A Smart Anchorage Management Method and System Based on Multi-Source Heterogeneous Data

Abstract

This invention discloses a smart anchorage management method and system based on multi-source heterogeneous data. Through the collaborative efforts of three types of acquisition terminals—airborne, shore-based, and shipborne—it comprehensively collects four categories of multi-source data—anchorage images, shore-based management images, ship parameters, and radar monitoring—in three stages: pre-event preparation, in-event management, and post-event verification, achieving full-dimensional coverage of the environment, management, and ships. Based on a comprehensive feature dataset, it sets quantitative standards for safety and size constraints, and uses a particle swarm optimization algorithm to iteratively select the optimal mooring position, avoiding the shortcomings of traditional algorithms, balancing safety distance and ship size, and maximizing the utilization of anchorage space. A two-way interactive link is constructed among the three types of terminals to achieve real-time data transmission and command closed-loop. By calculating the three-dimensional deviation factors of path node position, heading, and speed, it accurately generates adjustment commands and optimizes parameters, ensuring smooth and accurate deviation correction and guaranteeing efficient and orderly ship mooring and navigation management.

Description

Technical Field

[0001] This invention relates to the field of ship navigation technology, and in particular to a smart anchorage management method and system based on multi-source heterogeneous data. Background Technology

[0002] With the rapid development of smart ports and smart shipping industries, anchorages, as core hubs for ship navigation, berthing, and waiting, directly impact port operational efficiency and shipping safety in terms of management efficiency, safety, and accuracy. During anchorage management, the status monitoring and deviation correction of ships at various path nodes, as well as the data transmission and command interaction between data acquisition terminals, shore-based management platforms, and shipboard control terminals, are crucial for ensuring ships accurately enter virtual anchorages and safely berth. Traditional anchorage management models are no longer suitable for the trends of larger ships, more complex anchorage environments, and more refined management needs. There is an urgent need for technological innovation to build efficient interaction mechanisms and precise deviation monitoring and correction systems to overcome traditional management bottlenecks and promote the transformation of anchorage management towards intelligence and smart technology.

[0003] Existing technologies, upon detecting ship navigation deviations, cannot accurately generate corresponding adjustment commands based on the type of deviation, and lack scientific calculation formulas to optimize adjustment parameters. This can easily lead to problems such as over-adjustment, untimely adjustment, or insufficient adjustment force, making it impossible to achieve smooth and accurate correction of deviations. This increases safety risks such as loss of attitude control and deviation from the path during ship navigation, and affects the accurate docking of ships.

[0004] Meanwhile, the data interaction, deviation monitoring, command generation and execution are all independent of each other and lack a collaborative mechanism. This makes it impossible to achieve closed-loop management of the entire process of data collection, deviation calculation, command generation, execution feedback and re-monitoring. As a result, navigation deviations cannot be detected and corrected in a timely manner, which further exacerbates the safety hazards of anchorage management. Summary of the Invention

[0005] To achieve the above objectives, one of the technical solutions adopted by the present invention is: a smart anchorage management method based on multi-source heterogeneous data, the method comprising:

[0006] We collect multi-source datasets of anchorage areas, berthing port management, and the vessels themselves using a data collection strategy, and extract the corresponding effective features using an appropriate feature extraction algorithm. A weighted fusion strategy is used to weight each effective feature to obtain a comprehensive feature dataset; The optimal berthing position of the ship is determined by processing the comprehensive feature dataset through optimization strategies, and a target virtual anchorage area is generated. The optimal navigation path for a ship from entering the anchorage to the target virtual anchorage area is determined using a path planning algorithm. Based on the node setting strategy, set the navigation constraints for each path node on the optimal path; When a ship is traveling along the optimal navigation path, multi-source datasets are collected in real time at each path node through interactive and acquisition strategies. The comprehensive deviation factor between the ship's current navigation position and the optimal navigation path is calculated. When the comprehensive deviation factor is greater than the comprehensive deviation threshold at the current node, a corresponding adjustment command is generated to enable the ship to enter the virtual anchorage area and complete berthing according to the optimal navigation path.

[0007] Furthermore, the data collection strategy involves collecting data from the anchorage area, port control area, and the vessel itself via a data collection terminal according to a preset data collection sequence. This data collection strategy provides data support for determining the optimal berthing position of the vessel, generating the target virtual anchorage area, planning the optimal navigation path, and dynamically adjusting the vessel. The data acquisition terminals include aerial acquisition terminals, shore-based acquisition terminals, and shipborne acquisition terminals; The multi-source dataset includes anchorage area image data, port management image data, ship hull parameter information data, and radar monitoring data; The data acquisition sequence includes collecting overall image data of the anchorage area, image data of the anchorage entrance, and initial dynamic data within the anchorage when the ship sails to a preset distance before the anchorage entrance. The shipborne acquisition terminal collects basic ship parameters to ensure that the system obtains the initial environment and basic ship data of the anchorage, providing initial support for the planning of the optimal berthing position and the optimal navigation path. From the moment a vessel enters the anchorage entrance until the vessel completes berthing and comes to a stop after a preset time, the aerial acquisition terminal, shore-based acquisition terminal, and shipborne acquisition terminal collect multi-source datasets in real time at a preset frequency, directly supporting path guidance, comprehensive deviation factor calculation, and adjustment command generation. After the vessel completes its berthing, the aerial, shore-based, and shipborne data acquisition terminals continue to operate for a preset duration to collect additional data on the vessel's berthing posture and the degree of fit between the vessel and the virtual anchorage area. This data is used to verify the standardization of the vessel's berthing and the rationality of the virtual anchorage area planning.

[0008] Furthermore, the overall geographical boundary features of the anchorage, the location and size features of obstacles within the anchorage, the outline and location features of existing moored vessels within the anchorage, and the correlation features of anchorage water depth distribution are extracted from the image data of the anchorage area, providing a basis for obstacle avoidance planning of the optimal mooring position for vessels and analysis of anchorage space utilization. The boundary features of anchorage entrances / exits, the passage posture features of ships entering and leaving anchorages, and the environmental features of shore-controlled areas are extracted from the image data of anchorage port management. The image features related to ship heading and speed are extracted through the ship passage posture features to provide a basis for the path guidance and dynamic adjustment of ships entering and leaving anchorages. Extract ship size characteristics, ship type characteristics, and ship attitude correlation characteristics from ship hull parameter information data to provide a basis for size planning and path constraint setting of target virtual anchorage area; Extracting real-time ship position coordinates, speed, direction, and turning angle characteristics from radar monitoring data, as well as the real-time dynamic characteristics of obstacles and other ships within the anchorage, provides core data support for optimal navigation path planning, comprehensive deviation factor calculation, and adjustment command generation.

[0009] Furthermore, the weighted fusion strategy includes determining, based on the importance and data accuracy of each effective feature, that the weight coefficient of effective features of radar monitoring is greater than the weight coefficient of effective features of anchorage area images, which is greater than the weight coefficient of effective features of ship hull parameters, which is greater than the weight coefficient of effective features of port management images, and the sum of all weight coefficients is 1. Each valid feature is standardized to eliminate the dimensional differences between different types of features, ensuring the rationality of weighted fusion calculation, and the values ​​of all valid features are mapped to the [0,1] interval through standardization. A linear weighted summation algorithm is used to multiply each standardized effective feature by its corresponding weight coefficient and then sum them to obtain a comprehensive feature dataset.

[0010] Furthermore, the optimization strategy includes: A safety constraint is defined as ensuring that the distance between the optimal mooring position of a vessel and obstacles or other existing moored vessels within the anchorage is no less than a preset safety threshold. Define the length of the virtual anchorage area as a preset multiple of the ship's length and the width as a preset multiple of the ship's width as dimensional constraints; Each potential ship berthing position is iterated using particle swarm optimization with safety and size constraints. After the iteration is completed, the particle position corresponding to the global optimal fitness value is determined as the optimal mooring position of the ship. Combining the coordinate characteristics of the mooring position, the ship size characteristics, and the anchorage environment characteristics, a target virtual anchorage area is generated.

[0011] Furthermore, the node setting strategy includes: Based on the total length of the optimal navigation path, the complexity of the anchorage environment, and the preset node density, the optimal navigation path is divided into anchorage entrance or exit areas, obstacle-dense areas, and path segments close to the target virtual anchorage area. Each path segment filters path nodes according to point selection rules. Path nodes include anchorage entrance starting point, obstacle avoidance turning point, path curvature change point, starting point near virtual anchorage area, and optimal navigation path ending point. Based on the environment of the path segment where the node is located and the navigation requirements of the ship, the corresponding navigation constraints are matched for each set path node. The constraints include speed constraints, turning angle constraints and navigation direction constraints.

[0012] Furthermore, the interaction strategy includes: Establish a two-way interactive link between the data acquisition terminal, the shore-based management and control platform, and the shipborne control terminal; The data acquisition terminal is triggered by an interactive strategy to collect multi-source datasets in real time. The data acquisition terminal transmits the collected data on the ship's position, navigation attitude and anchorage environment to the shore-based control platform through a two-way interactive link for the calculation of the comprehensive deviation factor. The shore-based control platform transmits the calculated comprehensive deviation factor, deviation judgment result and corresponding adjustment instructions to the shipborne control terminal in real time through the interactive link.

[0013] Furthermore, the adjustment commands include: heading adjustment commands, speed adjustment commands, and comprehensive adjustment commands; When a ship deviates from the optimal path, a course adjustment command is generated. The course adjustment command includes adjusting the ship's turning angle and turning speed to guide the ship to gradually correct its course and return to the optimal path. The turning angle is adjusted according to the magnitude of the comprehensive deviation factor. The larger the comprehensive deviation factor, the closer the turning angle is to the upper limit of the constraint threshold. When a ship's speed is too fast or too slow, causing it to deviate from the optimal path, or when its speed does not meet the speed constraints of the current node, a speed adjustment command is generated. The heading adjustment command includes adjusting the ship's target speed and acceleration / deceleration rate to avoid exacerbating the deviation due to abnormal speed, while ensuring the stability of the ship's navigation. When a ship has deviations in both course and speed, or when the combined deviation factor exceeds the combined deviation threshold, a combined adjustment command is generated. The combined adjustment command first adjusts the course, then the speed, and simultaneously limits the turning angle and target speed to ensure that the ship returns to the optimal navigation path and avoids navigation risks.

