Integrated treatment method and system for river ecological restoration

Abstract

The application relates to the technical field of river ecological early warning, in particular to an integrated treatment method and system for river ecological restoration; the released water body of a dam is identified to obtain released water body parameters; at the starting moment of release, the water surface width data, water flow speed and water temperature downstream of the dam are identified to obtain segmented areas, and the segmented areas are numbered along the water flow direction; the first moment when the released water body reaches the segmented areas and the second moment when the released water body leaves the segmented areas are determined in combination with the water release parameters; initial water algae area at the first moment and water body change data during the first moment and the second moment are collected, the water body change data including water level change data, continuous illumination data, continuous water temperature data and nutrition change data; a water area anomaly prediction model is constructed to identify the initial water algae area and the water body change data to obtain a water body anomaly coefficient. The water body anomaly coefficient of the segmented areas is used to early warn the ecological anomaly downstream of the dam.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of river ecological early warning, in particular to an integrated processing method and system for river ecological restoration. BACKGROUND

[0002] After the dam intercepts the river to form a reservoir, the organic matter in the upstream converges at this place; with the dam flood discharge and periodic flow control, the eutrophic water body in the reservoir spreads to the downstream of the river; when encountering a steep terrain, the water flow can quickly take away the eutrophic substances released by the reservoir; and when encountering a slow water flow area, the eutrophic substances may stay there for a long time, causing the water body in this area to be seriously eutrophic, the water algae to be rampant, and the stability of the ecological system to be destroyed.

[0003] Therefore, it is necessary to analyze the influence of the eutrophic water body released by the reservoir on the downstream river during the dam opening, and timely and accurately warn and restore the ecological abnormalities of the downstream river.

[0004] In the prior art, the ecological abnormalities of the water body are identified mainly by monitoring the eutrophic indicators and the content of water algae in the water body; however, during the dam flood discharge, the water body environment in the downstream of the river changes complexly, and when directly detecting the content of water algae, there are many interference factors, which is easy to cause errors.

[0005] Therefore, an integrated processing method and system for river ecological restoration are proposed. SUMMARY

[0006] The purpose of the present application is to provide an integrated processing method and system for river ecological restoration, which identifies and warns the water body abnormalities downstream of the dam through the water body abnormality coefficient of the segmented area of the river.

[0007] To achieve the above purpose, the present application provides an integrated processing method for river ecological restoration, comprising:

[0008] identifying the released water body of the dam to obtain the released water body parameters; the released water body parameters include the release start time, the release end time, the release flow, and the release speed;

[0009] At the release start time, the water surface width data, the water flow rate and the water temperature downstream of the dam are identified to obtain the segmented area, and the segmented area is numbered along the water flow direction;

[0010] combined with the water release parameters, the first time when the released water body reaches the segmented area and the second time when the released water body leaves the segmented area are determined; the initial water algae area at the first time and the water body change data during the first time and the second time are collected, the water body change data including the water level change data, the continuous illumination data, the continuous water temperature data and the nutrition change data;

[0011] The water area anomaly prediction model is constructed to identify the initial water algae area and water body change data, obtain a water body anomaly coefficient, and perform early warning.

[0012] The method for segmenting the river to obtain the segmented area is as follows:

[0013] A unit distance is set, the riverbed is divided along the water flow direction to obtain a unit riverbed area, and the unit riverbed area is sequentially numbered along the water flow direction;

[0014] Data of the unit riverbed area is collected to obtain water surface width data, water flow speed, water temperature, and unit riverbed number;

[0015] The water surface width data, water flow speed, and water temperature are identified and clustered by a clustering model to obtain a plurality of clustering sets;

[0016] The unit riverbed areas with the same number and belonging to the same clustering set are divided into the same segmented area.

[0017] The water surface width data acquisition process includes:

[0018] The water level change data of the river after the water body is released is obtained by querying the historical release data of the dam according to the water body release parameter;

[0019] The water level change range of the unit riverbed area is determined according to the initial water level at the release start time and the water level change data;

[0020] The width change data of the water level change range of the unit riverbed area is obtained as the water surface width data.

