Intelligent ship cargo loading and unloading optimization method and system

Abstract

The invention discloses a ship intelligent cargo loading and unloading optimization method and system, and relates to the technical field of cargo loading and unloading optimization, and the method comprises the steps: processing real-time video data through a deep learning model, generating a cargo feature data set, and combining ship loading drawing information to construct a cargo optimization model; solving the cargo optimization model by utilizing an improved ant colony algorithm, generating a cargo loading and unloading operation instruction sequence, transmitting the cargo loading and unloading operation instruction sequence to automatic loading and unloading equipment of a wharf, executing loading and unloading operation, and simultaneously collecting operation state data frames in a loading and unloading process in real time; and analyzing the deviation between the operation state data frame and the optimal loading and unloading path planning, and when the deviation exceeds a preset threshold value, solving the cargo optimization model again, and generating an updated cargo loading and unloading operation instruction sequence. According to the method, cargo features are accurately identified through a deep learning model to form a data set, a multi-constraint optimization model is constructed in combination with a ship loading drawing, an optimal loading and unloading instruction sequence is generated by using an improved ant colony algorithm, and intelligent planning of loading and unloading operation is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of cargo loading and unloading optimization, and in particular to an intelligent cargo loading and unloading optimization method and system for ships. Background Art

[0002] As an important tool for global cargo transportation, the loading and unloading efficiency of ships is directly related to the port throughput capacity and logistics operation efficiency. With the development of automated ports and intelligent shipping systems, the traditional method of relying on manual experience and static loading drawings to make loading and unloading decisions has gradually exposed problems such as low operating efficiency, poor responsiveness, and insufficient dynamic adaptability. Although some existing ports have introduced IoT sensing devices, intelligent identification systems and other means to assist loading and unloading operations, the overall intelligence level of the system still needs to be improved due to the lag in responding to real-time operating status, the lack of a dynamic optimization mechanism for loading and unloading paths, and limited cargo identification accuracy. It is difficult to meet the efficient loading and unloading needs of multiple types and sizes of cargo in complex shipping scenarios. Therefore, how to achieve real-time closed-loop optimization of loading and unloading path planning and operation scheduling has become a key issue that needs to be urgently addressed in current intelligent loading and unloading technology for ships.

[0003] CN116468348A discloses a cargo loading and unloading method, an Internet of Things system, an electronic device, and a storage medium. The method obtains spatial modeling data of the cargo compartment and combines it with the storage information of the cargo to be transferred to determine the loading and unloading parameters and generate corresponding loading and unloading instructions, thereby realizing unmanned automated loading and unloading operations driven by the Internet of Things system. This solution has certain technical progress in improving loading and unloading efficiency, but the method relies on structured storage information input for cargo identification, lacks the ability to intelligently process unstructured data (such as video streams) in real-time complex environments, and cannot dynamically adjust the loading and unloading path according to the actual size and position of the cargo. In addition, once the loading and unloading strategy is generated, it lacks a feedback mechanism and cannot optimize the instructions in time according to the execution deviation, resulting in a high risk of errors and scheduling delays in actual operations.

[0004] CN117557109B proposes a method for the integrated and coordinated operation of digital and physical ship infrastructure. Through methods such as fuzzy clustering and gray correlation analysis, it optimizes and coordinates the ship's cargo loading and unloading infrastructure to build a more reasonable loading and unloading system architecture. This method has certain advantages in the deployment of loading and unloading facilities and system-level coordination, and is suitable for improving system stability and economic benefits. However, its loading and unloading strategy generation is still mainly based on static simulation and test data, lacking a response mechanism and path correction capability for real-time on-site operation status. It is difficult to achieve rapid updating and execution of optimal instructions in a dynamically changing loading and unloading environment, limiting its adaptability and intelligence level in complex port operations. Summary of the Invention

[0005] In view of the problems of insufficient recognition accuracy, delayed response of path optimization and low level of intelligent loading and unloading process in existing ship cargo loading and unloading methods, the present invention is proposed.

