Coastal fishery intelligent text and time sequence diagram generation algorithm and system thereof

Abstract

The invention relates to a coastal fishery intelligent text and time sequence diagram generation algorithm and system, and aims to improve the accuracy and practicability of coastal fishery wind forecasting. According to the system, sea level average wind and horizontal air pressure gradient data of data sources such as EC-thin are acquired, and a fine wind power series grid forecasting result is generated by applying a grid forecasting calculation technology and a dynamic threshold setting method. Furthermore, the system judges the dominant wind level by using a Marking Square algorithm, and particularly performs deep learning KAN convolution algorithm correction for the wind level caused by typhoon so as to improve the forecasting precision. Finally, the system screens out wind power and wind direction data in a key period, automatically generates an intelligent text description and a time sequence diagram, and provides visual and accurate forecast information for fishermen and related mechanisms. The method has the beneficial effects that the forecasting accuracy is improved, the real-time performance and pertinence are enhanced, intelligent and visual output is realized, the emergency response capability is improved, and the sustainable development of fishery is promoted.

Description

Technical Field

[0001] The invention relates to the technical field of meteorological monitoring, and in particular to an algorithm and system for generating intelligent text and time series diagrams of coastal fisheries. Background Art

[0002] At present, the forecast of extreme winds in the ocean is mainly based on contour distribution and single-point curves, while the forecast service for coastal fisheries often relies on text forecasts. Single-point curves cannot represent the entire sea area, and the existing correction methods cannot provide text and time-series forecast products for the region due to the lack of reliable observation data for offshore fisheries and the fact that their own algorithms mostly use multivariate linear regression modeling. In addition, since it involves production safety, the Meteorological Bureau needs to be responsible for the forecast results. The results of such algorithms ultimately need to be subjectively corrected by forecasters on this basis. Since the algorithm itself has low accuracy and deviates greatly from the forecaster's forecast of the outside world, in the actual service process, such objective algorithm results are difficult to apply to actual business. Coastal wind forecasts are still mainly based on subjective forecasts by forecasters. However, there are a large number of offshore fisheries, and obtaining reasonable and accurate forecasts often requires a lot of manpower to make forecasts, which puts great pressure on local meteorological stations.

[0003] Therefore, a type of objective algorithm is needed that can not only ensure that the output forecast results are basically consistent with those of the forecaster, thereby alleviating the forecaster's pressure, but also take into account the forecast service work of offshore safety precautions to ensure the safety of fishermen going out to sea. Summary of the invention

[0004] In order to solve the problems of the prior art, the present invention provides a coastal fishery intelligent text and time sequence diagram generation algorithm and a system thereof.

[0005] In order to solve the above technical problems, the present invention is implemented by the following technical solutions: In the first aspect, an algorithm and system for generating intelligent text and time series diagrams of coastal fisheries, comprising the following steps:

[0006] Step 1, data acquisition step: obtain the EC-thin sea level average wind and the maximum value of the sea level horizontal pressure gradient as input data;

[0007] Step 2, grid forecast calculation step: according to the input data, calculate the deviation between the actual wind force level and the corresponding average wind and the absolute value of the horizontal gradient of air pressure, set the dynamic judgment threshold, and calculate the grid forecast result of the wind force level of the specified fishing ground;

[0008] Step 3: Determine the dominant wind level: Use the Marching Square algorithm to calculate the position information of the contour lines of various wind level intervals, determine the dominant wind level of the target fishing ground, and determine whether the wind level is caused by a typhoon;

[0009] Step 4, deep learning correction step: when it is determined that the dominant wind level is caused by a typhoon, the deep learning KAN convolution algorithm is used to perform multi-layer convolution calculations on the absolute value of the horizontal pressure gradient in the target sea area, and the corrected wind force forecast value is output;

[0010] Step 5, smart text and time series diagram generation step: filter the correction results every six hours, generate wind speed and direction data for key periods of the day through iterative processing, and generate smart text and time series diagram based on this.

