Unmanned fleet intelligent operation method and system based on artificial intelligence, and storage medium

Abstract

The invention discloses an unmanned fleet intelligent operation method and system based on artificial intelligence, and a storage medium, and the method comprises the steps: collecting water area environment data and fishery resource data in a monitoring area, carrying out the regional clustering grouping of a preset water area, carrying out the water area feeding demand analysis based on the clustered region group, and carrying out the automatic feeding of the water area; the feeding demand degree of each monitoring area is obtained; feeding priority analysis is carried out based on the feeding demand degree, a first area and a second area are marked, based on the initial position information of the unmanned fleet, the first area is used as a feeding demand point, shortest path planning is carried out on the unmanned fleet through a Floyd path algorithm, and a shortest navigation path is obtained; according to the real-time path, ship behavior early warning judgment and secondary path planning of the unmanned fleet are carried out, optimized path information is generated, and the optimized path information is sent to a preset unmanned ship terminal. According to the invention, the unmanned fleet system with full automation, high early warning capability, high safety and high self-processing capability can be realized.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and more specifically, to an intelligent operation method, system and storage medium for an unmanned fleet based on artificial intelligence. Background Art

[0002] With the continuous development of technology, artificial intelligence technology has been more and more widely applied in various fields. In the field of water navigation, as a new type of transportation and fishery feeding method, the unmanned fleet has the advantages of high efficiency, safety, environmental protection, etc., and has been widely applied.

[0003] However, restricted by traditional technologies, there is currently a lack of joint scheduling technology for multiple unmanned boats, a lack of analysis of the efficient operation of the fleet in the water area, a lack of analysis of the efficient feeding of the unmanned fleet in the water area, etc. This has seriously hindered the application of unmanned boats. Therefore, there is an urgent need for an intelligent operation method, system and storage medium for an unmanned fleet based on artificial intelligence. Summary of the Invention

[0004] The present invention overcomes the defects of the prior art and provides an intelligent operation method, system and storage medium for an unmanned fleet based on artificial intelligence.

[0005] The first aspect of the present invention provides an intelligent operation method for an unmanned fleet based on artificial intelligence, including:

[0006] Obtain the target water area range information, and construct a water area map model based on the target water area range information;

[0007] Based on preset water area monitoring points, divide the water area through the water area map model to form multiple monitoring areas;

[0008] Collect the water area environment data and fishery resource data in the monitoring area, perform regional clustering analysis based on the water area environment data and fishery resource data, perform water area feeding demand analysis based on the clustered area groups, and obtain the feeding demand degree of each monitoring area;

[0009] Perform feeding priority analysis based on the feeding demand degree and mark the first area and the second area. Based on the initial position information of the unmanned fleet, use the first area as the feeding demand point, and perform the shortest path planning for the unmanned fleet through the Floyd path algorithm to obtain the shortest navigation path of each unmanned boat;

[0010] Perform navigation control of the unmanned fleet based on the shortest navigation path, obtain the path of the unmanned boat in real time, perform early warning judgment on the ship behavior according to the real-time path. If there is an abnormality in the unmanned boat, generate an early warning information, perform secondary route planning on the unmanned fleet through the early warning information, generate optimized path information, and send the optimized path information to a preset unmanned boat terminal.

[0011] In this solution, obtaining the information on the target water area range and constructing a water area map model based on the information on the target water area range are specifically as follows:

[0012] Obtain the information on the target water area range;

[0013] The information on the target water area range includes the water area area, the contour of the water area map, and the information on preset water area monitoring points;

[0014] Construct a water area map model based on three-dimensional visualization based on the information on the target water area range.

[0015] In this solution, based on the preset water area monitoring points, the water area is divided into multiple monitoring areas through the water area map model, specifically as follows:

[0016] Through the water area map model, divide the entire target water area into multiple monitoring areas based on the preset water area monitoring points;

[0017] One monitoring area corresponds to one water area monitoring point.

[0018] In this solution, collecting the water area environment data and fishery resource data in the monitoring areas, performing regional clustering analysis based on the water area environment data and fishery resource data, and performing water area feeding demand analysis based on the clustered regional groups, and obtaining the feeding demand degree of each monitoring area, including:

[0019] Within an analysis period, collect the corresponding water area environment data and underwater image data for each monitoring area;

[0020] The water area environment data includes water temperature, flow velocity, flow direction, and water quality data;

[0021] Perform fishery resource image recognition based on the image recognition model on the underwater image data, and perform resource statistics based on the recognition results to obtain the fishery resource data of each monitoring area;

[0022] Perform data cleaning, data standardization, and data integration on the water area environment data and fishery resource data of each monitoring area to form integrated monitoring data;

[0023] Use the integrated monitoring data of all monitoring areas as clustering sample data, perform clustering grouping based on the spectral clustering algorithm, and form multiple regional groups;

[0024] Each regional group includes at least one monitoring area.

[0025] In this solution, the water environment data and fishery resource data in the acquisition and monitoring area are used to perform regional clustering analysis based on the water environment data and fishery resource data. Then, based on the clustered regional groups, water feeding demand analysis is carried out, and the feeding demand degree of each monitoring area is obtained. Specifically:

[0026] Select all the monitoring areas in a regional group and mark all the monitoring areas as the current monitoring areas;

[0027] Integrate the fishery resource data in all the current monitoring areas to form the regional group fishery resource data;

[0028] Calculate and analyze the feeding demand degree of the regional group according to the regional group fishery resource data and fishery breeding requirements;

[0029] Take the feeding demand degree as the feeding demand degree of all the monitoring areas in a regional group;

[0030] Analyze the remaining regional groups and obtain the feeding demand degrees of all the monitoring areas.

[0031] In this solution, based on the feeding demand degree, feeding priority analysis is carried out and the first area and the second area are marked. Based on the initial position information of the unmanned fleet, taking the first area as the feeding demand point, the shortest path planning for the unmanned fleet is carried out through the Floyd path algorithm to obtain the shortest navigation path of each unmanned ship. Specifically:

[0032] Carry out feeding priority analysis according to the feeding demand degree. The analysis process is based on the judgment and comparison of the first threshold and the second threshold, and all the monitoring areas are marked to obtain the first area, the second area, and the third area;

[0033] Obtain the initial position information of each unmanned ship in the unmanned fleet;

[0034] Based on the initial position information, taking the initial position of the unmanned ship as the origin, mark all the first areas covered within the preset radius based on the origin as the feeding demand points;

[0035] Take the initial position of the unmanned ship as the starting point and the ending point, and take the feeding demand points as the passing points. Based on the Floyd path algorithm, carry out the shortest path planning for the unmanned ship to obtain multiple initial paths of the unmanned ship;

[0036] Based on the multiple initial paths, calculate and analyze the number of second areas passed by each path, and mark the path with the largest number of second areas passed as the shortest navigation path of the unmanned ship;

[0037] Analyze all the unmanned ships in the fleet to obtain the shortest navigation path of each unmanned ship.

