Remote sensing cooperation-based industrial zone water quality multi-parameter online monitoring method and system

Abstract

The present application relates to the field of data processing, more particularly, the present application relates to an industrial area water quality multi-parameter online monitoring method and system based on remote sensing cooperation, the method comprises: collecting multispectral remote sensing data and ground water quality parameters, constructing effective water grid and fitting benchmark calculation coefficient;Calculate the optical state gradient index of the effective water grid, generate a dynamic correction coefficient combined with the global transition scaling coefficient;The sensitivity of the benchmark calculation coefficient is adjusted by using the dynamic correction coefficient, the target water quality concentration is calculated, and the abnormality is judged.The present application adjusts the sensitivity of the model by the optical state gradient index, solves the error problem caused by the fixed coefficient model when the optical characteristics of the water body change suddenly, realizes the global dynamic monitoring and hierarchical alarm of the industrial water body, and improves the precision and real-time performance of the water quality inversion.

Description

Technical Field

[0001] This invention relates to the field of data processing. More specifically, this invention relates to a method and system for online monitoring of multiple parameters of water quality in industrial areas based on remote sensing collaboration. Background Technology

[0002] Industrial zone water bodies are characterized by dispersed pollution sources, frequent hydrological monitoring, and variable meteorological conditions, resulting in strong spatiotemporal fluctuations in water composition. Traditional fixed-point sampling methods, limited by station coverage and data transmission cycles, struggle to fully depict the migration trajectory of pollution fronts and concentration gradients in mixing areas. Online monitoring, combining satellite remote sensing and ground-based monitoring, can simultaneously acquire large-scale spatial spectral information and localized real-world water quality benchmarks. This approach combines wide-area coverage with high-frequency updates, providing fundamental data support for tracing industrial pollution sources, tracking water quality trends, and emergency response, and has become a crucial component of the environmental regulatory system.

[0003] However, existing collaborative monitoring methods mostly rely on historical samples to establish fixed-coefficient models, implicitly assuming that the optical properties of water bodies are spatiotemporally uniform and stable. But in actual industrial environments, factors such as sewage discharge, tides, and wind frequently alter the local optical characteristics of water bodies. Fixed-coefficient models cannot adjust synchronously, leading to a disconnect between the calculated baseline and reality. In areas with mixed sewage discharge or abrupt turbidity changes, concentration fluctuations and misjudgments of water conditions are prone to occur. Therefore, how to make water quality estimation models adapt to the dynamic changes in the optical properties of water bodies is a pressing technical problem that needs to be solved in this field. Summary of the Invention

[0004] To address the aforementioned technical problem of how to make water quality estimation models adapt to the dynamic changes in the optical properties of water bodies, this invention provides solutions in the following aspects.

[0005] In the first aspect, the online monitoring method for multiple parameters of water quality in industrial areas based on remote sensing collaboration includes: Multispectral data from optical satellite sensors and water quality parameters collected synchronously from ground automatic monitoring stations were acquired and preprocessed. Multiple effective water body grids were constructed based on the acquired data, and baseline extrapolation coefficients and intercept terms were fitted. Select any effective water body grid as the target grid, and calculate the optical state gradient index of the target grid. The optical state gradient index is used to reflect the magnitude of reflectance change of the target grid in the horizontal and vertical directions and the reliability of the spectral characteristics. Using the optical state gradient index as input, and combining the preset global transition scaling factor and optical state gradient threshold, a dynamic correction factor adapted to the current water disturbance intensity is generated. The sensitivity of the benchmark estimation coefficient is adjusted using the dynamic correction coefficient, and the target water quality concentration is calculated by combining the spectral characteristics of the target grid. A water quality parameter distribution map is generated based on the distribution characteristics of the target water quality concentration, and anomaly status is determined, thereby realizing dynamic monitoring of the entire water body in the industrial area.