[0014] Furthermore, at each path node, the shipborne control terminal feeds back the ship's current navigation status to the shore-based management platform through an interactive link. The shore-based management platform, in conjunction with the navigation constraints preset at that node, performs real-time verification of the ship's current navigation status. If it finds that the ship's navigation status violates the constraints, it generates an early warning command through an interactive strategy and sends it to the shipborne control terminal simultaneously.

[0015] Another technical solution adopted by the present invention is as follows: The system is used in the above-mentioned intelligent anchorage management method based on multi-source heterogeneous data. The system includes: a data acquisition module, which is used to acquire multi-source datasets of anchorage area, port of call management and the ship itself. The data acquisition module includes an aerial acquisition unit, a shore-based acquisition unit and a ship-borne acquisition unit, which correspond to the aerial acquisition end, the shore-based acquisition end and the ship-borne acquisition end, respectively. It can complete the initial acquisition, real-time acquisition and supplementary acquisition of multi-source datasets according to a preset acquisition sequence and preset frequency, so as to provide data support for subsequent feature extraction, deviation calculation and other links. The feature extraction module is used to receive multi-source datasets transmitted by the acquisition module and extract corresponding effective features through an adapted feature extraction algorithm, including effective features of anchorage area images, effective features of port management images, effective features of ship parameters, and effective features of radar monitoring. The weighted fusion module receives various effective features transmitted by the feature extraction module. The weighted fusion module is equipped with a weighted fusion strategy to complete the standardization processing and linear weighted summation calculation of effective features, generate a comprehensive feature dataset, eliminate the heterogeneity of multi-source features, and provide data support for the implementation of optimization strategies. The optimization module receives the comprehensive feature dataset transmitted by the weighted fusion module. The optimization module is equipped with an optimization strategy, which uses the particle swarm optimization algorithm combined with safety constraints and size constraints to iteratively select the optimal mooring position for the ship and generate the target virtual anchorage area. The path planning module receives relevant features of the target virtual anchorage area transmitted by the optimization module, and determines the optimal navigation path for the ship from the anchorage entrance to the target virtual anchorage area through the path planning algorithm, providing a basis for node setting and path guidance. The node setting module is used to receive the optimal navigation path transmitted by the path planning module. The node setting module is equipped with a node setting strategy, which divides the path segments, filters the path nodes, and matches the corresponding navigation constraints for each path node. The interaction module is used to realize two-way interaction between the data acquisition module, the shore-based control module, and the shipborne control module. The interaction module is equipped with an interaction strategy to build a two-way interaction link and complete the real-time transmission of multi-source datasets, comprehensive deviation factors, adjustment commands, early warning commands, and ship navigation status data. The deviation calculation and command generation module is used to calculate the comprehensive deviation factor between the ship's current navigation position and the optimal navigation path when the ship navigates to each path node, based on the multi-source dataset collected in real time by the acquisition module. When the comprehensive deviation factor is greater than the comprehensive deviation threshold of the current node, the module generates the corresponding heading adjustment command, speed adjustment command, or comprehensive adjustment command according to the type of adjustment command carried. At the same time, it works with the interaction module to complete the real-time verification of the ship's navigation status and generate a warning command when the ship violates the navigation constraints of the node. The shore-based control module, as the core control unit of the system, is used to receive data and parameters transmitted by each module, coordinate the operation of each module, and complete core control functions such as comprehensive deviation factor calculation, adjustment command and early warning command generation, and navigation status verification. It corresponds to the shore-based control platform in the interaction strategy. The shipboard control module, installed on the ship, is used to receive adjustment commands, early warning commands, optimal navigation paths, node constraints and other information transmitted by the interaction module, to provide feedback on the ship's current navigation status, and to trigger the ship control system to perform adjustment operations. It corresponds to the shipboard control terminal in the interaction strategy. The data storage module is used to retain multi-source datasets collected by the acquisition module, effective features extracted by the feature extraction module, comprehensive feature datasets generated by the weighted fusion module, operating parameters of each module, adjustment instructions, early warning instructions, ship navigation status data and interactive data, providing data support for subsequent data traceability, algorithm optimization and docking standardization verification.

[0016] Compared with existing technologies, the present invention has the following advantages: by coordinating three types of acquisition terminals—airborne, shore-based, and shipborne—and following a three-stage acquisition sequence of pre-event preparation, in-event control, and post-event verification, it comprehensively collects four major categories of multi-source datasets: anchorage area images, port management images, ship parameters, and radar monitoring data, thereby achieving full-dimensional data coverage of the environment, control, and ships.

[0017] By optimizing the strategy and using a comprehensive feature dataset as a foundation, quantitative standards for safety and size constraints are set. Combined with particle swarm optimization, potential mooring locations within the anchorage are iteratively screened. This effectively avoids the shortcomings of traditional algorithms, such as slow convergence and susceptibility to local optima. The selected optimal mooring locations not only meet the safety distance requirements from obstacles and other vessels but also adapt to the vessel size, maximizing the utilization of anchorage space. By constructing a two-way interactive link between the data acquisition terminal, the shore-based control platform, and the shipborne control terminal through an interactive strategy, real-time data transmission and command closed-loop are achieved. When the ship sails to each path node, the comprehensive deviation factor of position, heading, and speed is calculated by combining real-time multi-source data to quantify the degree of deviation. Based on the deviation type, three types of adjustment commands—heading, speed, and comprehensive—are accurately generated. The adjustment parameters are optimized through scientific formulas to ensure smooth and accurate deviation correction and avoid over-adjustment or untimely adjustment. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the intelligent anchorage management method based on multi-source heterogeneous data of the present invention.

[0019] Figure 2 This is a connection diagram of the ship navigation safety analysis system based on dynamic data control according to the present invention.

[0020] Figure 3 This is a satellite cloud image of a certain anchorage area.

[0021] Figure 4 This is a schematic diagram of the route nodes and target virtual anchorage areas.

[0022] Figure 5 To pass The ship planning result diagram obtained by the algorithm; Figure 6 A curve showing the planned arrival speed.

[0023] Figure 7 A curve for calculating the berthing speed. Detailed Implementation

[0024] The technical solutions of the intelligent anchorage management method and system based on multi-source heterogeneous data provided by the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0025] Example 1 like Figures 1-7 As shown, the intelligent anchorage management method based on multi-source heterogeneous data includes: collecting multi-source datasets of anchorage area, mooring port management and the vessel itself through a collection strategy, and extracting corresponding effective features through an adapted feature extraction algorithm.

[0026] Furthermore, the data collection strategy involves collecting data from the anchorage area, port control area, and the vessel itself via a data collection terminal according to a preset data collection sequence. This strategy provides data support for determining the optimal berthing position of the vessel, generating the target virtual anchorage area, planning the optimal navigation path, and dynamically adjusting the vessel.

[0027] The data acquisition terminals include aerial acquisition terminals, shore-based acquisition terminals, and shipborne acquisition terminals.

[0028] The multi-source dataset includes anchorage area image data, port management image data, ship hull parameter information data, and radar monitoring data.

[0029] The data collection sequence includes: when the vessel sails to a preset distance before anchorage, the aerial and shore-based acquisition terminals collect overall image data of the anchorage area, anchorage boundary image data, and initial dynamic data within the anchorage; the shipborne acquisition terminal collects basic hull parameters, ensuring the system obtains initial anchorage environment and basic vessel data, providing initial support for optimal berthing location planning and optimal navigation path planning; from the moment the vessel enters the anchorage until a preset time has elapsed after the vessel has completed berthing and come to a stop, the aerial, shore-based, and shipborne acquisition terminals collect multi-source datasets in real time at a preset frequency, directly supporting path guidance, comprehensive deviation factor calculation, and adjustment command generation; after the vessel completes berthing, the aerial, shore-based, and shipborne acquisition terminals continue to run for a preset time to supplement the collection of data on the vessel's berthing attitude and the vessel's fit with the virtual anchorage area, used to verify the standardization of vessel berthing and the rationality of the virtual anchorage area planning.

[0030] In this embodiment, the multi-source dataset covers four major categories of data: anchorage images, port management images, ship parameters, and radar monitoring data. This achieves full-dimensional data coverage of the environment, control, and ships. The various types of data complement each other, which can meet the needs of different control links and solve the pain point of traditional control data being singular. It provides a comprehensive data foundation for accurate feature extraction and support for subsequent control.

[0031] The three-stage data collection sequence forms a closed-loop process of pre-event preparation, in-event control, and post-event verification, ensuring that the entire control process is supported by data: Pre-event data collection provides advance protection for berthing location and route planning, avoiding initial environmental risks; in-event real-time data collection directly supports route guidance, deviation calculation, and adjustment command generation, ensuring the accuracy of ship navigation; and post-event supplementary data collection verifies the standardization of berthing and the rationality of virtual anchorage planning, promptly identifies deficiencies, optimizes control, and improves the accuracy and standardization of control.

[0032] Furthermore, the overall geographical boundary features of the anchorage, the location and size features of obstacles within the anchorage, the outline and location features of existing moored vessels within the anchorage, and the correlation features of anchorage water depth distribution are extracted from the anchorage area image data, providing a basis for obstacle avoidance planning for optimal vessel mooring positions and anchorage space utilization analysis.

[0033] Specifically, from visible light images, infrared images, and laser point cloud data of the anchorage area, the YOLO target detection algorithm and the laser point cloud segmentation algorithm are used to extract the overall geographical boundary features of the anchorage, the location and size features of obstacles within the anchorage, the outline and location features of existing moored vessels within the anchorage, and the correlation features of the anchorage water depth distribution.

[0034] Among them, the calculation formulas for extracting the size features of obstacles within the anchorage based on laser point cloud data are as follows: , , in, The length of the obstacle, i.e., the dimension of the obstacle along its longest axis; The width of the obstacle is the dimension of the obstacle along its shortest axis. For obstacle laser point cloud in Maximum and minimum coordinates along the axis; For obstacle laser point cloud in Maximum and minimum coordinates along the axis; Laser point cloud of the midpoint section of the obstacle in Maximum and minimum coordinates along the axis; Laser point cloud of the midpoint section of the obstacle in Maximum and minimum coordinates along the axis.

[0035] The correlation features of anchorage water depth distribution are extracted by combining laser point cloud data and water level sensor data. The calculation formula is as follows: , in, This represents the actual water depth at a certain point in the anchorage. The current water level height is collected by a shore-based water level sensor; For reference datum Axis coordinates; For the laser point cloud of a certain point in the anchorage Axis coordinates.