[0021] The determination process of the first time and the second time of the segmented area is as follows:

[0022] The riverbed distance of the segmented area from the dam and the released water body parameter are identified by a pre-trained neural network model to determine the expected arrival time and the expected departure time of the released water body passing through the segmented area;

[0023] The expected arrival time range is determined according to the expected arrival time; the water flow speed data, water temperature data, and water level data of the segmented area in the expected arrival time range are obtained, and the first time is determined according to the change amplitude of the data;

[0024] The expected departure time range is determined according to the expected departure time; the water flow speed data, water temperature data, and water level data of the segmented area in the expected departure time range are obtained, and the second time is determined according to the change amplitude of the data.

[0025] The water area anomaly prediction model comprises a water algae area prediction layer, a water algae area contrast layer and a regional anomaly early warning layer;

[0026] The water algae area prediction layer predicts the growth rate of water algae according to initial water algae area, water level change data in water body change data, continuous light data, continuous water temperature data and nutrient change data, to obtain final water algae area; the continuous light data is light data of a segmented region during the period when the release water body flows through; the continuous water temperature data is water temperature data of the segmented region during the period when the release water body flows through; and the initial water algae area is obtained by remote sensing technology.

[0027] The water algae area contrast layer obtains a water algae growth coefficient according to the final water algae area and the initial water algae area, and determines a water body anomaly coefficient according to the water algae growth coefficient.

[0028] The regional anomaly early warning layer performs identification and early warning according to the water body anomaly coefficient.

[0029] The nutrient change data comprises concentration change data of total nitrogen, total phosphorus, nitrate nitrogen and orthophosphate.

[0030] An integrated processing system for river ecological restoration comprises:

[0031] A release water body identification module identifies the release water body of a dam to obtain release water body parameters; the release water body parameters comprise release start time, release end time, release flow and release speed.

[0032] A river region segmentation module identifies water surface width data, water flow speed and water temperature downstream of the dam at the release start time to obtain segmented regions and number the segmented regions along the water flow direction.

[0033] A release water body tracking module determines the first time when the release water body reaches a segmented region and the second time when the release water body leaves the segmented region in combination with the water body release parameters; collects initial water algae area at the first time and water body change data during the first time and the second time, wherein the water body change data comprises water level change data, continuous light data, continuous water temperature data and nutrient change data.

[0034] A water area anomaly early warning module identifies initial water algae area and water body change data by constructing a water area anomaly prediction model to obtain a water body anomaly coefficient and perform early warning and restoration.

[0035] Compared with the prior art, the present application has the following beneficial effects:

[0036] 1、The application adopts a clustering model, divides the river section according to internal hydrological physical properties such as water surface width, flow rate and water temperature, ensures that the hydrological characteristics of each segmented region are highly homogeneous, improves the representativeness and accuracy of the analysis result, can effectively reduce the internal variance of the model input data, and significantly improves the stability and precision of the prediction model.

[0037] 2、The pre-trained neural network model can quickly and intelligently predict the expected time range of the water body influence based on limited input parameters, greatly reduces the time interval that needs to be monitored, and accurately determines the first time and the second time corresponding to the actual generation and extinction process of the hydrological pulse in the narrowed expected range, thereby providing data quality guarantee for the accuracy of the core prediction model.

[0038] 3、The application predicts the growth rate of algae according to the initial algae area of the river segmented region and the water body change data, obtains the predicted algae area, obtains the algae growth coefficient according to the predicted algae area and the initial algae area, and can accurately identify the water body ecological anomaly of the river segmented region according to the water body anomaly coefficient. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 It is a flowchart of an integrated processing method for river ecological restoration of the application;

[0040] Figure 2 It is a structural diagram of a water area anomaly prediction model of the application;

[0041] Figure 3 It is a structural diagram of an integrated processing system for river ecological restoration of the application. DETAILED DESCRIPTION

[0042] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0043] Embodiment one:

[0044] The application provides an integrated processing method for river ecological restoration, and the flowchart is as shown in Figure 1 The application provides an integrated processing method for river ecological restoration, and the flowchart is as shown in

[0045] The release water body of the dam is identified to obtain release water body parameters; the release water body parameters include release start time, release end time, release flow, and release speed.

[0046] At the release starting moment, the water surface width data, water flow rate and water temperature downstream of the dam are identified to obtain segmented areas, and the segmented areas are numbered along the water flow direction;

[0047] Combined with the water release parameters, the first moment when the released water reaches the segmented area and the second moment when the released water leaves the segmented area are determined; the initial water algae area at the first moment and the water change data during the first moment and the second moment are collected, and the water change data includes water level change data, continuous illumination data, continuous water temperature data and nutrient change data;

[0048] The water area anomaly prediction model is constructed to identify the initial water algae area and the water change data to obtain a water anomaly coefficient for early warning and repair.

