Integrated method and system for shipborne intelligent meteorological observation, detection, forecast and warning

Abstract

The present invention discloses a method and system for integrating shipborne intelligent meteorological observation, detection, forecast and warning, and relates to the technical field of route forecast and warning. The steps of the method include: during the travel of a fishing boat, the area to be fished is divided into fishing windows; and all fishing area panes are divided into fishing panes; the meteorological data of all fishing panes are obtained through a shipborne intelligent meteorological observation and detection station to obtain the meteorological impact value of the fish school; a fish school-meteorological prediction model is established to predict the distribution status of the fish school in all fishing panes in all fishing windows; the recommended value of the fishing window is obtained by calculating the distribution status of the fish school and the meteorological data, a regional recommendation strategy is established, the recommended direction of travel of the fishing boat is output, and it is determined whether to issue an early warning to the operator. The present invention solves the problem of low model generalization ability caused by different meteorological conditions and fish school distribution laws in different sea areas and seasons.

Description

Technical Field

[0001] The present invention relates to the technical field of route forecast and warning, and in particular to a method and system for integrating shipborne intelligent meteorological observation, detection, forecast and warning. Background Art

[0002] With the increasing shortage of global fishery resources and the complexity and variability of the marine environment, improving the efficiency and accuracy of fishing has become the key to the sustainable development of fisheries. Traditional fishing methods often rely on the experience and intuition of fishermen, lack of scientific data support and accurate prediction methods, resulting in low fishing efficiency, serious waste of resources, and even damage to the marine ecological environment. In recent years, with the development of intelligent technology and big data, using advanced meteorological observation technology, data analysis methods and prediction models to guide fishing has become a new research direction.

[0003] Marine meteorological conditions have an important impact on the distribution and migration of fish. Meteorological factors such as water temperature, dissolved oxygen content, light intensity, water flow speed and air pressure not only directly affect the survival and activities of fish, but also indirectly affect the distribution of fish through food chains, habitat selection and other means. Therefore, accurately understanding the relationship between marine meteorological conditions and fish distribution is of great significance for improving fishery efficiency and protecting the marine ecological environment.

[0004] For example, the existing Chinese patent with publication number CN118571070B discloses a method and system for automatic positioning and early warning management of fishing vessels based on Beidou navigation, including the following steps: collecting real-time meteorological data, identifying changes in wind speed, waves and current speed, performing time series analysis on meteorological data, using Beidou navigation, calculating the risk level of the sea area, and obtaining risk level assessment results. In the present invention, by performing time series analysis on meteorological data, potential dangers can be predicted more accurately. When the risk level exceeds the safety threshold, the speed and heading of the fishing boat are adjusted in time. The dynamic adjustment can reflect environmental changes in real time, ensuring that the fishing boat can avoid dangerous areas and adopt the safest navigation strategy. The synchronously updated navigation plan and position tracking further enhance the controllability and predictability of navigation, and timely early warning management not only reduces the incidence of accidents, but also optimizes the navigation efficiency and resource utilization of fishing boats.

[0005] However, it lacks the association between meteorological data and fish school relationships, and ignores the prediction of fish school distribution, which is extremely critical information in fishery fishing. It only focuses on meteorological data and risk levels, and does not mention the prediction of fish school distribution, which may limit its application effect in fishery fishing; and there may be significant differences in meteorological conditions and fish school distribution patterns in different sea areas and seasons. The existing model may lack sufficient adaptability and cannot maintain accurate prediction capabilities in different environments. To this end, the present invention provides a ship-borne intelligent meteorological observation and detection and forecast warning integrated method and system. Summary of the invention

[0006] The purpose of the present invention is to provide a method and system for integrating shipborne intelligent weather observation, detection, forecasting and early warning to solve the existing problems raised in the above-mentioned background technology.