[0036] The image data of the port of call is used to extract the characteristics of ship traffic at the anchorage boundary, the characteristics of ship movement when entering and leaving the anchorage, and the environmental characteristics of the shore control area. The image features of ship heading and speed are extracted from the ship movement characteristics to provide a basis for the path guidance and dynamic adjustment of ships entering and leaving the anchorage.

[0037] Specifically, CNN convolutional neural networks and optical flow methods are used to extract features from the image data of the port of call management, and the frame rate of feature extraction is consistent with the frame rate of image acquisition.

[0038] Extracting ship heading angle based on ship attitude characteristics The calculation formula is as follows: , in, The ship's heading angle, with a range of values ​​of... With due north as Increasing clockwise; The coordinates of the ship's bow in the image coordinate system; The coordinates of the ship's stern in the image coordinate system.

[0039] Extracting ship speed using optical flow method The calculation formula is as follows: , in, For the speed of passage of ships; For ships in two adjacent frames Displacement in the axial direction; For ships in two adjacent frames Displacement in the axial direction; The time interval between two adjacent frames; This is the calibration coefficient between image pixels and actual distance.

[0040] The ship's size, type, and attitude characteristics are extracted from the ship's hull parameter information data to provide a basis for the size planning and path constraint setting of the target virtual anchorage area. The real-time position coordinates, speed, direction, and turning angle characteristics of the ship, as well as the real-time dynamic characteristics of obstacles and other ships in the anchorage are extracted from the radar monitoring data to provide core data support for optimal navigation path planning, comprehensive deviation factor calculation, and adjustment command generation.

[0041] Specifically, data preprocessing and feature quantization algorithms are used to provide features for the ship's hull parameter information data.

[0042] The formula for calculating ship attitude-related characteristics is: , , in, This represents the actual heel angle of the vessel; a positive value indicates that the vessel is heeling to the right, and a negative value indicates that the vessel is heeling to the left. This represents the actual trim angle of the ship. A positive value indicates that the bow of the ship is tilted upwards, while a negative value indicates that the bow of the ship is tilted downwards. The original tilt angle acquired by the attitude sensor; The original pitch angle acquired by the attitude sensor; These are the calibration values ​​for the roll and pitch angles, respectively.

[0043] The coding method is used to quantify the ship types, such as cargo ships, passenger ships, and fishing vessels, into numerical features. The specific coding rules are as follows: cargo ship = 1, passenger ship = 2, fishing ship = 3, other ships = 4.

[0044] In this embodiment, the specific feature extraction algorithm and quantization calculation formula achieve standardization and precision in feature extraction, avoiding errors from manual extraction. At the same time, the calculation logic of each feature is clarified, realizing comprehensive and accurate data collection and standardized and precise feature extraction. The feature calculation logic is clarified, the reliability of features is improved, and the orderly implementation of the entire control process is supported, forming a closed-loop control process.

[0045] The method also includes: weighting each effective feature using a weighted fusion strategy to obtain a comprehensive feature dataset.

[0046] Furthermore, the weighted fusion strategy includes determining, based on the importance and data accuracy of each effective feature, that the weight coefficient of effective features of radar monitoring is greater than the weight coefficient of effective features of anchorage area images, which is greater than the weight coefficient of effective features of ship hull parameters, which is greater than the weight coefficient of effective features of port management images, and the sum of all weight coefficients is 1. Each valid feature is standardized to eliminate dimensional differences between different feature types, ensuring the rationality of weighted fusion calculations. The standardization process then maps the values ​​of all valid features to... interval; A linear weighted summation algorithm is used to multiply each standardized effective feature by its corresponding weight coefficient and then sum them to obtain a comprehensive feature dataset.

[0047] Specifically, the effective feature weights of radar monitoring Effective feature weights of anchorage area images Effective feature weights of hull parameters Weighting of effective features in images for port of call management The weighting coefficients must satisfy the following constraint formula: , .

[0048] Since the dimensions of various effective features differ significantly, direct weighted fusion will lead to the feature with the larger dimension dominating the calculation result, affecting the fusion accuracy. Therefore, it is necessary to standardize each effective feature to eliminate the dimensional differences between different types of features and ensure the rationality of weighted fusion calculation.

[0049] Standardized processing adopts The standardization algorithm maps the values ​​of all valid features to a unified value. The interval preserves the relative differences of the features themselves while unifying the feature values. The standardized calculation formula is as follows: , in, For the first The standardized feature values ​​of the effective features, with a range of values ​​being: ; For the first The original extracted values ​​of each effective feature; For the first The minimum value of the effective features of a class; For the first The maximum value of the effective features of a class.

[0050] For each standardized valid feature, it is multiplied by its corresponding weight coefficient, and then all products are summed to obtain a comprehensive feature dataset. This dataset integrates the core information of various valid features, highlighting the value of core features such as radar monitoring, while also taking into account the role of other auxiliary features. The linear weighted summation formula for the comprehensive feature dataset is: , in, This is the core fusion value of the comprehensive feature dataset; The total number of valid features; For the first Weight coefficients corresponding to effective features of a class; For the first The standardized feature values ​​of the class's effective features.

[0051] In this embodiment, the weighted fusion strategy integrates four types of effective features—radar monitoring, anchorage area images, ship parameters, and port management images—into a comprehensive feature dataset through scientific weight allocation, standardization, and linear weighted summation. This effectively solves the pain points of dimensional differences and uneven feature importance in the multi-source feature fusion process, providing high-quality and highly reliable feature support for subsequent optimal berthing location optimization, path planning, and dynamic adjustment in smart anchorage management.

[0052] The method also includes: processing the comprehensive feature dataset through optimization strategies to determine the optimal mooring position of the ship and generating a target virtual anchorage area.

[0053] Furthermore, the optimization strategy includes defining a safety constraint that the distance between the ship's optimal mooring position and obstacles and other existing moored ships within the anchorage is not less than a preset safety threshold.

[0054] Specifically, safety constraints are designed to mitigate the risk of collisions during vessel berthing. These constraints are based on a quantitative analysis of the location and size features of obstacles within the anchorage, the location features of existing moored vessels, and the real-time coordinates of the vessel, extracted from a comprehensive feature dataset. This ensures that vessels maintain a safe distance from surrounding hazardous targets after berthing, fundamentally preventing safety hazards such as collisions and grounding.

[0055] Let the coordinates of the ship's mooring position be... The center coordinates of an obstacle within the anchorage are: The center coordinates of a certain existing moored vessel are The preset safety threshold is Then the quantitative formula for safety constraints is: , , in, The coordinates of the ship's berthing position are consistent with the coordinate system in the radar monitoring data; The center plane coordinates of an obstacle within the anchorage are determined based on the obstacle's location features extracted from laser point cloud data. The center plane coordinates of a certain existing moored vessel within the anchorage are determined by fusing image data and radar monitoring data of the anchorage area. The preset safety threshold has a range of values. It can be adaptively adjusted according to the ship size and the complexity of the anchorage environment. The larger the ship size and the denser the obstacles in the anchorage, the larger the value.

[0056] Define the virtual anchorage area as having a length that is no less than a preset multiple of the ship's length and a width that is no less than a preset multiple of the ship's width as dimensional constraints.

[0057] Specifically, the size constraint ensures that the virtual anchorage area can fully accommodate the vessel to be moored, while reserving reasonable operational space. This constraint is based on the quantification of hull size features extracted from the comprehensive feature dataset, ensuring that the virtual anchorage area is compatible with the vessel size and avoiding situations where the vessel cannot moor properly or its mooring posture is abnormal due to an excessively small anchorage area.

[0058] Let the length of the ship to be moored be... The width of the ship is The length of the virtual anchorage area is Width is The preset length multiple is The preset multiple of the ship's width is The quantification formula for size constraints is as follows: , , in, The actual length of the vessel to be moored is determined based on the vessel size characteristics extracted from the hull parameter information data; The actual beam of the vessel to be moored is determined based on the vessel size characteristics extracted from the hull parameter information data; The length of the target virtual anchorage area, along the direction of the ship's mooring; The width of the target virtual anchorage area, perpendicular to the direction in which the ship is moored; Set a preset multiple for the ship's length, with a range of values ​​as follows: ; A preset multiple for the ship's width is provided, with a range of values. .

[0059] Each potential ship berthing location is iterated using particle swarm optimization with safety and size constraints.

[0060] After the iteration is completed, the particle position corresponding to the global optimal fitness value is determined as the optimal mooring position of the ship. Combining the coordinate characteristics of the mooring position, the ship size characteristics, and the anchorage environment characteristics, a target virtual anchorage area is generated.

[0061] Specifically, the particle swarm optimization algorithm, combined with the aforementioned safety and size constraints, is used to iteratively screen all potential ship berthing locations within the anchorage. The core is to gradually converge to the globally optimal berthing location by simulating the motion behavior of the particle swarm, thus solving the problems of "slow convergence and easy getting trapped in local optima" in traditional optimization algorithms. This ensures that the selected berthing locations not only meet the constraints but also maximize the use of anchorage space and minimize navigation costs.

[0062] The fitness function is the core criterion for evaluating the quality of potential berthing locations. Combined with safety constraints, dimensional constraints, and anchorage space utilization efficiency, it ensures that the selected berthing locations not only meet the constraints but also maximize anchorage space utilization and minimize the distance the ship must travel to that location. The formula is: , in, For the ship from entering the anchorage to the The distance to each potential parking location is calculated based on path features in the comprehensive feature dataset; The available anchorage space area around potential berthing locations is extracted based on anchorage area image data; The required area of ​​the virtual anchorage area, i.e. ; For the first Minimum distance between potential berthing locations and surrounding obstacles and existing moored vessels; These are the navigation distance weighting coefficient, space utilization weighting coefficient, and safety distance weighting coefficient, respectively, satisfying... .

[0063] The particle velocity update formula is used to control the adjustment step size and direction of the particles, and its update formula is as follows: , in, For the first The particle in the first Speed ​​during the next iteration; For the first The particle in the first Speed ​​during the next iteration; The inertial weight has a value range of [value range missing]. It is used to adjust the inertia of particles and control the iteration speed; The learning factor has a value range of 1. It is used to adjust the speed at which particles converge to their individual optimal or global optimal positions; It is a random number, and its value range is [value range missing]. This is used to increase the randomness of the iteration process and avoid getting trapped in local optima; For the first The optimal position of an individual particle during the iteration process; This represents the globally optimal position during the entire particle swarm iteration process; For the first The particle in the first The position at the next iteration.

