Artificial intelligence-based enteromorpha prevention and control early warning method and system

Abstract

The invention relates to the technical field of prevention and control early warning, and discloses an enteromorpha prevention and control early warning method and system based on artificial intelligence, and the method comprises the following steps: carrying out the independent region division of a range needing early warning, and carrying out the sorting through serial numbers; ocean comprehensive information is collected; encoding the collected ocean comprehensive information, inputting the encoded ocean comprehensive information into the ocean ecological coupling model, and outputting ocean data of each region in each hour in n hours in the future; obtaining first comprehensive information and second comprehensive information in combination with the ocean data; generating first plane data based on the first comprehensive information, and generating second plane data based on the second comprehensive information; inputting the first plane data and the second plane data into an artificial intelligence model, and outputting early warning information representing each region; and according to the information output by the artificial intelligence model, decoding and sending out early warning information and alarm reminding. According to the method, the influence of eddy current is considered in the model, so that accurate prevention and control early warning of enteromorpha is realized.

Description

Technical Field

[0001] The present invention relates to the field of prevention, control and early warning technology, and more specifically, to an artificial intelligence-based enteromorpha prevention, control and early warning method and system. Background Art

[0002] In coastal areas, especially in semi-enclosed or open seas, ocean energy power generation devices are increasingly being deployed for the development and utilization of renewable energy. For example, oscillating water column wave energy power generation devices are often arranged nearshore or offshore, generating electricity by capturing wave energy. Because these devices are usually located in offshore areas, the water currents are not as strong, and offshore areas are more likely to have land-based nutrient inputs, which easily form an environment suitable for the growth of enteromorpha. Therefore, these areas may also be susceptible to enteromorpha outbreaks. Enteromorpha may adhere to the surface or structure of the device, affecting its normal operation. A large amount of enteromorpha may change the hydrodynamic environment and affect the efficiency of the device's wave energy conversion. This in turn affects the power generation efficiency of the power generation device, resulting in a reduction in electricity supply. Therefore, prevention and control of enteromorpha is very necessary.

[0003] The oscillating water column wave energy power generation device relies on the up and down oscillation of the water column to compress the air. This process requires frequent flow of water inside and outside the device. Because the physical structure of the device will hinder and guide the flow of seawater, it may cause uneven flow speed and direction around it. Under the influence of waves, the local energy release and uneven force distribution of the oscillating water column will cause the rotational movement of the water body, especially at the edge and bottom of the device. This uneven energy distribution may lead to the formation of small-scale eddies. Once formed, the eddies can become areas where nutrients and tiny particles condense, providing rich nutrient conditions, which may accelerate the growth and aggregation of Enteromorpha. In addition, the eddies will provide a relatively closed environment, making it difficult for floating Enteromorpha to spread to other areas, thereby increasing the difficulty of predicting Enteromorpha.

[0004] Due to the accumulation of nutrients in a small area, the outbreak of Ulva in a small area may be accelerated, but there is no display and warning in the forecast, which affects the efficiency of the device's wave energy conversion, and then affects the power generation efficiency, resulting in a reduction in electricity supply. Summary of the Invention

[0005] The present invention provides an artificial intelligence-based enteromorpha prevention and control early warning method and system, which solves the technical problem in related technologies that eddy currents generated around power generation devices lead to accelerated growth of enteromorpha, which is difficult to predict and issue early warnings.