[0007] To achieve the above object, the present invention provides the following technical solution: a shipborne intelligent meteorological observation, detection, forecast and early warning integrated method, comprising the following steps:

[0008] S1. When the fishing boat is moving, the fishing area is divided into 9 fishing windows of size M×M with the hull as the center; and all fishing area panes are divided into N×N fishing panes;

[0009] S2. Obtain the meteorological data of all fishing windows through the ship-borne intelligent meteorological observation and detection station to obtain the meteorological impact value of the fish school;

[0010] S3, establishing a fish school-weather forecast model to predict the distribution status of fish schools in all fishing windows in all fishing windows;

[0011] S4. Calculate the recommended value of the fishing window through the fish distribution status and meteorological data, establish a regional recommendation strategy, output the recommended direction of travel for the fishing boat, and determine whether to issue an early warning to the operating personnel.

[0012] A further improvement of the present invention is that the specific construction process of the fish school-weather prediction model includes: taking the spring to winter of each year in the past as a recording cycle, calculating the sum of the actual fish school density in each season in each fishing window in each recording cycle, obtaining the value of the sum and the proportion of all data in the recording cycle of the regional fishing window, and adding one to obtain the seasonal cycle change factor of each seasonal node in each recording cycle, extracting the seasonal cycle change factor of each seasonal node and multiplying it with the fish school meteorological impact value of the corresponding season to obtain the meteorological impact fish school change value of the recording cycle.

[0013] A further improvement of the present invention is that the specific construction process of the fish school-weather prediction model also includes: calculating the weighted sum of the mean fish school density of the seasonal nodes in the recording cycle and the meteorological impact on the fish school change value of the corresponding season of the recording cycle as the historical fish school distribution state, using the historical fish school meteorological impact value and the historical fish school distribution state as training data, and automatically constructing a tree structure based on the importance of each feature of the fish school meteorological impact value, each node represents a decision rule, and the leaf node represents the prediction result.

[0014] The present invention is further improved in that the specific construction process of the tree structure includes: evaluating the predictive ability of each feature on the distribution state of fish schools by calculating the information gain of each feature in the historical meteorological impact value; setting the meteorological threshold of fish school impact, listing the data with meteorological impact value greater than the meteorological threshold of fish school impact as large impact meteorological value, and listing the data with meteorological impact value less than or equal to the meteorological threshold of fish school impact as small impact meteorological value; when the small impact meteorological value of a certain feature is greater than the set first quantity threshold, the feature is selected as the split feature, on the contrary, when the small impact meteorological value of a certain feature is less than or equal to the set first quantity threshold, the feature with the largest information gain is selected as the split feature of the current node; according to different values ​​of the selected features, the data set is divided into multiple subsets, each subset corresponds to a child node; for each child node, the model repeats the above process until there are no more features to split.

[0015] A further improvement of the present invention is that the specific construction process of the fish school-weather prediction model also includes: performing a post-pruning operation on the fish school-weather prediction model, including, for each leaf node that is not the fish school distribution, considering the impact of small-impact meteorological values ​​on model performance, calculating the error based on cross-validation after pruning, and according to the evaluation result, selecting the fish school distribution node with the largest error reduction after pruning for pruning; repeating the above process on the pruned tree structure until there are no more nodes to prune; if the accuracy of the pruned model is greater than the set accuracy threshold, accepting the pruning result; otherwise, retraining the model; and obtaining the fish school distribution status of each fishing window of the future timestamp TT.

[0016] A further improvement of the present invention is that the recommended value of the fishing area window is obtained by calculating the product of the meteorological stability index in the area and the fish distribution status in the area; and the fishing windows are numbered 1 to 9 from the upper left corner to the lower right corner; the fish distribution status in the area is obtained by calculating the average value of the fish distribution status of all fishing windows in the area.

[0017] A further improvement of the present invention is that the meteorological stability index in the area is obtained by setting a threshold value for meteorological impact of fish schools, extracting fishing panes whose meteorological impact values ​​of fish schools are less than the threshold value for meteorological impact of fish schools, merging the extracted adjacent area panes through the merge_intervals function, sorting the area panes using the sorted() function, then merging adjacent, overlapping or touching area panes through a loop and conditional judgment, and saving the merged area panes in a merged list, and finally, the function returns the merged list as a result, recorded as the first fishing partition in the fishing window, repeating the above steps until all fishing panes in all fishing windows are traversed to obtain the first fishing partition to the Kpth fishing partition, where Kp represents the number of partitions in the fishing window numbered p; and obtaining the meteorological stability index by weighted summation of the average stability within the window and the distance-weighted stability.