[0064] The particle position update formula is used to adjust the coordinates of potential mooring positions, and its update formula is as follows: , The iterative process is as follows: 1. Initialize the particle swarm: randomly generate Given several potential parking locations, set the number of iterations and inertia weights. Learning factors Parameters such as these.

[0065] 2. For each particle, determine whether it meets the safety constraints and size constraints. Particles that do not meet the constraints have a fitness value of 0. Particles that meet the constraints have a fitness value calculated according to the fitness function.

[0066] 3. Compare the current fitness value of each particle with its own historical best fitness value, and update the individual's optimal position. Compare the individual optimal fitness values ​​of all particles and update the global optimal position. .

[0067] 4. Update the velocity of each particle according to the above iterative formula. and location .

[0068] 5. If the preset number of iterations is reached, or the global optimal fitness value does not change significantly after a preset number of iterations, then stop the iteration; otherwise, return to step 2 and continue the iteration.

[0069] After the iteration is completed, the particle position corresponding to the global optimal fitness value is determined as the optimal mooring position of the ship. Combining the coordinate characteristics of the mooring position, the ship size characteristics, and the anchorage environment characteristics, a target virtual anchorage area is generated.

[0070] Specifically, combining the coordinate characteristics of the optimal mooring location, the ship's size characteristics, and the anchorage environment characteristics such as water depth and obstacle distribution, a target virtual anchorage area is generated according to the size constraint formula: centered on the optimal mooring location, extending along the ship's direction of travel. Extending perpendicular to the direction of navigation This creates a rectangular virtual anchorage area, and marks the coordinate range, length, width, and water depth requirements of this area. This provides a clear target benchmark for subsequent ship navigation path planning and dynamic adjustments, ensuring that ships can accurately and safely enter the virtual anchorage area and complete the berthing.

[0071] In this embodiment, the weighted fusion strategy and the optimal berthing position optimization strategy form a complete link of feature fusion, optimization screening and target generation. The former provides high-quality feature support for the latter, and the latter transforms the fused data into implementable management results. This effectively solves the pain points of chaotic multi-source features, inaccurate berthing position planning and high safety risks in smart anchorage management, and lays a solid foundation for subsequent path planning and dynamic adjustment. The two strategies work together.

[0072] The method also includes: determining the optimal navigation path for a ship from entering the anchorage to the target virtual anchorage area through a path planning algorithm.

[0073] Specifically, select As the core path planning algorithm, this algorithm has the advantages of high search efficiency, guaranteed path optimality, and easy integration with constraints. It can quickly converge to the optimal path and is also adaptable to dynamic anchorage environments.

[0074] The prerequisite for route planning is to build an anchorage environment model to quantify information such as navigable areas, risk areas, and anchorage boundaries within the anchorage. The environment modeling adopts the grid method, which divides the entire anchorage area into uniform square grids. Each grid corresponds to a state, and the grid size is set in combination with the ship size and the anchorage accuracy requirements.

[0075] The determination of grid status is based on a comprehensive feature dataset. Specifically, a navigable grid is defined as one free of obstacles and other moored vessels, with water depth meeting vessel draft requirements, and the grid center's distance from obstacles and other moored vessels exceeding a preset safety threshold. Among these, the ship's draft requirement is determined based on the anchorage water depth distribution correlation features in the comprehensive feature dataset. An unnavigable grid is defined as a grid containing obstacles, other moored vessels, or where the water depth is less than the vessel's draft, preventing passage. Obstacles are determined based on the location and size features of obstacles within the anchorage from the comprehensive feature dataset; other moored vessels are determined based on the outline and location features of existing moored vessels from the comprehensive feature dataset.

[0076] The risk warning grid is a navigable area within the grid itself, but the distance to obstacles and other moored vessels is within [a certain range]. Within the designated area, vessels are permitted to pass, but must reduce speed and undergo real-time monitoring. The navigable area is updated in real-time based on centralized radar monitoring data from a comprehensive feature dataset.

[0077] Let the raster map of the anchorage environment model be Where m is the number of raster rows and n is the number of raster columns, the coordinate system of the raster map is consistent with the coordinate system of the radar monitoring data and the optimal mooring position in the comprehensive feature dataset to ensure data compatibility; the coordinates of a single raster are... Corresponding to the actual plane coordinates of the anchorage The coordinate transformation formula is: , , in, The actual planar coordinates of the anchorage corresponding to the center of a single grid cell; The row and column coordinates of a single raster cell in a raster map; For the overall geographical boundary of the anchorage in The minimum coordinates along the axis are extracted based on the overall geographic boundary features of the anchorage in the comprehensive feature dataset. For the overall geographical boundary of the anchorage in The minimum coordinates along the axis are extracted based on the overall geographic boundary features of the anchorage in the comprehensive feature dataset. This is the side length of a single grid cell, i.e., the grid size.

[0078] Secondly, the route planning must meet both environmental and ship constraints. All constraints are quantified based on the core features in the comprehensive feature dataset to ensure the feasibility and safety of the route planning.

[0079] The primary function of environmental constraints is obstacle avoidance, and they are set based on the anchorage environmental features in the comprehensive feature dataset. The vessel's navigation path must remain entirely within the navigable grid, must not cross non-navigable grids, and should avoid risk warning grids as much as possible. If crossing is unavoidable, deceleration constraints must be met. The distance between any point on the path and obstacles or other existing moored vessels within the anchorage must be greater than a preset safety threshold. The constraint quantization formula is: , , in, The actual planar coordinates of any point on the ship's navigation path; This is a set of navigation paths for ships from entering the anchorage to the target virtual anchorage area. The actual coordinates of the center of an obstacle within the anchorage are determined based on the location features of the obstacle in the comprehensive feature dataset. The actual center coordinates of a certain existing moored vessel within the anchorage are determined based on the positional features of existing moored vessels in the comprehensive feature dataset. This is a preset safety threshold.

[0080] The purpose of ship constraints is to adapt to the ship's size and sailing attitude, based on the hull parameter features in the comprehensive feature dataset.

[0081] Among them, the turning radius constraint is the minimum turning radius that exists during the ship's navigation. Path planning must ensure that the radius of curvature at the turning point of the path is not less than the minimum turning radius of the ship to avoid the ship losing attitude control due to excessively sharp turns. The constraint quantification formula is as follows: , in, The radius of curvature at the turning point of the path; The minimum turning radius of the ship is calculated based on the hull parameter features in the comprehensive feature dataset. The calculation formula is as follows: ,in This is the turning radius coefficient, and its value range is... .

[0082] The draft constraint is that the anchorage water depth at any point along the path must be greater than the actual draft of the vessel to prevent it from running aground. The constraint quantification formula is as follows: , in, The actual water depth at a point on the path is determined based on the anchorage water depth distribution correlation features in the comprehensive feature dataset. The actual draft of the ship is extracted from the hull parameter information data in the comprehensive feature dataset. This refers to the water depth reserve, with a value range of [value missing]. It is used to address errors in water depth measurement.

[0083] The A* algorithm uses a heuristic function to guide the search direction and quickly find the optimal path from the starting point (entering the anchorage) to the end point (the center of the target virtual anchorage area). The algorithm introduces a cost function to quantify the search priority of each grid cell. The smaller the cost function value, the closer the grid cell is to the optimal path and the higher its search priority.

[0084] The cost function is calculated as the sum of the actual cost incurred and the estimated remaining cost. This ensures that the search process considers both the rationality of the already traveled path and accurately guides convergence towards the destination. The calculation formula is as follows: ; The smaller the value, the closer the grid is to the optimal path, and the higher the search priority; the algorithm selects from the open list in each iteration. The grid with the smallest value is expanded until the endpoint grid is found, ensuring that the final generated path is the optimal solution.

[0085] Actual cost This is used to quantify the actual path length of a ship from entering the anchorage (starting point) to the current grid n. Its calculation depends on the actual cost of the parent grid and the distance between adjacent grids, while also incorporating path constraints from the comprehensive feature dataset to ensure that the cost calculation is consistent with the actual navigation scenario. The specific calculation logic is as follows: If the current grid cell n is the starting point The actual cost (No initial driving distance); If the current grid n is not the starting grid, then the actual cost is equal to that of its parent grid. The actual cost, plus the distance between the parent grid and the current grid n, is calculated using the following formula: , During the calculation, if the current grid n is a risk warning grid, a small penalty value needs to be added to the actual cost. ( The algorithm is guided to avoid risky areas as much as possible; the penalty value is set based on the risk warning grid judgment standard in the comprehensive feature dataset, which does not affect the optimality of the path and can improve navigation safety.

[0086] Heuristic Cost It is used to predict the movement of ships from the current grid. Drive to the destination To determine the shortest distance, considering the accuracy requirements of anchorage route planning, a combination of Manhattan distance and Euclidean distance is used for calculation, balancing computational efficiency and prediction accuracy. The calculation formula is as follows: , in, The row and column coordinates of the current grid cell n in the grid map; End point Row and column coordinates in a raster map; The heuristic weighting coefficients range from 0.6 to 0.8 and are used to balance the weights of Euclidean distance and Manhattan distance. This is the grid side length, consistent with the grid size used in environment modeling.

[0087] The design of the heuristic cost function must satisfy admissibility, i.e. ,in, For the current grid The true shortest distance to the destination ensures that the algorithm will not miss the optimal path; combined with the grid distribution of the anchorage environment model, the above formula can fully meet the adoption requirements, while adapting to the straight-line and turning requirements of ship navigation.

[0088] The search process of the A* algorithm is as follows: Step 1: Initialize parameters and list. Set the starting grid. The actual cost Inspirational cost Calculated using the heuristic cost formula above, the cost function is... Add the starting grid to the open list. Close list Initialized to empty; the parent grid of the starting grid. .

[0089] Step 2: Check if the open list is empty. If An empty list indicates there are no feasible paths within the anchorage. In this case, an alarm mechanism should be triggered. Based on the anchorage environment features in the comprehensive feature dataset, the entry point to the anchorage should be replanned or the virtual anchorage area should be adjusted. The list is not empty. Proceed to the next step.

[0090] Step 3: Select the best raster and expand. From the open list Selected from The grid with the smallest value As the current extended raster, it is removed from Remove from list, add to close list If the current extended grid End grid If the optimal path has been found, proceed to the path backtracking step; if the destination has not been reached, continue to the next step.

[0091] Step 4: Filtering and updating adjacent grid cells. Traverse the current grid cells. 8 adjacent grids Each adjacent grid cell is filtered and processed.