[0006] Therefore, the problem to be solved by the present invention is how to achieve efficient cargo identification, real-time optimization of loading and unloading paths, and intelligent scheduling and correction updates of operation instructions in a complex and dynamic ship operation environment.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, an embodiment of the present invention provides a method for optimizing intelligent cargo loading and unloading of ships, which includes:

[0009] The deep learning model is used to process real-time video data collected from the ship's cargo loading and unloading area to identify cargo type, size, and current location information, generating a cargo feature dataset.

[0010] Building a cargo optimization model based on the cargo feature dataset and combined with ship loading drawing information;

[0011] Solving the cargo optimization model using an improved ant colony algorithm to generate a cargo loading and unloading operation instruction sequence, wherein the cargo loading and unloading operation instruction sequence includes cargo loading and unloading priorities, operation equipment allocation plans, and optimal loading and unloading path planning;

[0012] Transmitting the cargo loading and unloading operation instruction sequence to the terminal automated loading and unloading equipment to perform the loading and unloading operation, while simultaneously collecting operation status data frames during the loading and unloading process in real time;

[0013] The deviation between the operation status data frame and the optimal loading and unloading path planning is analyzed. When the deviation exceeds a preset threshold, the cargo optimization model is re-solved to generate an updated cargo loading and unloading operation instruction sequence.

[0014] As a preferred solution of the intelligent cargo loading and unloading optimization method of the present invention, the deviation between the operation status data frame and the optimal loading and unloading path planning is analyzed. When the deviation exceeds a preset threshold, the cargo optimization model is re-solved to generate an updated cargo loading and unloading operation instruction sequence, including:

[0015] Receive the processed operation status data frame and extract the actual operation parameter set, where the actual operation parameter set includes cargo grabbing coordinates, spreader posture angle and task execution timestamp;

[0016] Acquiring an expected parameter set from a loading and unloading operation instruction sequence, wherein the expected parameter set includes a planned grasping coordinate, an allowable posture deviation threshold, and a planned completion time;

[0017] Calculating a spatial position deviation and a temporal deviation based on the actual operation parameter set and the expected parameter set;

[0018] When the spatial position deviation is greater than the preset spatial position deviation or the time deviation is greater than the preset time deviation, the task is marked as an abnormal task node;

[0019] Performing root cause analysis on the marked abnormal task nodes and adjusting the constraints of the cargo optimization model;

[0020] An incremental recalculation method is used to solve the adjusted cargo optimization model and generate an updated sequence of loading and unloading operation instructions. A version identifier is added to the modified part, and the integrity of the instructions is ensured by a digital signature. At the same time, the original sequence is retained as a rollback backup.

[0021] As a preferred solution of the intelligent cargo loading and unloading optimization method of the present invention, the cargo loading and unloading operation instruction sequence is transmitted to the terminal automated loading and unloading equipment to perform the loading and unloading operation, and the operation status data frame of the loading and unloading process is collected in real time, including:

[0022] Sending the loading and unloading operation instruction sequence to the control cabinet of the port gantry crane via the industrial Ethernet protocol, wherein the loading and unloading operation instruction sequence is encapsulated into an ASN.1 data packet using the ASN.1 encoding format;

[0023] The control cabinet of the gantry crane parses the ASN.1 data packet, extracts the three-dimensional coordinate information in the cargo movement task list, and converts it into a target angle sequence of each joint motor through inverse kinematics calculation;

[0024] Executing the target angle sequence, the synchronous sensor group collects the operation status data frame;

[0025] Constructing a real-time data acquisition pipeline to convert the analog signals of the sensor group into analog signals, aligning them with the expected values ​​of the cargo movement task list, and generating a time-stamped operation status data frame;

[0026] A status monitoring process is run in the portal crane control cabinet to process the operation status data frame in real time.

