Methods, devices, equipment and media for identifying black and odorous water bodies

Abstract

This application relates to a method, apparatus, equipment, and medium for identifying black and odorous water bodies, belonging to the field of environmental monitoring technology. The method includes: extracting visual feature parameters from water body images; matching these visual feature parameters with visual feature parameters in a fusion-based black and odorous water body feature database; and identifying the monitored water body as a suspected pollution target area based on the matching results. It also involves collecting water gene data, spectral data, and real-time environmental parameters from water samples in the target area; determining the comprehensive similarity between the water sample and the fusion-based black and odorous water body feature database based on the water gene data, spectral data, real-time environmental parameters, and visual feature parameters; and determining the black and odorous water body level and core pollutant concentration of the water sample based on the comprehensive similarity. By integrating various monitoring data through the fusion-based black and odorous water body feature database, the method achieves the entire process from large-scale, efficient screening to precise identification to pollution source tracing, meeting the current practical needs for routine and refined supervision of black and odorous water bodies.

Description

Technical Field

[0001] This application relates to the field of environmental monitoring technology, specifically to a method, device, equipment, and medium for identifying black and odorous water bodies. Background Technology

[0002] Black and odorous water bodies are a key and challenging aspect of urban water environment management, and their accurate identification and effective monitoring are crucial for ensuring water ecological security. Currently, black and odorous water body identification technologies are developing in parallel with multiple technological approaches, which can be broadly categorized into the following three types: The first category is large-scale screening technology based on remote sensing imagery. This technology has evolved to the stage of high-resolution imagery combined with "algorithm-marker" collaborative identification. By combining algorithms such as band ratio method with identification marks such as water color and texture, it can better reflect the spatiotemporal variation characteristics of black and odorous water bodies and achieve targeted identification of small black and odorous water bodies. However, this technology is limited by atmospheric correction bias, mixed pixel effect, and interference from complex urban backgrounds (such as shadow occlusion and strong heterogeneity of water body optical properties). Its identification accuracy for small streams is limited, and it lacks real-time dynamic monitoring capabilities. It can usually only provide clue-based screening results and cannot achieve accurate determination of water bodies and pollution source tracing.

[0003] The second category is monitoring technology based on traditional water quality sensors. This technology determines the state of water bodies by detecting single physicochemical parameters such as dissolved oxygen, pH, and turbidity. However, this type of technology has inherent drawbacks such as limited detection capabilities and slow response times. In practical applications, it often heavily relies on manual sampling and subsequent laboratory analysis, which is not only time-consuming and labor-intensive but also has narrow coverage and poor timeliness, making it difficult to meet the needs of routine supervision and accurately match the diverse data requirements for pollution source tracing.

[0004] The third category is water gene spectral identification technology. Although this technology has the potential for specific identification, a standardized system adapted to the coexistence of multiple pollutants and large concentration fluctuations has not yet been established in the field of black and odorous water bodies. In addition, existing related sensors have problems such as weak anti-pollution capabilities and insufficient adaptability to complex environments, and an effective integration mechanism with visual recognition technology has not yet been formed.

[0005] In summary, the industry generally suffers from fragmented monitoring technologies. Existing single technologies are insufficient to meet the full-process requirements of large-scale, efficient screening, accurate identification, and pollution source tracing. There is a lack of a collaborative monitoring scheme that can adapt to the complex characteristics of black and odorous water bodies, making it difficult to meet the current practical needs for routine and refined supervision of black and odorous water bodies. Summary of the Invention

[0006] In view of this, this application provides a method, device, equipment and medium for identifying black and odorous water bodies. The main purpose is to solve the technical problems that the existing single technical means are difficult to meet the full-process requirements of large-scale and efficient screening, accurate identification and pollution source tracing, lack a set of collaborative monitoring schemes that can adapt to the complex characteristics of black and odorous water bodies, and are difficult to meet the current practical needs of normalized and refined supervision of black and odorous water bodies.

[0007] Firstly, this application provides a method for identifying black and odorous water bodies, including: A fusion-based black and odorous water body feature database is constructed, which includes standard visual feature parameters, standard spectral feature parameters, and standard microbial gene feature parameters of black and odorous water bodies. The system collects real-time images of the monitored water area using a camera, extracts visual feature parameters from the water area images using a pre-trained attention mechanism model, and performs feature matching between the visual feature parameters and the standard visual feature parameters in the fusion-type black and odorous water body feature database. Based on the matching results, the monitored water area is identified as a suspected pollution target area. Water samples were collected from the target area. Water gene data of the water samples were collected using a water gene sensor. Spectrometer data of the water samples were collected. Real-time environmental parameters were also collected simultaneously. The comprehensive similarity between the water sample and the fused black and odorous water body feature database is determined based on the water gene data, the spectral data, the real-time environmental parameters, and the visual feature parameters. The black and odorous water level of the water sample is determined based on the comprehensive similarity, and the concentration of the core pollutants in the water sample is quantitatively calculated by combining the mapping relationship in the fusion black and odorous water feature database.

[0008] Secondly, this application provides a black and odorous water body identification device, comprising: A construction module is used to construct a fusion-type black and odorous water body feature library, which includes standard visual feature parameters, standard spectral feature parameters, and standard microbial gene feature parameters of black and odorous water bodies. The matching module is used to collect images of the monitored water area in real time through a camera, extract visual feature parameters of the water area image using a pre-trained attention mechanism model, and perform feature matching between the visual feature parameters and the standard visual feature parameters in the fusion black and odorous water body feature database. Based on the matching result, the monitored water area is identified as a suspected pollution target area. The acquisition module is used to collect water samples from the target area, collect water gene data of the water samples using a water gene sensor, collect spectral data of the water samples using a spectrometer, and simultaneously collect real-time environmental parameters. The determination module is used to determine the comprehensive similarity between the water sample and the fused black and odorous water feature database based on the water gene data, the spectral data, the real-time environmental parameters, and the visual feature parameters. The determination module is used to determine the black and odorous water level of the water sample based on the comprehensive similarity, and to quantitatively calculate the concentration of the core pollutants in the water sample by combining the mapping relationship in the fusion black and odorous water feature database.

