A method and system for predicting cyanobacterial blooms based on satellite remote sensing and weather forecast

Abstract

The present invention proposes a method and system for predicting cyanobacteria blooms based on satellite remote sensing and weather forecasts. The method includes: selecting the optimal meteorological factor based on the different contributions of 1-7-day cumulative meteorological factors to the outbreak of cyanobacteria blooms; using the optimal meteorological factor and ATBI band as the input of the prediction model; and constructing the prediction model based on the principle of Bayesian theorem. The scheme proposed in the present invention can use the cyanobacteria bloom monitoring results of large lakes and meteorological factor data as the main data sources, and construct a short-term prediction model for cyanobacteria blooms in large lakes in China based on Bayesian theorem, so as to realize the pixel-scale prediction of cyanobacteria blooms in multiple days in the future.

Description

Technical Field

[0001] The invention belongs to the field of cyanobacteria bloom prediction, and in particular relates to a cyanobacteria bloom prediction method and system based on satellite remote sensing and weather forecast. Background Art

[0002] Under the background of intensified human activities and climate change, the eutrophication of inland water bodies has become a major challenge to the global lake water environment. Blue algae blooms are an extreme manifestation of eutrophication and one of the most serious inland water environmental problems. Therefore, if the short-term prediction of blue algae blooms in large lakes can be achieved, relevant management departments can issue early warnings for blue algae blooms and take emergency response measures to effectively curb the outbreak of blue algae blooms, thereby reducing the environmental damage caused by blue algae blooms and the economic burden of governance.

[0003] At present, the research on cyanobacterial bloom prediction is developing slowly. Most of the existing models are for single lakes (Chen Yuwei et al., 2001; Zhang Hujun et al., 2022), and are overly dependent on field sampling data and hydrodynamic factors (Kong Fanxiang et al., 2009; Chen Liqiong et al., 2016). Such models have many input parameters and are not convenient for business applications (Mu et al., 2021). In the absence of uninterrupted support for field sampling data, the application of existing prediction models is greatly limited (Shi et al., 2017; Yussof et al., 2021; Cao et al., 2022). In terms of model prediction results, few models can achieve cyanobacterial bloom prediction at the pixel scale, and the existing cyanobacterial bloom model construction is also mostly limited to a single specific water body. Therefore, in general, there is still a lack of pixel-scale short-term prediction models for cyanobacterial blooms for typical large lakes. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes a technical solution of a cyanobacteria bloom prediction method based on satellite remote sensing and weather forecast to solve the above technical problems.

[0005] The first aspect of the present invention discloses a method for predicting cyanobacteria blooms based on satellite remote sensing and weather forecasts, the method comprising:

[0006] Step S1, collecting MODIS satellite remote sensing image data, and screening the MODIS satellite remote sensing image data to obtain ATBI band and cyanobacteria bloom distribution results; collecting meteorological data, and applying the screening of optimal cumulative meteorological factors and multicollinearity test to obtain the screened meteorological data;

[0007] Step S2: Using box type Figure 4The ATBI band is discretized and normalized by the quantile method to obtain the normalized ATBI band; Figure 4 The quantile method is used to discretize the filtered meteorological data to obtain the discretized meteorological data; the prior probability is calculated by using the ratio of the frequency of blue algae bloom outbreaks to the frequency of effective pixels without cloud coverage in the study area.

[0008] Step S3, using the normalized ATBI band, the discretized meteorological data, the prior probability and the distribution results of cyanobacteria bloom to construct a training set and a test set of the prediction model; constructing a prediction model based on the principle of Bayes' theorem; using the normalized ATBI band, the discretized meteorological data and the prior probability of the training set as the prediction model input, and the cyanobacteria bloom distribution results as the label of the prediction model, training the prediction model, and obtaining a trained prediction model;

[0009] Step S4, performing discretization and normalization processing on the MODIS satellite remote sensing image data collected in real time to obtain normalized real-time ATBI bands; performing the screening and discretization processing on the meteorological data collected in real time to obtain discretized real-time meteorological data; inputting the normalized real-time ATBI band and the discretized real-time meteorological data into the trained prediction model to obtain the posterior probability of the outbreak of cyanobacteria bloom events.

[0010] According to the method of the first aspect of the present invention, in step S1, the screening of the optimal cumulative meteorological factor includes:

[0011] Seven days of meteorological factor data were collected, and the data of one day with the greatest contribution of each type of meteorological factor to the outbreak of cyanobacteria bloom was screened out from the seven days of data.

[0012] According to the method of the first aspect of the present invention, in step S2, the box-type Figure 4 The quantile method performs discretization and normalization on the ATBI band, and the normalized ATBI band includes:

[0013] The ATBI bands in the cyanobacteria bloom area and the ATBI bands in the non-cyanobacteria bloom area were grouped and box plots were drawn. According to the upper and lower quartiles, median and 0 value of the two groups of box plots, they were divided into 8 levels to obtain the discrete ATBI bands.

[0014] The discretized ATBI band is normalized to obtain a normalized ATBI band.