[0006] Optionally, multiple effective water body grids can be constructed based on the collected data, including: The raw digital quantization values ​​collected by the optical satellite sensor are converted to obtain green light reflectance, red light reflectance, and near-infrared reflectance; The inverse distance weighted interpolation method is used to diffuse the measured concentration values ​​collected by each monitoring station to each grid in the two-dimensional grid coordinate system to obtain the reference concentration of each grid. Grids with near-infrared reflectivity higher than a set threshold are removed, and only water pixels are retained as valid water grids.

[0007] Optionally, obtaining the benchmark extrapolation coefficients and intercept term includes: For water quality parameters, the red light band was selected as the characteristic band, and the green light band was selected as the reference band. For each effective water body grid, calculate the normalized band ratio of its characteristic band reflectance to the reference band reflectance; The normalized band ratio of each effective water body grid is paired with the reference concentration of that effective water body grid. The normalized band ratio is used as the independent variable and the reference concentration is used as the dependent variable. The least squares method is used to fit a straight line to obtain the static linear model parameters, namely the baseline extrapolation coefficients and the intercept term.

[0008] Optionally, the calculation of the optical state gradient index of the target mesh includes: For the target grid, the reflectance difference in the horizontal and vertical directions is calculated only when its left and right neighbors in the horizontal direction and its upper and lower neighbors in the vertical direction are all effective water grids. The composite change is calculated in combination with the directional validity indicator, and the ratio of the composite change to the reflectance of the reference band of the target grid is used as the local relative change rate. Calculate the absolute value of the difference between the characteristic band reflectance and the near-infrared reflectance of the target grid, and obtain the spectral confidence weight by exponential transformation of the absolute value of the difference; A target window of a preset size is obtained with the target grid as the center. The mean of the sum of the products of the local relative change rate and the spectral confidence weight of all effective water grids within the target window is calculated to obtain the optical state gradient index.

[0009] Optionally, the generation of the dynamic correction coefficient includes: The mean value of the absolute difference between the reference concentration and the measured concentration in the grid where all ground monitoring stations are located is obtained and used as the global transition scaling factor. An optical state gradient threshold is set. When the optical state gradient index of the target mesh is lower than the optical state gradient threshold, a dynamic correction coefficient smaller than the baseline calculated coefficient is generated using the hyperbolic tangent function. When the optical state gradient index of the target mesh is higher than the optical state gradient threshold, a dynamic correction coefficient larger than the baseline calculated coefficient is generated.

[0010] Optionally, the calculation of the target water quality concentration includes: Multiply the normalized band ratio of the target grid by the dynamic correction coefficient to obtain the corrected band ratio. The target water quality concentration of the target grid is calculated by adding the corrected band ratio to the pre-fitted intercept term.

[0011] Optionally, the abnormal state determination includes: A preset warning threshold is set. If the target water quality concentration of any effective water body grid exceeds the warning threshold, the effective water body grid is determined to be in a warning state, and an alarm work order containing the location and concentration value of the exceeding standard is generated and pushed to the monitoring platform.

[0012] Secondly, a remote sensing-based online monitoring system for multiple parameters of water quality in industrial areas includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the remote sensing-based online monitoring system for multiple parameters of water quality in industrial areas is implemented as described in any one of the above-mentioned methods.

[0013] The present invention has the following beneficial effects: 1. This invention solves the error problem caused by abrupt changes in the optical properties of water bodies in industrial areas when traditional fixed-coefficient models introduce a dynamic correction mechanism driven by optical state gradient indices. Before calculating water concentration, this invention first quantifies and evaluates the reflectance variation of the target grid in the horizontal and vertical directions, and then weights it using spectral confidence weights. This mechanism enables the inference model to adaptively adjust its sensitivity according to the intensity of local water disturbances, avoiding concentration misjudgments caused by optical heterogeneity in sewage mixing zones or water flow fronts, and significantly improving the inversion accuracy under complex conditions.

[0014] 2. This invention achieves smooth transition and intelligent adjustment of model parameters by constructing a dynamic correction coefficient model. Unlike simple threshold switching logic, this invention uses the optical state gradient index as an input variable and generates dynamic correction coefficients through the nonlinear mapping characteristics of the hyperbolic tangent function. This design allows the model to automatically reduce sensitivity in stable water regions to suppress background noise, and automatically increase sensitivity to capture concentration changes when pollution discharge or severe disturbances are detected, maintaining the model in its optimal operating state without manual intervention.