[0038] In this solution, for the navigation control of the unmanned fleet based on the shortest navigation path, the path of the unmanned ship is obtained in real time, and the ship behavior early warning judgment is carried out according to the real-time path. If there is an abnormality in the unmanned ship, an early warning information is generated, and the unmanned fleet is re-routed based on the early warning information to generate optimized path information, and the optimized path information is sent to the preset unmanned ship terminal, including:

[0039] Generate a fleet navigation plan based on the shortest navigation path of each unmanned ship, and carry out the navigation control of the unmanned fleet according to the fleet navigation plan;

[0040] Based on the navigation process, obtain the position information of the current unmanned ship in real time and get the real-time path;

[0041] Visualize the shortest path corresponding to the current unmanned ship to form a path image, extract the contour features of the path image, and obtain the first feature data;

[0042] Visualize and extract features from the real-time path to obtain the second feature data;

[0043] Based on the standard Euclidean distance, calculate the similarity between the first feature data and the second feature data, and judge whether the similarity is greater than the preset similarity threshold. If so, judge that the navigation behavior of the current unmanned ship is abnormal, and record the position point of the current unmanned ship;

[0044] Obtain the navigation status information of the current unmanned ship in real time, and integrate the navigation status information with the position point of the current unmanned ship to form an early warning information.

[0045] In this solution, for the navigation control of the unmanned fleet based on the shortest navigation path, the path of the unmanned ship is obtained in real time, and the ship behavior early warning judgment is carried out according to the real-time path. If there is an abnormality in the unmanned ship, an early warning information is generated, and the unmanned fleet is re-routed based on the early warning information to generate optimized path information, and the optimized path information is sent to the preset unmanned ship terminal, specifically:

[0046] Obtain the navigation status information in the early warning information for abnormal assessment and generate corresponding emergency control instructions;

[0047] Obtain the position point of the current unmanned ship in the early warning information, and obtain other unmanned ships within the preset range of the position point of the current unmanned ship through the water area map model, and mark them as neighboring unmanned ships;

[0048] Statistically integrate the feeding demand points of the current unmanned ship and the feeding demand points of the neighboring unmanned ships to obtain a set of feeding demand points;

[0049] Taking the real-time position of the neighboring unmanned ship as the starting point and the position points in the feeding demand point set as the waypoints, perform a secondary shortest path planning for the neighboring unmanned ship through the Floyd path algorithm, and generate optimized path information.

[0050] The second aspect of the present invention also provides an intelligent operation system for an unmanned ship fleet based on artificial intelligence. The system includes: a memory and a processor. The memory includes an intelligent operation program for an unmanned ship fleet based on artificial intelligence. When the intelligent operation program for an unmanned ship fleet based on artificial intelligence is executed by the processor, the following steps are implemented:

[0051] Obtain the target water area range information, and construct a water area map model based on the target water area range information;

[0052] Based on the preset water area monitoring points, divide the water area through the water area map model to form multiple monitoring areas;

[0053] Collect the water area environment data and fishery resource data in the monitoring area, perform regional clustering analysis based on the water area environment data and fishery resource data, perform water area feeding demand analysis based on the clustered area groups, and obtain the feeding demand degree of each monitoring area;

[0054] Perform feeding priority analysis based on the feeding demand degree and mark the first area and the second area. Based on the initial position information of the unmanned ship fleet, use the first area as the feeding demand point, and perform the shortest path planning for the unmanned ship fleet through the Floyd path algorithm to obtain the shortest navigation path of each unmanned ship;

[0055] Perform navigation control of the unmanned ship fleet based on the shortest navigation path, obtain the path of the unmanned ship in real time, perform a ship behavior warning judgment according to the real-time path. If there is an abnormality in the unmanned ship, generate a warning information, perform a secondary route planning for the unmanned ship fleet through the warning information, generate optimized path information, and send the optimized path information to the preset unmanned ship terminal.

[0056] The third aspect of the present invention also provides a computer-readable storage medium. The computer-readable storage medium includes an intelligent operation program for an unmanned ship fleet based on artificial intelligence. When the intelligent operation program for an unmanned ship fleet based on artificial intelligence is executed by a processor, the steps of the intelligent operation method for an unmanned ship fleet based on artificial intelligence as described in any one of the above are implemented.

[0057] The present invention discloses an intelligent operation method, system and storage medium for an unmanned fleet based on artificial intelligence. By collecting water area environment data and fishery resource data in a monitoring area, regional clustering and grouping of a preset water area are performed, and based on the clustered regional groups, an analysis of the feeding requirements of the water area is carried out, and the feeding requirement degree of each monitoring area is obtained; based on the feeding requirement degree, an analysis of the feeding priority is carried out and the first area and the second area are marked. Based on the initial position information of the unmanned fleet, with the first area as the feeding requirement point, the shortest path planning of the unmanned fleet is carried out through the Floyd path algorithm to obtain the shortest navigation path; according to the real-time path, a judgment on the ship behavior warning and a secondary route planning of the unmanned fleet are carried out to generate optimized path information, and the optimized path information is sent to a preset unmanned ship terminal. Through the present invention, it is helpful to realize an unmanned fleet system with full automation, high warning ability, high safety and high self-processing ability. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 FIG. shows a flowchart of an intelligent operation method for an unmanned fleet based on artificial intelligence according to the present invention;

[0059] Figure 2 FIG. shows a flowchart of constructing a water area map model according to the present invention;

[0060] Figure 3 FIG. shows a block diagram of an intelligent operation system for an unmanned fleet based on artificial intelligence according to the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0061] In order to more clearly understand the above objects, features and advantages of the present invention, the present invention will be further described in detail below with reference to the drawings and specific embodiments. It should be noted that, without conflict, the embodiments of the present application and the features in the embodiments may be combined with each other.

[0062] In the following description, many specific details are set forth in order to fully understand the present invention. However, the present invention may be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited by the specific embodiments disclosed below.