[0015] According to the method of the first aspect of the present invention, in step S2, normalizing the discretized ATBI band to obtain the normalized ATBI band includes:

[0016] The segmentation threshold of cyanobacteria bloom is calculated by iterative triangle method, and the segmentation threshold is subtracted from the discrete ATBI band of the whole scene image, and the discrete ATBI band of the cyanobacteria bloom area is normalized to a value above 0; for the discrete ATBI band of the image without bloom, the maximum value of the ATBI band on the image is directly subtracted from the discrete ATBI band of the whole scene image, and the discrete ATBI band of all non-cyanobacteria bloom areas is normalized to a value below 0;

[0017] The segmentation threshold of cyanobacteria blooms is calculated by iterative triangulation method, comprising:

[0018] First, the numerical distribution histogram of the discretized ATBI band is calculated, and an iterative triangle method is used to find a point on the histogram so that the distance from the point to the line connecting the peak and the tail point of the histogram is the maximum value. The numerical value of the ATBI band corresponding to the point is the required segmentation threshold.

[0019] According to the method of the first aspect of the present invention, in step S2, the box-type Figure 4 The quantile method discretizes the filtered meteorological data, and the discretized meteorological data include:

[0020] The optimal cumulative meteorological factors after screening for each lake were grouped according to cyanobacterial bloom days and non-cyanobacterial bloom days, and box plots were drawn; each type of meteorological factor was divided into 7 levels according to the upper and lower quartiles and median lines of the two groups of box plots of cyanobacterial bloom days and non-cyanobacterial bloom days, and the discrete meteorological data were obtained.

[0021] According to the method of the first aspect of the present invention, in step S2, the ratio of the frequency of blue algae bloom outbreaks per pixel in the application study area to the frequency of effective pixels without cloud coverage is used to calculate the prior probability, which includes:

[0022]

[0023] Among them, P(C1) represents the prior probability.

[0024] According to the method of the first aspect of the present invention, in step S3, constructing a prediction model based on the principle of Bayesian theorem includes:

[0025] After the prior probability is input into the prediction model, the normalized ATBI bands and discretized meteorological data of the training set are traversed, and the normalized ATBI bands and discretized meteorological data are used as characteristic variables to construct a multidimensional conditional probability table under each characteristic variable for the two situations of whether a cyanobacterial bloom event occurs or not; according to the normalized ATBI bands and discretized meteorological data input data, the conditional probability of the response is determined using the multidimensional conditional probability table, and then the posterior probability of the cyanobacterial bloom event occurs is calculated.

[0026] The second aspect of the present invention discloses a cyanobacteria bloom prediction system based on satellite remote sensing and weather forecast, the system comprising:

[0027] The first processing module is configured to collect MODIS satellite remote sensing image data, and screen the MODIS satellite remote sensing image data to obtain ATBI band and cyanobacteria bloom distribution results; collect meteorological data, and apply the screening of optimal cumulative meteorological factors and multicollinearity test to obtain screened meteorological data;

[0028] The second processing module is configured to utilize a box-type Figure 4 The ATBI band is discretized and normalized by the quantile method to obtain the normalized ATBI band; Figure 4 The quantile method is used to discretize the filtered meteorological data to obtain the discretized meteorological data; the prior probability is calculated by using the ratio of the frequency of blue algae bloom outbreaks to the frequency of effective pixels without cloud coverage in the study area.

[0029] The third processing module is configured to use the normalized ATBI band, the discretized meteorological data, the prior probability and the distribution results of the blue algae bloom to construct a training set and a test set of the prediction model; construct a prediction model based on the principle of Bayes' theorem; use the normalized ATBI band, the discretized meteorological data, and the prior probability of the training set as the prediction model input, and the blue algae bloom distribution results as the label of the prediction model, train the prediction model, and obtain the trained prediction model;

[0030] The fourth processing module is configured to perform discretization and normalization processing on the MODIS satellite remote sensing image data collected in real time to obtain normalized real-time ATBI bands; perform the screening and discretization processing on the meteorological data collected in real time to obtain discretized real-time meteorological data; input the normalized real-time ATBI band and the discretized real-time meteorological data into the trained prediction model to obtain the posterior probability of the outbreak of cyanobacteria bloom events.

[0031] The third aspect of the present invention discloses an electronic device. The electronic device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps in any one of the methods for predicting cyanobacteria blooms based on satellite remote sensing and weather forecast in the first aspect of the present disclosure are implemented.

[0032] The fourth aspect of the present invention discloses a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in any one of the methods for predicting cyanobacteria blooms based on satellite remote sensing and weather forecast in the first aspect of the present disclosure are implemented.

[0033] In summary, the scheme proposed in the present invention can use the monitoring results of cyanobacterial blooms in large lakes and meteorological factor data as the main data sources, construct a short-term prediction model for cyanobacterial blooms in China's large lakes based on Bayesian theorem, and realize pixel-scale prediction of cyanobacterial blooms for multiple days in the future. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0035] Figure 1 is a flow chart of a method for predicting cyanobacteria blooms based on satellite remote sensing and weather forecast according to an embodiment of the present invention;

[0036] Figure 2 A hierarchical box plot of input data of a prediction model for cyanobacteria bloom in the next day according to an embodiment of the present invention;

[0037] Figure 3 A schematic diagram of a prediction model for cyanobacteria bloom according to an embodiment of the present invention;

[0038] Figure 4 A structural diagram of a cyanobacteria bloom prediction system based on satellite remote sensing and weather forecast according to an embodiment of the present invention;