[0006] The present invention provides an artificial intelligence-based early warning method for enteromorpha prevention and control, comprising the following steps: Step 100: Divide the area requiring early warning into independent regions, ensuring that each region does not overlap, forming multiple adjacent and continuous polygonal regions, and sorting them by sequence number; Step 200, collecting comprehensive ocean information including: hydrological element data of each region, ecological element data, ocean stereo observation data, satellite observation data and reanalysis data; Step 300: Encode the collected comprehensive ocean information and input it into the constructed ocean ecological coupling model, and output the ocean data of each area every hour for the next n hours; Step 400: Acquire first and second comprehensive information based on the ocean data output by the marine ecological coupling model. The first comprehensive information includes ocean data, equipment data, and enteromorpha image data. The second comprehensive information includes ocean data, equipment data, enteromorpha image data, terrain structure data, and equipment structure data. Step 500, generating first plane data based on first comprehensive information, and generating second plane data based on second comprehensive information; Step 600: Input the first plane data and the second plane data into an artificial intelligence model. The artificial intelligence model includes a first encoder, a second encoder, a third encoder, a fourth encoder, a first feature fusion layer, a first fully connected layer, and a second fully connected layer. The first encoder is used to input the first plane data and output a first identification feature to the second encoder. The second encoder outputs a second identification feature to the first feature fusion layer. The third encoder is used to input the second plane data and output a third identification feature to the fourth encoder. The fourth encoder outputs the fourth identification feature to the first feature fusion layer and the second output layer. The second output layer outputs eddy current information representing each area. The first feature fusion layer outputs the first fusion feature to the first output layer. The first output layer outputs warning information representing each area. Step 700: Decode the information output by the artificial intelligence model and issue early warning information and alarm reminders.

[0007] In a preferred embodiment, the hydrological element data include: temperature, salinity, depth, and oxygen content of each region; Ecological factor data include: pH value, water temperature, salinity, dissolved oxygen, chemical oxygen demand, active phosphate content, ammonia nitrogen content, nitrate nitrogen content, nitrite nitrogen content, inorganic nitrogen content, chlorophyll a concentration, sediment sulfide content, organic carbon content, total phosphorus content, total nitrogen content, and the species, density, and biomass of marine phytoplankton and zooplankton in each area; Ocean stereo observation data include: air pressure, temperature and humidity, wind speed and direction, visibility, precipitation, water temperature, tide level, wave height, and salinity in each region; Satellite observation data include: sea surface height and sea surface temperature in various regions; Reanalysis data: density, current velocity, mixed layer depth, sea surface wind stress; Ocean data include: ocean temperature, salinity, density, current field, chlorophyll concentration, nitrate concentration, and phosphate concentration in each region; Equipment data includes: distribution location of power generation equipment, kinetic energy conversion efficiency, and load; Enteromorpha image data includes: high-resolution images of the water surface in various areas taken by drones; Topographic structure data include: topographic structure maps of each region; Equipment structure data includes: equipment structure diagram.

[0008] In a preferred embodiment, the first plane data includes a first bin for storing data, a first serial data stream indexed by each bin to a first data object, and a first relationship list for storing association relationships between the first data objects indexed by the first bin, wherein the element value in the i-th row and j-th column of the first relationship list represents a quantized value of the association relationship between the first data objects indexed by the i-th and j-th first bins. The correlation relationship includes high and low, corresponding to two quantitative values ​​respectively.

[0009] The first data object includes: each divided area; The data objects with high association relationship are defined as follows: if the regions are adjacent, the association relationship is high; A No. 1 serial data stream of a No. 1 data object includes n No. 1 data units sorted in time, and the t-th No. 1 data unit represents the No. 1 comprehensive information collected at the t-th moment of the No. 1 data object.

[0010] In a preferred embodiment, the second plane data includes a second bin for storing data, a second serial data stream indexing a second data object, and a second relationship list for storing association relationships between the second data objects indexed by the second bin, wherein the element value in the i-th row and j-th column of the second relationship list represents a quantized value of the association relationship between the second data objects indexed by the i-th and j-th second bins. The correlation relationship includes high and low, corresponding to two quantitative values ​​respectively.

[0011] The first data object includes: each divided area; The data objects with high association relationship are defined as follows: if the regions are adjacent, the association relationship is high; A No. 2 serial data stream of a No. 2 data object includes n No. 2 data units sorted in time, and the t-th No. 2 data unit represents the No. 2 comprehensive information collected at the t-th moment of the No. 2 data object.