[0049] Preferably, the method for segmenting the river to obtain the segmented areas is as follows:

[0050] A unit distance is set, and the riverbed is divided along the water flow direction to obtain unit riverbed areas, and the unit riverbed areas are numbered in sequence along the water flow direction;

[0051] Data collection is performed on the unit riverbed areas to obtain water surface width data, water flow rate, water temperature and unit riverbed number;

[0052] The water surface width data, water flow rate and water temperature are identified and clustered by a clustering model to obtain a plurality of clustering sets;

[0053] Unit riverbed areas with the same number and belonging to the same clustering set are divided into the same segmented area.

[0054] Further, in the clustering process, to avoid a certain data from dominating the clustering calculation, standardization processing is performed; preferably, the Z-score standardization method is used to convert the data of each feature dimension to a standard normal distribution with a mean of 0 and a standard deviation of 1.

[0055] The K-Means++ method is used to intelligently select the initial K clustering centers to improve the convergence speed and avoid falling into local optimum. The following two steps are repeated until the clustering centers no longer change:

[0056] Assignment step: for each unit riverbed area feature vector, calculate its Euclidean distance from the K clustering centers, and attribute it to the nearest clustering center.

[0057] Update step: for each cluster, recalculate its geometric center (i.e. the mean of all feature vectors in the cluster), and take the center point as the new clustering center.

[0058] After the clustering algorithm converges, each unit riverbed region is marked with a cluster label.

[0059] The present application adopts a clustering model, divides the river section according to internal hydrological physical properties such as water surface width, flow rate and water temperature, ensures the high homogeneity of the hydrological characteristics in each segmented region, improves the representativeness and accuracy of the analysis result, can effectively reduce the internal variance of the model input data, and significantly improves the stability and precision of the prediction model.

[0060] The water surface width data acquisition process comprises:

[0061] According to the water body release parameters, the historical release data of the dam are queried to obtain water level change data of the river after the water body is released;

[0062] According to the initial water level at the release starting moment and the water level change data, the water level change range of the unit riverbed region is determined.

[0063] The width change data of the water level change range of the unit riverbed region is obtained as the water surface width data.

[0064] Based on the release water body parameters of the current water release event, mode matching is performed in the historical hydrological database to query the water level change process data of each monitoring point downstream under similar water release conditions in the past.

[0065] Combined with the initial water level at the current release starting moment and the water level change data queried from the historical data, the maximum and minimum water levels that each unit riverbed region may experience under the influence of the current water release event are estimated, so as to determine a water level change range.

[0066] The present application can dynamically determine a more representative water level change range for the current specific water release event by associating the historical release data of the dam with the downstream water level change, and calculate the corresponding water surface width, thereby improving the accuracy of subsequent hydrological simulation and ecological evolution prediction.

[0067] The determination process of the first moment and the second moment of the segmented region is:

[0068] The riverbed distance of the segmented region and the release water body parameters are identified by the pre-trained neural network model to determine the expected arrival moment and the expected departure moment of the release water body passing through the segmented region;

[0069] The expected arrival time range is determined according to the expected arrival moment; the water body flow rate data, water body temperature data and water body water level data of the segmented region in the expected arrival time range are obtained, and the first moment is determined according to the change amplitude of the data.

[0070] According to the expected departure time, the expected departure time range is determined; the water flow rate data, water temperature data and water level data of the segmented area in the expected departure time range are obtained, and the second time is determined according to the change amplitude of the data.

[0071] The structure and training process of the neural network model are as follows:

[0072] Model architecture: a multi-layer perceptron (MLP) feedforward neural network is adopted.

[0073] Input layer: a plurality of neurons are constructed, which respectively receive the standardized distance between the segmented area and the dam, the released water body parameters, and preferably, the average river bed slope of the segmented area and the average river bed roughness of the segmented area.

[0074] Hidden layer: two hidden layers, the first layer has 32 neurons, and the second layer has 16 neurons. Leaky ReLU is used as the activation function to prevent gradient disappearance.

[0075] Output layer: 2 neurons, using linear activation function, directly outputting the expected arrival time and the expected departure time.

[0076] Model training:

[0077] Based on the historical water release events and the monitoring records of the downstream hydrological stations, the data pairs of “input features-real arrival / departure time” are constructed.