[0018] A further improvement of the present invention is that the calculation formula for the average stability within the window is: ,in, represents the mean value of meteorological impact in the ith region, represents the number of panes in the ith region; the distance weighted stability calculation formula is: , Indicates the distance between partitions.

[0019] A further improvement of the present invention is that the area recommendation strategy includes extracting the direction corresponding to the maximum value of the fishing window recommendation as the direction of travel of the fishing boat; when the window recommendation value of the fishing area where the fishing boat is located is the largest, the fishing boat exits the area to be fished in the direction corresponding to the other 8 fishing windows with the largest recommended values, and when the distance from the area to be fished is greater than 2M, the direction of travel of the fishing boat is calculated again; when the other 8 fishing window recommendation values ​​are all less than the set window recommendation threshold, a warning to stop fishing is issued to the fishermen.

[0020] A further improvement of the present invention is that the meteorological data includes water temperature, dissolved oxygen content, light intensity, water flow velocity and air pressure, and the water temperature impact value, dissolved oxygen content impact value, light intensity impact value, water flow velocity impact value and air pressure impact value are calculated, and the weighted summation after standardization is obtained to obtain the meteorological impact value of the fish school.

[0021] On the other hand, the present invention provides a shipborne intelligent weather observation, detection, forecast and early warning integrated system, comprising:

[0022] A fishing window partitioning module is used to divide the area to be fished into 9 fishing windows of size M×M; and to divide all fishing area windows into N×N fishing windows;

[0023] A meteorological data analysis module is used to obtain meteorological data of all fishing windows through a ship-borne intelligent meteorological observation and detection station to obtain meteorological impact values ​​of fish schools;

[0024] A fish school-weather forecast model building module is used to predict the distribution status of fish schools in all fishing windows in all fishing windows;

[0025] The early warning and forecasting module calculates the recommended value of the fishing window through the fish distribution status and meteorological data, establishes a regional recommendation strategy, outputs the recommended direction of travel for fishing vessels, and determines whether to issue an early warning to the operating personnel.

[0026] A further improvement of the present invention is that the fish school-meteorological prediction model establishment module includes a seasonal climate fluctuation impact reduction unit and a model training unit; the seasonal climate fluctuation impact reduction unit is used to obtain the seasonal cycle change factor of each seasonal node in each recording cycle, extract the seasonal cycle change factor of each seasonal node and multiply it by the fish school meteorological impact value of the corresponding season to obtain the meteorological impact fish school change value of the recording cycle; the model training unit is used to obtain the fish school distribution status of each fishing window of the future timestamp TT.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] By constructing a fish school-weather forecast model and combining historical data with real-time meteorological observations, the present invention can accurately predict the distribution status of fish schools in different fishing windows and panes, and provide scientific fishing guidance for fishermen;

[0029] Based on the comprehensive analysis of fish distribution status and meteorological data, the present invention can calculate the recommended value of the fishing window, and formulate a regional recommendation strategy based on it, so as to guide fishing boats to select the optimal travel direction and fishing area, thereby improving fishing efficiency and resource utilization;

[0030] By adopting a tree structure model and post-pruning operations, the present invention can automatically construct and optimize the fish school-weather prediction model, making it more adaptable to the meteorological conditions and fish school distribution patterns in different sea areas and seasons, thereby improving the accuracy and generalization ability of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is a flow chart of the integrated method for shipborne intelligent meteorological observation, detection, forecasting and early warning of the present invention;

[0032] Figure 2 A flow chart for constructing a fish school-weather prediction model for the integrated method for shipborne intelligent weather observation, detection, forecasting and early warning of the present invention;

[0033] Figure 3 This is a framework diagram of the shipborne intelligent meteorological observation, detection, forecasting and early warning integrated system of the present invention. DETAILED DESCRIPTION

[0034] The technical solution of the present invention is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations on the technical solution of the present invention. The embodiments of the present invention and the technical features in the embodiments may be combined with each other unless there is a conflict.