[0011] In this embodiment, the coastal fishery intelligent text and time series diagram generation algorithm and system proposed by the present invention aims to achieve accurate forecast and intuitive display of wind force and direction in coastal fishery by integrating meteorological data analysis, grid forecast, dominant wind level judgment, deep learning correction and intelligent visualization technology. Its core principle is:

[0012] Data-driven: The average sea level wind and the maximum sea level horizontal pressure gradient provided by EC-thin (a meteorological data model or data source) are used as the prediction basis. These data reflect the changes in wind and pressure on the ocean surface and are key inputs for predicting wind levels.

[0013] Grid forecast: By calculating the deviation between the actual wind force level and the theoretical prediction value (based on average wind and pressure gradient) and setting a dynamic threshold, a grid forecast of the wind force level in a designated fishing area can be achieved. This method can more carefully reflect the spatial distribution characteristics of wind force in the area.

[0014] Prevailing wind level judgment: The Marching Square algorithm, an algorithm commonly used for contour drawing, is used to accurately calculate and determine the prevailing wind level and its spatial distribution in the target fishing ground. At the same time, combined with the typhoon path and intensity information, it is determined whether the prevailing wind level is caused by the typhoon, providing a basis for subsequent deep learning corrections.

[0015] Deep learning correction: In response to the complex wind field changes caused by typhoons, deep learning KAN (which may be a specifically designed convolutional neural network architecture) is used to perform multi-layer convolution calculations on the horizontal pressure gradient to capture more complex nonlinear relationships, thereby outputting more accurate corrected wind forecast values.

[0016] Intelligent visualization: By filtering and iteratively processing the six-hour correction results, the wind speed and direction data for key periods of the day are generated, and intelligent text descriptions and time series diagrams are automatically generated based on this. This method not only improves the timeliness of the forecast, but also enhances the readability and intuitiveness of the forecast results.

[0017] In a specific implementation of the first aspect, the data in step 1 also needs to be preprocessed, including data cleaning and outlier removal.

[0018] In a specific implementation of the first aspect, in the grid forecast calculation step, the setting of the dynamic judgment threshold is based on the historical data verification of multiple stations along the Zhejiang coast, so that the deviation between the wind level correction result under the influence of non-typhoons and the forecast result of the forecaster is controlled within 1 level (about 3m / s).

[0019] In a specific implementation of the first aspect, in the dominant wind level judgment step, the wind level with the largest wind force level and an overlapping area with the target fishing ground exceeding 1 / 3 of the fishing ground itself is selected as the forecast result for the corresponding time of the fishing ground, and whether the typhoon affects the fishing ground is identified by judging whether the contour line position information of the wind level contains the contour line of wind level 13 and above in the Pacific Ocean.

[0020] In a specific implementation of the first aspect, in the deep learning correction step, the deep learning model is trained using forecaster correction results from multiple typhoon fishing grounds to improve the forecast accuracy of the algorithm during typhoons.

[0021] In a specific implementation of the first aspect, the key time points selected in step five are set as 02:00 the next day, 08:00, 14:00, and 20:00 on the same day.

[0022] In the second aspect, a coastal fishery intelligent text and time series diagram generation system includes:

[0023] Data acquisition module, used to obtain the maximum value of EC-thin sea level average wind and sea level horizontal pressure gradient;

[0024] A grid forecast calculation module is used to calculate the grid forecast result of wind force level according to input data;

[0025] The dominant wind level judgment module is used to determine the dominant wind level of the target fishing ground and determine whether it is caused by a typhoon;

[0026] A deep learning correction module is used to correct forecast results during typhoons;

[0027] Intelligent text and timing diagram generation module, used to generate intelligent text and timing diagram

[0028] The modules are interconnected through data interfaces to jointly realize the automatic generation of intelligent text and time-series diagrams of coastal fisheries.

[0029] In a specific implementation of the second aspect, it also includes a user interaction interface for displaying the generated intelligent text and timing diagram, and providing a user input interface for receiving user queries and setting parameters.

[0030] In a specific implementation of the second aspect, it also includes a data storage module for storing historical data, model parameters and generated forecast results for subsequent analysis and optimization of algorithms.

[0031] In a specific implementation of the second aspect, it also includes an early warning module, which automatically sends early warning information to fishermen and relevant agencies when specific conditions are detected to ensure the safety of fishermen going out to sea.