[0009] Thirdly, this application provides an electronic device, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor executes the computer program to implement the black and odorous water body identification method described in the first aspect.

[0010] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the black and odorous water body identification method described in the first aspect.

[0011] By employing the above technical solutions, this application provides a method, apparatus, equipment, and medium for identifying black and odorous water bodies. Compared with existing technologies, this application constructs a fusion-type black and odorous water body feature library, which includes standard visual feature parameters, standard spectral feature parameters, and standard microbial gene feature parameters of black and odorous water bodies. It acquires real-time images of the monitored water area using a camera, extracts visual feature parameters from the water image using a pre-trained attention mechanism model, and performs feature matching between the visual feature parameters and the standard visual feature parameters in the fusion-type black and odorous water body feature library. Based on the matching results, it identifies the monitored water area as a suspected pollution target area. It collects water samples from the target area, using a water gene sensor to collect water gene data and a spectrometer to collect spectral data, while simultaneously collecting real-time environmental parameters. Based on the water gene data, spectral data, real-time environmental parameters, and visual feature parameters, it determines the comprehensive similarity between the water sample and the fusion-type black and odorous water body feature library. Based on the comprehensive similarity, it determines the black and odorous water body level of the water sample and quantitatively calculates the core pollutant concentration of the water sample by combining the mapping relationship in the fusion-type black and odorous water body feature library.

[0012] By employing the above technical solution, this application does not rely solely on remote sensing imagery for screening, but rather constructs a fusion-based black and odorous water body feature database. Real-time images of the monitored water areas are acquired via cameras, and a pre-trained attention mechanism model is used to extract visual feature parameters from the water images. These parameters are then matched with standard visual feature parameters in the fusion-based black and odorous water body feature database. Based on the matching results, suspected pollution target areas are identified. This approach is not limited to remote sensing imagery, avoiding the interference problems of atmospheric correction bias, mixed pixel effects, and complex urban backgrounds (such as shadow occlusion and strong heterogeneity of water body optical properties) that affect remote sensing technology, thus improving the identification accuracy of small streams and canals.

[0013] This application uses a camera to collect water area images in real time, which can obtain visual information of the monitored water area in real time and perform feature matching to lock the target area. It has real-time dynamic monitoring capabilities, overcoming the shortcomings of remote sensing image screening technology, which can only provide clue-based screening results and cannot achieve real-time dynamic monitoring.

[0014] In addition to using visual feature parameters for preliminary screening, this application also collects water gene data, spectral data, and real-time environmental parameters from water samples in the target area. It then integrates these data to determine the comprehensive similarity between the water sample and a fusion-based black and odorous water body feature database, thereby classifying the water sample as a black and odorous water body. This multi-dimensional data analysis, compared to traditional water quality sensors that only detect single physicochemical parameters such as dissolved oxygen, pH, and turbidity, significantly expands the detection dimensions and can more comprehensively and accurately reflect the water body's condition.

[0015] The monitoring process described in this application covers real-time image acquisition, feature extraction, sample collection, and comprehensive analysis of multiple data. By optimizing the collaborative work of each step, compared with traditional water quality sensor monitoring technology that relies on manual sampling and subsequent laboratory analysis, the response speed is greatly improved, and monitoring results can be obtained more quickly to meet the needs of routine supervision.

[0016] The integrated black and odorous water body feature database constructed in this application includes standard microbial gene feature parameters. It combines water gene data, spectral data, real-time environmental parameters, and visual feature parameters to determine the comprehensive similarity, classify the black and odorous water body, and quantitatively calculate the concentration of core pollutants. This process provides a practical foundation for the application of water gene spectral identification technology in black and odorous water bodies to form a standardized system adapted to the coexistence of multiple pollutants and large concentration fluctuations, helping to solve the current problem of the lack of a standardized system for this technology in the application of black and odorous water bodies.

[0017] This application comprehensively utilizes multiple monitoring technologies and data, and reduces the impact of environmental interference on a single sensor through multi-parameter collaborative analysis. For example, when water gene sensors may have problems such as weak anti-pollution capabilities or insufficient adaptability to complex environments, combining visual feature parameters, spectral data, and real-time environmental parameters allows for the assessment of water conditions from different perspectives, improving the stability and reliability of the entire monitoring system in complex environments.

[0018] In summary, this application constructs a collaborative monitoring scheme integrating multiple technologies such as visual recognition, water gene spectral recognition, spectral analysis, and environmental parameter acquisition. By organically combining various monitoring data through an integrated black and odorous water body characteristic database, it achieves the entire process from large-scale, efficient screening to precise identification and pollution source tracing, meeting the current practical needs for routine and refined supervision of black and odorous water bodies.

[0019] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0020] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0021] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A flowchart illustrating a method for identifying black and odorous water bodies provided in an embodiment of this application; Figure 2 This is a schematic diagram of a black and odorous water body identification device provided in an embodiment of this application. Detailed Implementation

[0023] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of this application, including various details to aid understanding. These should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this application. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description. It should be noted that, unless otherwise specified, the embodiments and features described herein can be combined with each other.

[0024] The following description, with reference to the accompanying drawings, describes a method, apparatus, device, and medium for identifying black and odorous water bodies according to embodiments of this application.

[0025] This application provides a method, device, equipment, and medium for identifying black and odorous water bodies. The main objective is to address the technical problem that existing single technical means are insufficient to meet the full-process requirements of large-scale and efficient screening, accurate identification, and pollution source tracing. There is a lack of a collaborative monitoring scheme that can adapt to the complex characteristics of black and odorous water bodies, making it difficult to meet the current practical needs for routine and refined supervision of black and odorous water bodies.