[0039] Figure 5 The figure is a structural diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0041] The first aspect of the present invention discloses a method for predicting cyanobacteria blooms based on satellite remote sensing and weather forecasts. Figure 1 FIG. 4 is a flow chart of a method for predicting cyanobacteria blooms based on satellite remote sensing and weather forecast according to an embodiment of the present invention. Figure 1 As shown, the method includes:

[0042] Step S1, collecting MODIS satellite remote sensing image data, and screening the MODIS satellite remote sensing image data to obtain ATBI band and cyanobacteria bloom distribution results; collecting meteorological data, and applying the screening of optimal cumulative meteorological factors and multicollinearity test to obtain the screened meteorological data;

[0043] Step S2: Using box type Figure 4 The ATBI band is discretized and normalized by the quantile method to obtain the normalized ATBI band; Figure 4 The quantile method is used to discretize the filtered meteorological data to obtain the discretized meteorological data; the prior probability is calculated by using the ratio of the frequency of blue algae bloom outbreaks to the frequency of effective pixels without cloud coverage in the study area.

[0044] Step S3, using the normalized ATBI band, the discretized meteorological data, the prior probability and the distribution results of cyanobacteria bloom to construct a training set and a test set of the prediction model; constructing a prediction model based on the principle of Bayes' theorem; using the normalized ATBI band, the discretized meteorological data and the prior probability of the training set as the prediction model input, and the cyanobacteria bloom distribution results as the label of the prediction model, training the prediction model, and obtaining a trained prediction model;

[0045] Step S4, performing discretization and normalization processing on the MODIS satellite remote sensing image data collected in real time to obtain normalized real-time ATBI bands; performing the screening and discretization processing on the meteorological data collected in real time to obtain discretized real-time meteorological data; inputting the normalized real-time ATBI band and the discretized real-time meteorological data into the trained prediction model to obtain the posterior probability of the outbreak of cyanobacteria bloom events.

[0046] In step S1, MODIS satellite remote sensing image data is collected and screened to obtain ATBI band and cyanobacteria bloom distribution results; meteorological data is collected, and the screening of optimal cumulative meteorological factors and multicollinearity test are applied to obtain the screened meteorological data.

[0047] In some embodiments, in step S1, the screening of the optimal cumulative meteorological factor includes:

[0048] Seven days of meteorological factor data were collected, and the data of one day with the greatest contribution of each type of meteorological factor to the outbreak of cyanobacteria bloom was screened out from the seven days of data.

[0049] Specifically, the input data required for the cyanobacteria bloom prediction model is divided into two parts, namely MODIS satellite remote sensing image data and meteorological data.

[0050] Acquisition and preprocessing of MODIS satellite remote sensing image data: MODIS satellite data is acquired through the Google EarthEngine (GEE) platform. The preprocessing of satellite data mainly includes data resampling, water body mask, aquatic plant mask, cloud mask, cloud shadow mask, data source patch edge mask, ATBI band calculation and extraction of cyanobacteria bloom distribution results.

[0051] Acquisition and preprocessing of meteorological data: The meteorological factors used are daily average wind speed (WS), sunshine hours (SH), daily average atmospheric pressure (AP), daily average temperature (Tem), daily precipitation (Pre) and daily average relative humidity (RH). Meteorological data are obtained from the National Meteorological Information Center (https: / / data.cma.cn). The preprocessing of meteorological data includes two steps: screening of optimal cumulative meteorological factors and multicollinearity test. Since this prediction model takes into account the cumulative effect of meteorological factors from the previous 1 to 7 days on the outbreak of cyanobacterial blooms (for example, the temperature data of the day, this model takes into account 7 temperature data, namely the temperature of the day, the cumulative temperature of the previous 2 days, the previous 3 days...), it is necessary to screen the meteorological factor data required by the prediction model for the optimal cumulative meteorological factors, and screen out the data of the day with the largest contribution of each type of meteorological factor to the outbreak of cyanobacterial blooms. At the same time, in order to ensure that the meteorological data input into the model are independent of each other, the screened meteorological factors are tested for multicollinearity.

[0052] In step S2, using the box type Figure 4 The ATBI band is discretized and normalized by the quantile method to obtain the normalized ATBI band; Figure 4 The quantile method is used to discretize the screened meteorological data to obtain the discretized meteorological data; the prior probability is calculated by applying the ratio of the frequency of cyanobacterial bloom outbreaks to the frequency of effective pixels without cloud coverage in the study area.

[0053] In some embodiments, in step S2, the box type Figure 4 The quantile method performs discretization and normalization on the ATBI band, and the normalized ATBI band includes:

[0054] The ATBI bands in the cyanobacteria bloom area and the ATBI bands in the non-cyanobacteria bloom area were grouped and box plots were drawn. According to the upper and lower quartiles, median and 0 value of the two groups of box plots, they were divided into 8 levels to obtain the discrete ATBI bands.

[0055] In order to more accurately distinguish the values ​​of ATBI spectral index in cyanobacteria bloom area and non-cyanobacteria bloom area, ATBI was normalized, cyanobacteria bloom area was normalized to above 0, and non-cyanobacteria bloom area was normalized to below 0, and the normalized ATBI band was obtained. The specific classification rules are as follows: Figure 2 shown.