[0092] 1. Validity Screening. Based on the grid status determination in environmental modeling and combined with obstacle and water depth features in the comprehensive feature dataset, determine whether adjacent grids are navigable. If they are non-navigable, exceed the anchorage boundary, or are already on the closed list, the navigability is checked. In the middle, skip the adjacent grid cell directly; 2. Constraint Verification. For valid adjacent grids, further verification is performed to determine if they meet ship constraints. If not, they are considered invalid grids and the process is skipped. 3. Cost Calculation. Calculate the distance from the current grid cell n to its adjacent grid cell. And calculate the temporary actual cost of the adjacent grid. If the adjacent grid cell is a risk warning grid cell, a penalty value needs to be added. ; 4. Cost update. If the adjacent grid cell is not in the open list. In the middle, calculate its heuristic cost. Cost function Add it Create a list and set its parent grid to the current grid n; if the adjacent grid is already in the list... In the list, compare Compared to the original actual cost of the grid ,like This indicates that the cost of reaching the adjacent grid cell via the current path is lower, so its value should be updated. 、 and update its parent raster to the current raster. ; Step 5: Iterative search. Repeat steps 2 through 4 until the destination grid is found. Alternatively, if no feasible path is determined, the search process can ensure that no potentially feasible path is overlooked. At the same time, the search efficiency can be improved by guiding the search through a cost function, thus avoiding ineffective iterations.

[0093] Step 6: Path backtracking and generation. Once the destination grid is found, start from the destination grid... Start, through the parent grid Backtracking in reverse order, find its parent grid, grandparent grid, and so on, until you return to the starting grid. By reversing the raster sequence obtained from the backtracking, the initial navigation path from the starting point (entering the anchorage) to the ending point (the center of the virtual anchorage) can be obtained. The path sequence is represented by raster coordinates, which need to be converted to the actual plane coordinates of the anchorage later.

[0094] Step 7: Convert the raster coordinate sequence of the initial path using the coordinate transformation formula. 、 This is converted into a sequence of actual plane coordinates of the anchorage to obtain the initial actual path. ,in The actual coordinates of the starting point The actual coordinates of the destination are used. Redundancy and smoothing processes are performed on the path using redundant path elimination and quadratic Bézier curves. The smoothed curve path is then sampled to generate a uniform sequence of coordinate points according to the ship's navigation accuracy requirements, resulting in the final optimal navigation path. At the same time, it marks the water depth and navigation speed suggestions for each sampling point along the route, providing precise guidance for ship navigation.

[0095] In this embodiment, the path planning algorithm further improves the control link, realizing the accurate generation of the optimal path from entering the anchorage to the virtual anchorage. The algorithm boasts high search efficiency and guaranteed path optimization. Combined with grid-based environmental modeling, it accurately delineates navigable, non-navigable, and risk warning grids within the anchorage, adapting to the dynamic environment of the anchorage. The quantified setting of dual constraints ensures safe obstacle avoidance and adapts to the ship's navigation attitude, preventing issues such as loss of control during turns and grounding. The scientific design of the cost function and heuristic function balances search efficiency and path optimization. A penalty mechanism guides the ship to avoid risky areas, while path backtracking and smoothing processes improve navigation stability and accuracy.

[0096] The method also includes: setting navigation constraints for each path node on the optimal path based on the node setting strategy; Furthermore, the node setting strategy includes: dividing the optimal navigation path into a path segment for entering / leaving the anchorage area, a path segment for a densely obstructed area, and a path segment for approaching the target virtual anchorage area, based on the total length of the optimal navigation path, the complexity of the anchorage environment, and the preset node density.

[0097] Specifically, based on the total length of the optimal navigation path, the complexity of the anchorage environment, and the preset node density, the optimal navigation path is divided into three core path segments. The criteria for dividing each path segment are clearly quantified, and precise division is achieved through calculation formulas, ensuring reasonable node distribution and highlighting key control points. First, define the core parameters. The total length of the optimal navigation path. based on The final optimal navigation path output by the algorithm is determined and calculated by accumulating the distances between the centers of adjacent grid cells. The calculation formula is as follows: ， in, For the optimal path, the first The actual coordinates of the center of each grid cell. This represents the total number of grid cells along the optimal path.

[0098] Anchorage environment complexity coefficient Used to quantify environmental factors such as obstacle distribution and ship density within anchorages, with a value range of [value range missing]. The calculation is based on the number of obstacles, their distribution density, and the number of existing moored vessels in the comprehensive feature dataset. The calculation formula is as follows: , in, The number of obstacles within 50m of the optimal navigation path is calculated based on the location and size characteristics of obstacles in the anchorage area from the comprehensive feature dataset. The maximum number of obstacles per unit area within the anchorage is calculated based on historical data. The number of existing moored vessels within a 50m radius of the optimal navigation path is calculated based on the positional characteristics of existing moored vessels in the comprehensive feature dataset. This represents the maximum number of vessels that can be moored within a unit area of ​​the anchorage, calculated based on the anchorage's design capacity.

[0099] Preset node density The number of nodes set per meter of path, with a value range of [value missing]. The node density is positively correlated with the complexity of the anchorage environment; the more complex the environment, the greater the node density, ensuring comprehensive node coverage in key areas.

[0100] The total number of nodes on the path is calculated based on the total path length and node density. It is used to constrain the number of nodes in each path segment. The calculation formula is as follows: ,in This indicates rounding up to the nearest integer to ensure the number of nodes meets management requirements.

[0101] These are the lengths of the path segments entering / exiting the anchorage area, the path segments in areas with dense obstacles, and the path segments approaching the target virtual anchorage area, respectively, satisfying... .

[0102] Based on the above parameters and considering the characteristics of the anchorage environment, the length of each path segment is determined using a proportional division method, with the specific proportions and environmental complexity coefficients being determined. To ensure a more reasonable division of areas with dense obstacles and more comprehensive node coverage, the path segments are divided as follows: 1. Path section for entering / exiting the anchorage area : Covers the area from the entry into the anchorage to the point where obstacles begin to appear (or the exit area). This area has a relatively simple environment, with vessels primarily entering / exiting. The proportion of the path length is [missing information]. The higher the environmental complexity, the smaller the proportion (prioritizing the number of nodes in areas with dense obstacles). The calculation formula is: , Value constraints are If the calculation result exceeds the constraint range, the corresponding boundary value is taken.

[0103] 2. Path sections in areas with dense obstacles The coverage area should include areas with concentrated obstacles and high navigational risks within the anchorage. These areas require densely packed nodes to ensure precise obstacle avoidance, with the path segment length ratio being [missing information]. The higher the environmental complexity, the larger the proportion. The calculation formula is: , Value constraints are If the calculation result exceeds the constraint range, the corresponding boundary value is taken.

[0104] 3. Path segment near the target virtual anchorage area The path segment covers an area sparsely populated with obstacles up to the center of the target virtual anchorage area. In this area, vessels primarily decelerate and adjust their attitude. The proportion of the path segment length is... The higher the environmental complexity, the smaller the proportion. The calculation formula is: ， Value constraints are If the calculation result exceeds the constraint range, adjust... and The proportions are determined to ensure that the total length of the three components meets the constraint.

[0105] Each path segment filters path nodes according to point selection rules. Path nodes include the starting point of entering the anchorage, obstacle avoidance turning points, path curvature change points, starting points near the virtual anchorage area, and the end point of the optimal navigation path.

[0106] Specifically, each route segment selects route nodes according to point selection rules. Combining route characteristics, environmental features, and ship navigation requirements, clear selection criteria are established to ensure that nodes cover all critical navigation stages. Route nodes include the starting point for entering the anchorage, obstacle avoidance turning points, points of abrupt changes in route curvature, the starting point near the virtual anchorage area, and the endpoint of the optimal navigation path. The selection rules for each node are quantified and clearly defined, as follows: First, clarify the node allocation for each path segment, based on the total number of nodes. The number of nodes in each path segment is allocated according to the ratio of the path segment length to ensure that the number of nodes is maximized in areas with dense obstacles. The calculation formula is as follows: , , , in, These represent the number of nodes entering / leaving the anchorage area, the number of nodes in the obstacle-dense area, and the number of nodes in the virtual anchorage area near the target, respectively. All are positive integers. If the calculation result is 0, it is taken as 1.

[0107] The starting point of entering the anchorage is the starting point of the optimal navigation path. As the first node in the entire path, it requires no further filtering and its coordinates are... Corresponding raster map coordinates .

[0108] Obstacle avoidance turning points exist in path segments with dense obstacles. The selection criterion is that the distance from the center of the obstacle on the path is equal to a preset safety threshold. The selection formula for points whose path direction changes is: and , in, The actual coordinates of a point on the path; The coordinates of the center of the obstacles surrounding the path; Preset safety threshold; For the first on the path The point, the first The navigation direction angle at each point is calculated based on the coordinates of two adjacent points along the path. The calculation formula is as follows: ; The path curvature abrupt change point exists in all path segments; the selection criterion is: the curvature at a certain point on the path. The curvature difference with adjacent points is greater than or equal to the preset curvature abrupt change threshold. The curvature abrupt change threshold is set based on the ship's minimum turning radius, and the calculation formula is as follows: (That is, the minimum turning radius corresponds to 20% of the curvature), the node selection formula is: and ; in, The preset curvature change threshold is associated with the minimum turning radius of the ship. For the first on the path The curvature at each point is calculated using the following formula: ; in, For the first on the path The, the The curvature of each point.

[0109] The starting point of the path segment approaching the target virtual anchorage area is the boundary point between the entry / exit of the anchorage area, the area with dense obstacles, and this area. The coordinates are calculated from the path segment length using the following formula: , in, These are the actual coordinates of the node. As the starting coordinates, The coordinates are the endpoint coordinates.

[0110] End point of the optimal navigation path The coordinates of the center of the target virtual anchorage area are given, which will serve as the last node of the entire path. Corresponding raster map coordinates .

[0111] Based on the environment of the path segment where the node is located and the navigation requirements of the ship, the corresponding navigation constraints are matched for each set path node. The constraints include speed constraints, turning angle constraints and navigation direction constraints.

[0112] Specifically, based on the environment of the path segment where the node is located and the navigation requirements of the ship, corresponding navigation constraints are matched for each set path node. The constraints include speed constraints, turning angle constraints, and navigation direction constraints. All constraints are quantitatively set and calculated based on a comprehensive feature dataset.