[0063] Figure 1 FIG. shows a flowchart of an intelligent operation method for an unmanned fleet based on artificial intelligence according to the present invention.

[0064] As Figure 1 shown, in a first aspect of the present invention, an intelligent operation method for an unmanned fleet based on artificial intelligence is provided, including:

[0065] S102. Obtain the target water area range information, and construct a water area map model based on the target water area range information;

[0066] S104. Based on the preset water area monitoring points, divide the water area through the water area map model to form multiple monitoring areas;

[0067] S106. Collect the water area environment data and fishery resource data in the monitoring areas, conduct regional clustering analysis based on the water area environment data and fishery resource data, conduct water area feeding demand analysis based on the clustered area groups, and obtain the feeding demand degree of each monitoring area;

[0068] S108. Conduct feeding priority analysis based on the feeding demand degree and mark the first area and the second area. Based on the initial position information of the unmanned fleet, use the first area as the feeding demand point, and conduct the shortest path planning for the unmanned fleet through the Floyd path algorithm to obtain the shortest navigation path of each unmanned ship;

[0069] S110. Conduct the navigation control of the unmanned fleet based on the shortest navigation path, obtain the path of the unmanned ship in real time, conduct ship behavior early warning judgment according to the real-time path. If there is an abnormality in the unmanned ship, generate an early warning message, conduct secondary route planning for the unmanned fleet through the early warning message, generate optimized path information, and send the optimized path information to the preset unmanned ship terminal.

[0070] It should be noted that the unmanned fleet refers to all unmanned ships. Each unmanned ship includes hardware such as a drive module, an underwater monitoring module, a GPS communication module, a feeding device, and a display terminal.

[0071] Figure 2 Shows the flowchart of the construction of the water area map model of the present invention.

[0072] According to an embodiment of the present invention, the obtaining of the target water area range information and the construction of the water area map model based on the target water area range information are specifically as follows:

[0073] S202. Obtain the target water area range information;

[0074] S204. The target water area range information includes the water area area, the water area map contour, and the preset water area monitoring point information;

[0075] S206. Construct a three-dimensional visualization-based water area map model based on the target water area range information.

[0076] It should be noted that in the water area map model, subsequent division of relevant areas, analysis and visualization of fishery resources, and visualization of the feeding navigation of the unmanned fleet can be carried out.

[0077] According to an embodiment of the present invention, the dividing of the water area through the water area map model based on the preset water area monitoring points to form multiple monitoring areas is specifically as follows:

[0078] Through the water area map model, the entire target water area is divided into multiple monitoring areas based on preset water area monitoring points;

[0079] One monitoring area corresponds to one water area monitoring point.

[0080] It should be noted that each monitoring point is equipped with underwater environment monitoring equipment, which can monitor and obtain underwater environment data in real time and analyze fishery resource data. Based on actual needs, relevant underwater data can also be obtained through the underwater device in the unmanned boat. The preset water area monitoring points are multiple position points preset in the water area.

[0081] According to the embodiment of the present invention, collecting water area environment data and fishery resource data in the monitoring area, performing regional clustering analysis based on the water area environment data and fishery resource data, performing water area feeding demand analysis based on the clustered area groups, and obtaining the feeding demand degree of each monitoring area, including:

[0082] In an analysis period, collecting corresponding water area environment data and underwater image data for each monitoring area;

[0083] The water area environment data includes water temperature, flow velocity, flow direction, and water quality data;

[0084] Performing fishery resource image recognition based on the image recognition model on the underwater image data, and performing resource statistics based on the recognition result to obtain fishery resource data for each monitoring area;

[0085] Performing data cleaning, data standardization, and data integration on the water area environment data and fishery resource data of each monitoring area to form monitoring integrated data;

[0086] Using the monitoring integrated data of all monitoring areas as clustering sample data, performing clustering grouping based on the spectral clustering algorithm, and forming multiple area groups;

[0087] Each area group includes at least one monitoring area.

[0088] It should be noted that the image recognition model is used to recognize fishery resources in underwater images, such as the recognition of fish, crustaceans, and algae. In the embodiments of the present invention, specifically, the YOLO object detection model, the CNN image recognition model, etc. can be used. These models have the ability of deep learning, and through the existing fishery resource maps (i.e., existing images of fish, crustaceans, algae, etc.), feature extraction learning, model training, image recognition, and image classification are carried out. In addition, in the recognition of fishery resource images, other classification models can still be used for recognition and analysis. Any relevant models for fishery resource image recognition can be used in the embodiments of the present invention for recognition and statistics. The main purpose of the recognition process is to conduct resource statistics, that is, the statistical analysis of resources such as fish, crustaceans, and algae. Fishery resource data includes information such as the quantity and type of various resources.

[0089] It is worth mentioning that in the present invention, through the spectral clustering algorithm, based on the water area environment data and fishery resource data, regional grouping is carried out to form multiple regional groups. In each regional group, the corresponding monitoring areas have certain resource similarity and demand similarity. By calculating and analyzing the feeding demand degree in units of regional groups, and based on the feeding demand degree, the priority analysis and path planning of the regions are carried out. Thereby, it can effectively reduce the complexity of path analysis for large water areas. Especially in the route planning analysis of large-scale water areas and unmanned fleets, the present invention can reduce unnecessary repetitive calculation and analysis while ensuring the rationality of the planning, improve the planning efficiency, and provide technical support for the feeding planning of unmanned fleets.

[0090] According to the embodiments of the present invention, the water area environment data and fishery resource data in the collected monitoring areas are clustered and analyzed based on the water area environment data and fishery resource data, and the water area feeding demand analysis is carried out based on the regional groups after clustering, and the feeding demand degree of each monitoring area is obtained. Specifically:

[0091] Select all the monitoring areas in a regional group and mark all the monitoring areas as the current monitoring areas;

[0092] Integrate the fishery resource data in all the current monitoring areas to form the regional group fishery resource data;

[0093] Calculate and analyze the feeding demand degree of the regional group according to the regional group fishery resource data and fishery breeding requirements;

[0094] Take the feeding demand degree as the feeding demand degree of all the monitoring areas in a regional group;

[0095] Analyze the remaining regional groups and obtain the feeding demand degrees of all the monitoring areas.

