Hyperspectral-based near-sensitive lake and reservoir water quality and water bloom short-term prediction and early warning method and system

Abstract

The invention discloses a short-term prediction and early-warning method and system for water blooms of water quality of near-sensitive lakes and reservoirs based on hyperspectrum, and belongs to the field of water quality prediction and early-warning technical methods. Combining the preprocessed water body proximity hyperspectral reflectivity with the time-space attribute information to generate a water body proximity hyperspectral reflectivity-time sequence data set; constructing a preset water quality inversion model set according to a space-time matching principle in combination with synchronously measured water quality parameters; generating a multi-water quality parameter-time sequence data set by using a preset water quality inversion model set; and building a prediction and early warning model based on the multiple water quality parameter-time sequence data set, and outputting a water quality and algal bloom prediction and early warning result. Measurement data are transmitted to a cloud server in real time, hyperspectral reflectivity data are converted into water quality parameters easy to understand by a user by constructing a water quality inversion model set, and the data are integrated on a software platform based on multi-parameter water quality; and current water quality information, historical long-term change and future short-term algal bloom prediction and early warning information can be checked anytime and anywhere.

Description

Technical Field

[0001] The present invention belongs to the technical field of water environment monitoring and early warning, and specifically relates to a method and system for short-term prediction and early warning of water quality and algal bloom in lakes and reservoirs based on hyperspectral near-sensing. Background Art

[0002] High-frequency monitoring and accurate prediction of water quality and algal blooms are important bases and cornerstones for grasping the dynamic characteristics of lake and reservoir water environment, analyzing the changing trends, and conducting causal mechanism analysis and scientific prevention and control of cyanobacterial blooms. However, traditional manual inspections lag behind the needs of water environment management and decision-making departments in terms of data collection frequency, timeliness and representativeness; satellite and UAV remote sensing is limited by spatiotemporal and spectral resolution, cloud and rain weather and atmospheric correction, and is unable to solve the problem of continuous water quality and algal bloom monitoring; high-frequency underwater probes are limited by high cost, easy pollution, low precision and difficult calibration, and it is difficult to guarantee monitoring accuracy and continuity. Therefore, existing water quality and algal bloom monitoring methods face challenges in capturing sudden water pollution events and rapid water ecological disturbances caused by extreme weather. In addition, cyanobacterial blooms are the result of the coupling of physical, chemical and biological factors of water bodies with meteorological and hydrological factors. The mechanism is complex and there are many influencing factors. Traditional cyanobacterial bloom prediction models based on ecological dynamics have problems such as complex form, numerous parameters and unclear mechanisms. Statistical models based on water environment data have unsatisfactory prediction model accuracy and low levels of automation and intelligence due to their simple form and diverse parameters, as well as poor continuity of input water environment data and mainly daily manual and satellite inversion data.

[0003] In response to the problems existing in the monitoring and prediction of water quality and algal blooms in lakes and reservoirs under the dual influence of current high-intensity human activities and rapid climate change, we have developed and utilized original, minute-level hyperspectral near-sensing (non-contact near-ground sensing) water quality meters, coupled multi-scene measured water quality data, upgraded the high-precision inversion model of key water quality parameters of lakes and reservoirs, achieved high-frequency monitoring of water quality and algal blooms under complex weather and water conditions, introduced deep learning algorithms with powerful learning and nonlinear approximation capabilities, and constructed a real-time prediction and early warning model for water quality and algal blooms driven by big data, to support the precise prevention and control of cyanobacteria blooms and scientific supervision of water environment, and to improve the level of intelligent ecological environmental monitoring and disaster emergency response capabilities. Summary of the invention

[0004] In order to solve the technical problems existing in the above-mentioned background technology, the present invention provides a method and system for short-term prediction and early warning of algal bloom based on hyperspectral near-sensing lake water quality.

[0005] The present invention adopts the following technical scheme: a short-term prediction and early warning method for water quality and algal bloom in lakes and reservoirs based on hyperspectral near-sensing data, comprising the following steps: According to the spatiotemporal attribute information, the near-sensing hyperspectral reflectance of the water body is collected in real time and preprocessed to obtain the preprocessed near-sensing hyperspectral reflectance of the water body; the preprocessed near-sensing hyperspectral reflectance of the water body is combined with the spatiotemporal attribute information to generate a near-sensing hyperspectral reflectance-time series data set of the water body; the spatiotemporal attribute information at least includes: spatial information, time information and water body type of the monitoring point; According to the water body type at the monitoring point, the corresponding water body near-sensing hyperspectral reflectance is screened out, and combined with the synchronously measured water quality parameters, a preset water quality inversion model set related to the water body type at the monitoring point is constructed according to the principle of time-space matching; the water body near-sensing hyperspectral reflectance-time series data set is used as the input variable, and the measured water quality parameters are used as the output dependent variable, and the preset water quality inversion model set is used to generate a multi-water quality parameter-time series data set; Based on the multi-water quality parameter-time series data set, a multi-time scale statistical analysis is performed to screen out the key multi-water quality parameter-time series data set and build a prediction and early warning model; the pre-processed near-sensing hyperspectral reflectance of the water body is used as an input feature to output a water quality and algal bloom prediction and early warning result; the key multi-water quality parameter-time series data set is updated according to the water quality and algal bloom prediction and early warning result, and the prediction and early warning model is optimized; Receive user queries for forecast and warning requirements at different time scales on the web terminal, and repeat the above steps to update and display the output results in real time.