[0035] The term "and / or" is only a description of the association relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " generally indicates that the related objects are in an "or" relationship.

[0036] Figure 1 The flowchart of the integrated method of shipborne intelligent weather observation, detection, forecasting and warning disclosed in this embodiment is shown, and the steps are as follows:

[0037] S1. When the fishing boat is moving, the fishing area is divided into 9 fishing windows of size M×M with the hull as the center; and all fishing area panes are divided into N×N fishing panes;

[0038] S2. Obtain the meteorological data of all fishing windows through the ship-borne intelligent meteorological observation and detection station to obtain the meteorological impact value of the fish school; the meteorological data includes water temperature, dissolved oxygen content, light intensity, water flow speed and air pressure, calculate the water temperature impact value, dissolved oxygen content impact value, light intensity impact value, water flow speed impact value and air pressure impact value, and standardize and weighted sum to obtain the meteorological impact value of the fish school. The water temperature impact value calculation formula is: ,in, The most suitable temperature for fish. represents the standard deviation of temperature tolerance, T represents the actual water temperature; the dissolved oxygen content, light intensity, water flow velocity and air pressure, and the calculation principles for calculating the water temperature influence value, dissolved oxygen content influence value, light intensity influence value, water flow velocity influence value and air pressure influence value are the same as the calculation principle of the water temperature influence value.

[0039] S3, establishing a fish school-weather forecast model to predict the distribution status of fish schools in all fishing windows in all fishing windows;

[0040] S4. Calculate the recommended value of the fishing window through the fish distribution status and meteorological data, establish a regional recommendation strategy, output the recommended direction of travel for the fishing boat, and determine whether to issue an early warning to the operating personnel.

[0041] The recommended value of the fishing area window is obtained by calculating the product of the meteorological stability index in the area and the distribution state of the fish in the area; and the fishing windows are numbered from 1 to 9 from the upper left corner to the lower right corner; the fish distribution state in the area is obtained by calculating the fish distribution state of all fishing panes in the area. The meteorological stability index in the area is obtained by setting the fish meteorological impact threshold, extracting the fishing panes whose fish meteorological impact value is less than the fish meteorological impact threshold, merging the extracted adjacent area panes through the merge_intervals function, sorting the area panes using the sorted() function, and then merging the adjacent, overlapping or touching area panes through a loop and conditional judgment, and saving the merged area panes to the merged list. Finally, the function returns the merged list as the result, which is recorded as the first fishing partition in the fishing window, and repeats the above steps until all fishing panes in all fishing windows are traversed to obtain the first fishing partition to the Kpth fishing partition, where Kp represents the number of partitions in the fishing window numbered p; the meteorological stability index is obtained by weighted summation of the average stability and distance weighted stability in the window.

[0042] The calculation formula for the average stability within the window is: ,in, represents the mean value of meteorological impact in the ith region, represents the number of panes in the ith region; the distance weighted stability calculation formula is: , Indicates the distance between partitions.

[0043] The area recommendation strategy includes extracting the direction corresponding to the maximum value of the fishing window recommendation as the direction of travel of the fishing boat; when the window recommendation value of the fishing area where the fishing boat is located is the largest, the fishing boat sails out of the area to be fished in the direction corresponding to the other 8 fishing windows with the largest recommended values ​​of the fishing windows, and when the distance from the area to be fished is greater than 2M, the direction of travel of the fishing boat is calculated again; when the other 8 fishing window recommendation values ​​are all less than the set window recommendation threshold, a warning to stop fishing is issued to the fishermen.