[0026] like Figure 1 As shown, embodiments of this application provide a method for identifying black and odorous water bodies. This method constructs a multi-dimensional fusion feature database of black and odorous water bodies, including visual, spectral, and microbial genetic features. Combined with visual pre-screening, precise sampling, and multi-factor correction algorithms, it achieves efficient screening, accurate identification, and quantitative source tracing of black and odorous water bodies, including: Step 101: Construct a fusion-type black and odorous water body feature database, which includes standard visual feature parameters, standard spectral feature parameters, and standard microbial gene feature parameters of black and odorous water bodies.

[0027] For the embodiments of this application, constructing a fusion-type black and odorous water body feature database may specifically include: Simultaneously collect water body image samples, spectral samples, and water body samples under different levels of black and odorous water bodies (mild / severe / extremely severe), pollution types (organic pollution / inorganic pollution / compound pollution), regional matrix (acidic water bodies in the south / alkaline water bodies in the north / coastal brackish water), and seasonal conditions (dry season / wet season / freezing season), covering 100+ typical black and odorous scenarios; After preprocessing the water body image samples, features can be extracted using a convolutional neural network (CNN) model. In this embodiment, YOLOv8 combined with the CBAM (Convolutional Block Attention Module) attention mechanism model is preferably used for extraction. The extracted visual feature parameters may include the original eight visual feature parameters such as RGB mean, gray-level histogram peak value, gray-level co-occurrence matrix entropy value, and scum area ratio, as well as four additional feature parameters added to improve the sensitivity of small target recognition: water flow rate, bubble density, degree of withering of shoreline vegetation, and reflectivity of oil film on the water surface. Absorbance curves in the 200-1100 nm wavelength range were obtained using a UV-Vis-NIR spectrophotometer, and spectral characteristic parameters were extracted accordingly. The extracted spectral characteristic parameters may include the original 15 spectral characteristic parameters (such as characteristic peak wavelength, peak absorbance, and absorbance ratio at 400 nm / 600 nm), as well as 10 newly added near-infrared characteristic parameters (such as absorbance ratio at 850 nm / 950 nm and full width at half maximum (FWHM) of the characteristic peak). The 16S chromosome of specific microorganisms (such as desulfovibrio and methanogens) in black and odorous water samples can be detected using a water gene sensor. Microbial gene sequences were obtained, and three core microbial gene characteristic parameters were extracted from them. The extracted visual, spectral, and microbial genetic parameters can be Z-score standardized to eliminate the influence of dimensions. Subsequently, these parameters are bound to the corresponding black and odorous water body level, pollution type, geographical location, and seasonal labels for the same sample, constructing a fusion-based black and odorous water body feature database containing no fewer than 1 million entries. This fusion-based black and odorous water body feature database can be optimized using a dual mechanism of transfer learning and incremental learning, supporting cross-regional adaptive adjustment of water body matrix and synchronous updates in the cloud, thereby forming a standardized and scalable benchmark database.

[0028] This application is the first to construct a three-dimensional fusion feature library of black and odorous water bodies that integrates "visual-spectral-microbial" technology. It deeply integrates camera visual recognition and water gene spectral recognition technology, adds specific microbial gene features of black and odorous water bodies and cross-regional adaptive mechanism, breaks through the limitations of traditional two-dimensional features, and covers multiple scenarios and types of pollution samples.

[0029] Step 102: Collect real-time images of the monitored water area using a camera, extract visual feature parameters of the water area images using a pre-trained attention mechanism model, and perform feature matching between the visual feature parameters and the standard visual feature parameters in the fusion-type black and odorous water body feature database. Based on the matching results, identify the monitored water area as the target area of ​​suspected pollution.

[0030] To achieve "rapid visual pre-screening," efficient screening is used to avoid invalid sampling and improve overall monitoring efficiency.

[0031] In practical implementation, the hardware configuration can use a 4K ultra-high-definition CMOS camera (resolution 3840×2160, frame rate 30fps), paired with an infrared-visible light dual-mode switching module (infrared wavelength 850nm) and a laser rangefinder (measurement accuracy ±1cm).

[0032] The system can activate a camera to capture real-time images of the monitored water area. Image features (i.e., visual feature parameters) are extracted from these images using a pre-trained YOLOv8-CBAM model (an attention mechanism model), and then matched against standard visual feature parameters from a fusion feature library. The system also acquires real-time light intensity data from the camera's photosensor and dynamically adjusts a preset visual matching threshold based on this intensity. For example, the threshold is set to 75% in bright light and adjusted to 65% in low light conditions to adapt to different recognition needs under varying lighting environments.

[0033] The system calculates the matching degree between visual feature parameters and standard visual feature parameters in a fusion-based black and odorous water body feature database. When the matching degree is greater than or equal to the adjusted preset visual matching degree threshold, the system is considered to have achieved a satisfactory match. At this point, the coordinates of the target area in the monitored water body (including distance information obtained from the laser rangefinder) are locked, and subsequent sampling procedures are triggered. The system supports simultaneously locking up to 10 suspected polluted areas and generating a sampling priority list based on confidence level (≥80% priority). If the matching degree does not meet the standard, the water body is determined to be normal, and sampling is not performed to avoid invalid sampling.

[0034] Step 103: Collect water samples from the target area, collect water gene data from the water samples using a water gene sensor, collect spectral data from the water samples using a spectrometer, and simultaneously collect real-time environmental parameters.

[0035] This application can employ a multi-channel stratified sampling and intelligent preprocessing system, adaptable to complex terrains and extreme environments, to obtain accurate analytical samples.