[0012] In a preferred embodiment, the artificial intelligence model includes a first encoder, a second encoder, a third encoder, a fourth encoder, a first feature fusion layer, a first fully connected layer and a second fully connected layer, wherein the first encoder is used to input plane data No. 1 and output the first identification feature to the second encoder, the second encoder outputs the second identification feature to the first feature fusion layer, the third encoder is used to input plane data No. 2 and output the third identification feature to the fourth encoder, the fourth encoder outputs the fourth identification feature to the first feature fusion layer and the second output layer, the second output layer outputs eddy current information representing each area, the first feature fusion layer outputs the first fusion feature to the first output layer, and the first output layer outputs warning information representing each area.

[0013] In a preferred embodiment, the first comprehensive information and the second comprehensive information are input into the model after feature engineering, and the first comprehensive information and the second comprehensive information include image modalities, and the CNN algorithm is used as a feature engineering method for the image modality information.

[0014] In a preferred embodiment, the outbreak level of Enteromorpha is divided into three levels: high, medium and low according to the distribution density of Enteromorpha.

[0015] In a preferred embodiment, the loss function during artificial intelligence model training is as follows: ; in represents the first loss value, represents the total number of training samples, Indicates the The true distribution density and outbreak risk level of the Enteromorpha samples, The output of the model The distribution density and outbreak risk level of Enteromorpha in each sample.

[0016] In a preferred embodiment, the second plane data is input into the third encoder, which outputs the third identification feature to the fourth encoder. The fourth encoder outputs the fourth identification feature to the second fully connected layer, and the second fully connected layer outputs the eddy current information representing each area. This process is independently pre-trained, and the training loss function is as follows: ; in represents the second loss value, represents the total number of training samples, Indicates the The true position, intensity, and radius of the eddy current of each sample, The output of the model The location, intensity, and radius of the eddy current of each sample.

[0017] The beneficial effects of the present invention are: the present invention takes into account that some power generation devices may generate eddies during the power generation process, constructs a marine ecological model and performs predictions, and then by considering the influence of eddies in the artificial intelligence model, the prevention and control of the enteromorpha around the power generation device is more accurate and efficient, reducing the impact of the enteromorpha outbreak on power generation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flow chart of an artificial intelligence-based enteromorpha prevention and control early warning method of the present invention. DETAILED DESCRIPTION

[0019] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed solely to enable those skilled in the art to better understand and implement the subject matter described herein, and that the functions and arrangements of the elements discussed may be varied without departing from the scope of this specification. Various examples may omit, substitute, or add various processes or components as needed. Furthermore, features described in some examples may be combined in other examples.

[0020] At least one embodiment of the present invention discloses an artificial intelligence-based early warning method for preventing and controlling enteromorpha, such as Figure 1 As shown, the following steps are included: Step 100: Divide the area requiring early warning into independent regions, ensuring that each region does not overlap, forming multiple adjacent and continuous polygonal regions, and sorting them by sequence number; Step 200, collecting comprehensive ocean information including: hydrological element data of each region, ecological element data, ocean stereo observation data, satellite observation data and reanalysis data; Hydrological element data include: temperature, salinity, depth, and oxygen content in each area; Ecological factor data include: pH value, water temperature, salinity, dissolved oxygen, chemical oxygen demand, active phosphate content, ammonia nitrogen content, nitrate nitrogen content, nitrite nitrogen content, inorganic nitrogen content, chlorophyll a concentration, sediment sulfide content, organic carbon content, total phosphorus content, total nitrogen content, and the species, density, and biomass of marine phytoplankton and zooplankton in each area; Ocean stereo observation data include: air pressure, temperature and humidity, wind speed and direction, visibility, precipitation, water temperature, tide level, wave height, and salinity in each region; Satellite observation data include: sea surface height and sea surface temperature in various regions; Reanalysis data: density, current velocity, mixed layer depth, sea surface wind stress; In one embodiment of the present invention, data is collected once at a fixed time interval, and the default value of the time interval is 1 hour.