[0078] The data set is divided into training, validation and test sets according to the ratio of 8:1:1. Adam optimizer and root mean square error (RMSE) loss function are used for training. To prevent overfitting, Dropout (ratio 0.2) and L2 regularization are introduced. When the loss on the validation set no longer improves for multiple epochs, the training is terminated in advance.

[0079] The input parameters of the current event are fed into the trained model, and the “expected arrival time” and “expected departure time” of each segmented area are obtained.

[0080] The identification and determination process of the change amplitude includes: quantifying the fuzzy “change amplitude” into explicit and executable rules.

[0081] Taking the water flow rate as an example:

[0082] First time determination rule: within the time range before and after the expected arrival time, the flow rate of the upstream endpoint of the segmented area is monitored in real time. When the 10-minute moving average value of the flow rate at this point first exceeds 30% of the baseline average value at 1 hour before the expected arrival time, the time point is accurately marked as “first time”.

[0083] Second time determination rule: around the expected departure time, the flow rate of the downstream endpoint of the segmented area is monitored. When the flow rate falls from the peak value, and its 10-minute moving average value first falls below 40% of the peak flow rate and continues to fall thereafter, the time point is accurately marked as the "second time".

[0084] The system starts real-time monitoring, and continuously compares the data with the above rules. Once the conditions are met, the accurate "first time" and "second time" are immediately locked and recorded.

[0085] In addition, for the judgment threshold of the speed change, that is, the "baseline average value 30%", the release flow and release speed of the water body release parameter can be adjusted; when the water body release is turbulent and has a greater impact on the downstream, that is, the release flow and release speed are larger, the judgment threshold is increased; when the water body release is relatively calm and has a smaller impact on the downstream, that is, the release flow and release speed are smaller, the judgment threshold is reduced.

[0086] Preferably, the water body flow rate data, water body temperature data and water body water level data of the segmented area are comprehensively judged to improve the judgment accuracy of the first time and the second time.

[0087] The pre-trained neural network model can quickly and intelligently predict the expected time range of the water body impact based on limited input parameters, greatly reducing the time interval that needs to be monitored; within the narrowed expected range, the start and end time of the impact is accurately marked by monitoring the change amplitude of the real-time data, ensuring that the determined first time and second time accurately correspond to the actual generation and extinction process of the hydrological pulse, and providing data quality guarantee for the accuracy of the core prediction model.

[0088] The structure of the water area anomaly prediction model is as shown in Figure 2 The water area anomaly prediction model comprises a water algae area prediction layer, a water algae area comparison layer and a regional anomaly early warning layer.

[0089] The water algae area prediction layer predicts the growth rate of water algae according to the initial water algae area, the water level change data in the water body change data, the continuous illumination data, the continuous water temperature data and the nutrition change data, to obtain the final water algae area; wherein the continuous illumination data is the illumination data of the segmented area during the release of the water body; the continuous water temperature data is the water temperature data of the segmented area during the release of the water body; and the initial water algae area is obtained by remote sensing technology.

[0090] The water algae area comparison layer obtains the water algae growth coefficient according to the final water algae area and the initial water algae area, and determines the water body anomaly coefficient according to the water algae growth coefficient.

[0091] The regional anomaly early warning layer identifies and warns according to the water body anomaly coefficient.

[0092] A gradient boosting decision tree model, such as the LightGBM framework, is used to build a water area anomaly prediction model.

[0093] A rich feature vector is constructed, including initial algae area, time window length, mean and standard deviation of all environmental factors (light, temperature, nutrients, water level change), etc.

[0094] A regression model capable of predicting the final algae area is trained using historical data (containing features and real final algae area).

[0095] The feature vector of the current event is input into the trained model to obtain the prediction result directly.

[0096] The present application predicts the growth rate of algae according to the initial algae area of the river segmented area and the water body change data, obtains the predicted algae area, and obtains the algae growth coefficient according to the predicted algae area and the initial algae area. According to the water body anomaly coefficient, the water body ecological anomaly of the river segmented area can be accurately identified.

[0097] Preferably, the nutrient change data includes concentration change data of total nitrogen, total phosphorus, nitrate nitrogen and orthophosphate.

[0098] Deploy online multi-parameter water quality monitoring stations at key cross sections of downstream river segments or representative positions of each segmented area.