[0113] Speed ​​constraints Based on the environmental complexity and node type of the path segment where the node is located, minimum and maximum speeds are set for ships at the node to avoid excessive speed leading to untimely obstacle avoidance and excessive speed affecting navigation efficiency. The formula for calculating the speed constraint, i.e., the maximum speed, is as follows: , Minimum sailing speed is ,and To ensure that ships do not stop.

[0114] in, For the first The maximum speed of the ship at each node; For the first Minimum speed of ships at each node; The maximum design speed of the ship; For the first The local environment complexity coefficient of the location of each node.

[0115] Steering angle constraints Based on the node type and the ship's minimum turning radius, the maximum turning angle of the ship at the node is set to avoid excessive turning angles that could lead to loss of ship attitude control. The constraint formulas and rules are as follows: Formula for calculating maximum steering angle: , in, For the first The maximum turning angle of the ship at each node; The side length of a single grid cell; This is the minimum turning radius for a ship.

[0116] The navigation direction constraint is located at Based on the path direction of the path segment where the node is located and the location of the target virtual anchorage area, the range of the ship's navigation direction at the node is set to ensure that the ship travels along the optimal path.

[0117] Among them, the target orientation angle For the first The optimal path direction angle at each node, and the path direction angle The computational logic is consistent, that is ; Directional deviation threshold For the first The maximum deviation angle of the ship's navigation direction at each node is calculated using the following formula: The more complex the environment, the smaller the deviation threshold, ensuring accurate direction. The navigation direction constraint is that the ship is in the Actual navigation direction angle at each node , must meet .

[0118] In this embodiment, the node setting strategy further improves the control loop, achieving refined control of the optimal path. It precisely divides the three major path segments using a quantitative formula, dynamically allocating node density and quantity based on environmental complexity to ensure comprehensive node coverage in key areas with dense obstacles. The node selection rules are clearly quantified, covering all critical navigation stages, and the matched speed, steering, and direction constraints meet the navigation needs of different path segments, avoiding risks caused by abnormal speed or steering, and providing a precise benchmark for subsequent dynamic deviation monitoring and command adjustments.

[0119] The method includes: when the ship is running along the optimal navigation path, multi-source datasets are collected in real time at each path node through an interaction strategy and a collection strategy, and the comprehensive deviation factor between the ship's current navigation position and the optimal navigation path is calculated; when the comprehensive deviation factor is greater than the comprehensive deviation threshold at the current node, a corresponding adjustment command is generated to enable the ship to enter the virtual anchorage area and complete the docking according to the optimal navigation path.

[0120] Specifically, the comprehensive deviation factor is the core indicator for quantifying the degree to which a ship deviates from the optimal navigation path. Based on multi-source datasets collected in real time at nodes, combined with preset parameters of the nodes, it quantifies the degree of deviation of the ship in three dimensions: position, heading, and speed. It is calculated by weighted summation and has a value range of [0,1]. The larger the value, the more serious the deviation of the ship.

[0121] The formula for calculating the comprehensive deviation factor is: , in, This is the comprehensive deviation factor at the nth path node, with a value range of [0,1]. This indicates that the ship sailed exactly along the optimal path without any deviation. This indicates that the vessel has deviated significantly from the optimal path and is beyond the scope of safety control, requiring emergency adjustments. Here are the weighting coefficients for position deviation, heading deviation, and speed deviation, respectively. Based on the priority setting of node navigation constraints, in conventional anchorage control scenarios ; For the first The position deviation factor at each path node has a value range of [0,1], which quantifies the degree of deviation between the ship's current position and the preset position of the node. For the first The heading deviation factor at each path node has a value range of [0,1], which quantifies the degree of deviation between the ship's current heading and the target heading at the node. For the first The speed deviation factor at each path node, with a value range of [0,1], quantifies the degree of deviation between the ship's current speed and the node speed constraint.

[0122] The position deviation factor is obtained based on the distance between the ship's current coordinates and the preset coordinates of the node, combined with safety threshold normalization processing. The calculation formula is as follows: , , in, For the first The straight-line distance between the current position of the ship at each node and the preset position of the node; For the first The current actual position coordinates of the ship at each node are collected in real time by the shipborne GPS and radar monitoring terminal. For the first The preset position coordinates of each path node; This is a preset safety threshold.

[0123] when hour, This is considered a serious positional deviation, triggering an emergency adjustment warning.

[0124] Based on the deviation angle between the ship's current course and the course of the target node, and combined with the normalization process of the course deviation threshold, the course deviation factor is obtained. The calculation formula is as follows: , in, For the first The ship's current actual heading angle at each node is collected in real time by the ship's attitude sensor, and the range of values ​​is... . For the first The target heading angle at each node, i.e., the preset heading angle in the node setting strategy, is calculated using the following formula: , in, For the first Each node has preset coordinates. For the first Each node has preset coordinates to ensure that the target heading angle is consistent with the direction of the optimal navigation path. For the first The directional deviation threshold at each node is calculated using the following formula: , in, For the first The local environmental complexity coefficient of each node is consistent with the environmental complexity coefficient in the comprehensive feature extraction above, and is calculated from the obstacles within a 20m radius around the node and the number of existing moored ships.

[0125] when hour, This is considered a serious deviation from the course.

[0126] The speed deviation factor is obtained by normalizing the deviation between the ship's current speed and the nodal speed constraint. The calculation formula is as follows: , in, For the first The ship's current actual speed at each node; Ask each The maximum and minimum sailing speeds at each node are calculated using the following formulas: , , in, To ensure that the ship has sufficient maneuverability; To determine the maximum design speed of the ship, features of the ship's hull parameters are extracted from the comprehensive feature dataset.

[0127] When the ship's speed is Within the range, there is no speed deviation. When the speed is lower than or higher When the speed reaches the ship's maximum design speed, the deviation factor is calculated proportionally, with a value range of [0,1]. hour, This is considered a serious speed deviation.

[0128] Speed ​​deviation factor It is obtained by normalizing the deviation between the ship's current speed and the nodal speed constraint. When the ship deviates from the optimal navigation path, an adjustment command is generated; when When the ship is deemed to be sailing normally, no adjustment command is required; only real-time data collection and monitoring continue.

[0129] Furthermore, the interaction strategy includes: constructing a two-way interaction link between the acquisition terminal, the shore-based management and control platform, and the shipborne control terminal; triggering the acquisition terminal to collect multi-source datasets in real time through the interaction strategy; the acquisition terminal synchronously transmits the collected data on the ship's position, navigation attitude, and anchorage environment to the shore-based management and control platform through the two-way interaction link for the calculation of the comprehensive deviation factor; the shore-based management and control platform transmits the calculated comprehensive deviation factor, deviation judgment result, and corresponding adjustment instructions to the shipborne control terminal in real time through the interaction link.

[0130] Furthermore, the adjustment commands include: heading adjustment commands, speed adjustment commands, and comprehensive adjustment commands.

[0131] When a ship deviates from the optimal path, a course adjustment command is generated. The course adjustment command includes adjusting the ship's turning angle and turning speed to guide the ship to gradually correct its course and return to the optimal navigation path. The turning angle is adjusted according to the magnitude of the comprehensive deviation factor. The larger the comprehensive deviation factor, the closer the turning angle is to the upper limit of the constraint threshold.

[0132] Specifically, when a ship's course deviates from the optimal path, that is... ,and , If only the heading deviation exists, a heading adjustment command is generated. The heading adjustment command includes adjusting the ship's turning angle and turning speed to guide the ship to gradually correct its heading and return to the optimal navigation path; among them, the turning angle is adjusted according to the magnitude of the comprehensive deviation factor. The larger the comprehensive deviation factor, the closer the turning angle is to the upper limit of the constraint threshold.

[0133] The formula for adjusting the steering angle is as follows: , , , in, For the first The heading adjustment and steering angle at each node, with a range of values. Ensure that the steering angle does not exceed the ship's maneuverability limits; For the first The overall deviation factor of each node; For the first The maximum steering angle of each node; For the grid side length in path planning, The minimum turning radius of the ship. This is the turning radius coefficient.

[0134] Steering speed adjustment formula (to ensure smooth steering and avoid loss of control): , in, For the ship's turning speed, the range of values ​​is... The greater the deviation, the slower the turning speed, ensuring smooth adjustment and avoiding loss of control of the ship's attitude. This represents the ship's current actual speed.

[0135] The execution logic of the course adjustment command is to make gradual corrections. After each adjustment, the ship's course data is re-acquired, and the course deviation factor is calculated until... Stop adjusting the course to ensure accurate course correction and avoid over-adjustment.

[0136] When a ship's speed is too fast or too slow, causing it to deviate from the optimal path, or when its speed does not meet the speed constraints of the current node, a speed adjustment command is generated. The course adjustment command includes adjusting the ship's target speed and acceleration / deceleration rate to avoid exacerbating the deviation due to abnormal speed, while ensuring the stability of the ship's navigation.

[0137] Specifically, when a ship's speed is too fast or too slow, causing it to deviate from the optimal path, or when its speed does not meet the speed constraints of the current node, i.e. ,and , If only a speed deviation exists, a speed adjustment command is generated. The speed adjustment command includes adjusting the ship's target speed and acceleration / deceleration rate to avoid exacerbating the deviation due to speed abnormalities, while ensuring the stability of the ship's navigation.

[0138] The formula for adjusting the target speed is: , in, To ensure the adjusted target speed of the ship, Within the specified range, adapt to node velocity constraints. The greater the deviation, the closer the target velocity is to the specified value. or This ensures rapid correction of speed deviations while avoiding sudden speed changes.

[0139] The formula for adjusting acceleration / deceleration rates is: , in, For acceleration / deceleration rate, the value ranges from [0.05, 0.1] × The greater the deviation, the lower the speed, to avoid sudden speed changes that could lead to abnormal ship attitude and difficulty in control. This refers to the ship's maximum design speed.

[0140] When a ship has deviations in both course and speed, or when the combined deviation factor exceeds the combined deviation threshold, a combined adjustment command is generated. The combined adjustment command first adjusts the course, then the speed, and simultaneously limits the turning angle and target speed to ensure that the ship returns to the optimal navigation path and avoids navigation risks.

[0141] When a ship simultaneously has deviations in both course and speed, or when the combined deviation factor exceeds the combined deviation threshold, i.e. and ,or At that time, a comprehensive adjustment command is generated. The comprehensive adjustment command first adjusts the course, then the speed, and simultaneously limits the turning angle and target speed to ensure that the ship returns to the optimal navigation path and avoids navigation risks.