[0021] Step 300: Encode the collected comprehensive ocean information and input it into the constructed ocean ecological coupling model, and output the ocean data of each area every hour for the next n hours; Ocean data include: ocean temperature, salinity, density, current field, chlorophyll concentration, nitrate concentration, and phosphate concentration in each region; In one embodiment of the present invention, the FVCOM ocean dynamic model is used as a submodule of the coupling model. The model uses an unstructured triangular grid in the horizontal direction and a vertical grid. Coordinate transformation and the numerical method using the finite volume method can simulate the three-dimensional original control equations of the free surface. This finite volume method combines the advantages of the finite element method in dealing with the complex and tortuous boundaries of the bay shore with the simple discrete structure and high computational efficiency of the finite difference method. Moreover, for areas with complex terrain and shore boundaries near the coast and estuary, it better guarantees the conservation of mass, momentum, salinity and heat. The main control equations of the model include the momentum equation, the continuity equation, the salinity diffusion equation, the temperature diffusion equation, the state equation, etc. The tidal control equation and the turbulence closure equation are used to construct a complete three-dimensional shallow sea tidal wave well-solved equation group. The finite volume method is used to solve the control equations, the dry / wet grid module is used to control the moving boundary, and the internal and external modes are used for alternating calculations.

[0022] In one embodiment of the present invention, the basic equation of the model is: Vertical adoption Coordinate transformation, The coordinate transformation is defined as:

[0023] exist Under coordinate transformation, the control equations of the model can be written as:

[0024]

[0025]

[0026] The horizontal diffusion term is defined as:

[0027]

[0028] in is the density of seawater, is the Coriolis parameter, U and V are the horizontal velocity components, is the vertical flow velocity, is the time coordinate, is the acceleration due to gravity, is the average water depth, is the water level; on the sea surface, , on the seabed, ; is the horizontal eddy viscosity coefficient, is the vertical turbulent viscosity coefficient.

[0029] In one embodiment of the present invention, the numerical model of marine ecosystem dynamics is mainly affected by physical, biological, and chemical processes. The model can reflect the changes in biological and chemical parameters under different physical conditions and external forcing, and can be comprehensively expressed by the following equation:

[0030] Where B represents the state variables of the marine ecological model, such as phytoplankton biomass, zooplankton biomass, dissolved oxygen, etc. Indicates the change of the above state variables over time; u, v, w represent the flow velocity in the x, y, and z directions respectively; represents the diffusion coefficient in the horizontal direction; represents the diffusion coefficient in the vertical direction; is the impact of changes in biological and chemical processes on ecological variables, and S is the source term of nutrients.

[0031] In one embodiment of the present invention, the ecological dynamic model of the coupling model includes at least the following ecological dynamic processes of nine state variables, and the basic equation is:

[0032]

[0033]

[0034]

[0035]

[0036]

[0037]

[0038]

[0039]

[0040] in is the total amount of small phytoplankton, is the total amount of macrophytoplankton, is the total amount of small zooplankton, is the total amount of macrozooplankton, is the nitrate concentration, is the ammonium salt concentration, is the concentration of particle organic nitrogen, is the silicate concentration, is the concentration of example organic silicon, Ah is the thermal diffusion coefficient, and F1 to F17 are the fluxes of various ecological processes.

[0041] In one embodiment of the present invention, to achieve real-time transfer and synchronous integration of ocean model variables into the ecological model, the ecological model is embedded within the ocean model. The calculation of all ecological variables in the coupled model involves both physical transfer and ecological processes, as shown in the following formula:

[0042] Where i=1,…,9, for example, C1=NO3, C2=NH4, etc. It represents the impact of ocean physical processes (mainly convection diffusion and eddy dissipation processes) on ecological variables, specifically expressed as:

[0043] where w is the vertical velocity and ATv is the eddy diffusion coefficient.