[0027] As a preferred solution of the intelligent cargo loading and unloading optimization method of the present invention, the method for generating the cargo loading and unloading operation instruction sequence is as follows:

[0028] Initializing a parameter set of the improved ant colony algorithm, wherein the parameter set includes a pheromone concentration matrix, a heuristic factor matrix, and the number of ants;

[0029] Construct a solution space mapping relationship, mapping the effective grid cells of the cargo hold-cargo association matrix to nodes in the topological graph, where the edge weights between nodes include the crane movement distance and the cargo flipping difficulty coefficient;

[0030] Perform path building operations for each artificial ant;

[0031] Adopting an elite retention strategy to update the pheromone concentration matrix;

[0032] When the number of iterations reaches the maximum number of iterations or the optimal path fitness value has a multiple consecutive change rate less than the preset change rate, the algorithm is terminated and the loading and unloading operation instruction sequence corresponding to the global optimal path is output;

[0033] feasibility verification of the loading and unloading operation instruction sequence through discrete event simulation to detect conflicts with dynamic obstacles in the cargo optimization model;

[0034] When the conflict rate exceeds the preset conflict rate, the weight parameters in the heuristic factor matrix are adjusted and the path optimization solution process is re-executed until a loading and unloading operation instruction sequence that meets all constraints and can be directly executed is obtained.

[0035] As a preferred solution of the intelligent cargo loading and unloading optimization method of the present invention, the path construction operation is performed for each artificial ant, including:

[0036] Calculating the transition probability according to the pheromone concentration matrix and the heuristic factor matrix, and selecting the next visited node;

[0037] Dynamically update the taboo table to prohibit repeated visits to nodes that have already loaded or unloaded goods;

[0038] When three consecutive nodes fail to meet the stacking constraint rules, the backtracking mechanism is triggered to reselect the path;

[0039] After completing the path traversal of all ants, the fitness value of each path is calculated. The specific formula is as follows:

[0040]

[0041] Among them, F is the fitness value, ε is the space utilization weight coefficient, V i is the volume of the i-th cargo, V total is the total available volume of the cargo hold, σ is the penalty coefficient for center of gravity offset, Δx is the offset of the cargo center of gravity in the x-axis direction, Δy is the offset of the cargo center of gravity in the y-axis direction, and n is the total number of cargoes in the current cargo hold.

[0042] As a preferred solution of the intelligent cargo loading and unloading optimization method of the present invention, the cargo optimization model is constructed as follows:

[0043] Extracting ship loading plan information from a ship management database, wherein the ship loading plan information includes cargo capacity data of each cargo hold, a deck load distribution diagram, and the coordinates of securing point positions;

[0044] Spatially matching the cargo position coordinate data with the cabin boundary coordinates of the ship loading plan information to determine the target cargo hold number to which each cargo belongs;

[0045] According to the cargo type code and weight level, query the cargo attribute knowledge base to obtain the corresponding stacking constraint rules;

[0046] Using the target cargo hold number as an index, establish a cargo hold-cargo association matrix, where rows of the cargo hold-cargo association matrix correspond to cargo hold grid cells, and columns store cargo IDs, sizes, and stacking constraints;

[0047] Calculating the current cumulative load of each cargo hold grid cell based on the grid code of the deck load distribution diagram;

[0048] The cargo hold-cargo association matrix, stacking constraint rules, and securing point position coordinates are integrated to construct a cargo optimization model with cargo hold space utilization and center of gravity balance coefficient as optimization objectives;

[0049] The cargo optimization model is verified for solvability. When there are cargoes that cannot satisfy the stacking constraints, an adjustment suggestion list including the conflicting cargo IDs and constraint types is generated and fed back to the ship cargo management system.