[0099] Integrate special sensor probes for each index on the monitoring station:

[0100] Total nitrogen / total phosphorus: use ultraviolet-visible full spectrum method or more complex heating digestion-colorimetric method online analyzer.

[0101] Nitrate nitrogen: use ion selective electrode (ISE) method or ultraviolet absorption method probe.

[0102] Orthophosphate: use micro online chemical analyzer of phosphomolybdenum blue colorimetric method.

[0103] The monitoring station automatically collects and measures water samples at a preset frequency (such as once an hour), and attaches a time stamp to the measurement data (concentration value), and transmits it to the central database in real time through a wireless network.

[0104] The present application also proposes an integrated treatment system for river ecological restoration, the system structure is as shown in Figure 3 , comprising:

[0105] The release water body identification module identifies the release water body of the dam to obtain the release water body parameters; the release water body parameters include release start time, release end time, release flow, and release speed.

[0106] A river area segmentation module identifies water surface width data, water flow rate and water temperature downstream of the dam at a release start time to obtain segmented areas and number the segmented areas along a water flow direction;

[0107] A released water body tracking module determines a first time when the released water body reaches the segmented areas and a second time when the released water body leaves the segmented areas in combination with water release parameters; collects initial water algae area at the first time and water body change data during the first time and the second time, the water body change data including water level change data, continuous illumination data, continuous water temperature data and nutrient change data;

[0108] A water area anomaly early warning module constructs a water area anomaly prediction model to identify the initial water algae area and the water body change data to obtain a water body anomaly coefficient for early warning repair.

[0109] The released water body of the dam is identified to obtain released water body parameters; water surface width data, water flow rate and water temperature downstream of the dam are identified at a release start time to obtain segmented areas and number the segmented areas along a water flow direction; a first time when the released water body reaches the segmented areas and a second time when the released water body leaves the segmented areas are determined in combination with water release parameters; initial water algae area at the first time and water body change data during the first time and the second time are collected, the water body change data including water level change data, continuous illumination data, continuous water temperature data and nutrient change data; a water area anomaly prediction model is constructed to identify the initial water algae area and the water body change data to obtain a water body anomaly coefficient for accurate early warning of ecological anomalies downstream of the dam.

[0110] Embodiment two:

[0111] The application provides an integrated processing method for river ecological restoration, which comprises the following steps:

[0112] The released water body of the dam is identified to obtain released water body parameters; the released water body parameters include a release start time, a release end time, a release flow rate and a release speed;

[0113] Water surface width data, water flow rate and water temperature downstream of the dam are identified at a release start time to obtain segmented areas and number the segmented areas along a water flow direction;

[0114] A first time when the released water body reaches the segmented areas and a second time when the released water body leaves the segmented areas are determined in combination with water release parameters; initial water algae area at the first time and water body change data during the first time and the second time are collected, the water body change data including water level change data, continuous illumination data, continuous water temperature data and nutrient change data;

[0115] A water area anomaly prediction model is constructed to identify the initial water algae area and the water body change data to obtain a water body anomaly coefficient for early warning repair.

[0116] The method for segmenting the river into segmented areas is as follows:

[0117] A unit distance is set, and the riverbed is divided along the water flow direction to obtain unit riverbed areas, and the unit riverbed areas are sequentially numbered along the water flow direction;

[0118] Data collection is performed on the unit riverbed areas to obtain water surface width data, water flow velocity, water temperature, and unit riverbed numbers;

[0119] The water surface width data, water flow velocity, and water temperature are identified and clustered by a clustering model to obtain a plurality of clustering sets;

[0120] Unit riverbed areas with the same number and belonging to the same clustering set are divided into the same segmented area.

[0121] Preferably, at the release start time, the river segment 50 km downstream of the dam is segmented.

[0122] First, the riverbed is divided along the water flow direction at a unit distance of 100 meters, and the 50 km river segment is divided into 500 unit riverbed areas with consecutive numbers, numbered D001 to D500.

[0123] The digital elevation model (DEM) with a grid accuracy of 5 meters for this river segment is called, and the historical highest and lowest water levels of each unit riverbed area are determined based on similar water release events in the historical database. Subsequently, using hydrodynamic software such as HEC-RAS or the hydrological analysis tool of ArcGIS, the water surface line change from the lowest water level to the highest water level is simulated on the DEM, with a step size of 0.1 meters, to accurately calculate the water surface width of each unit riverbed area at different water levels, forming a detailed "water level-width" relationship data set as input for subsequent clustering of water surface width data.