[0044] Step 400: Acquire first and second comprehensive information based on the ocean data output by the marine ecological coupling model. The first comprehensive information includes ocean data, equipment data, and enteromorpha image data. The second comprehensive information includes ocean data, equipment data, enteromorpha image data, terrain structure data, and equipment structure data. Equipment data includes: distribution location of power generation equipment, kinetic energy conversion efficiency, and load; Enteromorpha image data includes: high-resolution images of the water surface in various areas taken by drones; Topographic structure data include: topographic structure maps of each region; Equipment structure data includes: equipment structure diagram; Step 500, generating first plane data based on first comprehensive information, and generating second plane data based on second comprehensive information; The first plane data includes a first bin for storing data, a first serial data stream indexed by each first bin to a first data object, and a first relationship list for storing association relationships between the first data objects indexed by the first bin, wherein the element value in the i-th row and j-th column of the first relationship list represents a quantized value of the association relationship between the first data objects indexed by the i-th and j-th first bins; The correlation relationship includes high and low, corresponding to two quantitative values ​​respectively.

[0045] The first data object includes: each divided area; The data objects with high association relationship are defined as follows: if the regions are adjacent, the association relationship is high; A serial data stream No. 1 of a data object No. 1 includes n No. 1 data units ordered by time, and the t-th No. 1 data unit represents the No. 1 comprehensive information collected at the t-th moment of the data object No. 1; The second plane data includes a second bin for storing data, a second serial data stream indexed by a second bin to a second data object, and a second relationship list for storing association relationships between the second data objects indexed by the second bin, wherein the element value in the i-th row and j-th column of the second relationship list represents a quantized value of the association relationship between the second data objects indexed by the i-th and j-th second bins; The correlation relationship includes high and low, corresponding to two quantitative values ​​respectively.

[0046] The first data object includes: each divided area; The data objects with high association relationship are defined as follows: if the regions are adjacent, the association relationship is high; A second serial data stream of a second data object includes n second data units sorted in time, and the t-th second data unit represents the second comprehensive information collected at the t-th moment of the second data object; Step 600: Input the first plane data and the second plane data into the artificial intelligence model, and output warning information representing each area; The artificial intelligence model includes a first encoder, a second encoder, a third encoder, a fourth encoder, a first feature fusion layer, a first fully connected layer, and a second fully connected layer, wherein the first encoder is used to input plane data No. 1 and output a first identification feature to the second encoder, the second encoder outputs a second identification feature to the first feature fusion layer, the third encoder is used to input plane data No. 2 and output a third identification feature to the fourth encoder, the fourth encoder outputs the fourth identification feature to the first feature fusion layer and the second output layer, the second output layer outputs eddy current information representing each area, the first feature fusion layer outputs the first fusion feature to the first output layer, and the first output layer outputs warning information representing each area; In one embodiment of the present invention, the first comprehensive information and the second comprehensive information are input into the model after feature engineering, and the first comprehensive information and the second comprehensive information include image modalities, and the CNN algorithm is used as a feature engineering method for the image modality information; In one embodiment of the present invention, the calculation formula of the first encoder is as follows: ; ; ; ; ; ; in represents the t-th data unit of the first serial data stream corresponding to the v-th bin of the first plane data, and They respectively represent the t-th and t-1-th first identification features of the No. 1 serial data stream corresponding to the v-th No. 1 bin of the No. 1 plane data, 、 、 、 and They represent the tth first, second, third, fourth and fifth intermediate features of the vth first bin of the first plane data, respectively. represents the t-1th fourth intermediate feature of the vth first surface element of the first plane data, → represents the cycle direction, 、 、 、 、 、 、 、 Represent the first, second, third, fourth, fifth, sixth, seventh, and eighth serial recognition weight parameters respectively, 、 、 、 Represents the first, second, third, and fourth serial identification bias parameters respectively represents the Hadamard product, , n represents the total number of data units in the first serial data stream, when t=1 , tanh is the tanh function, Represents the Sigmoid function.