[0036] In this embodiment, collecting water samples from the target area may specifically include: Based on the distance information obtained by the laser rangefinder, the automatic lifting sampling arm is controlled to adjust its travel (adjustable from 0-300cm) to the target area; A sampling device equipped with a three-channel sampling head can be activated to collect water samples from the target area in layers, obtaining surface, middle and bottom water samples. It supports synchronous or individual sampling and is suitable for water depth scenarios of 0.5-5m. Water samples are pretreated, and the filter membrane pore size can be automatically switched (5 Å) based on real-time turbidity data (combined with camera visual estimation and sensor detection). / 10 / 20 When the turbidity > 50 NTU, start the pre-sedimentation tank for sedimentation (sedimentation time adjustable from 1 to 3 minutes); the sample then undergoes two-stage filtration, first passing through a 10... Nylon filter membranes remove suspended particles, then pass through 0.45 A polyethersulfone (PES) membrane removes colloidal impurities, and the final product is introduced into a polytetrafluoroethylene (PTFE) buffer cell. The buffer cell incorporates a temperature sensor (accuracy ±0.1 ppm). ) and low-temperature preservation module (4 (Constant temperature), untested samples can be stored for 24 hours; a new automatic pH adjustment unit is added, which injects buffer solution through a micro peristaltic pump to stabilize the sample pH at 7.0±0.2, and is equipped with a dechlorination module to eliminate residual chlorine interference and ensure sample stability.

[0037] In this embodiment, the spectral data of the water sample is collected using a spectrometer, which may specifically include: A pulsed xenon lamp light source (pulse frequency 10Hz, light intensity stability ±0.5%) was used, along with a high-resolution fiber optic spectrometer (resolution 0.3nm, signal-to-noise ratio ≥1500:1) and a 10mm optical path quartz cuvette. The pretreated water sample is injected into the cuvette at a constant rate of 3 mL / s using a peristaltic pump; the pulsed xenon lamp emits a continuous spectrum of 200-1100 nm after preheating for 30 s; the spectrometer automatically adjusts the integration time (5-20 ms) according to the concentration of the water sample and acquires the spectral curve at a preset frequency (e.g., 10 Hz). Simultaneously acquire real-time environmental parameters across multiple dimensions, including temperature, pressure (accuracy ±0.001 MPa), humidity (accuracy ±5% RH), and turbidity (accuracy ±1 NTU). For each sample, 10 spectral curves are subjected to median filtering to remove noise, and the average of the five valid peak values ​​is taken as the final spectral data.

[0038] Step 104: Determine the comprehensive similarity between the water sample and the integrated black and odorous water body feature database based on water gene data, spectral data, real-time environmental parameters, and visual feature parameters.

[0039] This application achieves "precise detection of water genes" and "dynamic weighted fusion judgment" through multi-factor correction and dynamic weighted fusion, effectively eliminating environmental interference and improving the accuracy of judgment.

[0040] Specifically, determining the overall similarity involves the following process: Real-time environmental parameters can be used to perform multi-factor correction on spectral data to obtain corrected spectral characteristic parameters. Specifically, a piecewise nonlinear fitting algorithm can be used. For example: 5-15 Correction factor 0.0015 / 15-35 Correction factor 0.002 / 35-45 Correction factor 0.03 / A multivariate regression correction model was constructed by incorporating humidity, turbidity, and pH correction terms to eliminate interference from complex environmental factors. Simultaneously, 3... Criteria for filtering outlier data points; The spectral similarity is calculated by comparing the corrected spectral feature parameters with the standard spectral feature parameters in the fusion-type black and odorous water body feature database. The microbial gene characteristic parameters of the synchronously collected water gene data are compared with the standard microbial gene characteristic parameters in the fusion black and odorous water body characteristic database to calculate the microbial characteristic matching degree. Based on the comparison results between the visual feature parameters in step 102 and the standard visual feature parameters in the fusion-type black and odorous water body feature database, the visual similarity is determined. The feature reliability coefficient is calculated based on real-time environmental parameters, and the feature weights corresponding to spectral feature parameters, microbial gene feature parameters, and visual feature parameters are dynamically allocated based on the reliability coefficient. Wherein, the visual feature weights correspond to the visual feature parameters. : Dynamically adjust between 0.2 and 0.4; Spectral feature weights corresponding to spectral feature parameters Dynamically adapts between 0.6 and 0.8; Microbial feature weights corresponding to microbial gene feature parameters : Fixed at 0.1.

[0041] The formula "Overall Similarity =" is used to calculate the overall similarity. ×Visual Similarity+ ×Spectral similarity+ × Microbial characteristic matching degree ( The calculation yielded the overall similarity between the water sample and the integrated black and odorous water body feature database.

[0042] Step 105: Determine the black and odorous water level of the water sample based on the comprehensive similarity, and quantitatively calculate the concentration of core pollutants in the water sample by combining the mapping relationship in the integrated black and odorous water feature database.

[0043] For the embodiments of this application, determining the black and odorous water level of a water sample based on comprehensive similarity may specifically include: When the overall similarity is in the first similarity threshold range (≥92%), the water sample is determined to be extremely black and odorous. When the overall similarity is in the second similarity threshold range (85%-92%), the water sample is determined to be severely black and odorous. When the overall similarity is within the third similarity threshold range (80%-85%), the water sample is determined to be of mild black and odorous water quality. When the overall similarity is within the fourth similarity threshold range (75%-80%), the water sample is determined to be of the suspected black and odorous water level (at which point the secondary sampling process can be initiated). When the overall similarity is within the fifth similarity threshold range (<75%), the black and odorous water body of the water sample is determined to be normal water body.

[0044] Specifically, the quantitative calculation of the concentration of core pollutants may include: Based on the mapping relationship between "characteristic parameters - pollutant type - pollutant concentration" in the integrated black and odorous water body characteristic database, the concentrations of core pollutants such as ammonia nitrogen (detection limit 0.01 mg / L), sulfide (detection limit 0.005 mg / L), and COD (detection limit 1 mg / L) are quantitatively calculated using a partial least squares regression model, enabling pollution source tracing with an error controlled within 5%.

[0045] Furthermore, the embodiments of this application also include result output and feedback steps: Generates a comprehensive monitoring report including images, spectral curves, overall similarity, black and odor levels, pollutant concentration gradient maps, microbial community structure analysis, and pollution risk level assessment (low / medium / high / extremely high). Data is transmitted via dual-channel communication (RS485 and BeiDou short message + 5G dual-mode communication, with BeiDou communication latency ≤60s), and the push cycle is adjustable from 1 to 60 minutes. Simultaneously, it triggers tiered early warnings (mild yellow, severe orange, complex red, extremely high risk purple), with warning signals simultaneously output via local audible and visual alarms (volume adjustable from 85-110dB) and remote platform SMS + APP pop-up windows, achieving fully automated collaborative monitoring throughout the entire process.