[0050] As a preferred solution of the intelligent cargo loading and unloading optimization method of the present invention, the method for generating the cargo feature data set is:

[0051] Install multiple cameras in the cargo loading and unloading area of ​​the ship to collect real-time video data and multi-view image data of the operation area;

[0052] Inputting the real-time video data into a deep learning model to extract cargo image features, wherein the deep learning model is designed based on an improved ResNet-101 network structure;

[0053] Matching the cargo image features with a preset cargo type feature library to obtain cargo type information;

[0054] Based on the multi-view image data, obtaining cargo size data through a three-dimensional reconstruction algorithm;

[0055] Based on the feature point information extracted from the multi-view image data, a visual positioning algorithm is used to obtain the coordinate data of the cargo position;

[0056] The cargo type information, the cargo size data and the cargo location coordinate data are integrated into a cargo feature data set, wherein the cargo feature data set includes cargo location coordinate data, cargo type code and weight grade.

[0057] In a second aspect, an embodiment of the present invention provides an intelligent ship cargo loading and unloading optimization system, comprising: an identification module for processing real-time video data collected from a ship cargo loading and unloading operation area using a deep learning model to identify cargo type, size, and current location information, and generate a cargo feature dataset;

[0058] A construction module, which constructs a cargo optimization model based on the cargo feature dataset and the ship loading drawing information;

[0059] a generation module, configured to solve the cargo optimization model using an improved ant colony algorithm to generate a cargo loading and unloading operation instruction sequence, wherein the cargo loading and unloading operation instruction sequence includes cargo loading and unloading priorities, operation equipment allocation plans, and optimal loading and unloading path planning;

[0060] An execution module is used to transmit the cargo loading and unloading operation instruction sequence to the terminal automated loading and unloading equipment to execute the loading and unloading operation, and simultaneously collect the operation status data frames during the loading and unloading process in real time;

[0061] The updating module is used to analyze the deviation between the operation status data frame and the optimal loading and unloading path planning, and when the deviation exceeds a preset threshold, re-solve the cargo optimization model to generate an updated cargo loading and unloading operation instruction sequence.

[0062] In a third aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the processor executes the computer program, it implements any step of the above-mentioned method for optimizing intelligent cargo loading and unloading of ships.

[0063] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the above-mentioned method for optimizing intelligent cargo loading and unloading of ships is implemented.

[0064] Compared with the prior art, the present invention has the following beneficial effects:

[0065] 1. The system processes real-time video data collected from ship cargo loading and unloading areas through a deep learning model to identify cargo type, size, and current location information, generating a cargo feature dataset. Using an improved ResNet-101 network structure design, combined with multi-view image data for 3D reconstruction and visual positioning, it achieves precise cargo identification. Compared with traditional manual identification methods, this significantly improves recognition efficiency and accuracy, and reduces the human error rate.

[0066] 2. Based on the cargo feature dataset and combined with ship loading plan information, a cargo loading and unloading optimization model is constructed. Through spatial matching and association matrix construction, static plan data is organically combined with dynamic cargo information to form a multi-objective optimization model that includes space utilization and center of gravity balance coefficients. This modeling approach overcomes the shortcomings of traditional loading and unloading operation planning, which often lack consideration of cargo hold space characteristics, and enables accurate modeling of key constraints such as load distribution and securing point locations. By integrating stacking constraint rules, safety hazards caused by improper cargo stacking are prevented, significantly improving the safety and space utilization efficiency of ship cargo loading.

[0067] 3. Using an improved ant colony algorithm to solve the cargo optimization model, a cargo loading and unloading operation instruction sequence is generated. The cargo loading and unloading operation instruction sequence includes the cargo loading and unloading priorities, the operation equipment allocation plan, and the optimal loading and unloading path planning. This step solves the problem of traditional ant colony algorithms easily falling into local optimality through the elite retention strategy and dynamic tabu table mechanism. At the same time, the fitness function comprehensively considers space utilization and center of gravity offset factors, ensuring the global optimality of the loading and unloading plan.

[0068] 4. Transmit the cargo loading and unloading operation instruction sequence to the terminal's automated loading and unloading equipment to execute the loading and unloading operations, while simultaneously collecting operation status data frames during the loading and unloading process in real time. This step achieves secure and reliable instruction transmission through the Industrial Ethernet protocol and the ASN.1 encoding format, and accurately converts the abstract instructions into an angular sequence executable by the equipment through inverse kinematics calculations. Furthermore, the constructed real-time data acquisition pipeline ensures that the sensor data is time-aligned with the expected value, forming a time-stamped operation status data frame.