[0056] The step of normalizing the discrete ATBI band to obtain the normalized ATBI band includes:

[0057] The segmentation threshold of cyanobacteria bloom is calculated by iterative triangle method, and the segmentation threshold is subtracted from the discrete ATBI band of the whole scene image, and the discrete ATBI band of the cyanobacteria bloom area is normalized to a value above 0; for the discrete ATBI band of the image without bloom, the maximum value of the ATBI band on the image is directly subtracted from the discrete ATBI band of the whole scene image, and the discrete ATBI band of all non-cyanobacteria bloom areas is normalized to a value below 0;

[0058] The segmentation threshold of cyanobacteria blooms is calculated by iterative triangulation method, comprising:

[0059] First, the numerical distribution histogram of the discretized ATBI band is calculated, and an iterative triangle method is used to find a point on the histogram so that the distance from the point to the line connecting the peak and the tail point of the histogram is the maximum value. The numerical value of the ATBI band corresponding to the point is the required segmentation threshold.

[0060] The utilization box type Figure 4 The quantile method discretizes the filtered meteorological data, and the discretized meteorological data include:

[0061] The optimal cumulative meteorological factors after screening for each lake were grouped according to the cyanobacterial bloom days and non-cyanobacterial bloom days, and box plots were drawn; each type of meteorological factor was divided into 7 levels according to the upper and lower quartiles and median lines of the two groups of box plots of cyanobacterial bloom days and non-cyanobacterial bloom days, and the discrete meteorological data were obtained. The specific classification rules are as follows: Figure 2 shown.

[0062] The ratio of the frequency of blue algae blooms per pixel in the application study area to the frequency of effective pixels without cloud cover is used to calculate the prior probability, including:

[0063]

[0064] Among them, P(C1) represents the prior probability.

[0065] Specifically, for the monitoring results of cyanobacteria blooms in three large lakes, since the data from 2010 to 2018 were used as the model training data set, the monthly cyanobacteria bloom frequency of each lake from 2010 to 2018 was also calculated, that is, the ratio of the cyanobacteria bloom frequency per pixel in each study area to the frequency of valid images without cloud cover, as the prior probability P(C1). At the same time, the prior probability was Laplace corrected based on the above formula to correct the extreme cases where the prior probability was 0 or 1.

[0066] In step S3, the normalized ATBI bands, discretized meteorological data, prior probabilities and cyanobacteria bloom distribution results are used to construct a training set and a test set of a prediction model; a prediction model is constructed based on the principle of Bayes' theorem; the normalized ATBI bands, discretized meteorological data and prior probabilities of the training set are used as prediction model inputs, and the cyanobacteria bloom distribution results are used as labels of the prediction model, and the prediction model is trained to obtain a trained prediction model.

[0067] In some embodiments, in step S3, constructing a prediction model based on the principle of Bayesian theorem includes:

[0068] After the prior probability is input into the prediction model, the normalized ATBI bands and discretized meteorological data of the training set are traversed, and the normalized ATBI bands and discretized meteorological data are used as characteristic variables to construct a multidimensional conditional probability table under each characteristic variable for the two situations of whether a cyanobacterial bloom event occurs or not; according to the normalized ATBI bands and discretized meteorological data input data, the conditional probability of the response is determined using the multidimensional conditional probability table, and then the posterior probability of the cyanobacterial bloom event occurs is calculated.

[0069] Specifically, the training and testing data sets of the cyanobacteria bloom prediction model are composed of the current water surface state data set, the meteorological data set, and the future bloom state data set. The current water surface state data set is the ATBI band of the day calculated from the MODIS image. The meteorological data set is composed of the optimal meteorological factor data after the above preprocessing. According to the discretization and grading results of the meteorological factors, the meteorological factors are written as image band data, and one factor is written as a band image. The future bloom state data set is a binary map of the cyanobacteria bloom distribution results on the day to be predicted (pixels with bloom outbreaks are recorded as 1, and pixels without bloom outbreaks are recorded as 0).

[0070] The data set of 8 bands consisting of ATBI images (1 band), meteorological data (6 bands) and the bloom distribution map of the day to be predicted (1 band) is called a set of training data. The number of training data and test data for each study area predicting cyanobacteria blooms after 1, 4, and 7 days. It should be noted that when the constructed cyanobacteria bloom prediction model is applied in the future, it is necessary to use weather forecast data, that is, meteorological data between the day and the date to be predicted. However, in the process of building the prediction model, only the historical measured data of the meteorological station can be used. Therefore, in order to further explore the application effect of weather forecast data in the model constructed by the present invention, a prediction model test based on weather forecast meteorological data was also carried out. The weather forecast meteorological data of the three study areas from June to October 2023 were collected, as well as the ATBI images and bloom distribution maps calculated by the supporting MODIS images. The number of test data for cyanobacteria blooms predicted in each study area based on weather forecast meteorological data 1, 4, and 7 days later is shown in Table 1.

[0071] Table 1

[0072]

[0073] In step S4, the MODIS satellite remote sensing image data collected in real time is discretized and normalized to obtain normalized real-time ATBI bands; the meteorological data collected in real time is screened and discretized to obtain discretized real-time meteorological data; the normalized real-time ATBI band and the discretized real-time meteorological data are input into the trained prediction model to obtain the posterior probability of the outbreak of cyanobacteria bloom event.