[0046] Through the above embodiments, this application has achieved a comprehensive improvement in recognition accuracy (≥98%) and efficiency (visual pre-screening ≤200ms), solved the monitoring problem in complex environments, and significantly reduced manual operation and maintenance costs (60%).

[0047] In summary, the black and odorous water body identification method provided in this application, compared with the existing technology, can construct a fusion-type black and odorous water body feature library, which includes standard visual feature parameters, standard spectral feature parameters, and standard microbial gene feature parameters of black and odorous water bodies; collect water images of the monitored water area in real time through a camera, extract visual feature parameters of the water images using a pre-trained attention mechanism model, and perform feature matching between the visual feature parameters and the standard visual feature parameters in the fusion-type black and odorous water body feature library; based on the matching results, identify the monitored water area as a suspected pollution target area; collect water samples from the target area, collect water gene data of the water samples using a water gene sensor, collect spectral data of the water samples using a spectrometer, and simultaneously collect real-time environmental parameters; determine the comprehensive similarity between the water sample and the fusion-type black and odorous water body feature library based on the water gene data, spectral data, real-time environmental parameters, and visual feature parameters; determine the black and odorous water body level of the water sample based on the comprehensive similarity, and quantitatively calculate the core pollutant concentration of the water sample by combining the mapping relationship in the fusion-type black and odorous water body feature library.

[0048] By employing the above technical solution, this application does not rely solely on remote sensing imagery for screening, but rather constructs a fusion-based black and odorous water body feature database. Real-time images of the monitored water areas are acquired via cameras, and a pre-trained attention mechanism model is used to extract visual feature parameters from the water images. These parameters are then matched with standard visual feature parameters in the fusion-based black and odorous water body feature database. Based on the matching results, suspected pollution target areas are identified. This approach is not limited to remote sensing imagery, avoiding the interference problems of atmospheric correction bias, mixed pixel effects, and complex urban backgrounds (such as shadow occlusion and strong heterogeneity of water body optical properties) that affect remote sensing technology, thus improving the identification accuracy of small streams and canals.

[0049] This application uses a camera to collect water area images in real time, which can obtain visual information of the monitored water area in real time and perform feature matching to lock the target area. It has real-time dynamic monitoring capabilities, overcoming the shortcomings of remote sensing image screening technology, which can only provide clue-based screening results and cannot achieve real-time dynamic monitoring.

[0050] In addition to using visual feature parameters for preliminary screening, this application also collects water gene data, spectral data, and real-time environmental parameters from water samples in the target area. It then integrates these data to determine the comprehensive similarity between the water sample and a fusion-based black and odorous water body feature database, thereby classifying the water sample as a black and odorous water body. This multi-dimensional data analysis, compared to traditional water quality sensors that only detect single physicochemical parameters such as dissolved oxygen, pH, and turbidity, significantly expands the detection dimensions and can more comprehensively and accurately reflect the water body's condition.

[0051] The monitoring process described in this application covers real-time image acquisition, feature extraction, sample collection, and comprehensive analysis of multiple data. By optimizing the collaborative work of each step, compared with traditional water quality sensor monitoring technology that relies on manual sampling and subsequent laboratory analysis, the response speed is greatly improved, and monitoring results can be obtained more quickly to meet the needs of routine supervision.

[0052] The integrated black and odorous water body feature database constructed in this application includes standard microbial gene feature parameters. It combines water gene data, spectral data, real-time environmental parameters, and visual feature parameters to determine the comprehensive similarity, classify the black and odorous water body, and quantitatively calculate the concentration of core pollutants. This process provides a practical foundation for the application of water gene spectral identification technology in black and odorous water bodies to form a standardized system adapted to the coexistence of multiple pollutants and large concentration fluctuations, helping to solve the current problem of the lack of a standardized system for this technology in the application of black and odorous water bodies.

[0053] This application comprehensively utilizes multiple monitoring technologies and data, and reduces the impact of environmental interference on a single sensor through multi-parameter collaborative analysis. For example, when water gene sensors may have problems such as weak anti-pollution capabilities or insufficient adaptability to complex environments, combining visual feature parameters, spectral data, and real-time environmental parameters allows for the assessment of water conditions from different perspectives, improving the stability and reliability of the entire monitoring system in complex environments.

[0054] In summary, this application constructs a collaborative monitoring scheme integrating multiple technologies such as visual recognition, water gene spectral recognition, spectral analysis, and environmental parameter acquisition. By organically combining various monitoring data through an integrated black and odorous water body characteristic database, it achieves the entire process from large-scale, efficient screening to precise identification and pollution source tracing, meeting the current practical needs for routine and refined supervision of black and odorous water bodies.

[0055] Based on the above Figure 1 The specific implementation of the method shown in this embodiment provides a black and odorous water body identification device, such as... Figure 2 As shown, the device includes: a construction module 31, a matching module 32, a data acquisition module 33, a determination module 34, and a judgment module 35; Module 31 is used to construct a fusion-type black and odorous water body feature library, which includes standard visual feature parameters, standard spectral feature parameters and standard microbial gene feature parameters of black and odorous water bodies. The matching module 32 is used to collect water area images of the monitored water area in real time through a camera, extract the visual feature parameters of the water area images using a pre-trained attention mechanism model, and perform feature matching between the visual feature parameters and the standard visual feature parameters in the fusion black and odorous water body feature database. Based on the matching result, the monitored water area is identified as a suspected pollution target area. The acquisition module 33 is used to acquire water samples from the target area, acquire water gene data of the water samples using a water gene sensor, acquire spectral data of the water samples using a spectrometer, and simultaneously acquire real-time environmental parameters. The determination module 34 is used to determine the comprehensive similarity between the water sample and the fused black and odorous water feature database based on the water gene data, the spectral data, the real-time environmental parameters and the visual feature parameters. The determination module 35 is used to determine the black and odorous water level of the water sample based on the comprehensive similarity, and to quantitatively calculate the concentration of the core pollutants in the water sample by combining the mapping relationship in the fusion black and odorous water feature database.