[0124] To eliminate the dimension effect, the Z-score standardization method is used to process the three groups of data of water surface width, water flow velocity, and water temperature of all 500 unit areas, with a mean value of 0 and a standard deviation of 1.

[0125] Determine the optimal number of clusters K: The "Elbow Method" is used to determine the optimal number of clusters. By calculating the within-cluster sum of squares corresponding to K values from 2 to 20, it is found that when K=5, the descending slope of WCSS value appears a significant inflection point, and tends to be flat. Therefore, the optimal number of clusters for this segmentation is determined to be 5.

[0126] Clustering and region merging: K-Means++ algorithm is used to cluster the standardized data. After clustering, the system scans the cluster labels of all unit riverbed regions. For example, the 18 consecutive unit riverbed regions numbered from D035 to D052 are all divided into the same cluster set, which is characterized by “wide valley, low flow rate, high water temperature”. Therefore, the system merges these 18 geographically connected unit riverbed regions to form a homogeneous subsection region, and numbers it as “Subsection Region-03”.

[0127] After the above steps, the final 50-kilometer river section is divided into 15 subsection regions with different physical characteristics.

[0128] The starting position of “Subsection Region-03” is 4.2 kilometers away from the dam. The distance, the release parameters of this time, and the average riverbed slope of this region are input into the pre-trained neural network model.

[0129] The determination process of the first time and the second time of the subsection region is as follows:

[0130] The pre-trained neural network model is used to identify the riverbed distance of the subsection region from the dam and the release water parameters, and determine the expected arrival time and the expected departure time of the released water passing through the subsection region;

[0131] According to the expected arrival time, the expected arrival time range is determined; the water flow rate data, water temperature data and water level data of the subsection region in the expected arrival time range are obtained, and the first time is determined according to the variation amplitude of the data;

[0132] According to the expected departure time, the expected departure time range is determined; the water flow rate data, water temperature data and water level data of the subsection region in the expected departure time range are obtained, and the second time is determined according to the variation amplitude of the data.

[0133] The following comprehensive judgment rules are adopted in this embodiment:

[0134] The first time: around the expected arrival time, when the online monitoring instrument deployed at the upstream endpoint of “Subsection Region-03” simultaneously meets “10-minute moving average flow rate first exceeds 30% of the baseline value” and “10-minute moving average water level rises by more than 0.25 meters” and “10-minute moving average water temperature fluctuation exceeds 10%”, this time point is accurately marked as the first time.

[0135] The second time: around the expected departure time, when the flow rate at the downstream endpoint of the region falls from the peak value, and its 10-minute moving average first falls below 40% of the peak flow, this time point is marked as the second time.

[0136] The water area anomaly prediction model comprises a water algae area prediction layer, a water algae area contrast layer and a regional anomaly early warning layer;

[0137] The water algae area prediction layer predicts the growth rate of water algae according to initial water algae area, water level change data in water body change data, continuous light data, continuous water temperature data and nutrient change data, to obtain final water algae area; the continuous light data is light data of the segmented region during the release water body flows through; the continuous water temperature data is water temperature data of the segmented region during the release water body flows through; the initial water algae area is obtained through remote sensing technology;

[0138] The water algae area contrast layer obtains a water algae growth coefficient according to the final water algae area and the initial water algae area, and determines a water body anomaly coefficient according to the water algae growth coefficient;

[0139] The regional anomaly early warning layer performs identification and early warning according to the water body anomaly coefficient.

[0140] The initial water algae area of the segmented region-03 and a series of data collected during the release water body flows through, such as water level, light, water temperature and nutrient salt change, are input into a gradient boosting decision tree model pre-trained through a large amount of historical data, to obtain a water quality anomaly coefficient for early warning.

