds the safe range, an anomaly signal is triggered, and the marked anomaly data group is obtained. The distribution map generation module generates a dynamic distribution map of water quality monitoring based on the marked abnormal data groups, thereby determining the potential location and range of water pollution sources. The diffusion trend analysis module obtains preliminary judgment results on the diffusion trend of pollution sources based on the potential location range and outputs the distribution characteristics of the diffusion trend; The governance decision-making module acquires real-time data on the dynamic changes of pollution sources based on the distribution characteristics of the diffusion trend, and determines the priority areas for pollution control.

2. The water quality inspection system based on a full-spectrum UAV according to claim 1, characterized in that: The spectral acquisition module acquires a raw spectral dataset containing information related to pollutant types and concentrations, obtaining a preliminary spectral feature distribution. This includes initial spectral data acquisition of water samples using a pre-established spectral signal database, scanning the water samples with a spectrometer to acquire a raw spectral dataset containing information related to pollutant types and concentrations, resulting in a preliminary spectral feature distribution map. Based on the preliminary spectral feature distribution map, a data preprocessing tool is used to denoise and perform baseline correction on the raw spectral dataset. Key band features are extracted from the processed spectral data to determine the positions of spectral absorption peaks related to pollutant types. If the key band features do not match a preset threshold range, a spectral comparison tool is used to match the processed spectral data with standard spectral data in the database one by one to determine the specific classification and possible concentration range of the pollutant. For the matched pollutant types and concentration ranges, a regression analysis tool is used to quantify the intensity data of the spectral absorption peaks to obtain the precise value of the pollutant concentration, resulting in the final pollutant distribution result of the water sample.

3. The water quality inspection system based on a full-spectrum UAV according to claim 1, characterized in that: The data preprocessing module performs denoising on the original spectral dataset based on the preliminary spectral feature distribution to obtain a clean spectral signal set. This includes using a data preprocessing tool to perform preliminary noise filtering on the original spectral dataset based on the initially obtained spectral feature distribution data, and performing basic corrections for environmental noise and background interference to obtain a set of spectral signals that have undergone preliminary denoising. For the set of spectral signals that have undergone preliminary denoising, a band separation tool is used to perform feature decomposition on the data, extracting the main spectral features from the data while removing irrelevant interference to determine the separated feature band dataset. If residual noise is still detected in the separated feature band dataset, a signal enhancement tool is used to perform secondary correction on the feature band dataset, adjusting the waveform in conjunction with a preset threshold range to obtain a spectral signal group with higher purity. Finally, an interference removal tool is used to perform a final comparison of the spectral signal group with higher purity, and a comparative analysis method is used to determine whether the signal group meets the preset purity standard to obtain the final clean spectral signal set.

4. The water quality inspection system based on a full-spectrum UAV according to claim 1, characterized in that: The feature extraction module, for a pure spectral signal set, extracts signal features unique to pollutants, determines the spectral band range corresponding to each feature, and obtains a characteristic spectral signal subset. This includes processing the pure spectral signal set, employing the following process for feature extraction: Based on the pure spectral signal set, a signal decomposition tool is used to perform layered processing on the data, separating waveform segments related to pollutant features to obtain a preliminary decomposed signal dataset; for the preliminary decomposed signal dataset, a band segmentation tool is used to define the band range; if the signal purity within the band range is lower than a preset threshold, a waveform adjustment tool is used for optimization to obtain an adjusted band signal group; using the adjusted band signal group, a feature matching tool is used to compare and analyze the band signals, extracting portions that match the resonance characteristics from the data to determine a characteristic signal segment set; based on the characteristic signal segment set, a data integration tool is used to classify and organize the segments, ranking them according to their correlation with pollutant features to obtain the final characteristic spectral signal subset.

5. The water quality inspection system based on a full-spectrum UAV according to claim 1, characterized in that: The classification and judgment module performs preliminary classification and judgment of pollutant types using a featured subset of spectral signals. If the classification confidence level is higher than a preset threshold, it outputs the corresponding pollutant type identifier. This includes comparing the spectral features in the featured subset of signals one by one using a data comparison tool to extract feature matching information related to the pollutant class, thus obtaining a preliminary feature comparison dataset. For the preliminary feature comparison dataset, a signal processing tool is used to standardize the signal intensity of the features. If the standardized signal intensity is higher than a preset threshold, it is classified as a valid signal segment, determining a valid feature signal combination. Using the valid feature signal combination, a classification tool is used to classify the signal segment, and a classification judgment is made based on the confidence level. If the confidence level is higher than a preset threshold, the corresponding pollutant class judgment basis is obtained. Based on the pollutant class judgment basis, a category output tool is used to associate and map the classification result with the type identifier, generating the final category output information and determining the type identifier data corresponding to the pollutant class.