[0015] 3. This invention improves the reliability of abnormal state determination by integrating the dual criteria of local relative change rate and spectral confidence weight. When calculating the optical state gradient index, this invention not only considers the reflectance difference between grid neighbors (local relative change rate) but also introduces the exponential mapping result of the difference between red light and near-infrared reflectance (spectral confidence weight). This dual verification mechanism effectively identifies non-aquatic interference signals such as clouds and aerosols, ensuring that the final generated water quality parameters and early warning results have extremely high confidence, providing accurate decision-making basis for environmental supervision. Attached Figure Description

[0016] Figure 1 This is a flowchart of steps S1-S3 in the online monitoring method for multiple parameters of water quality in industrial areas based on remote sensing collaboration, according to an embodiment of the present invention.

[0017] Figure 2 This is a schematic diagram illustrating the calculation of the target water quality concentration in a single effective water body grid within the online multi-parameter monitoring method for industrial water quality based on remote sensing collaboration, as described in this embodiment of the invention.

[0018] Figure 3 This is a structural block diagram of the online monitoring system for multiple parameters of water quality in industrial areas based on remote sensing collaboration, according to an embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0020] This invention provides a method for online monitoring of multiple parameters of water quality in industrial areas based on remote sensing collaboration. It is applicable to water bodies with spatiotemporally dynamic optical characteristics, and is particularly suitable for complex scenarios such as industrial areas, sewage outlets, and tidal estuaries. Taking a typical industrial area water body as an example, this invention typically employs multispectral remote sensing images from optical satellite sensors, along with water quality parameters synchronously collected by ground-based automatic monitoring stations deployed in that area, to provide a detailed explanation of the technical solution.

[0021] Reference Figure 1 The online monitoring method for multiple parameters of water quality in industrial areas based on remote sensing collaboration includes steps S1-S3, as detailed below: S1: Collect and preprocess multispectral data from optical satellite sensors and water quality parameters collected synchronously from ground automatic monitoring stations. Based on the collected data, construct multiple effective water body grids and fit the baseline extrapolation coefficients and intercept terms.

[0022] Water quality inversion in industrial zone waters requires both extensive spectral information from satellite remote sensing and accurate concentration benchmarks from ground monitoring stations. However, raw remote sensing images contain interference from atmospheric scattering and cloud cover, resulting in temporal and spatial misalignments between ground data and imagery, and inconsistencies in the numerical dimensions between different sensors. Without preprocessing, a reliable optical-concentration mapping relationship cannot be established for subsequent calculations.

[0023] Therefore, it is necessary to preprocess the collected multi-source data and use the collected multi-source data to fit a baseline extrapolation coefficient and intercept term of a static linear model, which can serve as the starting point for subsequent dynamic adjustments.

[0024] The data collection process described above is as follows: The satellite uses optical sensors to read raw digital quantization values ​​in the green, red, and near-infrared bands at the moment of satellite transit; and ground-based automatic monitoring stations deployed around the water area collect water quality parameters, such as chlorophyll concentration and suspended solids concentration, synchronously with the satellite transit.

[0025] Then, establish the row and column numbers. The identified two-dimensional grid coordinate system, in which For line numbers, The column number is used to locate the spatial position of each pixel in the multispectral image.

[0026] In addition, by cross-checking the satellite and ground station clocks, only image-measured raw data pairs with time differences within the allowable range are retained. This operation ensures that the satellite spectral signals used for subsequent modeling and the ground water quality samples come from the same water body state at the same time, avoiding spectral-concentration mismatches caused by time asynchrony, thus ensuring that the subsequently fitted baseline coefficients are true and effective.

[0027] The specific preprocessing steps described above are as follows: The raw digital quantization values ​​collected by the optical satellite sensor are converted into reflectance using radiometric calibration, and then the influence of atmospheric aerosol scattering is subtracted using the dark pixel method to obtain green light reflectance, red light reflectance, and near-infrared reflectance.