[0032] The forecast results output by the intelligent algorithm and the subjective forecast results of the meteorological bureau forecasters tend to be consistent, which is specifically reflected in:

[0033] 1. The judgment of wind force 8, wind force 9 and gale force 9 or above for offshore fishing boats’ safety is relatively accurate;

[0034] 2. The algorithm's judgment on the trend of wind strength strengthening and weakening in the sea area is basically consistent with that of the forecaster;

[0035] 3. Automatically generate the maximum wind forecast results for 13 fishing grounds in the East China Sea in batches and output them to forecasters as objective results, thus improving the efficiency of forecasters’ subjective corrections;

[0036] 4. Capable of generating intelligent text and timing diagrams for extreme wind forces in fishing grounds;

[0037] It can identify the wind increase process caused by typhoons and make more accurate forecasts.

[0038] The beneficial effects of the present invention are:

[0039] 1. By introducing the grid forecast calculation steps, combining the deviation analysis between the actual wind force level and the corresponding average wind and the absolute value of the horizontal gradient of air pressure, and setting the dynamic judgment threshold, the present invention can more accurately predict the wind force level of coastal fishing grounds. Especially under extreme weather conditions such as typhoons, the deep learning KAN convolution algorithm is used to correct the forecast results, which significantly improves the accuracy and reliability of wind forecasts and provides fishermen and related institutions with a more scientific decision-making basis;

[0040] 2. This system can obtain EC-thin and other data sources in real time, and quickly process data and forecast calculations to generate wind speed and direction data for key periods of six hours. This high-frequency update and targeted forecasting allows fishermen to understand the wind speed changes in the fishing grounds in a timely manner, so as to reasonably arrange the time for sea operations, reduce safety risks, and improve fishing efficiency;

[0041] 3. This system not only generates accurate wind forecast data, but also further converts this data into intelligent text and time series diagrams. This intelligent output method enables fishermen and related personnel to understand the forecast information intuitively and quickly, making it easier to make more accurate decisions. At the same time, the display of time series diagrams also makes it easier to compare historical data and future trends, providing a reference for long-term planning;

[0042] 4. Through the dominant wind level judgment step, the system can automatically identify and judge the impact of extreme weather such as typhoons on fishing grounds. Once it is confirmed that the dominant wind level is caused by a typhoon, the system will immediately start the deep learning correction process to generate the corrected wind force forecast value at the fastest speed. This efficient emergency response mechanism helps fishermen and relevant institutions to take preventive measures in advance and reduce disaster losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a schematic diagram of the overall algorithm flow of the present invention.

[0044] Figure 2 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions 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 creative work are within the scope of protection of the present invention.

[0046] like Figure 1-2 The invention discloses an algorithm and system for generating intelligent text and time series diagram of a coastal fishery, comprising the following steps: step 1, data acquisition step: obtaining the maximum value of the EC-thin sea level average wind and the sea level horizontal pressure gradient as input data; step 2, grid forecast calculation step: calculating the deviation between the actual wind force level and the corresponding average wind and the absolute value of the horizontal pressure gradient according to the input data, setting a dynamic judgment threshold, and calculating the grid forecast result of the wind force level of the specified fishery; step 3, dominant wind level judgment step: using the Marching Square algorithm, calculating the position information of the contour lines of various wind level intervals, determining the dominant wind level of the target fishery, and determining whether the wind level is caused by a typhoon; step 4, deep learning correction step: when it is determined that the dominant wind level is caused by a typhoon, using the deep learning KAN convolution algorithm to perform multi-layer convolution calculation on the absolute value of the horizontal pressure gradient of the target sea area, and outputting the corrected wind force forecast value; step 5, intelligent text and time series diagram generation step: screening the correction results for six hours at a time, generating the wind force and direction data for the key time period of a day through iterative processing, and generating intelligent text and time series diagram accordingly.

[0047] The algorithm is mainly used for forecasting services in 13 fishing grounds in the East China Sea. In order to make the algorithm output results basically consistent with those of forecasters, the data source selection is consistent with the forecaster's offshore wind reference data, mainly the EC-thin sea level average wind and the maximum value of the sea level horizontal pressure gradient. In terms of business, forecasters make empirical judgments based on the average wind and the corresponding gradient size.