[0047] In one embodiment of the present invention, the calculation formula of the second encoder is as follows: ; in represents the first-surface feature of the vth first-surface in the lth layer, represents the set of first-surface elements with high correlation with first-surface element v, represents the first plane recognition weight matrix of layer l, is the attention weight of the first surfel u on the first surfel v in the lth layer, which is calculated by the following formula: ; in is a learnable attention vector, Represents vector concatenation operation, represents the second plane recognition weight matrix of layer l, represents transpose, yes Activation function; and They represent the first-surface features of the vth first-surface element in the l-1th layer, , E represents the total number of layers, when l=1 , , when l=E Equal to the identification feature of the first bin v; and They respectively represent the t-th first identification feature of the No. 1 serial data stream corresponding to the v-th and u-th No. 1 bins of the No. 1 plane data, and n represents the total number of No. 1 data units in the No. 1 serial data stream.

[0048] In one embodiment of the present invention, the calculation formula of the third encoder is as follows: ; ; ; in The t-th third identification feature of the second serial data stream corresponding to the v-th second bin of the second plane data, represents the t-1th third identification feature of the second serial data stream corresponding to the vth second bin, and They represent the first and second intermediate features of the vth second bin, represents the tth second data unit of the second serial data stream of the vth second bin, and is the serial recognition weight matrix, is the serial identification bias parameter, represents the Sigmoid function, represents the Hadamard product, and tanh is the tanh function.

[0049] In one embodiment of the present invention, the calculation formula of the fourth encoder is as follows: ; ; in represents the aggregate representation of the vth second-order facet at layer l, and They represent the third identification features of the vth second-order facet at the lth layer and the l-1th layer, respectively. represents the set of second facets that have a high correlation with the second facet v; Indicates the universal quantifier symbol; represents the facet feature of the u-th second facet at the l-1th layer, , E represents the total number of layers, when l=1 , , 、 Respectively represent the t-th and t-1-th third identification features of the second serial data stream corresponding to the v-th and u-th second bins of the second plane data, n represents the total number of second data units in the second serial data stream, and l=E Equal to the third identification feature of the second bin v, Represents the aggregation function of layer l, such as average or maximum pooling function. represents the plane recognition weight matrix of the lth layer, σ represents the Sigmoid function, Represents the concatenation function.

[0050] In one embodiment of the present invention, the calculation formula of the second fully connected layer is as follows: ; in Represents a vector, the i-th component of the vector represents the eddy current information of the i-th region, one component represents a vector, and the value of the vector represents the position, intensity, and radius of the eddy current at the future time point. represents the fourth identification feature of the vth second bin in the lth layer, represents the set of second facets that have a high correlation with the second facet v, represents the second fully connected function.

[0051] In one embodiment of the present invention, the calculation formula of the first feature fusion layer is as follows: ; in, represents the first fusion feature, represents the concatenation function, represents the sum weight matrix, represents the summation bias parameter, represents the second identification feature of the vth first bin in the lth layer, represents the set of first-surface elements with high correlation with first-surface element v, represents the fourth identification feature of the vth second bin in the lth layer, It represents the set of second-surface elements that have a high correlation with the second-surface element v.

[0052] In one embodiment of the present invention, the calculation formula of the first output layer is as follows: ; in Represents a vector, the i-th component of the vector represents the warning information of the i-th area, and one component represents a vector. The values ​​of the vector represent the distribution density and outbreak risk level of Enteromorpha at future time points. represents the first fusion feature, represents the first fully connected function.

[0053] In one embodiment of the present invention, the outbreak level of Enteromorpha is divided into three levels: high, medium, and low according to the distribution density of Enteromorpha.

[0054] Step 700: Decode the information output by the artificial intelligence model and issue early warning information and alarm reminders.