[0006] In a further embodiment, the collection process of collecting the near-sensing hyperspectral reflectance of the water body is as follows: Using the transmittance The transmission plate obtains the downward irradiance on the water surface : ;in, is the irradiance of the transmission plate; The following formula is used to obtain the near-sensing hyperspectral reflectance of the collected water body: : ; In the formula, is the water-leaving irradiance of the transmission plate.

[0007] In a further embodiment, the preprocessing steps of the preprocessed water body near-sensing hyperspectral reflectance are as follows: Draw a spectral curve based on the near-sensing hyperspectral reflectance of the water body, regard the spectral curve as a vector, and calculate the modulus length of each vector : ;in, , To collect near-sensing hyperspectral reflectance of water bodies in real time based on spatiotemporal attribute information; The normalized near-sensitivity hyperspectral reflectance is obtained by calculating the ratio of each wavelength to the mode length using the following formula: ; The normalized near-sensing hyperspectral reflectance is calculated using the following formula: To perform a linear correction: ;in, is the near-sensing hyperspectral reflectance after radiation correction, represents the radiation correction factor, is the radiation correction slope.

[0008] In a further embodiment, the types of water bodies at the monitoring points include at least: clean reservoirs, natural rivers, urban rivers and lakes; The synchronously measured water quality parameters include at least: total nitrogen, total phosphorus, chlorophyll a, transparency, water temperature, dissolved oxygen, permanganate index, turbidity and total suspended solids concentration; The preset water quality inversion model set includes at least: a near-sensing hyperspectral total phosphorus inversion model, a near-sensing hyperspectral total nitrogen inversion model, a near-sensing hyperspectral transparency inversion model, a near-sensing hyperspectral total suspended matter concentration inversion model, a near-sensing hyperspectral water temperature inversion model and a near-sensing hyperspectral dissolved oxygen inversion model.

[0009] In a further embodiment, the time-space matching principle includes at least: a time matching principle and a space matching principle; The time matching principle is expressed as follows: The collection time point of the near-sensing hyperspectral reflectance of the water body is defined as , the detection time point of synchronous water quality parameters is , then the collection time point and detection time point The relationship should be as follows: ; The spatial matching principle includes: longitude and latitude matching principle and distance matching principle; the longitude and latitude matching principle is expressed as follows: ,in, is the longitude and latitude of the collection point of the near-sensing hyperspectral reflectance of the water body, Represents the dimension reduction of the detection points of simultaneous water quality parameters; The distance matching principle is expressed as follows: , where is the standard observation range radius, Indicates the horizontal distance between the water sample collection point and the near-sensing hyperspectral instrument monitoring center; , The near-sensing hyperspectral instrument is placed at a height above the water surface. Indicates the field of view.

[0010] In a further embodiment, the multiple time scales include at least: an hourly scale and a daily scale, and the key multiple water quality parameter-time series data set includes: hourly water quality parameters and daily water quality parameters; the screening process of the key multiple water quality parameter-time series data set is as follows: introducing the correlation between the hourly water quality parameters and the synchronous chlorophyll a concentration, and the daily water quality parameters and the synchronous chlorophyll a concentration , through the correlation The value of is used to determine whether the corresponding water quality parameter is a key parameter: ; In the formula, is the Pearson correlation coefficient, is the significance level; If the correlation The value of is 1, which means that the multi-water quality parameter is significantly correlated with the synchronous chlorophyll a concentration and is a key water quality parameter and is updated to the key multi-water quality parameter-time series data set; if the correlation If the value of is 0, it means that the multi-water quality parameter is weakly correlated with the synchronous chlorophyll a concentration and is not a key water quality parameter.