[0048] The algorithm implementation is divided into the following steps:

[0049] According to the calculated average wind and absolute values ​​of pressure gradient, according to the conditions at different times and locations, the deviation between the actual wind force level and the corresponding average wind and absolute values ​​of pressure horizontal gradient is calculated, and a dynamic judgment threshold is set, and finally the grid forecast result of the wind force level of the specified fishing ground is calculated. This empirical formula has been tested at multiple stations along the Zhejiang coast. The deviation between the wind force correction result and the forecast result of the forecaster can be controlled within 1 level (about 3m / s) when there is no typhoon influence (below level 10 wind), and because the data source selection is consistent with the forecaster, the trend of increasing and decreasing wind shown in the time series diagram of the correction result is basically consistent with the forecaster's judgment.

[0050] The same fishing ground at the same time is often affected by multiple levels of wind. In order to determine the main influencing wind level at the current moment that represents the entire fishing ground area, and whether the fishing ground is affected by a tropical cyclone (typhoon), it is necessary to identify the dominant wind level in the target fishing ground area to determine whether the wind is caused by a typhoon. On the basis of the first step, the MarchingSquare algorithm is used to calculate the position information of the contour lines of various wind level intervals, and the overlapping area of ​​the target fishing ground and various wind levels is calculated. The wind level with the largest wind level and an overlapping area exceeding 3 / 1 of the fishing ground itself is selected as the forecast result for the corresponding moment of the fishing ground. At the same time, it is calculated whether the position information of the contour line of the wind level contains the contour line of wind level 13 and above in the Pacific Ocean, so as to determine whether it is affected by a typhoon. This step can well determine the dominant wind force in a region, and can identify whether this wind force is caused by a typhoon, laying a good foundation for the third step.

[0051] In order to improve the applicability of the algorithm during typhoons, when step 2 determines that the main wind force in the target fishing ground is caused by the typhoon, the deep learning technology KAN convolution algorithm is used to perform multi-layer convolution calculations on the absolute value of the horizontal pressure gradient in the target sea area, and the output is the difference between the forecaster's subjective correction wind level and the wind level calculated in step 2. After training and reasoning with the forecaster's correction results in typhoon fishing grounds for many times, the wind force forecast value of the fishing ground is corrected. This step better solves the bottleneck of the first step based on the dynamic threshold setting algorithm facing the small results of typhoon forecasts.

[0052] In order to generate a time series diagram of wind force levels and corresponding text forecast results for marine fisheries. The six-hour correction results are screened, where 02:00 the next day, 08:00 the same day, 14:00, and 20:00 represent the forecast results for the second half of the night, morning, afternoon, and evening, respectively. The average wind direction at the corresponding time and the wind force level after algorithm correction are calculated as input data; determine whether the wind force level and wind direction angle between adjacent moments are similar. If the wind force level is similar, only the wind force of the earliest period is retained. If the wind direction intersection angle is less than 67.5, the wind direction of the earliest period is retained. In this way, the wind force and direction data of 4 moments of the day are continuously iterated to finally generate intelligent text. This text has been tested by the business and is basically consistent with the forecaster's forecast.

[0053] The above four steps achieve the effect that the objective algorithm correction results are basically consistent with the subjective forecaster's forecast results, effectively alleviating the forecaster's work pressure on coastal fishery forecast services.

[0054] The data in step one also needs to be preprocessed, including data cleaning and outlier removal. In the grid forecast calculation step, the setting of the dynamic judgment threshold is based on the historical data verification of multiple stations along the coast of Zhejiang, so that the deviation between the wind level correction result under the influence of non-typhoons and the forecaster's forecast result is controlled within 1 level (about 3m / s). In the dominant wind level judgment step, the wind level with the largest wind force level and an overlapping area with the target fishing ground exceeding 1 / 3 of the fishing ground itself is selected as the forecast result of the fishing ground at the corresponding time, and the wind level is judged by judging whether the contour position information of the wind level contains the Pacific Ocean Sea 13 and above wind level contour line to identify whether it is affected by the typhoon. In the deep learning correction step, the deep learning model is trained using the forecaster's correction results of multiple typhoon fishing grounds to improve the forecast accuracy of the algorithm during typhoons. In step five, the key time points are selected and set to 02:00 the next day, 08:00, 14:00, and 20:00 on the same day.