[0056] In specific application scenarios, module 31 is used to simultaneously collect water image samples, spectral samples, and water samples under different black and odorous water body levels, pollution types, regional matrix, and seasonal conditions; extract visual feature parameters of the water body image samples through a convolutional neural network model; obtain the absorbance curve of the spectral samples through a spectrophotometer and extract spectral feature parameters based on the absorbance curve; detect the microbial gene sequence of the water samples through a water gene sensor and extract microbial gene feature parameters based on the microbial gene sequence; bind the visual feature parameters, the spectral feature parameters, and the microbial gene feature parameters with the corresponding black and odorous water body level and pollution type labels, and perform standardization processing to construct the fusion-type black and odorous water body feature library that supports incremental updates.

[0057] In specific application scenarios, the matching module 32 is specifically used to obtain the real-time light intensity of the photosensitive sensor of the camera, and dynamically adjust the preset visual matching degree threshold according to the light intensity; when the matching degree value of the visual feature parameter and the standard visual feature parameter in the fusion black and odorous water body feature database is greater than or equal to the adjusted preset visual matching degree threshold, it is determined that the matching standard is met, and the target area of ​​the monitored water area is locked.

[0058] In a specific application scenario, the acquisition module 33 is specifically used to control the automatic lifting sampling arm to adjust its travel to the target area based on the distance information obtained by the laser rangefinder; to start the sampling device equipped with a three-channel sampling head, to collect water samples from the target area in layers, to obtain water samples from the target area, and to preprocess the water samples, wherein the layers include surface, middle and bottom water.

[0059] In a specific application scenario, the acquisition module 33 is specifically used to inject the pretreated water sample into a cuvette at a constant rate, emit a continuous spectrum using a pulsed xenon lamp light source, and acquire a spectral curve at a preset frequency by adjusting the integration time according to the sample concentration of the water sample using the spectrometer; the spectral curve is then subjected to median filtering and noise reduction processing, and the average value of the effective peaks is taken as the spectral data.

[0060] In a specific application scenario, module 34 is specifically used to perform multi-factor correction processing on the spectral data using the real-time environmental parameters to obtain corrected spectral feature parameters; compare the corrected spectral feature parameters with the standard spectral feature parameters in the integrated black and odorous water body feature database to calculate spectral similarity; compare the microbial gene feature parameters of the synchronously collected water gene data with the standard microbial gene feature parameters in the integrated black and odorous water body feature database to calculate microbial feature matching degree; determine visual similarity based on the comparison result of the visual feature parameters with the standard visual feature parameters in the integrated black and odorous water body feature database; calculate the feature weights corresponding to the spectral feature parameters, the microbial gene feature parameters, and the visual feature parameters respectively based on the real-time environmental parameters; and determine the comprehensive similarity between the water sample and the integrated black and odorous water body feature database based on the spectral similarity, the microbial feature matching degree, the visual similarity, and the feature weights corresponding to the spectral feature parameters, the microbial gene feature parameters, and the visual feature parameters.

[0061] In specific application scenarios, the determination module 35 is specifically used to determine the black and odorous water body level of the water sample as extremely severe black and odorous when the comprehensive similarity is in the first similarity threshold range; to determine the black and odorous water body level of the water sample as severely black and odorous when the comprehensive similarity is in the second similarity threshold range; to determine the black and odorous water body level of the water sample as mild black and odorous when the comprehensive similarity is in the third similarity threshold range; to determine the black and odorous water body level of the water sample as suspected black and odorous when the comprehensive similarity is in the fourth similarity threshold range; and to determine the black and odorous water body level of the water sample as normal water when the comprehensive similarity is in the fifth similarity threshold range. The similarity thresholds for the first, second, third, fourth, and fifth similarity threshold ranges decrease sequentially. Based on the mapping relationship between feature parameters and pollutant types and concentrations in the fused black and odorous water body feature database, the concentration of the core pollutant is calculated using a partial least squares regression model.

[0062] It should be noted that other corresponding descriptions of the functional units involved in the black and odorous water body identification device provided in this embodiment can be found in [reference needed]. Figure 1The corresponding description in [the document] will not be repeated here.

[0063] Based on the above, Figure 1 Accordingly, this embodiment also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method. Figure 1 The method shown.

[0064] Based on this understanding, the technical solution of this application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as CD-ROM, USB flash drive, mobile hard drive, etc.) and includes several instructions to cause a computer device (such as personal computer, server, or network device, etc.) to execute the methods of various implementation scenarios of this application.

[0065] Based on the above, Figure 1 The method shown, and Figure 2 To achieve the above objectives, the present application also provides an electronic device, comprising a storage medium and a processor; the storage medium for storing a computer program; and the processor for executing the computer program to implement the above-described virtual device embodiments. Figure 1 The method shown.

[0066] Optionally, the aforementioned physical devices may also include a user interface, a network interface, a camera, radio frequency (RF) circuitry, sensors, audio circuitry, a Wi-Fi module, etc. The user interface may include a display screen, input units such as a keyboard, etc., and optional user interfaces may also include USB interfaces, card reader interfaces, etc. The network interface may optionally include standard wired interfaces, wireless interfaces (such as Wi-Fi interfaces), etc.

[0067] Those skilled in the art will understand that the physical device structure provided in this embodiment does not constitute a limitation on the physical device, and may include more or fewer components, or combine certain components, or have different component arrangements.

[0068] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the hardware and software resources of the aforementioned physical device, supporting the operation of the black and odorous water body identification program and other software and / or programs. The network communication module is used to enable communication between the various components within the storage medium, as well as communication with other hardware and software in the black and odorous water body identification physical device.