[0096] It should be noted that the feeding demand degree reflects the feeding demand level of a monitoring area and is related to the current fishery resource situation and demand situation. The current fishery resource situation is obtained based on the fishery resource data of the regional group. The larger the current fishery resource quantity, the lower the feeding demand degree; the higher the fishery farming demand, the higher the corresponding feeding demand degree. Therefore, the feeding demand degree is affected by two aspects of data (the fishery resource data of the regional group and the fishery farming demand). And the higher the feeding demand degree, the higher the feeding priority of the corresponding area.

[0097] According to an embodiment of the present invention, based on the feeding demand degree, feeding priority analysis is performed to mark the first area and the second area. Based on the initial position information of the unmanned fleet, with the first area as the feeding demand point, the shortest path planning for the unmanned fleet is carried out through the Floyd path algorithm to obtain the shortest navigation path of each unmanned ship. Specifically:

[0098] Feeding priority analysis is performed according to the feeding demand degree. The analysis process is based on the judgment and comparison of the first threshold and the second threshold. All monitoring areas are marked to obtain the first area, the second area, and the third area;

[0099] Obtain the initial position information of each unmanned ship in the unmanned fleet;

[0100] Based on the initial position information, with the initial position where the unmanned ship is located as the origin, all the first areas covered within the preset radius based on the origin are marked as feeding demand points;

[0101] Taking the initial position where the unmanned ship is located as the starting point and the ending point, and the feeding demand points as the passing points, the shortest path planning for the unmanned ship is carried out based on the Floyd path algorithm to obtain multiple initial paths of the unmanned ship;

[0102] Based on the multiple initial paths, calculate and analyze the number of second areas passed by each path, and mark the path with the largest number of second areas passed as the shortest navigation path of the unmanned ship;

[0103] Analyze all the unmanned ships in the fleet to obtain the shortest navigation path of each unmanned ship.

[0104] It should be noted that in the judgment and comparison based on the first threshold and the second threshold, all monitoring areas are marked to obtain the first area, the second area, and the third area. The condition for marking the first area is that the feeding demand degree of the corresponding monitoring area is greater than the first threshold. The condition for marking the second area is that the feeding demand degree of the corresponding monitoring area is less than or equal to the first threshold and greater than or equal to the second threshold. The condition for marking the third area is that the feeding demand degree of the corresponding monitoring area is less than the third threshold. The first threshold is greater than the second threshold. The feeding priorities of the first, second, and third areas decrease gradually. The first area is the necessary feeding area and needs to be analyzed as a feeding demand point during subsequent route planning. Among the set starting point and ending point, based on the analysis requirements of the unmanned ship, the ending point can be not set, and the route planning is carried out with the purpose of completing the passing area. The multiple initial paths are obtained by screening from multiple legal paths planned by an algorithm. The screening principle is to select the shortest several paths. Each initial path passes through the corresponding feeding demand point.

[0105] The value range of the feeding demand degree is [0 - 100]. The feeding demand degree is related to the current fishery resource situation and demand situation. It can be scored by analyzing the resources and demand, and the score value is mapped to the [0 - 100] interval for representing the feeding demand degree value. For general waters, the first threshold, the second threshold, and the third threshold can be respectively set as 80, 50, and 30. Additionally, based on the research and analysis requirements, the numerical interval and the three thresholds can all be dynamically adjusted.

[0106] Among the marking of all first areas covered within the preset radius based on the origin as feeding demand points, the middle position point of the first area is used as the feeding demand point, and the feeding demand point corresponds to the first area one by one.

[0107] In the acquisition of the initial position information of each unmanned ship in the unmanned ship fleet, the initial positions of the unmanned ship fleet are generally scattered. In this embodiment, by analyzing the position ranges of each scattered unmanned ship, the feeding points are allocated, so as to maximize the feeding efficiency of the unmanned ship fleet.

[0108] According to the embodiment of the present invention, for the navigation control of the unmanned ship fleet based on the shortest navigation path, the path of the unmanned ship is obtained in real time, and the ship behavior early warning judgment is carried out according to the real-time path. If there is an abnormality in the unmanned ship, an early warning information is generated, and a secondary route planning is carried out for the unmanned ship fleet through the early warning information to generate optimized path information, and the optimized path information is sent to the preset unmanned ship terminal, including:

[0109] Generate a fleet navigation plan based on the shortest navigation path of each unmanned ship, and carry out the navigation control of the unmanned ship fleet according to the fleet navigation plan;

[0110] Based on the navigation process, obtain the position information of the current unmanned ship in real time and obtain the real-time path;

[0111] Visualize the shortest path corresponding to the current unmanned ship to form a path image, extract contour features from the path image to obtain first feature data;

[0112] Visualize and extract features from the real-time path to obtain second feature data;

[0113] Based on the standard Euclidean distance, calculate the similarity between the first feature data and the second feature data, and determine whether the similarity is greater than a preset similarity threshold. If so, determine that the navigation behavior of the current unmanned ship is abnormal and record the position point of the current unmanned ship;

[0114] Real-time obtain the navigation status information of the current unmanned ship, and integrate the navigation status information with the position point of the current unmanned ship to form a warning message.

[0115] It is worth mentioning that in the present invention, by extracting features in the form of a path image and using the form of feature comparison to judge route deviation, a judgment result consistent with the actual situation can be achieved. Since the shortest navigation path is an ideal path, there will be certain deviations in reality, and there will be certain reasonable deviations in actual navigation. Within the reasonable deviation range, the navigation tasks of the unmanned ship fleet are not affected. Through the contour feature extraction judgment of the present invention, reasonable deviation regulation can be carried out based on the preset similarity threshold. The higher the similarity threshold is set, the higher the range of reasonable deviation becomes, so it can be set according to the actual situation to adapt to different unmanned ship fleets and different water area environments. In the traditional technology, often through simple route comparison, the error is large, the controllable range of the judgment standard is low, and the judgment result is prone to be inconsistent with the actual navigation. The process of visualizing and extracting features from the real-time path is the same as the analysis process of the first feature data.

[0116] In the first feature data and the second feature data, specifically, the shortest path or the real-time path is visualized in the map to generate a path image, and the first feature data and the second feature data are obtained through image contour feature extraction. The contour extraction can be based on edge detection operators, etc. to extract contour features. Here, the similarity between the two paths is mainly compared in the form of visualization. Both the first feature data and the second feature data are contour feature data.