[0055] In one embodiment of the present invention, when the outbreak risk in all areas is low, only a warning message is issued without an alarm reminder; otherwise, a warning message and an alarm reminder are issued.

[0056] In one embodiment of the present invention, the loss function during artificial intelligence model training is as follows: ; in represents the first loss value, represents the total number of training samples, Indicates the The true distribution density and outbreak risk level of the Enteromorpha samples, The output of the model The distribution density and outbreak risk level of Enteromorpha in each sample.

[0057] In one embodiment of the present invention, the second plane data is input into the third encoder, which outputs the third identification feature to the fourth encoder. The fourth encoder outputs the fourth identification feature to the second fully connected layer, and the second fully connected layer outputs eddy current information representing each region. This process is independently pre-trained, and the training loss function is as follows: ; in represents the second loss value, represents the total number of training samples, Indicates the The true position, intensity, and radius of the eddy current of each sample, The output of the model The location, intensity, and radius of the eddy current of each sample.

[0058] The above describes an embodiment of the present invention, but this embodiment is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Ordinary technicians in this field can also make more forms of equivalent embodiments based on the inspiration of this embodiment, all of which are protected by this embodiment.

Claims

1. An artificial intelligence-based early warning method for Enteromorpha control, characterized in that: The following steps are involved: Step 100, the range that needs to be warned is divided into independent regions, ensuring that each region does not overlap, forming multiple adjacent and continuous polygonal regions, and sorting them by sequence number; Step 200, collecting comprehensive ocean information including: hydrological element data of each region, ecological element data, ocean stereo observation data, satellite observation data and reanalysis data; Step 300, encode the collected comprehensive ocean information and input it into the constructed marine ecological coupling model, and output the ocean data of each area every hour for the next n hours; Step 400, combining the ocean data output by the marine ecological coupling model to obtain No. 1 comprehensive information and No. 2 comprehensive information, the No. 1 comprehensive information includes: ocean data, equipment data, Enteromorpha image data, No. 2 comprehensive information includes: ocean data, equipment data, Enteromorpha image data, terrain structure data, equipment structure data; Step 500, generating first plane data based on first comprehensive information, and generating second plane data based on second comprehensive information; Step 600, inputting the No. 1 plane data and the No. 2 plane data into an artificial intelligence model, the artificial intelligence model comprising a first encoder, a second encoder, a third encoder, a fourth encoder, a first feature fusion layer, a first fully connected layer, and a second fully connected layer, wherein the first encoder is used to input the No. 1 plane data and output a first identification feature to the second encoder, the second encoder outputs a second identification feature to the first feature fusion layer, the third encoder is used to input the No. 2 plane data and output a third identification feature to the fourth encoder, the fourth encoder outputs the fourth identification feature to the first feature fusion layer and the second output layer, the second output layer outputs eddy current information representing each area, the first feature fusion layer outputs the first fusion feature to the first output layer, and the first output layer outputs warning information representing each area; Step 700, decode and issue warning information and alarm reminders based on the information output by the artificial intelligence model.

2. The artificial intelligence-based early warning method for Enteromorpha control according to claim 1, characterized in that: The hydrological element data include: temperature, salinity, depth, and oxygen content in each area; The ecological factor data include: pH value, water temperature, salinity, dissolved oxygen, chemical oxygen demand, active phosphate content, ammonia nitrogen content, nitrate nitrogen content, nitrite nitrogen content, inorganic nitrogen content, chlorophyll a concentration, sediment sulfide content, organic carbon content, total phosphorus content, total nitrogen content, and the types, density and biomass of marine phytoplankton and zooplankton in each area; Ocean stereoscopic observation data include: air pressure, temperature and humidity, wind speed and direction, visibility, precipitation, water temperature, tide level, wave height, and salinity in each region; Satellite observation data include: sea surface height and sea surface temperature in various regions; Reanalysis data: density, current velocity, mixed layer depth, sea surface wind stress; Ocean data include: ocean temperature, salinity, density, current field, chlorophyll concentration, nitrate concentration, and phosphate concentration in each region; Equipment data includes: the distribution location of power generation equipment, kinetic energy conversion efficiency, and load; Enteromorpha image data include: high-resolution images of water surfaces in various regions taken by drones; The terrain structure data include: terrain structure maps of each region; Equipment structure data includes: equipment structure diagram.