6. The water quality inspection system based on a full-spectrum UAV according to claim 1, characterized in that: The concentration analysis module performs multidimensional analysis on a characteristic subset of spectral signals based on pollutant type identification to obtain the predicted range of pollutant concentration. This includes matching features within the spectral signal subset one by one using a data comparison tool to obtain feature information related to the pollutant type, resulting in a preliminary feature comparison set. For this preliminary feature comparison set, a signal processing tool is used to standardize the feature information. If the standardized signal value is higher than a preset threshold, it is classified as a valid signal segment, determining the valid feature combination for concentration prediction. Using the valid feature combination, a pre-established concentration prediction tool is invoked, and multidimensional analysis methods are combined to perform in-depth processing on the valid signal segment to obtain prediction range data related to pollutant concentration. Based on the prediction range data, a result mapping tool is used to associate and bind the prediction range with the pollutant type identification, generating the final concentration prediction output information and determining the concentration range corresponding to the pollutant type.

7. The water quality inspection system based on a full-spectrum UAV according to claim 1, characterized in that: The anomaly detection module dynamically updates a subset of spectral signals for a predicted value range. If the predicted value range exceeds the safe range, anomaly signal marking is triggered, resulting in a marked anomalous data group. This includes comparing the predicted range with a preset safe range using a range judgment tool; if the predicted range exceeds the safe range, an anomaly marking process is triggered, resulting in a marked preliminary anomalous data fragment. For this preliminary anomalous data fragment, a signal adjustment tool dynamically updates the subset of spectral signals, combining the latest signal features obtained by the real-time demand module to determine the updated signal subset content. Based on the updated signal subset, a data grouping tool classifies and organizes the signal features. If the classified signal features meet the anomalous data criteria, they are assigned to the anomalous data group, resulting in a classified anomalous data set. Using the classified anomalous data set, a marking processing tool performs a secondary verification on the anomalous data group, combining real-time feedback from the demand module to determine the accuracy of the final marked anomalous data group, thus obtaining the final processing result.

8. The water quality inspection system based on a full-spectrum UAV according to claim 1, characterized in that: The distribution map generation module generates a dynamic distribution map of water quality monitoring based on marked abnormal data groups. This process determines the potential location range of water pollution sources by: classifying the marked abnormal data groups using a data integration tool; extracting water quality monitoring-related feature information from the data groups to obtain a classified pollution feature set; identifying key data points in the set corresponding to the water body range; generating a dynamic distribution map using a graphics drawing tool based on the pollution feature set; mapping and comparing the key data points in the map with the monitoring area; marking data points exceeding a preset threshold range as high-risk areas to obtain an annotated distribution map; continuously updating the abnormal data groups in the high-risk areas using a data acquisition interface based on the annotated distribution map; obtaining updated pollution feature data by combining real-time information within the water body range; determining whether the data meets the conditions for calculating potential locations; and using a location estimation tool to perform multi-dimensional comparisons of the high-risk areas based on the updated pollution feature data, extracting spatial information related to pollution source location from the data, and determining the potential location range of water pollution sources.

9. The water quality inspection system based on a full-spectrum UAV according to claim 1, characterized in that: The diffusion trend analysis module, based on the potential location range, obtains a preliminary judgment result of the pollution source diffusion trend and outputs the distribution characteristics of the diffusion trend. This includes: using data integration tools to initially screen the spatial information related to the pollution source location based on the abnormal data; extracting key indicators related to the diffusion trend from the potential range to obtain a pre-organized data set; using data comparison tools to perform multi-dimensional comparisons on the pre-organized data set; combining the region mapping and preset threshold rules; if the indicators of a certain region exceed the threshold range, it is marked as a high-risk diffusion region, and the distribution characteristics of the high-risk region are determined; using the distribution characteristics of the high-risk region, a data update interface is used to obtain new information related to data depth; spatial data matching the diffusion trend is extracted from the new information to determine the preliminary direction of the diffusion trend; based on the preliminary direction, a graphics drawing tool is used to generate a visualization map corresponding to the distribution characteristics; and a secondary comparison of spatial data is performed on the high-risk regions in the map to obtain the final distribution characteristics of the pollution source diffusion trend.

10. The water quality inspection system based on a full-spectrum UAV according to claim 1, characterized in that: The governance decision-making module, based on the distribution characteristics of the diffusion trend, acquires real-time data on the dynamic changes of pollution sources and determines the final priority areas for pollution control. This includes extracting signal data related to the pollution sources from a pre-established database using spectral signal acquisition tools based on the distribution characteristics of the diffusion trend, performing preliminary screening on the signal data, obtaining a set of spectral signals that match the dynamic changes, and obtaining a basic information set for subsequent comparison. For the aforementioned basic information set, a data comparison tool is used for continuous tracking and analysis. Real-time data related to distribution characteristics is extracted from the spectral signal set. If the fluctuation of the real-time data exceeds a preset threshold range, it is marked as an abnormal signal region, and the preliminary distribution range of the abnormal signal region is determined. Based on the preliminary distribution range of the abnormal signal region, the latest dynamic change information related to the pollution source is obtained using a data update interface. The latest information is matched and compared with the regional division to determine whether it is consistent with the priority target of governance, and the distribution information of high-priority governance regions is obtained. Based on the distribution information of the high-priority governance regions, a visualization distribution map corresponding to the diffusion trend is generated using a graphics drawing tool. Signal matching verification is performed on the abnormal regions in the distribution map to obtain the final regional division data related to governance priority and determine the priority ranking basis for pollution governance.

Images