[0117] According to an embodiment of the present invention, for the navigation control of the unmanned ship fleet based on the shortest navigation path, the path of the unmanned ship is obtained in real time, and the behavior warning of the ship is judged according to the real-time path. If there is an abnormality in the unmanned ship, a warning message is generated, and the secondary route planning of the unmanned ship fleet is carried out through the warning message to generate optimized path information, and the optimized path information is sent to a preset unmanned ship terminal, specifically:

[0118] Obtain the navigation status information in the early warning information for anomaly assessment and generate corresponding emergency control instructions;

[0119] Obtain the position points of the current unmanned ship in the early warning information, and through the water area map model, obtain other unmanned ships within the preset range of the position points of the current unmanned ship, and mark them as neighboring unmanned ships;

[0120] Statistically integrate the feeding demand points of the current unmanned ship and the feeding demand points of neighboring unmanned ships to obtain a set of feeding demand points;

[0121] Based on the real-time position of the neighboring unmanned ship as the starting point and the position points in the set of feeding demand points as the waypoints, perform a secondary shortest path planning for the neighboring unmanned ships through the Floyd path algorithm, and generate optimized path information.

[0122] It should be noted that the navigation status information includes the navigation status of the unmanned ship, fuel situation, communication signal status, etc. Due to the complexity of the water area environment and the uncertainty of the unmanned ship fleet navigation, when carrying out large-scale unmanned ship fleet navigation tasks, there may be some unexpected situations, such as the route deviation, the behavior characteristics do not match the expectations, etc., and the reasons for the abnormal route may be insufficient fuel, GPS signal problems, water area route congestion, etc. Therefore, corresponding emergency controls need to be made, such as staying on standby in place, returning, etc.

[0123] The optimized path information is analyzed for neighboring unmanned ships. There is at least one neighboring unmanned ship.

[0124] In the present invention, by marking the abnormal unmanned ship, integrating the remaining waypoint feeding points of the current unmanned ship and the remaining waypoint feeding points of neighboring ships, performing secondary planning based on the path algorithm, generating new path information in real time and sending it to neighboring ships, thus effectively realizing real-time early warning regulation in case of emergencies under the condition of large-scale fleet operation, greatly improving the environmental adaptability of the unmanned ship fleet, reducing manual intervention, improving the informatization level of the unmanned ship fleet, and helping to realize an unmanned ship fleet system with full automation, high early warning ability, high safety, and high self-processing ability. Compared with the prior art, the present invention can realize dynamic route regulation of the unmanned ship fleet, and can also meet the feeding requirements of the unmanned ship fleet and plan the optimal path.

[0125] Figure 3 Shows a block diagram of an intelligent operation system for an unmanned ship fleet based on artificial intelligence according to the present invention.

[0126] In the second aspect of the present invention, there is also provided an intelligent operation system 3 for an unmanned fleet based on artificial intelligence. The system includes: a memory 31 and a processor 32. The memory includes an intelligent operation program for an unmanned fleet based on artificial intelligence. When the intelligent operation program for an unmanned fleet based on artificial intelligence is executed by the processor, the following steps are implemented:

[0127] Obtain the target water area range information, and construct a water area map model based on the target water area range information;

[0128] Based on preset water area monitoring points, divide the water area through the water area map model to form multiple monitoring areas;

[0129] Collect water area environment data and fishery resource data in the monitoring areas, perform regional clustering analysis based on the water area environment data and fishery resource data, perform water area feeding demand analysis based on the clustered regional groups, and obtain the feeding demand degree of each monitoring area;

[0130] Perform feeding priority analysis based on the feeding demand degree and mark the first area and the second area. Based on the initial position information of the unmanned fleet, use the first area as the feeding demand point, and perform the shortest path planning for the unmanned fleet through the Floyd path algorithm to obtain the shortest navigation path of each unmanned ship;

[0131] Perform navigation control of the unmanned fleet based on the shortest navigation path, obtain the path of the unmanned ship in real time, perform early warning judgment on the ship behavior according to the real-time path. If there is an abnormality in the unmanned ship, generate an early warning message, perform secondary route planning on the unmanned fleet through the early warning message, generate optimized path information, and send the optimized path information to the preset unmanned ship terminal.

[0132] It should be noted that the unmanned fleet refers to all unmanned ships. Each unmanned ship includes hardware such as a drive module, an underwater monitoring module, a GPS communication module, a feeding device, and a display terminal.

[0133] According to an embodiment of the present invention, the obtaining of the target water area range information and constructing a water area map model based on the target water area range information are specifically as follows:

[0134] Obtain the target water area range information;

[0135] The target water area range information includes water area area, water area map contour, and preset water area monitoring point information;

[0136] Construct a three-dimensional visualization-based water area map model based on the target water area range information.

[0137] It should be noted that in the water area map model, subsequent division of relevant areas, analysis and visualization of fishery resources, and visualization of feeding voyages of unmanned fleets can be carried out.

[0138] According to an embodiment of the present invention, based on preset water area monitoring points, the water area is divided through a water area map model to form multiple monitoring areas, specifically as follows:

[0139] Through the water area map model, the entire target water area is divided based on preset water area monitoring points to form multiple monitoring areas;

[0140] One monitoring area corresponds to one water area monitoring point.

[0141] It should be noted that each monitoring point is equipped with underwater environment monitoring equipment, which can monitor and obtain underwater environment data and analyze fishery resource data in real time. Based on actual needs, relevant underwater data can also be obtained through the underwater device in the unmanned ship. The preset water area monitoring points are multiple position points preset in the water area.

[0142] According to an embodiment of the present invention, collecting water area environment data and fishery resource data in the monitoring area, performing regional clustering analysis based on the water area environment data and fishery resource data, performing water area feeding demand analysis based on the clustered area groups, and obtaining the feeding demand degree of each monitoring area, including:

[0143] Within an analysis period, collecting corresponding water area environment data and underwater image data for each monitoring area;

[0144] The water area environment data includes water temperature, flow rate, flow direction, and water quality data;

[0145] Performing fishery resource image recognition based on the image recognition model on the underwater image data, and performing resource statistics based on the recognition results to obtain fishery resource data for each monitoring area;

[0146] Clearing, standardizing, and integrating the water area environment data and fishery resource data of each monitoring area to form monitoring integrated data;

[0147] Using the monitoring integrated data of all monitoring areas as clustering sample data, performing clustering grouping based on the spectral clustering algorithm, and forming multiple area groups;

[0148] Each area group includes at least one monitoring area.