[0069] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware platform, or it can be implemented by hardware. By applying the solution of this embodiment, compared with the existing technology, this application can construct a fusion-type black and odorous water body feature database, which includes standard visual feature parameters, standard spectral feature parameters, and standard microbial gene feature parameters of black and odorous water bodies; real-time acquisition of water images of the monitored water area through cameras; extraction of visual feature parameters of the water images using a pre-trained attention mechanism model; feature matching of the visual feature parameters with the standard visual feature parameters in the fusion-type black and odorous water body feature database; identification of the monitored water area as a suspected pollution target area based on the matching results; collection of water samples from the target area; acquisition of water gene data of the water samples using a water gene sensor; acquisition of spectral data of the water samples using a spectrometer; and simultaneous acquisition of real-time environmental parameters; determination of the comprehensive similarity between the water sample and the fusion-type black and odorous water body feature database based on the water gene data, spectral data, real-time environmental parameters, and visual feature parameters; determination of the black and odorous water body level of the water sample based on the comprehensive similarity; and quantitative calculation of the core pollutant concentration of the water sample based on the mapping relationship in the fusion-type black and odorous water body feature database.

[0070] By employing the above technical solution, this application does not rely solely on remote sensing imagery for screening, but rather constructs a fusion-based black and odorous water body feature database. Real-time images of the monitored water areas are acquired via cameras, and a pre-trained attention mechanism model is used to extract visual feature parameters from the water images. These parameters are then matched with standard visual feature parameters in the fusion-based black and odorous water body feature database. Based on the matching results, suspected pollution target areas are identified. This approach is not limited to remote sensing imagery, avoiding the interference problems of atmospheric correction bias, mixed pixel effects, and complex urban backgrounds (such as shadow occlusion and strong heterogeneity of water body optical properties) that affect remote sensing technology, thus improving the identification accuracy of small streams and canals.

[0071] This application uses a camera to collect water area images in real time, which can obtain visual information of the monitored water area in real time and perform feature matching to lock the target area. It has real-time dynamic monitoring capabilities, overcoming the shortcomings of remote sensing image screening technology, which can only provide clue-based screening results and cannot achieve real-time dynamic monitoring.

[0072] In addition to using visual feature parameters for preliminary screening, this application also collects water gene data, spectral data, and real-time environmental parameters from water samples in the target area. It then integrates these data to determine the comprehensive similarity between the water sample and a fusion-based black and odorous water body feature database, thereby classifying the water sample as a black and odorous water body. This multi-dimensional data analysis, compared to traditional water quality sensors that only detect single physicochemical parameters such as dissolved oxygen, pH, and turbidity, significantly expands the detection dimensions and can more comprehensively and accurately reflect the water body's condition.

[0073] The monitoring process described in this application covers real-time image acquisition, feature extraction, sample collection, and comprehensive analysis of multiple data. By optimizing the collaborative work of each step, compared with traditional water quality sensor monitoring technology that relies on manual sampling and subsequent laboratory analysis, the response speed is greatly improved, and monitoring results can be obtained more quickly to meet the needs of routine supervision.

[0074] The integrated black and odorous water body feature database constructed in this application includes standard microbial gene feature parameters. It combines water gene data, spectral data, real-time environmental parameters, and visual feature parameters to determine the comprehensive similarity, classify the black and odorous water body, and quantitatively calculate the concentration of core pollutants. This process provides a practical foundation for the application of water gene spectral identification technology in black and odorous water bodies to form a standardized system adapted to the coexistence of multiple pollutants and large concentration fluctuations, helping to solve the current problem of the lack of a standardized system for this technology in the application of black and odorous water bodies.

[0075] This application comprehensively utilizes multiple monitoring technologies and data, and reduces the impact of environmental interference on a single sensor through multi-parameter collaborative analysis. For example, when water gene sensors may have problems such as weak anti-pollution capabilities or insufficient adaptability to complex environments, combining visual feature parameters, spectral data, and real-time environmental parameters allows for the assessment of water conditions from different perspectives, improving the stability and reliability of the entire monitoring system in complex environments.

[0076] In summary, this application constructs a collaborative monitoring scheme integrating multiple technologies such as visual recognition, water gene spectral recognition, spectral analysis, and environmental parameter acquisition. By organically combining various monitoring data through an integrated black and odorous water body characteristic database, it achieves the entire process from large-scale, efficient screening to precise identification and pollution source tracing, meeting the current practical needs for routine and refined supervision of black and odorous water bodies.