[0028] The inverse distance weighted interpolation method is used to diffuse the measured concentration values ​​collected by each monitoring station into each grid in the two-dimensional grid coordinate system to obtain the reference concentration of each grid. At the same time, grids with near-infrared reflectance higher than a set threshold (e.g., 0.15) are identified as land or clouds and removed, and only water pixels are retained as effective water grids.

[0029] Furthermore, for water quality parameters such as chlorophyll concentration, red light bands are selected as characteristic bands, and green light bands are selected as reference bands. For each effective water body grid, its normalized band ratio is calculated. That is, the difference between the reflectance of the characteristic band and the reflectance of the reference band is used as the first parameter, and the sum of the reflectance of the characteristic band and the reflectance of the reference band is used as the second parameter. The ratio of the first parameter to the second parameter is calculated to obtain the normalized band ratio of the effective water body grid.

[0030] Furthermore, the normalized band ratio of each effective water body grid is paired with the reference concentration of that effective water body grid. Using the normalized band ratio as the independent variable and the reference concentration as the dependent variable, a straight line is fitted using the least squares method, satisfying the following relationship: In the formula, For reference concentration, The reference coefficient represents the magnitude of the change in water concentration when the normalized band ratio changes by one unit. The normalized band ratio of the effective water body grid. For the intercept term, it represents when When the value is 0, it means that the reflectance of the characteristic band and the reference band are equal, which represents the basic background concentration in the water body.

[0031] In summary, the parameters of the static linear model are finally obtained. and It outputs the preprocessed reflectance grid, that is, each effective water body grid corresponds to a set of green light reflectance, red light reflectance and near-infrared reflectance.

[0032] S2: Calculate the target water quality concentration for each effective water body grid based on the baseline extrapolation coefficient and intercept term.

[0033] Because the static parameters fitted by S1 above—the baseline extrapolation coefficients and intercept terms—are global average mapping relationships obtained by least-squares fitting across all effective water body grids in the entire image, they cannot reflect the spatial variations and instantaneous disturbances in the optical properties of water bodies in industrial areas. In local areas such as sewage outlets and tidal fronts, the fixed-coefficient model will produce significant extrapolation errors. Therefore, a dynamic adjustment mechanism is needed to enable the extrapolation model to adaptively adjust its sensitivity based on the actual optical state of each grid, thereby accurately calculating the target water quality concentration for each effective water body grid.

[0034] Reference Figure 2 The calculation process for the target water quality concentration of a single effective water body grid includes steps S20-S22, and the specific calculation process is as follows: S20: Select any effective water body grid as the target grid and calculate the optical state gradient index of the target grid.

[0035] To assess the true disturbance intensity of the effective water grid, the local relative change rate is first calculated to detect the spatial abrupt change in reflectance, while spectral confidence weights are calculated to identify interference from clouds, aerosols, and other sources. Then, the optical state gradient index of the effective water grid is calculated by combining the local relative change rate and the spectral confidence weights.

[0036] First, any valid water body grid is taken as the target grid. Since non-water body grids such as land and clouds were removed in S1 above, the neighbors of some water body edge grids may be marked as invalid. Therefore, for the target grid, only when both its horizontal left and right neighbors are valid water body grids, the difference between the green light reflectance of the target grid's right neighbor and the green light reflectance of its left neighbor is calculated to obtain the horizontal reflectance difference of the target grid. Similarly, the vertical reflectance difference of the target grid is calculated.

[0037] Furthermore, the horizontal and vertical reflectance differences of the target mesh are assigned their own validity flags: when both horizontal neighbors of the target mesh are valid water bodies, its horizontal validity flag is defined as 1; otherwise, it is 0. Similarly, the vertical validity flag is either 0 or 1. This validity flag is used to avoid invalid differences caused by land cover or cloud cover.

[0038] Furthermore, by comprehensively evaluating the magnitude of reflectance changes in both the horizontal and vertical directions, a composite change is calculated. The ratio of this composite change to the green light reflectance of the target mesh is taken as the local relative change rate of the target mesh.