[0011] In a further embodiment, the analysis process of the prediction and early warning model is as follows: Define the sliding window as , the sliding step length is , respectively calculate the average values ​​of multiple water quality parameters within the optimal sliding window , Standard Deviation and median , analyze the water quality parameters at the center of the optimal sliding window : like , then the water quality and algal bloom forecast warning result indicates that the water quality is normal and no warning treatment is performed; like , then the water quality and algal bloom prediction and warning results indicate that the water quality is abnormal, and an alarm message is issued. In a further embodiment, the matching logic relationship of the time-space matching principle is as follows: Assign the logical judgment result of the time-space matching principle to , when the time matching principle and the space matching principle are satisfied at the same time, then , then the corresponding synchronously measured water quality parameters and water body near-sensing hyperspectral reflectance are added to the data for model construction; On the contrary, when at least one of the time matching principle and the space matching principle is not satisfied, 0 is assigned to , , then delete the corresponding synchronously measured water quality parameters and water body near-sensing hyperspectral reflectance; .

[0012] In a further embodiment, it also includes: Corresponding water quality parameters Remove from the key multi-water quality parameter-time series data set and use the median Replacement; The multiple water quality parameter-time series data set was smoothed using the Savitzky-Golay filter to reduce the noise in the acquisition process, and a fifth-order polynomial curve was fitted and filled using the nearest neighbor interpolation method to obtain the updated key multiple water quality parameter-time series data set.

[0013] The hyperspectral near-sensing lake water quality and algae bloom short-term prediction and early warning system is used to implement the water quality and algae bloom short-term prediction and early warning method as described above, including: The first module is configured to collect the near-sensing hyperspectral reflectance of the water body in real time according to the spatiotemporal attribute information and obtain the preprocessed near-sensing hyperspectral reflectance of the water body through preprocessing; the preprocessed near-sensing hyperspectral reflectance of the water body is combined with the spatiotemporal attribute information to generate a near-sensing hyperspectral reflectance-time series data set of the water body; the spatiotemporal attribute information at least includes: spatial information, time information and water body type of the monitoring point; The second module is configured to screen out the corresponding near-sensing hyperspectral reflectance of the water body according to the water body type of the monitoring point, and to build a preset water quality inversion model set related to the water body type of the monitoring point according to the principle of time-space matching in combination with the synchronously measured water quality parameters; the near-sensing hyperspectral reflectance-time series data set of the water body is used as the input variable, and the measured water quality parameters are used as the output dependent variable, and the preset water quality inversion model set is used to generate a multi-water quality parameter-time series data set; The third module is configured to perform multi-time scale statistical analysis based on the multi-water quality parameter-time series data set, screen out key multi-water quality parameter-time series data sets, and build a prediction and early warning model; use the pre-processed near-sensing hyperspectral reflectance of the water body as an input feature, and output a water quality and algal bloom prediction and early warning result; update the key multi-water quality parameter-time series data set according to the water quality and algal bloom prediction and early warning result, and optimize the prediction and early warning model; The fourth module is configured to receive user queries on different time scales of forecast and warning requirements on the web terminal, and repeat the above steps to update and display the output results in real time.

[0014] Beneficial effects of the present invention: The present invention uses a near-sensing hyperspectral instrument to achieve real-time high-frequency acquisition of water body hyperspectral data, and transmits the measured data to a cloud server in real time for processing; by constructing a water quality inversion model set, the hyperspectral reflectance data is converted into water quality parameters (such as total nitrogen, chlorophyll content, dissolved oxygen, etc.) that are easy for users to understand, based on multi-parameter water quality and integrating these data on a software platform; Users only need to install the software platform on their terminals to view current water quality information, historical long-term changes, and future short-term algal bloom forecast and warning information anytime and anywhere, without having to obtain information through multiple channels or complex processes. This water quality monitoring method not only improves the efficiency and accuracy of information acquisition, but also reduces monitoring costs, expands the monitoring scope, and provides strong technical support for water quality monitoring and environmental protection work. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flow chart of the short-term prediction and early warning method for algal bloom in lakes and reservoirs based on hyperspectral near-sensing water quality. DETAILED DESCRIPTION

[0016] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0017] Example 1 Blue algae blooms are the result of the coupling of physical, chemical, biological, meteorological and hydrological factors in water bodies. The mechanism is complex and there are many influencing factors. The traditional blue algae bloom prediction model based on ecological dynamics has problems such as complex form, numerous parameters and unclear mechanisms. The statistical model based on water environment data has unsatisfactory prediction accuracy due to its simple form and diverse parameters, and the poor continuity of input water environment data, and the low level of automation and intelligence.