[0149] It is worth mentioning that in the present invention, through the spectral clustering algorithm, regional grouping is performed based on water area environment data and fishery resource data to form multiple regional groups. In each regional group, the corresponding monitoring areas have certain resource similarities and demand similarities. By calculating and analyzing the feeding demand degree in units of regional groups, and based on the feeding demand degree, the priority analysis and path planning of the regions are carried out. Thereby, it can effectively reduce the complexity of path analysis for large water areas. Especially in the route planning analysis of large-scale water areas and unmanned fleets, the present invention can reduce unnecessary repetitive calculation and analysis while ensuring the rationality of the planning, improve the planning efficiency, and provide technical support for the feeding planning of unmanned fleets.

[0150] According to an embodiment of the present invention, the water area environment data and fishery resource data in the collected monitoring areas are collected, regional clustering analysis is performed based on the water area environment data and fishery resource data, water area feeding demand analysis is performed based on the clustered regional groups, and the feeding demand degree of each monitoring area is obtained. Specifically:

[0151] All monitoring areas in a regional group are selected, and all the monitoring areas are marked as the current monitoring areas;

[0152] The fishery resource data in all the current monitoring areas are integrated to form regional group fishery resource data;

[0153] According to the regional group fishery resource data and fishery breeding requirements, the feeding demand degree of the regional group is calculated and analyzed;

[0154] The feeding demand degree is used as the feeding demand degree of all monitoring areas in a regional group;

[0155] The remaining regional groups are analyzed, and the feeding demand degrees of all monitoring areas are obtained.

[0156] It should be noted that the feeding demand degree reflects the feeding demand degree of a monitoring area and is related to the current fishery resource situation and demand situation. The current fishery resource situation is obtained based on the regional group fishery resource data. The larger the current fishery resource quantity, the lower the feeding demand degree, and the higher the fishery breeding demand, the higher the corresponding feeding demand degree. Therefore, the feeding demand degree is affected by two aspects of data (regional group fishery resource data and fishery breeding requirements). And the higher the feeding demand degree, the higher the feeding priority of the corresponding area.

[0157] According to an embodiment of the present invention, based on the feeding demand degree, feeding priority analysis is performed and the first area and the second area are marked. Based on the initial position information of the unmanned fleet, with the first area as the feeding demand point, the shortest path planning of the unmanned fleet is carried out through the Floyd path algorithm, and the shortest navigation path of each unmanned ship is obtained. Specifically:

[0158] Perform feeding priority analysis according to the feeding demand degree. The analysis process is based on the judgment and comparison of the first threshold and the second threshold. All monitored areas are marked to obtain the first area, the second area, and the third area;

[0159] Obtain the initial position information of each unmanned ship in the unmanned ship fleet;

[0160] Based on the initial position information, with the initial position of the unmanned ship as the origin, all first areas covered within the preset radius based on the origin are marked as feeding demand points;

[0161] Take the initial position of the unmanned ship as the starting point and the ending point, and take the feeding demand points as the passing points. Based on the Floyd path algorithm, perform the shortest path planning for the unmanned ship to obtain multiple initial paths of the unmanned ship;

[0162] Based on the multiple initial paths, calculate and analyze the number of second areas passed by each path, and mark the path with the largest number of second areas passed as the shortest navigation path of the unmanned ship;

[0163] Analyze all unmanned ships in the fleet to obtain the shortest navigation path of each unmanned ship.

[0164] It should be noted that in the marking of all monitored areas based on the judgment and comparison of the first threshold and the second threshold to obtain the first area, the second area, and the third area, the condition for marking the first area is that the feeding demand degree of the corresponding monitored area is greater than the first threshold, the condition for marking the second area is that the feeding demand degree of the corresponding monitored area is less than or equal to the first threshold and greater than or equal to the second threshold, and the condition for marking the third area is that the feeding demand degree of the corresponding monitored area is less than the third threshold. The first threshold is greater than the second threshold. The feeding priorities of the first, second, and third areas decrease gradually. The first area is the necessary feeding area and needs to be analyzed as a feeding demand point in subsequent route planning. In setting the starting point and the ending point, based on the analysis requirements of the unmanned ship, the ending point can be not set, and the path planning is carried out with the purpose of completing the passing area. The multiple initial paths are obtained by screening from multiple legal paths planned by the algorithm. The screening principle is to select the shortest several paths. Each initial path passes through the corresponding feeding demand point.

[0165] In the marking of all first areas covered within the preset radius based on the origin as feeding demand points, that is, take the middle position point of the first area as the feeding demand point, and the feeding demand point corresponds to the first area one by one.

[0166] In the obtaining of the initial position information of each unmanned ship in the unmanned ship fleet, the initial positions of the unmanned ship fleet are generally scattered. In this embodiment, by analyzing the position range of each scattered unmanned ship, the feeding points are allocated, so as to maximize the feeding efficiency of the unmanned ship fleet.

[0167] According to an embodiment of the present invention, for the navigation control of an unmanned fleet based on the shortest navigation path, the path of the unmanned ship is obtained in real time, and the behavior warning of the ship is judged according to the real-time path. If there is an abnormality in the unmanned ship, a warning message is generated, and a secondary route planning is performed on the unmanned fleet through the warning message to generate optimized path information, and the optimized path information is sent to a preset unmanned ship terminal, including:

[0168] Generate a fleet navigation plan based on the shortest navigation path of each unmanned ship, and perform the navigation control of the unmanned fleet according to the fleet navigation plan;

[0169] Based on the navigation process, obtain the position information of the current unmanned ship in real time and obtain the real-time path;

[0170] Visualize the shortest path corresponding to the current unmanned ship to form a path image, extract the contour features of the path image, and obtain the first feature data;

[0171] Visualize and extract features from the real-time path to obtain second feature data;

[0172] Based on the standard Euclidean distance, calculate the similarity between the first feature data and the second feature data, and judge whether the similarity is greater than a preset similarity threshold. If so, judge that the navigation behavior of the current unmanned ship is abnormal, and record the position point of the current unmanned ship;

[0173] Obtain the navigation state information of the current unmanned ship in real time, and integrate the navigation state information with the position point of the current unmanned ship to form a warning message.

[0174] It is worth mentioning that in the present invention, feature extraction is performed in the form of a path image, and route deviation judgment is performed in the form of feature comparison, so as to obtain a judgment result consistent with the actual situation. Since the shortest navigation path is an ideal path, there will be a certain deviation in reality, and there will be a certain reasonable deviation in actual navigation. Within the reasonable deviation range, the navigation task of the unmanned fleet is not affected. Through the contour feature extraction judgment of the present invention, the reasonable deviation can be regulated based on the preset similarity threshold. The higher the similarity threshold is set, the higher the range of the reasonable deviation becomes, so it can be set according to the actual situation to adapt to different unmanned fleets and different water area environments. In the traditional technology, simple route comparison is often used, with a large error, a low controllable range of the judgment standard, and the judgment result is prone to be inconsistent with the actual navigation. The process of visualizing and extracting features from the real-time path is the same as the analysis process of the first feature data.