[0044] Figure 2 The flow chart of constructing a fish school-weather prediction model of the integrated method for shipborne intelligent weather observation, detection, forecasting and early warning disclosed in this embodiment is shown, including:

[0045] The specific construction process of the fish school-weather prediction model includes: taking the spring to winter of each year in the past as a recording cycle, calculating the sum of the actual fish school density in each fishing window in each recording cycle, obtaining the value of the sum and the proportion of all data in the recording cycle of the regional fishing window, and adding one to obtain the seasonal cycle change factor of each seasonal node in each recording cycle, extracting the seasonal cycle change factor of each seasonal node and multiplying it with the meteorological impact value of the fish school in the corresponding season to obtain the meteorological impact fish school change value of the recording cycle, so as to eliminate the influence of seasonal climate fluctuations on the model accuracy;

[0046] A decision tree algorithm is selected, and the weighted sum of the mean fish density of the seasonal nodes in the historical record cycle and the meteorological impact of the fish change value in the corresponding season of the record cycle is used as training data to train the decision tree model. In this process, since the distribution of fish density in different seasons is relatively extreme and extreme meteorological conditions may occur, the size of the fish density will change the distribution characteristics of the data set. If the fish density is very large or the meteorological conditions are relatively extreme, it may become an extreme value in the data set, causing the data distribution to become unbalanced or deviate from the normal form. The adaptability of the model depends on its ability to handle data distribution; in order to improve this problem, the algorithm in the present invention will automatically construct a tree structure based on the importance of each feature in the meteorological data, each node represents a decision rule, and the leaf node represents the prediction result; specifically including:

[0047] The information gain of each feature in the historical meteorological impact value data set is calculated to evaluate the predictive ability of each feature on the distribution status of fish schools; the meteorological threshold for fish school impact is set, and the data with meteorological impact values ​​greater than the meteorological threshold for fish school impact are listed as large-impact meteorological values, and the data with meteorological impact values ​​less than or equal to the meteorological threshold for fish school impact are listed as small-impact meteorological values.

[0048] If a feature contains a large number of small-impact meteorological values, and these meteorological impact values ​​have a significant impact on the prediction results (that is, they can distinguish samples of different categories well), then this feature may obtain a high information gain and be selected as a split feature. On the contrary, if the small-impact meteorological values ​​have little impact on the classification results, or they are only a few extreme values, then the impact of these small-impact meteorological values ​​in the information gain calculation may be diluted by other samples with more large-impact meteorological values; therefore, the feature with the largest information gain is selected as the split feature of the current node; according to the different values ​​of the selected features, the data set is divided into multiple subsets. Each subset corresponds to a child node; for each child node, the model repeats the above process until there are no more features to split;

[0049] When the recursive process stops, the nodes obtained are called fish distribution leaf nodes. Fish distribution leaf nodes are no longer split and they represent the final prediction results.

[0050] Since the fish distribution density tends to be consistent in the same season, in order to prevent the model from overfitting, the fish-weather forecast model is post-pruned, including for each non-fish distribution leaf node, considering the impact of small-impact meteorological change values ​​on model performance, calculating the error obtained based on cross-validation after pruning it (i.e. replacing its subtree with the fish distribution leaf node), and selecting the fish distribution node with the largest error reduction after pruning for pruning according to the evaluation result; repeat the above process for the pruned tree structure until there are no more nodes to prune. If the model accuracy after pruning is greater than the set accuracy threshold, the pruning result is accepted; otherwise, the model is retrained.

[0051] The fish school distribution state of each window pane area at the future time stamp TT is obtained through the fish school-weather prediction model.

[0052] Figure 3 The framework diagram of the shipborne intelligent meteorological observation, detection, forecast and warning integrated system of the present invention is shown. Based on the same inventive concept as that of Embodiment 1 and Embodiment 2, the present invention provides a shipborne intelligent meteorological observation, detection, forecast and warning integrated system, including:

[0053] A fishing window partitioning module is used to divide the area to be fished into 9 fishing windows of size M×M; and to divide all fishing area windows into N×N fishing windows;

[0054] A meteorological data analysis module is used to obtain meteorological data of all fishing windows through a ship-borne intelligent meteorological observation and detection station to obtain meteorological impact values ​​of fish schools;

[0055] The fish school-weather prediction model establishment module is used to predict the distribution status of fish schools in all fishing windows in all fishing windows; it includes a seasonal climate fluctuation impact reduction unit and a model training unit; the seasonal climate fluctuation impact reduction unit is used to obtain the seasonal cycle change factor of each seasonal node in each recording cycle, extract the seasonal cycle change factor of each seasonal node and multiply it with the fish school meteorological impact value of the corresponding season to obtain the meteorological impact fish school change value of the recording cycle; the model training unit is used to obtain the fish school distribution status of each fishing window of the future timestamp TT.