3. The artificial intelligence-based enteromorpha prevention and control early warning method according to claim 1, characterized in that: The first plane data includes a first bin for storing data, a first serial data stream indexed by a first bin to a first data object, and the first plane data also includes a first relationship list for storing association relationships between the first data objects indexed by the first bin, wherein the element value of the i-th row and j-th column of the first relationship list represents the quantitative value of the association relationship between the first data objects indexed by the i-th and j-th first bins; The association relationships include high and low, corresponding to two quantitative values; The first data object includes: each divided area; The data objects with high association relationship are defined as follows: if the regions are adjacent, the association relationship is high; A No. 1 serial data stream of a No. 1 data object includes n No. 1 data units sorted in time, and the t-th No. 1 data unit represents the No. 1 comprehensive information collected at the t-th moment of the No. 1 data object.

4. The artificial intelligence-based enteromorpha prevention and control early warning method according to claim 1, characterized in that: The second plane data includes a second bin for storing data, a second serial data stream indexed by a second bin to a second data object, and the second plane data also includes a second relationship list for storing association relationships between the second data objects indexed by the second bin, the element value of the i-th row and j-th column of the second relationship list representing the quantized value of the association relationship between the second data objects indexed by the i-th and j-th second bins; The association relationships include high and low, corresponding to two quantitative values; The first data object includes: each divided area; The data objects with high association relationship are defined as follows: if the regions are adjacent, the association relationship is high; A No. 2 serial data stream of a No. 2 data object includes n No. 2 data units ordered in time, and the t-th No. 2 data unit represents the No. 2 comprehensive information collected at the t-th moment of the No. 2 data object.

5. The artificial intelligence-based enteromorpha prevention and control early warning method according to claim 1, characterized in that: After feature engineering, the first comprehensive information and the second comprehensive information are input into the model. The first comprehensive information and the second comprehensive information include image modalities. The CNN algorithm is used as a feature engineering method for the information of the image modality.

6. The artificial intelligence-based enteromorpha prevention and control early warning method according to claim 1, characterized in that: The outbreak level of Enteromorpha is divided into three levels: high, medium and low according to the distribution density of Enteromorpha.

7. The artificial intelligence-based enteromorpha prevention and control early warning method according to claim 1, characterized in that: The loss function when training the artificial intelligence model is as follows: ； in represents the first loss value, represents the total number of training samples, Indicates The real distribution density and outbreak risk level of the Enteromorpha samples, The output of the model is The distribution density and outbreak risk level of Enteromorpha in each sample.

8. The artificial intelligence-based enteromorpha prevention and control early warning method according to claim 1, characterized in that: The second plane data is input into the third encoder and outputs the third identification feature to the fourth encoder. The fourth encoder outputs the fourth identification feature to the second fully connected layer. The second fully connected layer outputs the eddy current information representing each area. This process is independently pre-trained, and the training loss function is as follows: ； in represents the second loss value, represents the total number of training samples, Indicates The true position, strength, and radius of the vortex of each sample, The output of the model is The location, strength, and radius of the eddy currents of each sample.

9. The artificial intelligence-based enteromorpha prevention and control early warning method according to claim 1, characterized in that: The early warning information for each region includes: the distribution density of Enteromorpha and the outbreak risk level.

10. An artificial intelligence-based enteromorpha prevention and control early warning system, characterized in that: It is used to execute an artificial intelligence-based enteromorpha prevention and control early warning method as described in any one of claims 1-9.

Images