[0074] Specifically, the input data required for the prediction model application is divided into two parts, namely, meteorological factor data (6 meteorological factors) and MODIS image data (1 ATBI band), that is, a 7-band data set. The 7-band data set is input into the trained prediction model to obtain the posterior probability of the outbreak of cyanobacteria bloom events.

[0075] In general, the input data required for model application are meteorological factor data and current MODIS image data. Based on the input of the application data of the above prediction model, the final output of the model is the pixel-by-pixel probability distribution map of cyanobacteria bloom outbreak in the target study area.

[0076] Here we take the model of predicting the next day for Taihu Lake, the next 4 days for Chaohu Lake and the next 7 days for Hulun Lake as examples to output the prediction model results, such as Figure 3 shown. Figure 3(a), (e) and (i) are the ATBI images of the day; (b), (f) and (j) are the probability maps of cyanobacteria bloom outbreaks on the day to be predicted; (c), (g) and (k) are the binary distribution maps of cyanobacteria blooms on the day to be predicted; (d), (h) and (l) are the distribution results of cyanobacteria blooms on the day to be predicted based on MODIS monitoring. Based on the binary distribution map of cyanobacteria blooms, the bloom area can be calculated, and then the bloom area proportion and cyanobacteria bloom outbreak level can be calculated. In addition, Figure 3 It can also be seen that the model prediction results are well consistent with the MODIS image monitoring results in terms of the distribution of cyanobacterial blooms, the area of ​​cyanobacterial blooms and regional bloom outbreaks.

[0077] In summary, the scheme proposed in the present invention can use the monitoring results of cyanobacterial blooms in large lakes and meteorological factor data as the main data sources, construct a short-term prediction model for cyanobacterial blooms in China's large lakes based on Bayesian theorem, and realize pixel-scale prediction of cyanobacterial blooms in the next 1, 4, and 7 days.

[0078] The second aspect of the present invention discloses a cyanobacteria bloom prediction system based on satellite remote sensing and weather forecast. Figure 4 FIG. 4 is a structural diagram of a cyanobacteria bloom prediction system based on satellite remote sensing and weather forecast according to an embodiment of the present invention; Figure 4 As shown, the system 100 includes:

[0079] The first processing module 101 is configured to collect MODIS satellite remote sensing image data, and screen the MODIS satellite remote sensing image data to obtain ATBI band and cyanobacteria bloom distribution results; collect meteorological data, and apply the screening of optimal cumulative meteorological factors and multicollinearity test to obtain screened meteorological data;

[0080] The second processing module 102 is configured to utilize a box-type Figure 4 The ATBI band is discretized and normalized by the quantile method to obtain the normalized ATBI band; Figure 4 The quantile method is used to discretize the filtered meteorological data to obtain the discretized meteorological data; the prior probability is calculated by using the ratio of the frequency of blue algae bloom outbreaks to the frequency of effective pixels without cloud coverage in the study area.

[0081] The third processing module 103 is configured to use the normalized ATBI band, the discretized meteorological data, the prior probability and the distribution results of the blue algae bloom to construct a training set and a test set of the prediction model; construct a prediction model based on the principle of Bayes' theorem; use the normalized ATBI band, the discretized meteorological data, and the prior probability of the training set as the prediction model input, and the blue algae bloom distribution results as the label of the prediction model, train the prediction model, and obtain the trained prediction model;

[0082] The fourth processing module 104 is configured to perform discretization and normalization processing on the MODIS satellite remote sensing image data collected in real time to obtain normalized real-time ATBI bands; perform the screening and discretization processing on the meteorological data collected in real time to obtain discretized real-time meteorological data; input the normalized real-time ATBI bands and the discretized real-time meteorological data into the trained prediction model to obtain the posterior probability of the outbreak of cyanobacteria bloom events.

[0083] According to the system of the second aspect of the present invention, the first processing module 101 is specifically configured such that the screening of the optimal cumulative meteorological factor includes:

[0084] Seven days of meteorological factor data were collected, and the data of one day with the greatest contribution of each type of meteorological factor to the outbreak of cyanobacteria bloom was screened out from the seven days of data.

[0085] Specifically, the input data required for the cyanobacteria bloom prediction model is divided into two parts, namely MODIS satellite remote sensing image data and meteorological data.

[0086] Acquisition and preprocessing of MODIS satellite remote sensing image data: MODIS satellite data is acquired through the Google EarthEngine (GEE) platform. The preprocessing of satellite data mainly includes data resampling, water body mask, aquatic plant mask, cloud mask, cloud shadow mask, data source patch edge mask, ATBI band calculation and extraction of cyanobacteria bloom distribution results.

[0087] Acquisition and preprocessing of meteorological data: The meteorological factors used are daily average wind speed (WS), sunshine hours (SH), daily average atmospheric pressure (AP), daily average temperature (Tem), daily precipitation (Pre) and daily average relative humidity (RH). Meteorological data are obtained from the National Meteorological Information Center (https: / / data.cma.cn). The preprocessing of meteorological data includes two steps: screening of optimal cumulative meteorological factors and multicollinearity test. Since this prediction model takes into account the cumulative effect of meteorological factors from the previous 1 to 7 days on the outbreak of cyanobacterial blooms (for example, the temperature data of the day, this model takes into account 7 temperature data, namely the temperature of the day, the cumulative temperature of the previous 2 days, the previous 3 days...), it is necessary to screen the meteorological factor data required by the prediction model for the optimal cumulative meteorological factors, and screen out the data of the day with the largest contribution of each type of meteorological factor to the outbreak of cyanobacterial blooms. At the same time, in order to ensure that the meteorological data input into the model are independent of each other, the screened meteorological factors are tested for multicollinearity.