[0018] In order to solve this problem, this embodiment provides a short-term prediction and early warning method for algal bloom based on hyperspectral near-sensing lake water quality, such as Figure 1 The following steps are shown: Based on multiple platforms including land-based, ship-based and tower-based, a near-sensing spectral monitor is placed on the water surface, and the near-sensing hyperspectral reflectivity of the water body is collected in real time according to the spatiotemporal attribute information and preprocessed to obtain the preprocessed near-sensing hyperspectral reflectivity of the water body; the preprocessed near-sensing hyperspectral reflectivity of the water body is combined with the spatiotemporal attribute information to generate a near-sensing hyperspectral reflectivity-time series data set of the water body; the spatiotemporal attribute information at least includes: spatial information, time information and the type of water body at the monitoring point. In a further embodiment, the spatial information can be understood as the geographical longitude and latitude information, and the time information can be the year, month, day, hour, minute and second when the near-sensing hyperspectral reflectivity is collected. In order to obtain stable and reliable near-sensing hyperspectral reflectance of water bodies, strict requirements are imposed on the installation environment of the near-sensing hyperspectral spectrum: 1) There are no trees, buildings or shadows around the equipment; 2) There are no tall white walls, reflective glass or other objects that may change due to its spectrum within a horizontal range of 2 meters; 3) The water depth of the spectral collection area is greater than the transparency of the water body, and the bottom cannot be seen; 4) There are no white waves, foam, aquatic vegetation or other non-water objects that are fixed in the spectral collection area for a long time; 5) The near-sensing hyperspectral lens should be at an angle of 135°~180° to the plane where the sun is located; 6) The near-sensing hyperspectral instrument is placed at an observation distance of about 2~10 meters above the water surface.

[0019] According to the water body type at the monitoring point, the corresponding water body near-sensing hyperspectral reflectance is screened out, and combined with the synchronously measured water quality parameters, a preset water quality inversion model set related to the water body type at the monitoring point is constructed according to the principle of time-space matching; the water body near-sensing hyperspectral reflectance-time series data set is used as the input variable, and the measured water quality parameters are used as the output dependent variable, and the preset water quality inversion model set is used to generate a multi-water quality parameter-time series data set; Based on the multi-water quality parameter-time series data set, a multi-time scale statistical analysis is performed to screen out the key multi-water quality parameter-time series data set and build a prediction and early warning model; the pre-processed near-sensing hyperspectral reflectance of the water body is used as an input feature to output a water quality and algal bloom prediction and early warning result; the key multi-water quality parameter-time series data set is updated according to the water quality and algal bloom prediction and early warning result, and the prediction and early warning model is optimized; Receive user queries for forecast and warning requirements at different time scales on the web terminal, and repeat the above steps to update and display the output results in real time.

[0020] Based on the above description, the collection process of water body near-sensing hyperspectral reflectance is as follows: Using the transmittance The transmission plate obtains the downward irradiance on the water surface : ;in, is the irradiance of the transmission plate; The following formula is used to obtain the near-sensing hyperspectral reflectance of the collected water body: : ; In the formula, is the water-leaving irradiance of the transmission plate.

[0021] It is worth mentioning that the hyperspectrometer is tilted at an angle of 45° to the horizontal direction, with an azimuth angle between 90 and 135°. The lens is loaded with a polarizer to filter out stray light from other directions of the environment and the water quality, thereby improving the signal-to-noise ratio. The high-definition video camera is a circular camera that can achieve 360° horizontal rotation and 180° vertical rotation, which can capture and record the current water environment conditions and query historical water surface conditions, and assist in confirming and judging abnormal water environment conditions.

[0022] Considering that the surface optical characteristics of outdoor water bodies change with the lighting conditions and are affected by wind and waves, solar altitude angle, light intensity and other factors, it is necessary to perform preprocessing such as normalization and radiation correction on the near-sensing hyperspectral water quality reflectance data to eliminate the influence of different weather, wind speed and light intensity on the amplitude of the spectral curve. The preprocessing includes at least normalization and radiation correction; the normalization adopts the ratio normalization method; the radiation correction adopts the cross calibration method and uses a linear function for calibration.

[0023] Furthermore, the preprocessing steps of the preprocessed water body near-sensing hyperspectral reflectance are as follows: Draw a spectral curve based on the near-sensing hyperspectral reflectance of the water body, regard the spectral curve as a vector, and calculate the modulus length of each vector : ;in, , To collect near-sensing hyperspectral reflectance of water bodies in real time based on spatiotemporal attribute information; The normalized near-sensitivity hyperspectral reflectance is obtained by calculating the ratio of each wavelength to the mode length using the following formula: : ; The normalized near-sensing hyperspectral reflectance is calculated using the following formula: To perform a linear correction: ;in, is the near-sensing hyperspectral reflectance after radiation correction, represents the radiation correction factor, is the radiation correction slope. In this embodiment, The value of is 1.086, The value of is 0.0054.

[0024] In another embodiment, the water body types at the monitoring points include at least: clean reservoirs, natural rivers, urban rivers and lakes. Correspondingly, the synchronously measured water quality parameters include at least: total nitrogen, total phosphorus, chlorophyll a, transparency, water temperature, dissolved oxygen, permanganate index, turbidity and total suspended solids concentration. Correspondingly, the multi-water quality parameter-time series data set is a parameter-time series data set about nitrogen, total phosphorus, chlorophyll a, transparency, water temperature, dissolved oxygen, permanganate index, turbidity and total suspended solids concentration.