[0069] 5. Analyze the deviation between the operation status data frame and the optimal loading and unloading path plan. When the deviation exceeds a preset threshold, re-solve the cargo optimization model and generate an updated cargo loading and unloading operation instruction sequence. This step accurately quantifies spatial and temporal deviations by extracting and comparing actual operation parameters with expected parameters, and adjusts the optimization model constraints through root cause analysis. In particular, the use of an incremental recalculation method significantly reduces the computational complexity of replanning. Digital signatures ensure the integrity of instructions, and the original sequence is retained as a rollback backup. This overcomes the lack of real-time adjustment capabilities in traditional loading and unloading methods and provides an effective mechanism for responding to dynamic changes in the loading and unloading environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0071] Figure 1 Flowchart of the intelligent cargo loading and unloading optimization method for ships. DETAILED DESCRIPTION

[0072] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0073] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0074] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0075] The present invention is described in detail with reference to schematic diagrams. For ease of illustration, cross-sectional views of device structures may be partially enlarged and not to scale when describing embodiments of the present invention. Furthermore, the schematic diagrams are merely illustrative and should not limit the scope of the present invention. Furthermore, in actual production, the three-dimensional dimensions of length, width, and depth should be included.

[0076] In the description of the present invention, it should be noted that the terms "upper, lower, inner, and outer" and other references to orientations or positional relationships are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate and simplify the description of the present invention. They are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation, and therefore should not be construed as limitations on the present invention. Furthermore, the terms "first, second, or third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0077] In this disclosure, unless otherwise specified or limited, the terms "mounted," "connected," and "connected" should be interpreted broadly. For example, they may refer to fixed, removable, or integral connections. They may also refer to mechanical, electrical, or direct connections, indirect connections through an intermediary, or internal communication between two components. Those skilled in the art will understand the specific meanings of these terms in this disclosure.

[0078] Example 1

[0079] Reference Figure 1 , which is the first embodiment of the present invention, provides a method for optimizing intelligent cargo loading and unloading of ships, comprising:

[0080] S1: Real-time video data collected from the ship's cargo loading and unloading area is processed through a deep learning model to identify the cargo type, size, and current location information, and generate a cargo feature dataset.

[0081] S1.1: Install multiple cameras in the cargo loading and unloading area of ​​the ship to collect real-time video data and multi-view image data of the operation area.

[0082] It should be noted that the installation positions of the multi-view camera group cover the key locations of the entire operation area, including the dock loading and unloading area, the deck storage area and the cargo area in the cabin, and the overlap rate of the field of view of adjacent cameras is not less than 15%; the acquisition frequency is not less than 24 frames per second, and the video resolution reaches 1920×1080 pixels.

[0083] S1.2: Input real-time video data into a deep learning model to extract cargo image features. The deep learning model is designed based on an improved ResNet-101 network structure.

[0084] It should be noted that the improved ResNet-101 network structure adds a spatial attention mechanism and skip connection layer on the basis of the standard ResNet to improve the model's ability to recognize irregularly shaped cargo; the deep learning model has been pre-trained on more than 10,000 cargo image samples, covering ship cargo types including containers, bulk cargo and large machinery and equipment.

[0085] Preferably, the deep learning model analyzes the real-time video data frame by frame, separates the cargo target from the background in each frame of the image through the instance segmentation algorithm, and extracts the cargo image features; obtains the three-dimensional size parameter data of the cargo target, including length, width, height and volume estimation, and the accuracy error is controlled within the range of ±3%.

[0086] S1.3: Match the cargo image features with the preset cargo type feature library to obtain cargo type information.