[0088] According to the system of the second aspect of the present invention, the second processing module 102 is specifically configured as follows: Figure 4 The quantile method performs discretization and normalization on the ATBI band, and the normalized ATBI band includes:

[0089] The ATBI bands in the cyanobacteria bloom area and the ATBI bands in the non-cyanobacteria bloom area were grouped and box plots were drawn. According to the upper and lower quartiles, median and 0 value of the two groups of box plots, they were divided into 8 levels to obtain the discrete ATBI bands.

[0090] In order to more accurately distinguish the values ​​of ATBI spectral index in cyanobacteria bloom area and non-cyanobacteria bloom area, ATBI was normalized, cyanobacteria bloom area was normalized to above 0, and non-cyanobacteria bloom area was normalized to below 0, and the normalized ATBI band was obtained. The specific classification rules are as follows: Figure 2 shown.

[0091] The step of normalizing the discrete ATBI band to obtain the normalized ATBI band includes:

[0092] The segmentation threshold of cyanobacteria bloom is calculated by iterative triangle method, and the segmentation threshold is subtracted from the discrete ATBI band of the whole scene image, and the discrete ATBI band of the cyanobacteria bloom area is normalized to a value above 0; for the discrete ATBI band of the image without bloom, the maximum value of the ATBI band on the image is directly subtracted from the discrete ATBI band of the whole scene image, and the discrete ATBI band of all non-cyanobacteria bloom areas is normalized to a value below 0;

[0093] The segmentation threshold of cyanobacteria blooms is calculated by iterative triangulation method, comprising:

[0094] First, the numerical distribution histogram of the discretized ATBI band is calculated, and an iterative triangle method is used to find a point on the histogram so that the distance from the point to the line connecting the peak and the tail point of the histogram is the maximum value. The numerical value of the ATBI band corresponding to the point is the required segmentation threshold.

[0095] The utilization box type Figure 4 The quantile method discretizes the filtered meteorological data, and the discretized meteorological data include:

[0096] The optimal cumulative meteorological factors after screening for each lake were grouped according to the cyanobacterial bloom days and non-cyanobacterial bloom days, and box plots were drawn; each type of meteorological factor was divided into 7 levels according to the upper and lower quartiles and median lines of the two groups of box plots of cyanobacterial bloom days and non-cyanobacterial bloom days, and the discrete meteorological data were obtained. The specific classification rules are as follows: Figure 2 shown.

[0097] The ratio of the frequency of blue algae blooms per pixel in the application study area to the frequency of effective pixels without cloud cover is used to calculate the prior probability, including:

[0098]

[0099] Among them, P(C1) represents the prior probability.

[0100] Specifically, for the monitoring results of cyanobacteria blooms in three large lakes, since the data from 2010 to 2018 were used as the model training data set, the monthly cyanobacteria bloom frequency of each lake from 2010 to 2018 was also calculated, that is, the ratio of the cyanobacteria bloom frequency per pixel in each study area to the frequency of valid images without cloud cover, as the prior probability P(C1). At the same time, the prior probability was Laplace corrected based on the above formula to correct the extreme cases where the prior probability was 0 or 1.

[0101] According to the system of the second aspect of the present invention, the third processing module 103 is specifically configured as follows: the construction of the prediction model based on the principle of Bayes' theorem includes:

[0102] After the prior probability is input into the prediction model, the normalized ATBI bands and discretized meteorological data of the training set are traversed, and the normalized ATBI bands and discretized meteorological data are used as characteristic variables to construct a multidimensional conditional probability table under each characteristic variable for the two situations of whether a cyanobacterial bloom event occurs or not; according to the normalized ATBI bands and discretized meteorological data input data, the conditional probability of the response is determined using the multidimensional conditional probability table, and then the posterior probability of the cyanobacterial bloom event occurs is calculated.

[0103] Specifically, the training and testing data sets of the cyanobacteria bloom prediction model are composed of the current water surface state data set, the meteorological data set, and the future bloom state data set. The current water surface state data set is the ATBI band of the day calculated from the MODIS image. The meteorological data set is composed of the optimal meteorological factor data after the above preprocessing. According to the discretization and grading results of the meteorological factors, the meteorological factors are written as image band data, and one factor is written as a band image. The future bloom state data set is a binary map of the cyanobacteria bloom distribution results on the day to be predicted (pixels with bloom outbreaks are recorded as 1, and pixels without bloom outbreaks are recorded as 0).