[0025] The preset water quality inversion model set includes at least: a near-sensing hyperspectral total phosphorus inversion model, a near-sensing hyperspectral total nitrogen inversion model, a near-sensing hyperspectral transparency inversion model, a near-sensing hyperspectral total suspended solids concentration inversion model, a near-sensing hyperspectral water temperature inversion model, and a near-sensing hyperspectral dissolved oxygen inversion model. Among them, the models in the preset water quality inversion model set are based on attribute information to screen out near-sensing hyperspectral reflectance data of urban rivers, natural rivers, reservoirs and natural lakes in the near-sensing hyperspectral time series data set of water bodies, and add water body attribute type labels; according to the principle of time-space matching, the synchronous measured water quality parameters are searched to generate a hyperspectral water quality quasi-synchronous data set; according to the 3:1 distribution principle, the hyperspectral in the generated modeling data set and the verification data set is used as the input feature, and the synchronous water quality parameters are used as the output feature, and are input into the machine learning algorithm to construct a multi-parameter water quality inversion model. The machine learning algorithm in this embodiment can be an extreme gradient boosting tree algorithm, a support vector machine, a random forest algorithm, a deep neural network or others.

[0026] In another embodiment, the time-space matching principle includes at least: a time matching principle and a space matching principle; The time matching principle is expressed as follows: The collection time point of the near-sensing hyperspectral reflectance of the water body is defined as , the detection time point of synchronous water quality parameters is , then the collection time point and detection time point The relationship should be as follows: ; The spatial matching principle includes: longitude and latitude matching principle and distance matching principle; the longitude and latitude matching principle is expressed as follows: ,in, is the longitude and latitude of the collection point of the near-sensing hyperspectral reflectance of the water body, Represents the dimension reduction of the detection points of simultaneous water quality parameters; The distance matching principle is expressed as follows: , where is the standard observation range radius, Indicates the horizontal distance between the water sample collection point and the near-sensing hyperspectral instrument monitoring center; , The near-sensing hyperspectral instrument is placed at a height above the water surface. Indicates the field of view.

[0027] Based on the above description, the matching logic relationship of the time-space matching principle is as follows: Assign the logical judgment result of the time-space matching principle to , when the time matching principle and the space matching principle are satisfied at the same time, then , then the corresponding synchronously measured water quality parameters and water body near-sensing hyperspectral reflectance are added to the data for model construction; On the contrary, when at least one of the time matching principle and the space matching principle is not satisfied, 0 is assigned to , , then delete the corresponding synchronously measured water quality parameters and water body near-sensing hyperspectral reflectance; .

[0028] In a further embodiment, the multiple time scales include at least: hourly scale and daily scale, and the key multiple water quality parameter-time series data set includes: hourly water quality parameters and daily water quality parameters. In view of the fact that the hourly multi-parameter water quality time series data set is continuous data, in this embodiment, the outlier processing adopts the 3 times standard deviation criterion, the missing value interpolation selects the median substitution method, or the smoothing processing adopts the Savitzky-Golay filter for smoothing processing to reduce the data anomalies and noise caused by wind and waves, solar altitude and flares.

[0029] The screening process of the key multi-water quality parameter-time series data set is as follows: Introduce the correlation between hourly water quality parameters and synchronous chlorophyll a concentration, daily water quality parameters and synchronous chlorophyll a concentration , through the correlation The value of is used to determine whether the corresponding water quality parameter is a key parameter: ; In the formula, is the Pearson correlation coefficient, is the significance level; If the correlation The value of is 1, which means that the multi-water quality parameter is significantly correlated with the synchronous chlorophyll a concentration and is a key water quality parameter and is updated to the key multi-water quality parameter-time series data set; if the correlation If the value of is 0, it means that the multi-water quality parameter is weakly correlated with the synchronous chlorophyll a concentration and is not a key water quality parameter.

[0030] It should be noted that since the growth and reproduction of phytoplankton is closely related to the water environment, the relationship between transparency, total nitrogen, total phosphorus, permanganate index and temperature and chlorophyll a based on the inversion of near-sensing hyperspectral data was analyzed. The results showed that both daily and hourly water quality parameters were significantly correlated with synchronous chlorophyll, among which the correlation with the permanganate index was the highest, followed by total nitrogen and then total phosphorus. However, the correlation between the hourly and daily inverted chlorophyll a concentration and other water quality parameters was greater than 0.3, creating a correlation between water quality parameters and synchronous chlorophyll a concentration. Based on the screening of the above key multi-water quality parameter-time series data set, the analysis process of the prediction and early warning model is as follows: Define the sliding window as , the sliding step length is ,like , . Calculate the average values ​​of multiple water quality parameters within the optimal sliding window , Standard Deviation and median , analyze the water quality parameters at the center of the optimal sliding window : like , then the water quality and algal bloom forecast warning result indicates that the water quality is normal and no warning treatment is performed; like , the water quality and algal bloom prediction and warning results indicate that the water quality is abnormal and an alarm message is issued.