[0056] The early warning and forecasting module calculates the recommended value of the fishing window through the fish distribution status and meteorological data, establishes a regional recommendation strategy, outputs the recommended direction of travel for fishing vessels, and determines whether to issue an early warning to the operating personnel.

[0057] The threshold and weight settings can be set by the operator.

[0058] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0059] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0060] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0061] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0062] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the enlightenment of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present invention and the claims, which all fall within the protection of the present invention.

Claims

1. The integrated method of shipborne intelligent meteorological observation, detection, forecast and early warning is characterized by: The following steps are involved: S1. When the fishing boat is moving, the fishing area is divided into 9 fishing windows of size M×M with the hull as the center; And divide all fishing area panes into N×N fishing panes; S2. Obtaining meteorological data of all fishing windows through a ship-borne intelligent meteorological observation and detection station to obtain a meteorological impact value of the fish school; the meteorological data includes water temperature, dissolved oxygen content, light intensity, water flow velocity and air pressure, calculating the water temperature impact value, dissolved oxygen content impact value, light intensity impact value, water flow velocity impact value and air pressure impact value, and obtaining the fish school meteorological impact value by weighted summation after standardization; S3, establishing a fish school-weather forecast model to predict the distribution status of fish schools in all fishing windows in all fishing windows; The specific construction process of the fish school-weather prediction model includes: taking the spring to winter of each year in the past as a recording cycle, calculating the sum of the actual fish school density in each fishing window in each recording cycle, obtaining the value of the sum and the proportion of all data in the recording cycle of the regional fishing window, and adding one to obtain the seasonal cycle change factor of each seasonal node in each recording cycle, extracting the seasonal cycle change factor of each seasonal node and multiplying it with the meteorological impact value of the fish school in the corresponding season to obtain the meteorological impact fish school change value of the recording cycle; The specific construction process of the fish school-weather prediction model also includes: calculating the weighted sum of the fish school density mean value of the seasonal node in the record cycle and the meteorological impact on the fish school change value of the corresponding season of the record cycle as the historical fish school distribution state, using the historical fish school meteorological impact value and the historical fish school distribution state as training data, and automatically constructing a tree structure based on the importance of each feature of the fish school meteorological impact value, each node represents a decision rule, and the leaf node represents the prediction result; The specific construction process of the tree structure includes: evaluating the prediction ability of each feature on the distribution state of fish schools by calculating the information gain of each feature in the historical meteorological impact value; setting the meteorological threshold of fish school impact, listing the data with the meteorological impact value greater than the meteorological threshold of fish school impact as the meteorological value with large impact, and listing the data with the meteorological impact value less than or equal to the meteorological threshold of fish school impact as the meteorological value with small impact; when the meteorological value of a certain feature with small impact is greater than the set first quantity threshold, the feature is selected as the split feature, on the contrary, when the meteorological value of a certain feature with small impact is less than or equal to the set first quantity threshold, the feature with the largest information gain is selected as the split feature of the current node; according to different values ​​of the selected features, the data set is divided into multiple subsets, each subset corresponds to a child node; for each child node, the model repeats the above process until there are no more features to split; The specific construction process of the fish school-weather prediction model also includes: performing a post-pruning operation on the fish school-weather prediction model, including, for each leaf node that is not the fish school distribution, considering the influence of the small-impact meteorological value on the model performance, calculating the error obtained based on cross-validation after pruning, and selecting the fish school distribution node with the largest error reduction after pruning for pruning according to the evaluation result; repeating the above process for the pruned tree structure until there are no more nodes to prune; if the accuracy of the pruned model is greater than the set accuracy threshold, accepting the pruning result; otherwise, retraining the model; obtaining the fish school distribution status of each fishing window of the future timestamp TT; S4. Calculate the recommended value of the fishing window through the fish distribution status and meteorological data, establish a regional recommendation strategy, output the recommended direction of travel for the fishing boat, and determine whether to issue an early warning to the operating personnel.