[0142] The first step is to calculate the turning angle and turning speed according to the heading adjustment command formula, and to correct the heading deviation first, until... The second step is to calculate the target speed and acceleration / deceleration rate according to the speed adjustment command formula, and correct the speed deviation until... ; Real-time monitoring of position deviation throughout the process to ensure If the positional deviation continues to increase, the adjustment should be immediately suspended, triggering an emergency warning and being fed back to the shore-based control platform to avoid the risk of collision.

[0143] The supplementary formula for the comprehensive adjustment constraints is as follows: ， The sum of the heading adjustment angle and the acceleration / deceleration rate shall not exceed 50% of the sum of the maximum turning angle and the maximum speed threshold at the node, so as to avoid the double adjustment causing the ship's attitude to become uncontrollable.

[0144] After all adjustment instructions are generated, they are sent to the shipborne control terminal in real time via a two-way interactive link. The shipborne control terminal automatically executes the adjustment operation and simultaneously collects the ship's real-time status once according to a preset collection frequency, synchronously feeding it back to the shore-based management platform. The shore-based management platform updates the comprehensive deviation factor in real time based on the feedback data. If the deviation is corrected to within the threshold range, then... If the deviation continues to exceed the threshold, the adjustment command parameters will be adjusted accordingly. Incremental synchronous optimization is performed until the deviation correction is completed, ensuring that the ship accurately enters the target virtual anchorage area along the optimal path and completes the docking.

[0145] In this embodiment, the real-time deviation monitoring and dynamic adjustment mechanism strengthens the navigation safety defense line. By quantifying the comprehensive deviation factor and calculating the sub-deviations, the degree of ship deviation is accurately determined. Combined with the two-way interactive link, real-time data transmission and command closed loop are realized. Three types of adjustment commands are accurately matched with the deviation type. Based on scientific formulas, the adjustment parameters are optimized to ensure that the deviation is corrected in a timely manner and the adjustment is stable and controllable. This effectively avoids risks such as attitude loss of control and collisions, and ensures that the ship accurately sails to the virtual anchorage and completes the docking.

[0146] Furthermore, at each path node, the shipborne control terminal feeds back the ship's current navigation status to the shore-based management platform through an interactive link. The shore-based management platform, in conjunction with the navigation constraints preset at that node, performs real-time verification of the ship's current navigation status. If it finds that the ship's navigation status violates the constraints, it generates an early warning command through an interactive strategy and sends it to the shipborne control terminal simultaneously.

[0147] Specifically, this applies to situations involving serious violations of constraints and significant safety risks, including instances where the vessel's position is incorrect and the straight-line distance between the vessel's current position and the preset node position is [not specified]. Greater than 1.2 times The minimum distance between the ship and the obstacle is less than 0.8 Draft greater than The steering angle exceeds the maximum steering angle at the node. 1.2 times or more. Such warning instructions must clearly indicate an emergency and immediately reduce the sailing speed to [a certain value]. The following actions are hereby suspended: the shipboard operators shall immediately take over manual control, while the shore-based control platform shall track the ship's dynamics in real time and continuously issue auxiliary guidance until the violation is resolved.

[0148] In this embodiment, the real-time node verification and hierarchical early warning mechanism relies on a two-way interactive link to realize real-time feedback and verification of the ship's navigation status. Combined with the node's preset constraints, it identifies violations. Clear emergency early warning standards and handling procedures are formulated for major safety risks. Through hierarchical control and early warning, it can quickly respond to major violations, enforce standardized ship operations, guide human intervention, and effectively avoid serious safety hazards such as collisions, groundings, and loss of attitude control.

[0149] Example 2 like Figure 2As shown, this system is used in the smart anchorage management method based on multi-source heterogeneous data described in Embodiment 1 above. The system includes: a data acquisition module, used to acquire multi-source datasets of anchorage area, port of call management, and the ship itself. The data acquisition module includes an aerial acquisition unit, a shore-based acquisition unit, and a ship-borne acquisition unit, corresponding to the aerial acquisition terminal, shore-based acquisition terminal, and ship-borne acquisition terminal, respectively. It can complete the initial acquisition, real-time acquisition, and supplementary acquisition of multi-source datasets according to a preset acquisition sequence and preset frequency, providing data support for subsequent feature extraction, deviation calculation, and other steps.

[0150] The feature extraction module receives multi-source datasets transmitted by the acquisition module and extracts corresponding effective features through an adapted feature extraction algorithm, including effective features of anchorage area images, effective features of port management images, effective features of ship hull parameters, and effective features of radar monitoring.

[0151] The weighted fusion module receives various effective features transmitted by the feature extraction module. The weighted fusion module is equipped with a weighted fusion strategy to complete the standardization processing of effective features, linear weighted summation calculation, generate a comprehensive feature dataset, eliminate the heterogeneity of multi-source features, and provide data support for the implementation of optimization strategies.

[0152] The optimization module receives the comprehensive feature dataset transmitted by the weighted fusion module. The optimization module is equipped with an optimization strategy, which uses the particle swarm optimization algorithm combined with safety constraints and size constraints to iteratively select the optimal mooring position for ships and generate the target virtual anchorage area.

[0153] The path planning module receives relevant features of the target virtual anchorage area transmitted by the optimization module. Through path planning algorithms, it determines the optimal navigation path for the ship from entering the anchorage to the target virtual anchorage area, providing a basis for node setting and path guidance.

[0154] The node setting module is used to receive the optimal navigation path transmitted by the path planning module. The node setting module is equipped with a node setting strategy, which divides the path into segments, filters path nodes, and matches corresponding navigation constraints for each path node.

[0155] The interaction module is used to realize two-way interaction between the data acquisition module, the shore-based control module, and the shipborne control module. The interaction module is equipped with an interaction strategy to build a two-way interaction link and complete the real-time transmission of multi-source datasets, comprehensive deviation factors, adjustment commands, early warning commands, and ship navigation status data.

[0156] The deviation calculation and command generation module is used to calculate the comprehensive deviation factor between the ship's current navigation position and the optimal navigation path when the ship navigates to each path node, based on the multi-source dataset collected in real time by the acquisition module. When the comprehensive deviation factor is greater than the comprehensive deviation threshold of the current node, the module generates the corresponding heading adjustment command, speed adjustment command, or comprehensive adjustment command according to the type of adjustment command carried. At the same time, in conjunction with the interaction module, the module completes the real-time verification of the ship's navigation status and generates an early warning command when the ship violates the navigation constraints of the node.

[0157] The shore-based control module, as the core control unit of the system, is used to receive data and parameters transmitted by each module, coordinate the operation of each module, and complete core control functions such as comprehensive deviation factor calculation, adjustment command and early warning command generation, and navigation status verification. It corresponds to the shore-based control platform in the interaction strategy.

[0158] The shipboard control module, installed on the ship, is used to receive adjustment commands, early warning commands, optimal navigation paths, node constraints, and other information transmitted by the interaction module, to provide feedback on the ship's current navigation status, and to trigger the ship's control system to perform adjustment operations. It corresponds to the shipboard control terminal in the interaction strategy.

[0159] The data storage module is used to retain multi-source datasets collected by the acquisition module, effective features extracted by the feature extraction module, comprehensive feature datasets generated by the weighted fusion module, operating parameters of each module, adjustment instructions, early warning instructions, ship navigation status data and interactive data, providing data support for subsequent data traceability, algorithm optimization and docking standardization verification.

[0160] In this embodiment, the acquisition module, through multi-unit and multi-terminal collaborative acquisition, completes comprehensive acquisition of multi-source data according to a preset time sequence and frequency, ensuring data comprehensiveness and real-time performance, and laying a solid data foundation for subsequent stages. The feature extraction and weighted fusion module works in tandem to accurately extract effective features, eliminate heterogeneity of multi-source data, and generate a high-quality comprehensive feature dataset, solving the problem of disorganized multi-source data and ensuring the accuracy of subsequent optimization and planning stages. The optimization module, path planning module, and node setting module are progressively integrated, relying on corresponding algorithms and strategies to accurately achieve optimal berthing location selection, optimal navigation path planning, and node constraint matching, providing clear guidance for ship navigation. The deviation calculation and command generation module, combined with the interaction module, achieves accurate deviation calculation, accurate command generation, and real-time warning of violations, quickly responding to navigation anomalies and avoiding safety risks. The shore-based control module coordinates the operation of various modules to achieve the integration of core control functions; the shipborne control module receives commands and provides feedback on status in real time, ensuring that commands are executed, and, together with the bidirectional link of the interaction module, achieves real-time transmission of data and commands, breaking down control barriers. The data storage module retains data from the entire process, providing support for subsequent traceability and optimization, and improving system scalability.

[0161] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A smart anchorage management method based on multi-source heterogeneous data, characterized in that, The method includes: We collect multi-source datasets of anchorage areas, berthing port management, and the vessels themselves using a data collection strategy, and extract the corresponding effective features using an appropriate feature extraction algorithm. A weighted fusion strategy is used to weight each effective feature to obtain a comprehensive feature dataset; The optimal berthing position of the ship is determined by processing the comprehensive feature dataset through optimization strategies, and a target virtual anchorage area is generated. The optimal navigation path for a ship from entering the anchorage to the target virtual anchorage area is determined using a path planning algorithm. Based on the node setting strategy, set the navigation constraints for each path node on the optimal path; When a ship is traveling along the optimal navigation path, multi-source datasets are collected in real time at each path node through interactive and acquisition strategies. The comprehensive deviation factor between the ship's current navigation position and the optimal navigation path is calculated. When the comprehensive deviation factor is greater than the comprehensive deviation threshold at the current node, a corresponding adjustment command is generated to enable the ship to enter the virtual anchorage area and complete berthing according to the optimal navigation path.