[0031] Further, Corresponding water quality parameters Remove from the key multi-water quality parameter-time series data set and use the median Replacement; The multiple water quality parameter-time series data set was smoothed using the Savitzky-Golay filter to reduce the noise in the acquisition process, and a fifth-order polynomial curve was fitted and filled using the nearest neighbor interpolation method to obtain the updated key multiple water quality parameter-time series data set.

[0032] It is worth mentioning that the prediction and warning model in this embodiment is a 4-layer long short-term memory network model, namely an input layer, a hidden layer of two memory modules and a fully connected output layer, using 90% of the input key multi-water quality parameter time series data set for modeling, and the remaining 10% of the input key multi-water quality parameter time series data set for verification. Among them, the time series of key water quality parameters such as total nitrogen, total phosphorus, chlorophyll a, transparency, suspended solids, permanganate index, turbidity and extinction coefficient in the past time step are input features.

[0033] The chlorophyll a concentration time series of the predicted time step is used as the output feature, and the root mean square error and mean absolute percentage error between the output chlorophyll a and the measured chlorophyll a are calculated to evaluate the accuracy and robustness of the model. In addition, the super parameters involved in the long short-term memory time network model are enumerated and compared using the grid search method, and the model is evaluated using the 5-fold cross validation method to screen out the optimal super parameters: the number of 2-layer LSTM neurons is 100 and 150 respectively, the optimizer selects adam, the activation function is sigmoid, and the number of model iterations is set to 60 times. The prediction time step is 1h, 2h, 3h, and 1d, 2d, and 3d in the future.

[0034] The past time step is between 2 and 30, and the optimal step refers to the past time step when the error of the long short-term memory network model based on the past time step of 2 to 30 is the smallest. At this time, the prediction and early warning model is also the optimal model.

[0035] The water quality data displayed to the user by this method is the water quality data collected and updated in real time at the current monitoring point, and at the same time provides rapid changes in water quality at different time scales during historical operation, ensuring that the user can fully understand the current and past water quality change information.

[0036] This method supports interaction through web pages, and users can select query content according to their needs and concerns; they can intuitively view the water quality status of the current point of interest, and can also conveniently view the historical water quality time series changes and future short-term algal bloom changes, which greatly improves the query and monitoring efficiency and improves the level of auxiliary decision-making.

[0037] This method can achieve real-time and high-frequency monitoring, display, and forecasting of water quality information, ensuring that users obtain the latest and most accurate information and assisting users in making decisions. Compared with the information lag problem of traditional manual monitoring, this method provides higher real-time and forward-looking performance.

[0038] This method displays water quality information that the user is interested in based on the user's needs and concerns; this means that the user can intuitively obtain effective information without being disturbed by other information.

[0039] Example 2 This embodiment provides a short-term prediction and early warning system for water quality and algal bloom in lakes and reservoirs based on hyperspectral near-sensing, which is used to implement the short-term prediction and early warning method for water quality and algal bloom as described in Example 1, including: The first module is configured to collect the near-sensing hyperspectral reflectance of the water body in real time according to the spatiotemporal attribute information and obtain the preprocessed near-sensing hyperspectral reflectance of the water body through preprocessing; the preprocessed near-sensing hyperspectral reflectance of the water body is combined with the spatiotemporal attribute information to generate a near-sensing hyperspectral reflectance-time series data set of the water body; the spatiotemporal attribute information at least includes: spatial information, time information and water body type of the monitoring point; The second module is configured to screen out the corresponding near-sensing hyperspectral reflectance of the water body according to the water body type of the monitoring point, and to build a preset water quality inversion model set related to the water body type of the monitoring point according to the principle of time-space matching in combination with the synchronously measured water quality parameters; the near-sensing hyperspectral reflectance-time series data set of the water body is used as the input variable, and the measured water quality parameters are used as the output dependent variable, and the preset water quality inversion model set is used to generate a multi-water quality parameter-time series data set; The third module is configured to perform multi-time scale statistical analysis based on the multi-water quality parameter-time series data set, screen out key multi-water quality parameter-time series data sets, and build a prediction and early warning model; use the pre-processed near-sensing hyperspectral reflectance of the water body as an input feature, and output a water quality and algal bloom prediction and early warning result; update the key multi-water quality parameter-time series data set according to the water quality and algal bloom prediction and early warning result, and optimize the prediction and early warning model; The fourth module is configured to receive user queries on different time scales of forecast and warning requirements on the web terminal, and repeat the above steps to update and display the output results in real time.