[0104] The data set of 8 bands consisting of ATBI images (1 band), meteorological data (6 bands) and the bloom distribution map of the day to be predicted (1 band) is called a set of training data. The number of training data and test data for each study area predicting cyanobacteria blooms after 1, 4, and 7 days. It should be noted that when the constructed cyanobacteria bloom prediction model is applied in the future, it is necessary to use weather forecast data, that is, meteorological data between the day and the date to be predicted. However, in the process of building the prediction model, only the historical measured data of the meteorological station can be used. Therefore, in order to further explore the application effect of weather forecast data in the model constructed by the present invention, a prediction model test based on weather forecast meteorological data was also carried out. The weather forecast meteorological data of the three study areas from June to October 2023 were collected, as well as the ATBI images and bloom distribution maps calculated by the supporting MODIS images. The number of test data for cyanobacteria blooms predicted in each study area based on weather forecast meteorological data 1, 4, and 7 days later is shown in Table 1.

[0105] According to the system of the second aspect of the present invention, the fourth processing module 104 is specifically configured as follows: the input data required for the application of the prediction model is divided into two parts, namely, meteorological factor data (6 meteorological factors) and MODIS image data (1 ATBI band), that is, a 7-band data set. The 7-band data set is input into the trained prediction model to obtain the posterior probability of the outbreak of the cyanobacteria bloom event.

[0106] In general, the input data required for model application are meteorological factor data and current MODIS image data. Based on the input of the application data of the above prediction model, the final output of the model is the pixel-by-pixel probability distribution map of cyanobacteria bloom outbreak in the target study area.

[0107] Here we take the model of predicting the next day for Taihu Lake, the next 4 days for Chaohu Lake and the next 7 days for Hulun Lake as examples to output the prediction model results, such as Figure 3 shown. Figure 3 (a), (e) and (i) are the ATBI images of the day; (b), (f) and (j) are the probability maps of cyanobacteria bloom outbreaks on the day to be predicted; (c), (g) and (k) are the binary distribution maps of cyanobacteria blooms on the day to be predicted; (d), (h) and (l) are the distribution results of cyanobacteria blooms on the day to be predicted based on MODIS monitoring. Based on the binary distribution map of cyanobacteria blooms, the bloom area can be calculated, and then the bloom area proportion and cyanobacteria bloom outbreak level can be calculated. In addition, Figure 3 It can also be seen that the model prediction results are well consistent with the MODIS image monitoring results in terms of the distribution of cyanobacterial blooms, the area of ​​cyanobacterial blooms and regional bloom outbreaks.

[0108] The third aspect of the present invention discloses an electronic device. The electronic device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of any one of the methods for predicting cyanobacteria blooms based on satellite remote sensing and weather forecasts disclosed in the first aspect of the present invention are implemented.

[0109] Figure 5 is a structural diagram of an electronic device according to an embodiment of the present invention, such as Figure 5 As shown, the electronic device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. Among them, the processor of the electronic device is used to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the electronic device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, near field communication (NFC) or other technologies. The display screen of the electronic device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the electronic device can be a touch layer covered on the display screen, or a button, a trackball or a touch pad set on the housing of the electronic device, or an external keyboard, touch pad or mouse, etc.

[0110] Those skilled in the art will understand that Figure 5The structure shown in the figure is only a structural diagram of the part related to the technical solution of the present disclosure, and does not constitute a limitation on the electronic device to which the technical solution of the present application is applied. The specific electronic device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0111] The fourth aspect of the present invention discloses a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the method for predicting cyanobacteria blooms based on satellite remote sensing and weather forecast in any one of the first aspects of the present invention are implemented.

[0112] Please note that the technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification. The above embodiments only express several implementation methods of the present application, and their descriptions are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that for ordinary technicians in this field, without departing from the concept of the present application, several variations and improvements can be made, which all belong to the scope of protection of the present application. Therefore, the scope of protection of the patent in this application shall be based on the attached claims.

Claims

1. A method for predicting cyanobacteria blooms based on satellite remote sensing and weather forecast, characterized in that: The method comprises: Step S1, collecting MODIS satellite remote sensing image data, and screening the MODIS satellite remote sensing image data to obtain ATBI band and cyanobacteria bloom distribution results; collecting meteorological data, and applying the screening of optimal cumulative meteorological factors and multicollinearity test to obtain the screened meteorological data; Step S2, using the box plot quartile method to discretize and normalize the ATBI band to obtain a normalized ATBI band; using the box plot quartile method to discretize the screened meteorological data to obtain discretized meteorological data; applying the ratio of the frequency of blue algae bloom outbreaks pixel by pixel in the study area to the frequency of effective pixels without cloud coverage to calculate the prior probability; Step S3, using the normalized ATBI band, the discretized meteorological data, the prior probability and the distribution results of cyanobacteria bloom to construct a training set and a test set of the prediction model; constructing a prediction model based on the principle of Bayes' theorem; using the normalized ATBI band, the discretized meteorological data and the prior probability of the training set as the prediction model input, and the cyanobacteria bloom distribution results as the label of the prediction model, training the prediction model, and obtaining a trained prediction model; Step S4, performing discretization and normalization processing on the MODIS satellite remote sensing image data collected in real time to obtain normalized real-time ATBI bands; performing the screening and discretization processing on the meteorological data collected in real time to obtain discretized real-time meteorological data; inputting the normalized real-time ATBI band and the discretized real-time meteorological data into the trained prediction model to obtain the posterior probability of the outbreak of cyanobacteria bloom events.

2. The method for predicting cyanobacteria blooms based on satellite remote sensing and weather forecast according to claim 1, characterized in that: In step S1, the screening of the optimal cumulative meteorological factor includes: Seven days of meteorological factor data were collected, and the data of one day with the greatest contribution of each type of meteorological factor to the outbreak of cyanobacteria bloom was screened out from the seven days of data.