2. The integrated method of shipborne intelligent weather observation, detection, forecasting and early warning according to claim 1 is characterized by: The recommended value of the fishing area window is obtained by calculating the product of the meteorological stability index in the area and the fish distribution status in the area; and the fishing windows are numbered 1 to 9 from the upper left corner to the lower right corner; the fish distribution status in the area is obtained by calculating the average value of the fish distribution status of all fishing windows in the area.

3. The integrated method of shipborne intelligent weather observation, detection, forecasting and early warning according to claim 2 is characterized by: The meteorological stability index in the area is obtained by setting a threshold value of meteorological influence of fish schools, and the fishing panes whose meteorological influence values ​​of fish schools are less than the threshold value of meteorological influence of fish schools are extracted. The extracted adjacent area panes are merged by the merge_intervals function, and the area panes are sorted using the sorted() function. Then, adjacent, overlapping or touching area panes are merged through a loop and conditional judgment, and the merged area panes are saved in a merged list. Finally, the function returns the merged list as a result, which is recorded as the first fishing partition in the fishing window. The above steps are repeated until all fishing panes in all fishing windows are traversed to obtain the first fishing partition to the Kpth fishing partition, where Kp represents the number of partitions in the fishing window numbered p. The meteorological stability index is obtained by weighted summing the average stability within the window and the distance-weighted stability.

4. The integrated method of shipborne intelligent weather observation, detection, forecasting and early warning according to claim 3 is characterized by: The calculation formula for the average stability within the window is: ,in, represents the mean value of meteorological impact in the ith region, represents the number of panes in the ith region; the distance weighted stability calculation formula is: , Indicates the distance between partitions.

5. The integrated method of shipborne intelligent weather observation, detection, forecasting and early warning according to claim 4 is characterized by: The area recommendation strategy includes extracting the direction corresponding to the maximum value of the fishing window recommendation as the direction of travel of the fishing boat; when the window recommendation value of the fishing area where the fishing boat is located is the largest, the fishing boat sails out of the area to be fished in the direction corresponding to the other 8 fishing windows with the largest recommended values ​​of the fishing windows, and when the distance from the area to be fished is greater than 2M, the direction of travel of the fishing boat is calculated again; when the other 8 fishing window recommendation values ​​are all less than the set window recommendation threshold, a warning to stop fishing is issued to the fishermen.

6. A shipborne intelligent meteorological observation, detection, forecast and warning integrated system, used to execute the shipborne intelligent meteorological observation, detection, forecast and warning integrated method as claimed in any one of claims 1 to 5, characterized in that: include: A fishing window division module is used to divide the area to be fished into 9 fishing windows of size M×M; And divide all fishing area panes into N×N fishing panes; A meteorological data analysis module is used to obtain meteorological data of all fishing windows through a ship-borne intelligent meteorological observation and detection station to obtain meteorological impact values ​​of fish schools; A fish school-weather forecast model building module is used to predict the distribution status of fish schools in all fishing windows in all fishing windows; The early warning and forecasting module calculates the recommended value of the fishing window through the distribution status of fish schools and meteorological data, establishes a regional recommendation strategy, outputs the recommended direction of travel for fishing vessels, and determines whether to issue an early warning to operating personnel.

7. The shipborne intelligent weather observation, detection, forecast and early warning integrated system according to claim 6 is characterized by: The fish school-weather prediction model establishment module includes a seasonal climate fluctuation impact reduction unit and a model training unit; the seasonal climate fluctuation impact reduction unit is used to obtain the seasonal cycle change factor of each seasonal node in each recording cycle, extract the seasonal cycle change factor of each seasonal node and multiply it by the fish school meteorological impact value of the corresponding season to obtain the meteorological impact fish school change value of the recording cycle; the model training unit is used to obtain the fish school distribution status of each fishing window of the future timestamp TT.

Images