[0175] According to an embodiment of the present invention, for the navigation control of an unmanned fleet based on the shortest navigation path, the path of the unmanned ship is obtained in real time, and the behavior warning of the ship is judged according to the real-time path. If there is an abnormality in the unmanned ship, a warning message is generated, and a secondary route planning is carried out for the unmanned fleet through the warning message to generate optimized path information, and the optimized path information is sent to a preset unmanned ship terminal, specifically as follows:

[0176] Obtain the navigation status information in the warning message for abnormal assessment and generate corresponding emergency control instructions;

[0177] Obtain the position point of the current unmanned ship in the warning message, and obtain other unmanned ships within a preset range of the position point of the current unmanned ship through the water area map model, and mark them as neighboring unmanned ships;

[0178] Statistically integrate the feeding demand points of the current unmanned ship and the feeding demand points of neighboring unmanned ships to obtain a set of feeding demand points;

[0179] Based on the real-time position of the neighboring unmanned ship as the starting point and the position points in the set of feeding demand points as the passing points, perform a secondary shortest path planning for the neighboring unmanned ships through the Floyd path algorithm and generate optimized path information.

[0180] It should be noted that the navigation status information includes the navigation status of the unmanned ship, fuel situation, communication signal status, etc. Due to the complexity of the water area environment and the uncertainty of the navigation of the unmanned fleet, there may be some unexpected situations during large-scale unmanned fleet navigation tasks, such as deviation of the route, inconsistent behavior characteristics with expectations, etc., and the reasons for the abnormal route may be insufficient fuel, GPS signal problems, water area route congestion, etc. Therefore, corresponding emergency controls need to be made, such as staying on standby in place, returning, etc.

[0181] The optimized path information is analyzed for neighboring unmanned ships. There is at least one neighboring unmanned ship.

[0182] In the present invention, by marking the abnormal unmanned ship, integrating the remaining passing feeding points of the current unmanned ship with the remaining passing feeding points of neighboring ships, and performing secondary planning based on the path algorithm, new path information is generated in real time and sent to neighboring ships, thereby effectively realizing real-time warning regulation in case of emergencies in the case of large-scale fleet operation, greatly improving the environmental adaptability of the unmanned fleet, reducing manual intervention, improving the informatization level of the unmanned fleet, and helping to realize an unmanned fleet system with full automation, high warning ability, high safety, and high self-processing ability. Compared with the prior art, the present invention can realize dynamic route regulation of the unmanned fleet, and can also meet the feeding requirements of the unmanned fleet and plan the optimal path.

[0183] The third aspect of the present invention further provides a computer-readable storage medium, which includes an intelligent operation program for an unmanned fleet based on artificial intelligence. When the intelligent operation program for the unmanned fleet based on artificial intelligence is executed by a processor, the steps of the intelligent operation method for the unmanned fleet based on artificial intelligence as described in any one of the above are implemented.

[0184] The present invention discloses an intelligent operation method, system and storage medium for an unmanned fleet based on artificial intelligence. By collecting water area environment data and fishery resource data in a monitoring area, performing regional clustering and grouping of a preset water area, analyzing the water area feeding requirements based on the clustered area groups, and obtaining the feeding requirement degree of each monitoring area; analyzing the feeding priority based on the feeding requirement degree and marking the first area and the second area, and taking the first area as the feeding requirement point based on the initial position information of the unmanned fleet, performing the shortest path planning for the unmanned fleet through the Floyd path algorithm to obtain the shortest navigation path; performing ship behavior early warning judgment and secondary route planning of the unmanned fleet according to the real-time path, generating optimized path information, and sending the optimized path information to a preset unmanned ship terminal. Through the present invention, it is helpful to realize an unmanned fleet system with full automation, high early warning ability, high safety and high self-processing ability.

[0185] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling or communication connection between the components shown or discussed with each other can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0186] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units; they can be located in one place or distributed to multiple network units; some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0187] In addition, each functional unit in the embodiments of the present invention can be all integrated in one processing unit, or each unit can be separately used as a unit, or two or more units can be integrated in one unit; the above integrated units can be implemented in the form of hardware, or in the form of a hardware plus software functional unit.

[0188] Those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions. The foregoing program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps including those of the above method embodiments. The foregoing storage medium includes: various media that can store program codes, such as removable storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0189] Alternatively, if the above integrated units of the present invention are implemented in the form of software function modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the embodiments of the present invention, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in the various embodiments of the present invention. The foregoing storage medium includes: various media that can store program codes, such as removable storage devices, ROM, RAM, magnetic disks, or optical discs.

[0190] As described above, the above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. An artificial intelligence-based unmanned fleet intelligent operation method, characterized in that: include: Acquire target water area range information, and construct a water area map model based on the target water area range information; Based on the preset water area monitoring points, the water area is divided through the water area map model to form multiple monitoring areas; Collecting water environment data and fishery resource data in the monitoring area, performing regional cluster analysis based on the water environment data and fishery resource data, performing water feeding demand analysis based on the clustered regional groups, and obtaining the feeding demand of each monitoring area; Based on the feeding demand, a feeding priority analysis is performed and a first area and a second area are marked. Based on the initial position information of the unmanned fleet, the first area is used as the feeding demand point, and the shortest path planning is performed on the unmanned fleet through the Floyd path algorithm to obtain the shortest navigation path for each unmanned ship; The navigation control of the unmanned fleet is carried out based on the shortest navigation path, the path of the unmanned ship is obtained in real time, and the ship behavior warning is judged according to the real-time path. If there is any abnormality of the unmanned ship, a warning message is generated, and the unmanned fleet is secondary route planned based on the warning information to generate optimized path information, which is sent to the preset unmanned ship terminal.

2. The method for intelligent operation of an unmanned fleet based on artificial intelligence according to claim 1, characterized in that: The acquiring of target water area range information and the building of a water area map model based on the target water area range information are specifically as follows: Obtain information on the scope of the target water area; The target water area information includes water area, water map outline, and preset water monitoring point information; A water area map model based on three-dimensional visualization is constructed based on the target water area range information.