[0077] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the term "comprising" or any other variations thereof is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0078] The above are merely specific embodiments of this application, enabling those skilled in the art to understand or implement this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to these embodiments, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method for identifying black and odorous water, characterized in that, The application relates to a black and odorous water body monitoring method and device. The method comprises the following steps: a fusion type black and odorous water body characteristic library is constructed, wherein the fusion type black and odorous water body characteristic library comprises standard visual characteristic parameters, standard spectral characteristic parameters and standard microbial gene characteristic parameters of black and odorous water bodies; a water area image of a monitoring water area is collected in real time through a camera, visual characteristic parameters of the water area image are extracted by using a pre-trained attention mechanism model, and the visual characteristic parameters are matched with the standard visual characteristic parameters in the fusion type black and odorous water body characteristic library; and according to a matching result, the monitoring water area is locked as a suspected pollution target area; a water body sample of the target area is collected, water gene data of the water body sample are collected by using a water gene sensor, spectral data of the water body sample are collected by using a spectrometer, and real-time environmental parameters are synchronously collected; comprehensive similarity of the water body sample and the fusion type black and odorous water body characteristic library is determined according to the water gene data, the spectral data, the real-time environmental parameters and the visual characteristic parameters; 2. The method of claim 1, wherein, a black and odorous water body grade of the water body sample is judged according to the comprehensive similarity, and a core pollutant concentration of the water body sample is quantitatively calculated in combination with a mapping relationship in the fusion type black and odorous water body characteristic library. The construction of the fusion type black and odorous water body characteristic library specifically comprises the following steps: water body image samples, spectral samples and water body samples under different black and odorous water body grades, pollution types, regional substrates and seasonal working conditions are synchronously collected; visual characteristic parameters of the water body image samples are extracted by using a convolutional neural network model, absorbance curves of the spectral samples are obtained by using a spectrophotometer, spectral characteristic parameters are extracted according to the absorbance curves, and microbial gene sequences of the water body samples are detected by using a water gene sensor, and microbial gene characteristic parameters are extracted according to the microbial gene sequences; 3. The method of claim 1, wherein, the visual characteristic parameters, the spectral characteristic parameters, the microbial gene characteristic parameters, corresponding black and odorous water body grades and pollution type labels are bound, and are subjected to standardization treatment, so that the fusion type black and odorous water body characteristic library supporting incremental updating is constructed. The matching of the visual characteristic parameters with the standard visual characteristic parameters in the fusion type black and odorous water body characteristic library and the locking of the monitoring water area as the suspected pollution target area according to a matching result specifically comprise the following steps: a real-time illumination intensity of a photosensitive sensor of the camera is obtained, and a preset visual matching degree threshold value is dynamically adjusted according to the illumination intensity; 4. The method of claim 1, wherein, when a matching degree value of the visual characteristic parameters and the standard visual characteristic parameters in the fusion type black and odorous water body characteristic library is greater than or equal to the adjusted preset visual matching degree threshold value, it is determined that the matching is up to standard, and the target area of the monitoring water area is locked. The collection of the water body sample of the target area specifically comprises the following steps: distance information obtained by a laser ranging sensor is used to control an automatic lifting sampling arm to adjust a stroke thereof to the target area; a sampling device provided with a three-channel sampling head is started, and water bodies of the target area are collected in layers respectively, so that the water body sample of the target area is obtained, and the water body sample is pretreated, wherein the layers include surface layer water bodies, middle layer water bodies and bottom layer water bodies.

5. The method of claim 4, wherein, The acquisition of spectral data from the water sample using a spectrometer specifically includes: The pretreated water sample is injected into a cuvette at a constant rate, and a pulsed xenon lamp light source is used to emit a continuous spectrum. The spectrometer adjusts the integration time according to the sample concentration of the water sample and collects the spectral curve at a preset frequency. The spectral curve is subjected to median filtering for noise reduction, and the average value of the effective peaks is taken as the spectral data.

6. The method of claim 1, wherein, The determination of the comprehensive similarity between the water sample and the integrated black and odorous water body feature database based on the water gene data, the spectral data, the real-time environmental parameters, and the visual feature parameters specifically includes: The spectral data is subjected to multi-factor correction processing using the real-time environmental parameters to obtain the corrected spectral characteristic parameters. The spectral similarity is calculated by comparing the corrected spectral feature parameters with the standard spectral feature parameters in the fusion-type black and odorous water body feature database. The microbial gene feature parameters of the synchronously collected water gene data are compared with the standard microbial gene feature parameters in the fusion-type black and odorous water body feature database to calculate the microbial feature matching degree. The visual similarity is determined based on the comparison results between the visual feature parameters and the standard visual feature parameters in the fusion-type black and odorous water body feature database. Calculate the feature weights corresponding to the spectral feature parameters, the microbial gene feature parameters, and the visual feature parameters based on the real-time environmental parameters; The comprehensive similarity between the water sample and the integrated black and odorous water feature database is determined based on the spectral similarity, the microbial feature matching degree, the visual similarity, and the feature weights corresponding to the spectral feature parameters, the microbial gene feature parameters, and the visual feature parameters.

7. The method of claim 1, wherein, The process of determining the black and odorous water level of the water sample based on the comprehensive similarity and quantitatively calculating the core pollutant concentration of the water sample by combining the mapping relationship in the fused black and odorous water feature database specifically includes: When the overall similarity is within the first similarity threshold range, the water body sample is determined to be severely black and odorous. When the overall similarity is within the second similarity threshold range, the water body sample is determined to be severely black and odorous. When the overall similarity is within the third similarity threshold range, the water body sample is determined to be of mild black and odorous water quality. When the overall similarity is within the fourth similarity threshold range, the water sample is determined to be of the suspected black and odorous water level. When the overall similarity is within the fifth similarity threshold range, the black and odorous water body level of the water body sample is determined to be normal water body, wherein the similarity thresholds of the first similarity threshold range, the second similarity threshold range, the third similarity threshold range, the fourth similarity threshold range, and the fifth similarity threshold range decrease sequentially. Based on the mapping relationship between characteristic parameters and pollutant types and concentrations in the integrated black and odorous water body characteristic database, the concentration of the core pollutants is calculated using a partial least squares regression model. 8.A black and odorous water body identification device, characterized in that, include: A construction module is used to construct a fusion-type black and odorous water body feature library, which includes standard visual feature parameters, standard spectral feature parameters, and standard microbial gene feature parameters of black and odorous water bodies. The matching module is used to collect images of the monitored water area in real time through a camera, extract visual feature parameters of the water area image using a pre-trained attention mechanism model, and perform feature matching between the visual feature parameters and the standard visual feature parameters in the fusion black and odorous water body feature database. Based on the matching result, the monitored water area is identified as a suspected pollution target area. The acquisition module is used to collect water samples from the target area, collect water gene data of the water samples using a water gene sensor, collect spectral data of the water samples using a spectrometer, and simultaneously collect real-time environmental parameters. The determination module is used to determine the comprehensive similarity between the water sample and the fused black and odorous water feature database based on the water gene data, the spectral data, the real-time environmental parameters, and the visual feature parameters. The determination module is used to determine the black and odorous water level of the water sample based on the comprehensive similarity, and to quantitatively calculate the concentration of the core pollutants in the water sample by combining the mapping relationship in the fusion black and odorous water feature database.

9. An electronic device comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 7.

Images