[0141] Although the embodiments of the present application have been shown and described, it can be understood by those skilled in the art that various changes, modifications, replacements and variations can be made to the embodiments without departing from the principles and spirits of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. An integrated treatment method for river ecological restoration, characterized in that, The method comprises the following steps: identifying the release water body of the dam to obtain release water body parameters; the release water body parameters include release start time, release end time, release flow, and release speed; at the release start time, identifying the water surface width data, water flow speed, and water temperature downstream of the dam to obtain segmented areas and number the segmented areas along the water flow direction; combining the water release parameters to determine the first time when the release water body reaches the segmented area and the second time when the release water body leaves the segmented area; collecting the initial water algae area at the first time and the water body change data during the first time and the second time, wherein the water body change data includes water level change data, continuous illumination data, continuous water temperature data, and nutrition change data; the determination process of the first time and the second time of the segmented area is as follows: identifying the riverbed distance of the segmented area from the dam and the release water body parameters through a pre-trained neural network model to determine the expected arrival time and the expected leaving time of the release water body through the segmented area; determining the expected arrival time range according to the expected arrival time; obtaining the water flow speed data, water temperature data, and water level data of the segmented area within the expected arrival time range, and determining the first time according to the change range of the data; determining the expected leaving time range according to the expected leaving time; obtaining the water flow speed data, water temperature data, and water level data of the segmented area within the expected leaving time range, and determining the second time according to the change range of the data; constructing a water area anomaly prediction model to identify the initial water algae area and the water body change data to obtain a water body anomaly coefficient and perform early warning.

2. The integrated processing method for river ecological restoration according to claim 1, wherein: the method for segmenting the river to obtain segmented areas is as follows: setting a unit distance, dividing the riverbed along the water flow direction to obtain unit riverbed areas, and numbering the unit riverbed areas in sequence along the water flow direction; collecting data of the unit riverbed areas to obtain water surface width data, water flow speed, water temperature, and unit riverbed numbers; identifying and clustering the water surface width data, water flow speed, and water temperature through a clustering model to obtain a plurality of clustering sets; dividing the unit riverbed areas with consecutive numbers in the same clustering set into the same segmented area.

3. The integrated processing method for river ecological restoration according to claim 2, wherein: the process for obtaining the water surface width data comprises: querying historical release data of the dam according to the water release parameters to obtain water level change data of the river after water release; determining the water level change range of the unit riverbed area according to the initial water level at the release start time and the water level change data; obtaining the width change data of the water level change range of the unit riverbed area as the water surface width data.

4. The integrated processing method for river ecological restoration according to claim 1, wherein: the water area anomaly prediction model comprises a water algae area prediction layer, a water algae area comparison layer, and a regional anomaly early warning layer. The water algae area prediction layer predicts the growth rate of the water algae according to the initial water algae area, water level change data in the water body change data, continuous light data, continuous water temperature data and nutrient change data, and obtains the final water algae area; The continuous light data is light data of the segmented area during the period when the release water body flows through, and the continuous water temperature data is water temperature data of the segmented area during the period when the release water body flows through; the initial water algae area is obtained by remote sensing technology; The water algae area comparison layer obtains the water algae growth coefficient according to the final water algae area and the initial water algae area, and determines the water body abnormality coefficient according to the water algae growth coefficient; The area abnormality early warning layer identifies and warns according to the water body abnormality coefficient.

5. The integrated treatment method for river ecological restoration according to claim 1, characterized in that: The nutrient change data includes concentration change data of total nitrogen, total phosphorus, nitrate nitrogen and orthophosphate.

6. An integrated treatment system for river ecological restoration, characterized in that, It comprises: A release water body identification module identifies the release water body of the dam to obtain release water body parameters; the release water body parameters include release start time, release end time, release flow, and release speed; A river area segmentation module identifies the water surface width data, water flow rate and water temperature downstream of the dam at the release start time to obtain segmented areas and number them along the water flow direction; A release water body tracking module determines the first time when the release water body reaches the segmented area and the second time when the release water body leaves the segmented area in combination with the water body release parameters; collects the initial water algae area at the first time and the water body change data during the first time and the second time, which includes water level change data, continuous light data, continuous water temperature data and nutrient change data; The determination process of the first time and the second time of the segmented area is as follows: A pre-trained neural network model is used to identify the river bed distance of the segmented area from the dam and the release water body parameters to determine the expected arrival time and the expected leaving time of the release water body through the segmented area; An expected arrival time range is determined according to the expected arrival time; water flow rate data, water temperature data and water level data of the segmented area in the expected arrival time range are obtained, and the first time is determined according to the change amplitude of the data; An expected leaving time range is determined according to the expected leaving time; water flow rate data, water temperature data and water level data of the segmented area in the expected leaving time range are obtained, and the second time is determined according to the change amplitude of the data; A water area abnormality early warning module constructs a water area abnormality prediction model to identify the initial water algae area and the water body change data to obtain the water body abnormality coefficient and perform early warning.

Images