3. The method for intelligent operation of an unmanned fleet based on artificial intelligence according to claim 1, characterized in that: Based on the preset water area monitoring points, the water area is divided through the water area map model to form multiple monitoring areas, specifically: Through the water area map model, the entire target water area is divided into multiple monitoring areas based on the preset water area monitoring points; One monitoring area corresponds to one water area monitoring point.

4. The method for intelligent operation of an unmanned fleet based on artificial intelligence according to claim 1, characterized in that: The water environment data and fishery resource data in the monitoring area are collected, regional cluster analysis is performed based on the water environment data and fishery resource data, water feeding demand analysis is performed based on the clustered regional groups, and the feeding demand of each monitoring area is obtained, including: In one analysis cycle, the corresponding water environment data and underwater image data are collected for each monitoring area; The water area environment data include water temperature, flow velocity, flow direction, and water quality data; Performing fishery resource image recognition based on an image recognition model based on the underwater image data, and performing resource statistics based on the recognition results to obtain fishery resource data for each monitoring area; Clarify, standardize and integrate the water environment data and fishery resource data of each monitoring area to form monitoring integrated data; The monitoring integrated data of all monitoring areas are used as cluster sample data, and clustering grouping based on spectral clustering algorithm is performed to form multiple regional groups; Each area group includes at least one monitoring area.

5. The method for intelligent operation of an unmanned fleet based on artificial intelligence according to claim 4, characterized in that: The water environment data and fishery resource data in the monitoring area are collected, regional cluster analysis is performed based on the water environment data and fishery resource data, water feeding demand analysis is performed based on the clustered regional groups, and the feeding demand of each monitoring area is obtained, specifically: Select all monitoring areas in a region group and mark all monitoring areas as current monitoring areas; Integrate the fishery resource data in all currently monitored areas to form regional group fishery resource data; Calculating and analyzing the feeding requirement of the regional group according to the fishery resource data of the regional group and the fishery breeding demand; Taking the feeding requirement as the feeding requirement of all monitoring areas in one area group; The remaining area groups were analyzed and the feeding requirements of all monitored areas were obtained.

6. The method for intelligent operation of an unmanned fleet based on artificial intelligence according to claim 5, characterized in that: The feeding priority analysis is performed based on the feeding demand and the first area and the second area are marked. Based on the initial position information of the unmanned fleet, the first area is used as the feeding demand point, and the shortest path planning is performed on the unmanned fleet through the Floyd path algorithm to obtain the shortest navigation path of each unmanned ship, which is specifically: Performing a feeding priority analysis according to the feeding demand, wherein the analysis process is based on a comparison between a first threshold and a second threshold, marking all monitoring areas to obtain a first area, a second area, and a third area; Obtain the initial position information of each unmanned ship in the unmanned fleet; Based on the initial position information, taking the initial position of the unmanned boat as the origin, all first areas covered within a preset radius based on the origin are marked as feeding demand points; The initial position of the unmanned boat is taken as the starting point and the end point, and the feeding demand point is taken as the passing point. The shortest path of the unmanned boat is planned based on the Floyd path algorithm to obtain multiple initial paths of the unmanned boat. Calculating and analyzing the number of second areas passed by each path based on the multiple initial paths, and marking the path passing the largest number of second areas as the shortest navigation path for the unmanned ship; Analyze all the unmanned ships in the fleet and obtain the shortest navigation path for each unmanned ship.

7. The method for intelligent operation of an unmanned fleet based on artificial intelligence according to claim 6, characterized in that: The method of controlling the navigation of the unmanned fleet based on the shortest navigation path, obtaining the path of the unmanned ship in real time, making early warning judgments on the behavior of the ship based on the real-time path, generating early warning information if there is an abnormality in the unmanned ship, performing secondary route planning on the unmanned fleet based on the early warning information, generating optimized path information, and sending the optimized path information to a preset unmanned ship terminal includes: Generate a fleet navigation plan based on the shortest navigation path of each unmanned ship, and perform navigation control of the unmanned fleet according to the fleet navigation plan; Based on the navigation process, the current position information of the unmanned ship is obtained in real time, and the real-time path is obtained; The shortest path corresponding to the current unmanned ship is visualized to form a path image, and contour features are extracted from the path image to obtain first feature data; Imaging and feature extracting the real-time path to obtain second feature data; Based on the standard Euclidean distance, the similarity between the first feature data and the second feature data is calculated to determine whether the similarity is greater than a preset similarity threshold. If so, it is determined that the navigation behavior of the current unmanned ship is abnormal, and the current position of the unmanned ship is recorded; The navigation status information of the current unmanned ship is obtained in real time, and the navigation status information is integrated with the current position point of the unmanned ship to form early warning information.

8. The method for intelligent operation of an unmanned fleet based on artificial intelligence according to claim 7, characterized in that: The navigation control of the unmanned fleet is performed based on the shortest navigation path, the path of the unmanned ship is obtained in real time, and the ship behavior warning judgment is performed according to the real-time path. If there is an abnormal unmanned ship, a warning information is generated, and the unmanned fleet is secondary route planned based on the warning information to generate optimized path information, and the optimized path information is sent to the preset unmanned ship terminal, specifically: Obtain the navigation status information in the warning information to conduct abnormality assessment and generate corresponding emergency control instructions; Obtain the location of the current unmanned ship in the warning information, obtain other unmanned ships within a preset range of the location of the current unmanned ship through the water map model, and mark them as neighboring unmanned ships; The feeding demand points of the current unmanned boat are statistically integrated with the feeding demand points of the neighboring unmanned boats to obtain a set of feeding demand points; Based on the real-time position of the neighboring unmanned boat as the starting point and the location point in the feeding demand point set as the passing point, the Floyd path algorithm is used to perform secondary shortest path planning for the neighboring unmanned boat and generate optimized path information.

9. An unmanned fleet intelligent operation system based on artificial intelligence, characterized in that: The system includes: a memory and a processor, wherein the memory includes an unmanned fleet intelligent operation program based on artificial intelligence, and when the unmanned fleet intelligent operation program based on artificial intelligence is executed by the processor, the steps of the unmanned fleet intelligent operation method based on artificial intelligence as described in any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes an unmanned fleet intelligent operation program based on artificial intelligence. When the unmanned fleet intelligent operation program based on artificial intelligence is executed by the processor, the steps of the unmanned fleet intelligent operation method based on artificial intelligence as described in any one of claims 1 to 8 are implemented.

Images