[0087] In an optional embodiment, constructing a preset cargo type feature library includes: the preset cargo type feature library includes feature templates of 50 standard cargo types commonly used in ship transportation, and each cargo type feature template consists of four parts: a shape descriptor, a texture feature vector, a color histogram, and an identification symbol pattern; the shape descriptor is represented by combining a Fourier shape descriptor with shape factors such as rectangularity, circularity, and eccentricity; the texture feature vector uses energy, contrast, homogeneity, and entropy values ​​extracted by local binary patterns (LBP) and gray-level co-occurrence matrices (GLCM); the color histogram uses a 64-dimensional quantized histogram of the HSV color space; and the identification symbol pattern is represented by geometric features extracted by Hough transform and contour detection.

[0088] In an optional embodiment, the cargo image feature matching process adopts a hierarchical cascade matching strategy, including coarse matching and fine matching; the coarse matching stage uses shape descriptors and color histograms for rapid screening, calculates the Euclidean distance between the cargo image features and each template in the preset cargo type feature library, and selects the top 10 candidate cargo types with the smallest distance; the coarse matching uses a KD tree index structure to accelerate the feature retrieval process, and the retrieval time complexity is controlled at the O(log n) level.

[0089] In an optional embodiment, during the fine matching stage, a weighted feature fusion method is used to calculate the comprehensive similarity of the candidate cargo types screened out by the rough matching. The specific formula is as follows:

[0090]

[0091] Among them, S is the comprehensive similarity score, α i is the weight coefficient of the i-th category feature, F i is the i-th type feature vector of the goods to be matched, T iis the corresponding i-th type feature vector in the feature library template, β is the scale adjustment factor, V i is the characteristic variance of the i-th category of goods to be matched, M i is the i-th feature variance of the template feature.

[0092] Preferably, after the comprehensive similarity calculation of all candidate cargo types is completed, the candidate cargo type with the highest comprehensive similarity that exceeds the preset comprehensive similarity is selected as the final matching result; if the highest comprehensive similarity does not reach the preset comprehensive similarity, the cargo is marked as an unknown type and the manual review process is triggered; if the identification result is obtained, the uncertainty estimate of the 95% confidence interval is calculated.

[0093] For example, the value range of the comprehensive similarity S is [0, 1], where 0 indicates complete dissimilarity and 1 indicates complete similarity. When the comprehensive similarity S>0.85, the match is considered successful, and the candidate cargo type with the highest comprehensive similarity that exceeds the preset comprehensive similarity is selected as the final matching result. When 0.7≤comprehensive similarity S≤0.85, manual review is triggered. When the comprehensive similarity S<0.7, the match is considered unsuccessful.

[0094] S1.4: Based on the multi-view image data, obtain the cargo size data through the 3D reconstruction algorithm.

[0095] S1.5: Based on the feature point information extracted from the multi-view image data, a visual positioning algorithm is used to obtain the cargo location coordinate data.

[0096] S1.6: Integrate cargo type information, cargo size data, and cargo location coordinate data into a cargo feature dataset.

[0097] In an optional embodiment, the cargo feature dataset includes cargo location coordinate data, cargo type code and weight grade. The cargo feature dataset is organized in JSON format, and each data entry contains cargo ID, type code, length, width and height dimensions, weight grade and six-degree-of-freedom posture information.

[0098] S2: Based on the cargo feature dataset and combined with the ship loading drawing information, a cargo optimization model is constructed.

[0099] S2.1: Extract ship loading plan information from the ship management database, where the ship loading plan information includes cargo capacity data of each cargo hold, deck load distribution diagram, and securing point position coordinates.

[0100] It should be noted that the deck load distribution diagram uses grid coding to mark the unit area load limits of different areas.

[0101] S2.2: Spatially match the cargo location coordinate data in the cargo feature dataset with the cabin boundary coordinates in the ship loading plan information to determine the target cargo hold number to which each cargo belongs.

[0102] It should be noted that the R*-tree spatial index is used to accelerate the spatial matching process, and the matching error tolerance is set to 5 cm.