3. The method for predicting cyanobacteria blooms based on satellite remote sensing and weather forecast according to claim 1, characterized in that: In step S2, the ATBI band is discretized and normalized by using the box plot quartile method to obtain the normalized ATBI band, which includes: The ATBI bands in the cyanobacteria bloom area and the ATBI bands in the non-cyanobacteria bloom area were grouped and box plots were drawn. According to the upper and lower quartiles, median and 0 value of the two groups of box plots, they were divided into 8 levels to obtain the discrete ATBI bands. The discretized ATBI band is normalized to obtain a normalized ATBI band.

4. The method for predicting cyanobacteria blooms based on satellite remote sensing and weather forecast according to claim 3, characterized in that: In the step S2, the discretized ATBI band is normalized to obtain the normalized ATBI band, which includes: The segmentation threshold of cyanobacteria bloom is calculated by iterative triangle method, and the segmentation threshold is subtracted from the discrete ATBI band of the whole scene image, and the discrete ATBI band of the cyanobacteria bloom area is normalized to a value above 0; for the discrete ATBI band of the image without bloom, the maximum value of the ATBI band on the image is directly subtracted from the discrete ATBI band of the whole scene image, and the discrete ATBI band of all non-cyanobacteria bloom areas is normalized to a value below 0; The segmentation threshold of cyanobacteria blooms is calculated by iterative triangulation method, comprising: First, the numerical distribution histogram of the discretized ATBI band is calculated, and an iterative triangle method is used to find a point on the histogram so that the distance from the point to the line connecting the peak and the tail point of the histogram is the maximum value. The numerical value of the ATBI band corresponding to the point is the required segmentation threshold.

5. The method for predicting cyanobacteria blooms based on satellite remote sensing and weather forecast according to claim 1, characterized in that: In step S2, the meteorological data after screening are discretized by using the box plot quartile method, and the discretized meteorological data obtained include: The optimal cumulative meteorological factors after screening for each lake were grouped according to cyanobacterial bloom days and non-cyanobacterial bloom days, and box plots were drawn; each type of meteorological factor was divided into 7 levels according to the upper and lower quartiles and median lines of the two groups of box plots of cyanobacterial bloom days and non-cyanobacterial bloom days, and the discrete meteorological data were obtained.

6. The method for predicting cyanobacteria blooms based on satellite remote sensing and weather forecast according to claim 1, characterized in that: In step S2, the ratio of the frequency of blue algae bloom outbreaks per pixel in the application study area to the frequency of effective pixels without cloud cover is used to calculate the prior probability, which includes: Among them, P(C1) represents the prior probability.

7. The method for predicting cyanobacteria blooms based on satellite remote sensing and weather forecast according to claim 1, characterized in that: In step S3, constructing a prediction model based on the principle of Bayesian theorem includes: After the prior probability is input into the prediction model, the normalized ATBI bands and discretized meteorological data of the training set are traversed, and the normalized ATBI bands and discretized meteorological data are used as characteristic variables to construct a multidimensional conditional probability table under each characteristic variable for the two situations of whether a cyanobacterial bloom event occurs or not; according to the normalized ATBI bands and discretized meteorological data input data, the conditional probability of the response is determined using the multidimensional conditional probability table, and then the posterior probability of the cyanobacterial bloom event occurs is calculated.

8. A system for predicting blue algae blooms based on satellite remote sensing and weather forecast, characterized in that: The system comprises: The first processing module is configured to collect MODIS satellite remote sensing image data, and screen the MODIS satellite remote sensing image data to obtain ATBI band and cyanobacteria bloom distribution results; collect meteorological data, and apply the screening of optimal cumulative meteorological factors and multicollinearity test to obtain screened meteorological data; The second processing module is configured to discretize and normalize the ATBI band using the box plot quartile method to obtain the normalized ATBI band; discretize the filtered meteorological data using the box plot quartile method to obtain the discretized meteorological data; and calculate the prior probability by using the ratio of the frequency of blue algae bloom outbreaks per pixel in the study area to the frequency of effective pixels without cloud coverage; The third processing module is configured to use the normalized ATBI band, the discretized meteorological data, the prior probability and the distribution results of the blue algae bloom to construct a training set and a test set of the prediction model; construct a prediction model based on the principle of Bayes' theorem; use the normalized ATBI band, the discretized meteorological data, and the prior probability of the training set as the prediction model input, and the blue algae bloom distribution results as the label of the prediction model, train the prediction model, and obtain the trained prediction model; The fourth processing module is configured to perform discretization and normalization processing on the MODIS satellite remote sensing image data collected in real time to obtain normalized real-time ATBI bands; perform the screening and discretization processing on the meteorological data collected in real time to obtain discretized real-time meteorological data; input the normalized real-time ATBI band and the discretized real-time meteorological data into the trained prediction model to obtain the posterior probability of the outbreak of cyanobacteria bloom event.

9. An electronic device, characterized in that: The electronic device includes a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the steps in the method for predicting cyanobacteria blooms based on satellite remote sensing and weather forecast described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the steps in the method for predicting cyanobacterial blooms based on satellite remote sensing and weather forecast described in any one of claims 1 to 7 are implemented.

Images