Claims

1. A short-term prediction and early warning method for algal bloom based on hyperspectral near-sensing lake water quality is characterized by: The following steps are involved: According to the spatiotemporal attribute information, the near-sensing hyperspectral reflectance of the water body is collected in real time and preprocessed to obtain the preprocessed near-sensing hyperspectral reflectance of the water body; Combining the preprocessed water body near-sensing hyperspectral reflectance with the spatiotemporal attribute information to generate a water body near-sensing hyperspectral reflectance-time series data set; The spatiotemporal attribute information at least includes: spatial information, time information and water body type of the monitoring point; According to the water body type at the monitoring point, the corresponding water body near-sensing hyperspectral reflectance is screened out, and combined with the synchronously measured water quality parameters, a preset water quality inversion model set related to the water body type at the monitoring point is constructed according to the principle of time-space matching; the water body near-sensing hyperspectral reflectance-time series data set is used as the input variable, and the measured water quality parameters are used as the output dependent variable, and the preset water quality inversion model set is used to generate a multi-water quality parameter-time series data set; Based on the multi-water quality parameter-time series data set, a multi-time scale statistical analysis is performed to screen out the key multi-water quality parameter-time series data set and build a prediction and early warning model; the pre-processed near-sensing hyperspectral reflectance of the water body is used as an input feature to output a water quality and algal bloom prediction and early warning result; the key multi-water quality parameter-time series data set is updated according to the water quality and algal bloom prediction and early warning result, and the prediction and early warning model is optimized; Receive user queries for forecast and warning requirements at different time scales on the web terminal, and repeat the above steps to update and display the output results in real time.

2. The method for short-term prediction and early warning of water quality and algae bloom based on hyperspectral near-sensing lake and reservoir water quality according to claim 1 is characterized in that: The collection process of collecting the near-sensing hyperspectral reflectance of the water body is as follows: Using the transmittance The transmission plate obtains the downward irradiance on the water surface : ;in, is the irradiance of the transmission plate; The following formula is used to obtain the near-sensing hyperspectral reflectance of the collected water body: : ; In the formula, is the water-leaving irradiance of the transmission plate.

3. The method for short-term prediction and early warning of water quality and algae bloom based on hyperspectral near-sensing lake and reservoir water quality according to claim 1 is characterized in that: The preprocessing steps of the preprocessed water body near-sensing hyperspectral reflectance are as follows: Draw a spectral curve based on the near-sensing hyperspectral reflectance of the water body, regard the spectral curve as a vector, and calculate the modulus length of each vector : ;in, , To collect near-sensing hyperspectral reflectance of water bodies in real time based on spatiotemporal attribute information; The normalized near-sensitivity hyperspectral reflectance is obtained by calculating the ratio of each wavelength to the mode length using the following formula: : ； The normalized near-sensing hyperspectral reflectance is calculated using the following formula: To perform a linear correction: ;in, is the near-sensing hyperspectral reflectance after radiation correction, represents the radiation correction factor, is the radiation correction slope.

4. The method for short-term prediction and early warning of water quality and algae bloom based on hyperspectral near-sensing lake and reservoir water quality according to claim 1 is characterized in that: The types of water bodies at the monitoring points include at least: clean reservoirs, natural rivers, urban rivers and lakes; The synchronously measured water quality parameters include at least: total nitrogen, total phosphorus, chlorophyll a, transparency, water temperature, dissolved oxygen, permanganate index, turbidity and total suspended solids concentration; The preset water quality inversion model set includes at least: a near-sensing hyperspectral total phosphorus inversion model, a near-sensing hyperspectral total nitrogen inversion model, a near-sensing hyperspectral transparency inversion model, a near-sensing hyperspectral total suspended matter concentration inversion model, a near-sensing hyperspectral water temperature inversion model and a near-sensing hyperspectral dissolved oxygen inversion model.

5. The method for short-term prediction and early warning of water quality and algal bloom based on hyperspectral near-sensing lake and reservoir water quality according to claim 1 is characterized in that: The time-space matching principle at least includes: time matching principle and space matching principle; The time matching principle is expressed as follows: The collection time point of the near-sensing hyperspectral reflectance of the water body is defined as , the detection time point of synchronous water quality parameters is , then the collection time point and detection time point The relationship should be as follows: ; The spatial matching principle includes: longitude and latitude matching principle and distance matching principle; the longitude and latitude matching principle is expressed as follows: ,in, is the longitude and latitude of the collection point of the near-sensing hyperspectral reflectance of the water body, Represents the dimension reduction of the detection points of simultaneous water quality parameters; The distance matching principle is expressed as follows: , where is the standard observation range radius, Indicates the horizontal distance between the water sample collection point and the near-sensing hyperspectral instrument monitoring center; , The near-sensing hyperspectral instrument is placed at a height above the water surface. Indicates the field of view.