[0055] A coastal fishery intelligent text and time series diagram generation system comprises: a data acquisition module for acquiring EC-thin sea level average wind and sea level horizontal pressure gradient maximum value; a grid forecast calculation module for calculating the grid forecast result of wind force level according to input data; a dominant wind level judgment module for determining the dominant wind level of a target fishery and judging whether it is caused by a typhoon; a deep learning correction module for correcting the forecast result during a typhoon; an intelligent text and time series diagram generation module for generating intelligent text and time series diagram; the modules are interconnected through data interfaces to jointly realize the automatic generation of intelligent text and time series diagram of coastal fisheries.

[0056] It also includes a user interaction interface for displaying the generated intelligent text and time series diagrams, and provides a user input interface for receiving user queries and setting parameters. It also includes a data storage module for storing historical data, model parameters and generated forecast results for subsequent analysis and algorithm optimization. It also includes an early warning module, which automatically sends early warning information to fishermen and relevant agencies when specific conditions are detected to ensure the safety of fishermen going out to sea.

[0057] The coastal fishery intelligent text and time series diagram generation algorithm and system proposed in this patent realize the accurate forecast and intuitive display of wind force and direction in coastal fishery based on advanced data processing, grid forecast, dominant wind level judgment, deep learning correction and intelligent visualization technology. The specific working principle is as follows:

[0058] Data acquisition and preprocessing:

[0059] The system uses the data acquisition module to obtain the average sea level wind and the maximum sea level horizontal pressure gradient from data sources such as EC-thin in real time.

[0060] Subsequently, data preprocessing, including data cleaning and outlier removal, is performed to ensure the accuracy and reliability of the input data.

[0061] Grid forecast calculation:

[0062] Based on the preprocessed input data, the grid forecast calculation module calculates the deviation between the actual wind force level and the absolute value of the horizontal gradient of the corresponding average wind and air pressure.

[0063] Based on the historical data of multiple stations along the coast of Zhejiang Province, a dynamic judgment threshold is set to control the wind level correction deviation under non-typhoon conditions within level 1 (about 3 m / s), thereby calculating the grid forecast results of the wind force level in the specified fishing ground.

[0064] Dominant wind level judgment:

[0065] Using the Marching Square algorithm, the dominant wind level judgment module calculates the position information of the contour lines of various wind level intervals.

[0066] The wind level with the largest wind force and an overlapping area with the target fishing ground exceeding 1 / 3 of the fishing ground itself is selected as the forecast result for the fishing ground at the corresponding time.

[0067] By judging whether the wind level is caused by a typhoon, it is possible to determine whether the wind level is located at the contour line of wind level 13 or above in the Pacific Ocean.

[0068] Deep Learning Correction:

[0069] When it is judged that the dominant wind level is caused by a typhoon, the deep learning correction module uses the deep learning KAN convolution algorithm to perform multi-layer convolution calculations on the absolute value of the horizontal pressure gradient in the target sea area.

[0070] This module uses the forecaster correction results of multiple typhoon fishing grounds to train the model to improve the forecast accuracy of the algorithm during typhoons and output the corrected wind forecast value.

[0071] Intelligent text and timing diagram generation:

[0072] The correction results are filtered every six hours, and the intelligent text and time series diagram generation module generates wind speed and direction data for key periods of the day (such as 02:00 the next day, 08:00, 14:00, and 20:00 on the same day) through iterative processing.

[0073] Based on these data, intelligent text descriptions and time series diagrams containing key information such as wind force and wind direction are automatically generated.

[0074] User interaction and data storage:

[0075] The user interaction interface displays the generated intelligent text and timing diagram, and provides a user input interface to receive user queries and set parameters.

[0076] The data storage module stores historical data, model parameters and generated forecast results to provide support for subsequent analysis and optimization algorithms.

[0077] Early warning function:

[0078] The early warning module monitors the forecast results in real time. When it detects certain conditions (such as an approaching typhoon, a sharp increase in wind speed, etc.), it automatically sends early warning information to fishermen and relevant agencies to ensure their safety at sea.