[0039] The above-mentioned synthetic changes satisfy the following relationship: In the formula, The synthetic variation of the target mesh. For horizontal validity identification, The difference in horizontal reflectivity of the target mesh. For vertical validity identification, This represents the difference in vertical reflectivity of the target mesh. and Ensure that contributions are only counted when neighbors are valid, to prevent spurious gradients from occurring across invalid regions.

[0040] Then, the absolute value of the difference between the red light reflectance and the near-infrared reflectance of the target grid is calculated as the third parameter. This third parameter is divided by a pre-defined reflectance difference attenuation parameter to obtain the fourth parameter. This fourth parameter is then mapped to obtain the spectral confidence weight. When the red light and near-infrared reflectances are very close, the spectral confidence weight is close to 1, indicating that the spectral characteristics of the target grid conform to typical water bodies, and its gradient changes are reliable. When the difference between the two is large, the spectral confidence weight rapidly approaches 0, thus effectively suppressing interference signals such as clouds and aerosols. The reflectance difference attenuation parameter can be determined based on historical statistics of pure water bodies.

[0041] The above spectral confidence weights satisfy the following relationship: In the formula, For spectral confidence weights, The fourth parameter, It is an exponential function with the natural constant e as its base.

[0042] Finally, a 3×3 target window is obtained centered on the target grid. The mean of the sum of the products of the local relative change rates and spectral confidence weights of all effective water grids within the target window is calculated to obtain the optical state gradient index of the target grid. The smaller the optical state gradient index, the more stable the water body.

[0043] S21: Using the optical state gradient index of the target mesh as known input, generate the dynamic correction coefficients of the target mesh.

[0044] The optical state gradient index calculated by S20 above reflects the intensity of local water disturbance, but it cannot be directly used for concentration estimation. A global transition scaling parameter needs to be introduced to control how quickly the sensitivity changes with the optical state gradient index.

[0045] Then, based on the global transition scaling parameters and combined with the optical state gradient index, the dynamic correction coefficient of the target mesh is calculated.

[0046] For each ground monitoring station, obtain the reference concentration of the effective water grid where the station is located and the measured concentration of the station. Then, calculate the absolute value of the difference between the reference concentration and the measured concentration of the station, and average the absolute values ​​of the differences between the reference concentration and the measured concentration of all stations to obtain a global transition scaling factor.

[0047] Furthermore, the optical state gradient threshold can be set by taking a relatively high quantile, such as the 90th percentile, based on the statistical distribution of the optical state gradient during historical normal periods.

[0048] For the target grid, when its optical state gradient index is lower than the set optical state gradient threshold, the water body is considered to be stable, and the calculation sensitivity should be reduced to avoid overreacting to normal fluctuations; when its optical state gradient index is higher than the set optical state gradient threshold, the water body is considered to have abrupt changes such as a sewage front, and the calculation sensitivity should be increased to accurately capture concentration changes; when its optical state gradient index is equal to the set optical state gradient threshold, the model can maintain the baseline sensitivity.

[0049] The dynamic correction coefficients described above satisfy the following relationship: In the formula, For the target mesh, (This refers to the dynamic correction coefficients.) The coefficients are calculated based on the baseline. This is the global transition scaling factor. The optical state gradient index of the target mesh. The optical state gradient threshold. It is the hyperbolic tangent function.

[0050] S22: Calculate the target water quality concentration of the target grid based on the dynamic correction coefficient of the target grid.

[0051] The aforementioned S21 has generated dynamic correction coefficients for the target grid. These coefficients adaptively adjust the estimation sensitivity based on the intensity of local optical disturbances in the water body. However, the dynamic correction coefficients themselves are merely scaling factors and need to be combined with spectral characteristics that reflect water quality components to ultimately calculate the water concentration. Furthermore, the raw reflectance data is affected by external factors such as sunlight, atmospheric transmission, and water surface reflection; using it directly can lead to unstable concentration estimations.

[0052] Therefore, it is necessary to use the normalized band ratio in S1 above to eliminate these external interferences, and then multiply it by the dynamic correction coefficient and add the background concentration, i.e., the intercept term calculated in S1, to obtain the accurate target water quality concentration.