6. The method for short-term prediction and early warning of water quality and algae bloom based on hyperspectral near-sensing lake and reservoir water quality according to claim 1 is characterized in that: The multi-time scale includes at least: hourly scale and daily scale, and the key multi-water quality parameter-time series data set includes: hourly water quality parameters and daily water quality parameters; the screening process of the key multi-water quality parameter-time series data set is as follows: introduce the correlation between hourly water quality parameters and synchronous chlorophyll a concentration, daily water quality parameters and synchronous chlorophyll a concentration , through the correlation The value of is used to determine whether the corresponding water quality parameter is a key parameter: ; In the formula, is the Pearson correlation coefficient, is the significance level; If the correlation The value of is 1, which means that the multi-water quality parameter is significantly correlated with the synchronous chlorophyll a concentration and is a key water quality parameter and is updated to the key multi-water quality parameter-time series data set; if the correlation If the value of is 0, it means that the multi-water quality parameter is weakly correlated with the synchronous chlorophyll a concentration and is not a key water quality parameter.

7. The method for short-term prediction and early warning of water quality and algal bloom based on hyperspectral near-sensing lake and reservoir water quality according to claim 1 is characterized in that: The analysis process of the prediction and early warning model is as follows: Define the sliding window as , the sliding step length is , respectively calculate the average values ​​of multiple water quality parameters within the optimal sliding window , Standard Deviation and median , analyze the water quality parameters at the center of the optimal sliding window : like , then the water quality and algal bloom forecast warning result indicates that the water quality is normal and no warning treatment is performed; like , the water quality and algal bloom prediction and warning results indicate that the water quality is abnormal and an alarm message is issued.

8. The method for short-term prediction and early warning of water quality and algal bloom based on hyperspectral near-sensing lake and reservoir water quality according to claim 5 is characterized in that: The matching logic relationship of the time-space matching principle is as follows: Assign the logical judgment result of the time-space matching principle to , when the time matching principle and the space matching principle are satisfied at the same time, then , then the corresponding synchronously measured water quality parameters and water body near-sensing hyperspectral reflectance are added to the data for model construction; On the contrary, when at least one of the time matching principle and the space matching principle is not satisfied, 0 is assigned to , , then delete the corresponding synchronously measured water quality parameters and water body near-sensing hyperspectral reflectance; 。 9. The method for short-term prediction and early warning of water quality and algal bloom based on hyperspectral near-sensing lake and reservoir water quality according to claim 7 is characterized in that: Also includes: Will Corresponding water quality parameters Remove from the key multi-water quality parameter-time series data set and use the median Replacement; The multiple water quality parameter-time series data set was smoothed using the Savitzky-Golay filter to reduce the noise in the acquisition process, and a fifth-order polynomial curve was fitted and filled using the nearest neighbor interpolation method to obtain the updated key multiple water quality parameter-time series data set.

10. A short-term prediction and early warning system for water quality and algae bloom in lakes and reservoirs based on hyperspectral near-sensing is used to implement a short-term prediction and early warning method for water quality and algae bloom as claimed in any one of claims 1 to 9, characterized in that: include: The first module is configured to collect the near-sensing hyperspectral reflectance of the water body in real time according to the spatiotemporal attribute information and obtain the preprocessed near-sensing hyperspectral reflectance of the water body through preprocessing; Combining the preprocessed water body near-sensing hyperspectral reflectance with the spatiotemporal attribute information to generate a water body near-sensing hyperspectral reflectance-time series data set; The spatiotemporal attribute information at least includes: spatial information, time information and water body type of the monitoring point; The second module is configured to screen out the corresponding near-sensing hyperspectral reflectance of the water body according to the water body type of the monitoring point, and to build a preset water quality inversion model set related to the water body type of the monitoring point according to the principle of time-space matching in combination with the synchronously measured water quality parameters; the near-sensing hyperspectral reflectance-time series data set of the water body is used as the input variable, and the measured water quality parameters are used as the output dependent variable, and the preset water quality inversion model set is used to generate a multi-water quality parameter-time series data set; The third module is configured to perform multi-time scale statistical analysis based on the multi-water quality parameter-time series data set, screen out key multi-water quality parameter-time series data sets, and build a prediction and early warning model; use the pre-processed near-sensing hyperspectral reflectance of the water body as an input feature, and output a water quality and algal bloom prediction and early warning result; update the key multi-water quality parameter-time series data set according to the water quality and algal bloom prediction and early warning result, and optimize the prediction and early warning model; The fourth module is configured to receive user inquiries on different time scales of forecast and warning requirements on the web terminal, and repeat the above steps to update and display the output results in real time.

Images