[0079] Embodiment: In fishery production, accurate prediction of wind force and direction in coastal fishing grounds is crucial to ensure the safety of fishermen and improve fishing efficiency. However, traditional forecasting methods often have problems such as insufficient accuracy and delayed warning. This embodiment proposes an intelligent text and time series diagram generation algorithm and system for coastal fishing grounds, which aims to achieve accurate forecast and intuitive display of wind force and direction in coastal fishing grounds by integrating advanced data processing, grid forecasting, deep learning correction and intelligent visualization technology.

[0080] Description of Embodiments

[0081] System Architecture

[0082] The coastal fishery intelligent text and time sequence diagram generation system of this embodiment mainly includes the following modules: data acquisition module, grid forecast calculation module, dominant wind level judgment module, deep learning correction module, intelligent text and time sequence diagram generation module, user interaction interface, data storage module and early warning module. The modules are interconnected through data interfaces to jointly realize the automatic generation and display of coastal fishery intelligent text and time sequence diagram.

[0083] Data acquisition and preprocessing

[0084] Data acquisition: The data acquisition module obtains the average wind at sea level and the maximum value of the horizontal pressure gradient at sea level in real time from data sources such as EC-thin.

[0085] Data preprocessing: Clean the acquired data and remove outliers to ensure the accuracy and reliability of the input data. For example, remove or correct data points that are obviously deviated from the normal range.

[0086] Grid forecast calculation

[0087] Input data processing: input the preprocessed data into the grid forecast calculation module.

[0088] Deviation calculation and threshold setting: Based on the input data, the deviation between the actual wind force level and the corresponding average wind and the absolute value of the horizontal gradient of air pressure is calculated. Based on the historical data of multiple stations along the Zhejiang coast, the dynamic judgment threshold is set to control the wind force correction deviation under non-typhoon influence within 1 level (about 3m / s).

[0089] Grid forecast result generation: Generate grid forecast results of wind force levels in specified fishing grounds based on the set thresholds and deviation calculation results.

[0090] Prevailing wind level judgment

[0091] Contour line calculation: Using the Marching Square algorithm, the dominant wind level judgment module calculates the location information of the contour lines of various wind level intervals.

[0092] Determination of dominant wind level: The wind level with the largest wind force level and the overlapping area with the target fishing ground exceeding 1 / 3 of the fishing ground itself is selected as the forecast result for the fishing ground at the corresponding time.

[0093] Typhoon identification: By judging whether the wind level is caused by a typhoon, the location information of the contour line of the wind level contains the contour line of wind level 13 and above in the Pacific Ocean, so as to identify whether the wind level is caused by a typhoon.

[0094] Deep Learning Correction

[0095] Model training: The deep learning KAN convolutional algorithm model is trained using forecaster correction results from multiple typhoon fishing grounds to improve the algorithm’s forecast accuracy during typhoons.

[0096] Correction calculation: When it is determined that the dominant wind level is caused by a typhoon, the deep learning correction module performs multi-layer convolution calculations on the absolute value of the horizontal pressure gradient in the target sea area and outputs the corrected wind forecast value.

[0097] Intelligent text and timing diagram generation

[0098] Data processing: Filter the six-hourly correction results and generate wind force and direction data for key periods of the day (such as 02:00 the next day, 08:00, 14:00, and 20:00 on the same day) through iterative processing.

[0099] Intelligent text and timing diagram generation: Based on the processed data, the intelligent text and timing diagram generation module automatically generates intelligent text descriptions and timing diagrams containing key information such as wind force and wind direction.

[0100] User interaction and data storage

[0101] User interaction interface: Provides a user input interface to receive user queries and set parameters, and displays the generated smart text and timing diagrams.

[0102] Data storage: The data storage module is used to store historical data, model parameters and generated forecast results for subsequent analysis and algorithm optimization.

[0103] Early warning function

[0104] Warning condition setting: The warning module sets warning thresholds based on preset conditions (such as approaching typhoon, rapid increase in wind speed, etc.).

[0105] Sending early warning information: When a situation that meets the early warning conditions is detected, early warning information will be automatically sent to fishermen and relevant agencies to ensure the safety of fishermen going out to sea.

[0106] Implementation Effect

[0107] The coastal fishery intelligent text and time series diagram generation algorithm and system of this embodiment integrates multiple advanced technologies to achieve accurate forecast and intuitive display of wind force and direction in coastal fisheries. The system can not only improve forecast accuracy and early warning capabilities, but also enhance user experience and support decision-making, providing strong support for fishery production and disaster prevention and mitigation.