[0053] The above multiplication of the normalized band ratio of the target grid with the dynamic correction coefficient of the target grid can be scaled according to the current optical disturbance intensity to achieve dynamic sensitivity adjustment; further, the product is added to the intercept term to compensate for the basic background concentration that still exists in the water body when the normalized band ratio approaches 0.

[0054] The target water quality concentration of the final output target grid is not affected by external light. In the sewage mixing zone, the response is improved due to the increase of the dynamic correction coefficient, while in the calm water area, it remains stable due to the decrease of the dynamic correction coefficient.

[0055] Then, based on the operations in S20-S22 above, the target water quality concentrations for all effective water grids can be obtained.

[0056] S3: Spatial reconstruction and filtering of the target concentration of each effective water body grid are performed to monitor the water quality of the industrial area.

[0057] The target water quality concentration of each effective water body grid is arranged into a two-dimensional matrix according to the original row and column positions of the two-dimensional grid coordinate system established in S1 above. Outliers are removed by applying a 3×3 window median filter to this two-dimensional matrix. After filtering, a water quality parameter distribution map is generated and transmitted to the monitoring platform.

[0058] Furthermore, based on the accumulated ground monitoring data of the industrial zone's waters, such as the 75th percentile of chlorophyll concentration or suspended solids concentration records from the past year, a concern threshold is set; then, an early warning threshold is determined based on the limits for the corresponding water quality category in the surface water environmental quality standards.

[0059] Furthermore, the comparison is performed grid-by-grid based on the set attention thresholds, warning thresholds, and target water quality concentrations: If the target water quality concentration of an effective water body grid is less than the threshold of interest, the effective water body grid is determined to be in a normal state, and only data archiving and recording are performed.

[0060] If the target water quality concentration of an effective water body grid is greater than or equal to the attention threshold but less than the warning threshold, the effective water body grid is determined to be in a state of attention, the highlighted area is marked in the distribution map, and the monitoring frequency of the grids adjacent to the location of the effective water body grid is automatically increased.

[0061] If the target water quality concentration of an effective water body grid is greater than or equal to the warning threshold, the effective water body grid is determined to be in a warning state, and a graded alarm work order is generated. The work order includes the coordinates of the location of the exceedance, the concentration value, the type of the exceedance parameter, and the suggested on-site verification time. The work order is then pushed to the terminal and the monitoring screen via API (Application Programming Interface).

[0062] This completes the online monitoring of multiple water quality parameters in the industrial area.

[0063] This invention also provides an online monitoring system for multiple parameters of water quality in industrial areas based on remote sensing collaboration. For example... Figure 3 As shown, the system includes a processor and a memory. The memory stores computer program instructions. When the computer program instructions are executed by the processor, the method for online monitoring of multiple parameters of water quality in industrial areas based on remote sensing collaboration according to the first aspect of the present invention is implemented.

[0064] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0065] It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept, and these all fall within the scope of protection of this invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for online monitoring of multiple parameters of water quality in industrial areas based on remote sensing collaboration, characterized in that, include: Multispectral data from optical satellite sensors and water quality parameters collected synchronously from ground automatic monitoring stations were acquired and preprocessed. Multiple effective water body grids were constructed based on the acquired data, and baseline extrapolation coefficients and intercept terms were fitted. Select any effective water body grid as the target grid, and calculate the optical state gradient index of the target grid. The optical state gradient index is used to reflect the magnitude of reflectance change of the target grid in the horizontal and vertical directions and the reliability of the spectral characteristics. Using the optical state gradient index as input, and combining the preset global transition scaling factor and optical state gradient threshold, a dynamic correction factor adapted to the current water disturbance intensity is generated. The sensitivity of the benchmark estimation coefficient is adjusted using the dynamic correction coefficient, and the target water quality concentration is calculated by combining the spectral characteristics of the target grid. A water quality parameter distribution map is generated based on the distribution characteristics of the target water quality concentration, and anomaly status is determined, thereby realizing dynamic monitoring of the entire water body in the industrial area.