2. The intelligent anchorage management method based on multi-source heterogeneous data according to claim 1, characterized in that, The data collection strategy involves collecting data from the anchorage area, port control area, and the vessel itself via a data collection terminal according to a preset data collection sequence. This strategy provides data support for determining the optimal berthing position of the vessel, generating the target virtual anchorage area, planning the optimal navigation path, and dynamically adjusting the vessel. The data acquisition terminals include aerial acquisition terminals, shore-based acquisition terminals, and shipborne acquisition terminals; The multi-source dataset includes anchorage area image data, port management image data, ship hull parameter information data, and radar monitoring data; The data acquisition sequence includes collecting overall image data of the anchorage area, image data of the anchorage entrance, and initial dynamic data within the anchorage when the ship sails to a preset distance before the anchorage entrance. The shipborne acquisition terminal collects basic ship parameters to ensure that the system obtains the initial environment and basic ship data of the anchorage, providing initial support for the planning of the optimal berthing position and the optimal navigation path. From the moment a vessel enters the anchorage entrance until the vessel completes berthing and comes to a stop after a preset time, the aerial acquisition terminal, shore-based acquisition terminal, and shipborne acquisition terminal collect multi-source datasets in real time at a preset frequency, directly supporting path guidance, comprehensive deviation factor calculation, and adjustment command generation. After the vessel completes its berthing, the aerial, shore-based, and shipborne data acquisition terminals continue to operate for a preset duration to collect additional data on the vessel's berthing posture and the degree of fit between the vessel and the virtual anchorage area. This data is used to verify the standardization of the vessel's berthing and the rationality of the virtual anchorage area planning.

3. The intelligent anchorage management method based on multi-source heterogeneous data according to claim 2, characterized in that, The overall geographical boundary features of the anchorage, the location and size features of obstacles within the anchorage, the outline and location features of existing moored vessels within the anchorage, and the correlation features of anchorage water depth distribution are extracted from the image data of the anchorage area to provide a basis for obstacle avoidance planning of the optimal mooring position of vessels and analysis of anchorage space utilization. The image data of the port of call is used to extract the characteristics of ship traffic at the anchorage boundary, the characteristics of ship movement when entering and leaving the anchorage, and the environmental characteristics of the shore control area. The image features of ship heading and speed are extracted from the ship movement characteristics to provide a basis for the path guidance and dynamic adjustment of ships entering and leaving the anchorage. Extract ship size characteristics, ship type characteristics, and ship attitude correlation characteristics from ship hull parameter information data to provide a basis for size planning and path constraint setting of target virtual anchorage area; Extracting real-time ship position coordinates, speed, direction, and turning angle characteristics from radar monitoring data, as well as the real-time dynamic characteristics of obstacles and other ships within the anchorage, provides core data support for optimal navigation path planning, comprehensive deviation factor calculation, and adjustment command generation.

4. The intelligent anchorage management method based on multi-source heterogeneous data according to claim 3, characterized in that, The weighted fusion strategy includes determining the weight coefficient of effective features of radar monitoring based on the importance and data accuracy of each effective feature. This weight coefficient is greater than the weight coefficient of effective features of anchorage area images, which is greater than the weight coefficient of effective features of ship hull parameters, which is greater than the weight coefficient of effective features of port management images, and the sum of all weight coefficients is 1. Each valid feature is standardized to eliminate the dimensional differences between different types of features, ensuring the rationality of weighted fusion calculation, and the values ​​of all valid features are mapped to the [0,1] interval through standardization. A linear weighted summation algorithm is used to multiply each standardized effective feature by its corresponding weight coefficient and then sum them to obtain a comprehensive feature dataset.

5. The intelligent anchorage management method based on multi-source heterogeneous data according to claim 4, characterized in that, The optimization strategy includes: A safety constraint is defined as ensuring that the distance between the optimal mooring position of a vessel and obstacles or other existing moored vessels within the anchorage is no less than a preset safety threshold. Define the length of the virtual anchorage area as a preset multiple of the ship's length and the width as a preset multiple of the ship's width as dimensional constraints; Each potential ship berthing position is iterated using particle swarm optimization with safety and size constraints. After the iteration is completed, the particle position corresponding to the global optimal fitness value is determined as the optimal mooring position of the ship. Combining the coordinate characteristics of the mooring position, the ship size characteristics, and the anchorage environment characteristics, a target virtual anchorage area is generated.

6. The intelligent anchorage management method based on multi-source heterogeneous data according to claim 5, characterized in that, The node setting strategy includes: Based on the total length of the optimal navigation path, the complexity of the anchorage environment, and the preset node density, the optimal navigation path is divided into a path segment for entering / exiting the anchorage area, a path segment for areas with dense obstacles, and a path segment for approaching the target virtual anchorage area. Each path segment filters path nodes according to point selection rules. Path nodes include the starting point of entering the anchorage, obstacle avoidance turning point, path curvature change point, starting point near the virtual anchorage area, and the end point of the optimal navigation path. Based on the environment of the path segment where the node is located and the navigation requirements of the ship, the corresponding navigation constraints are matched for each set path node. The constraints include speed constraints, turning angle constraints and navigation direction constraints.

7. The intelligent anchorage management method based on multi-source heterogeneous data according to claim 6, characterized in that, The interaction strategy includes: Establish a two-way interactive link between the data acquisition terminal, the shore-based management and control platform, and the shipborne control terminal; The data acquisition terminal is triggered by an interactive strategy to collect multi-source datasets in real time. The data acquisition terminal transmits the collected data on the ship's position, navigation attitude and anchorage environment to the shore-based control platform through a two-way interactive link for the calculation of the comprehensive deviation factor. The shore-based control platform transmits the calculated comprehensive deviation factor, deviation judgment result and corresponding adjustment instructions to the shipborne control terminal in real time through the interactive link.

8. The intelligent anchorage management method based on multi-source heterogeneous data according to claim 7, characterized in that, The adjustment commands include: heading adjustment commands, speed adjustment commands, and comprehensive adjustment commands; When a ship deviates from the optimal path, a course adjustment command is generated. The course adjustment command includes adjusting the ship's turning angle and turning speed to guide the ship to gradually correct its course and return to the optimal path. The turning angle is adjusted according to the magnitude of the comprehensive deviation factor. The larger the comprehensive deviation factor, the closer the turning angle is to the upper limit of the constraint threshold. When a ship's speed is too fast or too slow, causing it to deviate from the optimal path, or when its speed does not meet the speed constraints of the current node, a speed adjustment command is generated. The heading adjustment command includes adjusting the ship's target speed and acceleration / deceleration rate to avoid exacerbating the deviation due to abnormal speed, while ensuring the stability of the ship's navigation. When a ship has deviations in both course and speed, or when the combined deviation factor exceeds the combined deviation threshold, a combined adjustment command is generated. The combined adjustment command first adjusts the course, then the speed, and simultaneously limits the turning angle and target speed to ensure that the ship returns to the optimal navigation path and avoids navigation risks.

9. The intelligent anchorage management method based on multi-source heterogeneous data according to claim 8, characterized in that, At each path node, the shipborne control terminal feeds back the ship's current navigation status to the shore-based management platform through an interactive link. The shore-based management platform, in conjunction with the navigation constraints preset for that node, performs real-time verification of the ship's current navigation status. If the ship's navigation status is found to violate the constraints, an early warning command is generated through an interactive strategy and simultaneously sent to the shipborne control terminal.

10. A smart anchorage management and control system based on multi-source heterogeneous data, characterized in that, This system is used in the intelligent anchorage management method based on multi-source heterogeneous data as described in claims 1-9, and the system includes: The data acquisition module is used to collect multi-source datasets from anchorage areas, port management, and the vessels themselves. The data acquisition module includes an aerial data acquisition unit, a shore-based data acquisition unit, and a shipborne data acquisition unit, which correspond to the aerial data acquisition terminal, the shore-based data acquisition terminal, and the shipborne data acquisition terminal, respectively. It can complete the initial data acquisition, real-time data acquisition, and supplementary data acquisition of multi-source datasets according to a preset data acquisition sequence and preset frequency, providing data support for subsequent feature extraction, deviation calculation, and other processes. The feature extraction module is used to receive multi-source datasets transmitted by the acquisition module and extract corresponding effective features through an adapted feature extraction algorithm, including effective features of anchorage area images, effective features of port management images, effective features of ship parameters, and effective features of radar monitoring. The weighted fusion module receives various effective features transmitted by the feature extraction module. The weighted fusion module is equipped with a weighted fusion strategy to complete the standardization processing and linear weighted summation calculation of effective features, generate a comprehensive feature dataset, eliminate the heterogeneity of multi-source features, and provide data support for the implementation of optimization strategies. The optimization module receives the comprehensive feature dataset transmitted by the weighted fusion module. The optimization module is equipped with an optimization strategy, which uses the particle swarm optimization algorithm combined with safety constraints and size constraints to iteratively select the optimal mooring position for the ship and generate the target virtual anchorage area. The path planning module receives relevant features of the target virtual anchorage area transmitted by the optimization module, and determines the optimal navigation path for the ship from the anchorage entrance to the target virtual anchorage area through the path planning algorithm, providing a basis for node setting and path guidance. The node setting module is used to receive the optimal navigation path transmitted by the path planning module. The node setting module is equipped with a node setting strategy, which divides the path segments, filters the path nodes, and matches the corresponding navigation constraints for each path node. The interaction module is used to realize two-way interaction between the data acquisition module, the shore-based control module, and the shipborne control module. The interaction module is equipped with an interaction strategy to build a two-way interaction link and complete the real-time transmission of multi-source datasets, comprehensive deviation factors, adjustment commands, early warning commands, and ship navigation status data. The deviation calculation and command generation module is used to calculate the comprehensive deviation factor between the ship's current navigation position and the optimal navigation path when the ship navigates to each path node, based on the multi-source dataset collected in real time by the acquisition module. When the comprehensive deviation factor is greater than the comprehensive deviation threshold of the current node, the module generates the corresponding heading adjustment command, speed adjustment command, or comprehensive adjustment command according to the type of adjustment command carried. At the same time, it works with the interaction module to complete the real-time verification of the ship's navigation status and generate a warning command when the ship violates the navigation constraints of the node. The shore-based control module, as the core control unit of the system, is used to receive data and parameters transmitted by each module, coordinate the operation of each module, and complete core control functions such as comprehensive deviation factor calculation, adjustment command and early warning command generation, and navigation status verification. It corresponds to the shore-based control platform in the interaction strategy. The shipboard control module, installed on the ship, is used to receive adjustment commands, early warning commands, optimal navigation paths, node constraints and other information transmitted by the interaction module, to provide feedback on the ship's current navigation status, and to trigger the ship control system to perform adjustment operations. It corresponds to the shipboard control terminal in the interaction strategy. The data storage module is used to retain multi-source datasets collected by the acquisition module, effective features extracted by the feature extraction module, comprehensive feature datasets generated by the weighted fusion module, operating parameters of each module, adjustment instructions, early warning instructions, ship navigation status data and interactive data, providing data support for subsequent data traceability, algorithm optimization and docking standardization verification.

Images