[0108] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent text and time series diagram generation algorithm and system for coastal fisheries, characterized in that: The following steps are involved: Step 1, data acquisition step: obtain the EC-thin sea level average wind and the maximum value of the sea level horizontal pressure gradient as input data; Step 2, grid forecast calculation step: according to the input data, calculate the deviation between the actual wind force level and the corresponding average wind and the absolute value of the horizontal gradient of air pressure, set the dynamic judgment threshold, and calculate the grid forecast result of the wind force level of the specified fishing ground; Step 3: Determine the dominant wind level: Use the Marching Square algorithm to calculate the position information of the contour lines of various wind level intervals, determine the dominant wind level of the target fishing ground, and determine whether the wind level is caused by a typhoon; Step 4, deep learning correction step: when it is determined that the dominant wind level is caused by a typhoon, the deep learning KAN convolution algorithm is used to perform multi-layer convolution calculations on the absolute value of the horizontal pressure gradient in the target sea area, and the corrected wind force forecast value is output; Step 5, smart text and time series diagram generation step: filter the correction results every six hours, generate wind speed and direction data for key periods of the day through iterative processing, and generate smart text and time series diagram based on this.

2. According to claim 1, a coastal fishery intelligent text and time sequence diagram generation algorithm and system thereof are characterized in that: The data in step 1 also needs to be preprocessed, including data cleaning and outlier removal.

3. According to claim 1, a coastal fishery intelligent text and time sequence diagram generation algorithm and system thereof are characterized in that: In the grid forecast calculation step, the setting of the dynamic judgment threshold is based on the historical data verification of multiple stations along the Zhejiang coast, so that the deviation between the wind level correction result under non-typhoon influence and the forecast result of the forecaster is controlled within 1 level (about 3m / s).

4. According to claim 1, a coastal fishery intelligent text and time sequence diagram generation algorithm and system thereof are characterized in that: In the dominant wind level judgment step, the wind level with the largest wind force level and an overlapping area with the target fishing ground exceeding 1 / 3 of the fishing ground itself is selected as the forecast result for the corresponding time of the fishing ground, and whether the typhoon affects the fishing ground is identified by judging whether the contour line position information of the wind level contains the contour line of wind level 13 and above in the Pacific Ocean.

5. According to claim 1, a coastal fishery intelligent text and time sequence diagram generation algorithm and system thereof are characterized in that: In the deep learning correction step, the deep learning model is trained using the forecaster correction results of multiple typhoon fishing grounds to improve the forecast accuracy of the algorithm during typhoons.

6. According to claim 1, a coastal fishery intelligent text and time sequence diagram generation algorithm and system thereof are characterized by: In step 5, the key time points are selected and set as 02:00 the next day, 08:00, 14:00 and 20:00 on the same day.

7. A coastal fishery intelligent text and time series diagram generation system, characterized in that include: Data acquisition module, used to obtain the maximum value of EC-thin sea level average wind and sea level horizontal pressure gradient; A grid forecast calculation module is used to calculate the grid forecast result of wind force level according to input data; The dominant wind level judgment module is used to determine the dominant wind level of the target fishing ground and determine whether it is caused by a typhoon; A deep learning correction module is used to correct forecast results during typhoons; Intelligent text and timing diagram generation module, used to generate intelligent text and timing diagram; The modules are interconnected through data interfaces to jointly realize the automatic generation of intelligent text and time-series diagrams of coastal fisheries.

8. The coastal fishery intelligent text and time sequence diagram generation system according to claim 6 is characterized by: It also includes a user interaction interface for displaying the generated intelligent text and timing diagram, and provides a user input interface for receiving user queries and setting parameters.

9. The coastal fishery intelligent text and time sequence diagram generation system according to claim 6 is characterized by: It also includes a data storage module for storing historical data, model parameters and generated forecast results for subsequent analysis and optimization of algorithms.

10. The coastal fishery intelligent text and time sequence diagram generation system according to claim 6, characterized in that: It also includes an early warning module, which automatically sends early warning information to fishermen and relevant agencies when specific conditions are detected to ensure the safety of fishermen going out to sea.

Images