2. The method for online monitoring of multiple parameters of water quality in industrial areas based on remote sensing collaboration as described in claim 1, characterized in that, Multiple effective water body grids were constructed based on the collected data, including: The raw digital quantization values ​​collected by the optical satellite sensor are converted to obtain green light reflectance, red light reflectance, and near-infrared reflectance; The inverse distance weighted interpolation method is used to diffuse the measured concentration values ​​collected by each monitoring station to each grid in the two-dimensional grid coordinate system to obtain the reference concentration of each grid. Grids with near-infrared reflectivity higher than a set threshold are removed, and only water pixels are retained as valid water grids.

3. The method for online monitoring of multiple parameters of water quality in industrial areas based on remote sensing collaboration as described in claim 2, characterized in that, The acquisition of the baseline extrapolation coefficients and intercept terms includes: For water quality parameters, the red light band was selected as the characteristic band, and the green light band was selected as the reference band. For each effective water body grid, calculate the normalized band ratio of its characteristic band reflectance to the reference band reflectance; The normalized band ratio of each effective water body grid is paired with the reference concentration of that effective water body grid. The normalized band ratio is used as the independent variable and the reference concentration is used as the dependent variable. The least squares method is used to fit a straight line to obtain the static linear model parameters, namely the baseline extrapolation coefficients and the intercept term.

4. The method for online monitoring of multiple parameters of water quality in industrial areas based on remote sensing collaboration as described in claim 3, characterized in that, The calculation of the optical state gradient index of the target mesh includes: For the target grid, the reflectance difference in the horizontal and vertical directions is calculated only when its left and right neighbors in the horizontal direction and its upper and lower neighbors in the vertical direction are all effective water grids. The composite change is calculated in combination with the directional validity indicator, and the ratio of the composite change to the reflectance of the reference band of the target grid is used as the local relative change rate. Calculate the absolute value of the difference between the characteristic band reflectance and the near-infrared reflectance of the target grid, and obtain the spectral confidence weight by exponential transformation of the absolute value of the difference; A target window of a preset size is obtained with the target grid as the center. The mean of the sum of the products of the local relative change rate and the spectral confidence weight of all effective water grids within the target window is calculated to obtain the optical state gradient index.

5. The method for online monitoring of multiple parameters of water quality in industrial areas based on remote sensing collaboration according to claim 4, characterized in that, The generation of the dynamic correction coefficient includes: The mean value of the absolute difference between the reference concentration and the measured concentration in the grid where all ground monitoring stations are located is obtained and used as the global transition scaling factor. An optical state gradient threshold is set. When the optical state gradient index of the target mesh is lower than the optical state gradient threshold, a dynamic correction coefficient smaller than the baseline calculated coefficient is generated using the hyperbolic tangent function. When the optical state gradient index of the target mesh is higher than the optical state gradient threshold, a dynamic correction coefficient larger than the baseline calculated coefficient is generated.

6. The method for online monitoring of multiple parameters of water quality in industrial areas based on remote sensing collaboration as described in claim 5, characterized in that, The calculation of the target water quality concentration includes: Multiply the normalized band ratio of the target grid by the dynamic correction coefficient to obtain the corrected band ratio. The target water quality concentration of the target grid is calculated by adding the corrected band ratio to the pre-fitted intercept term.

7. The method for online monitoring of multiple parameters of water quality in industrial areas based on remote sensing collaboration according to claim 6, characterized in that, The abnormal state determination includes: A preset warning threshold is set. If the target water quality concentration of any effective water body grid exceeds the warning threshold, the effective water body grid is determined to be in a warning state, and an alarm work order containing the location and concentration value of the exceeding standard is generated and pushed to the monitoring platform.

8. A multi-parameter online monitoring system for water quality in industrial areas based on remote sensing collaboration, characterized in that, include: The processor and memory, wherein the memory stores computer program instructions that, when executed by the processor, implement the online multi-parameter monitoring method for water quality in industrial areas based on remote sensing collaboration as described in any one of claims 1-7.

Images