[0103] Specifically, when the cargo location coordinates fall within the boundary coordinate range of a cargo hold, the cargo hold number is marked as the target cargo hold number; when the cargo location coordinates fall into the overlapping area of ​​multiple cargo holds at the same time, the following arbitration strategy is executed: give priority to the upper cargo hold with a larger z-coordinate value; if the z-coordinates are the same, select the cargo hold with a larger hold capacity margin; if the hold capacity margins are the same, select the cargo hold closer to the securing point.

[0104] Furthermore, when the closest distance between the cargo location coordinates and all cargo hold boundaries is greater than the preset distance, this processing flow is triggered: perform a kNN search (k=3) in the R*-tree spatial index; take the number of the nearest neighbor cargo hold as the target cargo hold number; and record the matching deviation value to the system log.

[0105] Furthermore, when the stacking state of cargo is detected during the matching process, the total height of the current cargo in the stacking structure is calculated based on the six-degree-of-freedom posture information in the cargo feature data set; if the sum of the cargo's own height and the height of the stacking structure exceeds the remaining vertical space height of the current target cargo hold, that is, the overall stacking height of the cargo is higher than the cargo hold can accommodate, the adjacent cargo hold is automatically selected for re-matching, and the target cargo hold number is updated; when all cargo matching is completed, a matching result table is generated.

[0106] It should be noted that the matching result table includes the cargo ID (derived from the cargo feature dataset), the target cargo hold number, and the matching confidence (the Sigmoid value calculated based on the boundary distance).

[0107] And the space exception flag (0 / 1 indicates whether the arbitration strategy is triggered).

[0108] S2.3: Based on the cargo type code and weight level, query the cargo attribute knowledge base to obtain the corresponding stacking constraint rules.

[0109] It should be noted that the stacking constraint rules include the maximum number of stacking layers, a list of prohibited mixed loading types, and a center of gravity offset threshold.

[0110] S2.4: Using the target cargo hold number as the index, establish a cargo hold-cargo association matrix, where the rows of the cargo hold-cargo association matrix correspond to cargo hold grid cells, and the columns store cargo IDs, sizes, and stacking constraints.

[0111] It should be noted that the division size of the cargo hold grid unit is the same as the minimum operating spacing of the crane gripper.

[0112] S2.5: Based on the grid code of the deck load distribution diagram, calculate the current cumulative load of each cargo hold grid cell.

[0113] Preferably, when the current cumulative load exceeds 90% of the load limit per unit area, the grid is marked as unavailable in the cargo hold-cargo association matrix.

[0114] S2.6: Integrate the cargo hold-cargo association matrix, stacking constraints, and securing point coordinates to construct a cargo optimization model with cargo hold space utilization and center of gravity balance coefficient as optimization objectives.

[0115] It should be noted that the center of gravity balance coefficient is calculated by the vector synthesis result of the moments of various cargoes in the cargo hold.

[0116] S2.7: Verify the solvability of the cargo optimization model. When there are cargoes that cannot meet the stacking constraints, generate an adjustment suggestion list containing the conflicting cargo ID and constraint type and feed it back to the ship cargo management system.

[0117] S3: The cargo optimization model is solved using the improved ant colony algorithm to generate a cargo loading and unloading operation instruction sequence, where the cargo loading and unloading operation instruction sequence includes the cargo loading and unloading priority, operation equipment allocation plan and optimal loading and unloading path planning.

[0118] S3.1: Initialize the parameter set of the improved ant colony algorithm, where the parameter set includes the pheromone concentration matrix, the heuristic factor matrix and the number of ants.

[0119] It should be noted that the dimension of the pheromone concentration matrix is ​​the same as the cargo hold-cargo association matrix in the cargo optimization model, and the initial value is set to a gradient distribution according to the cargo priority.

[0120] S3.2: Construct a solution space mapping relationship to map the valid grid cells of the cargo hold-cargo association matrix to nodes in the topological graph, where the edge weights between nodes include the crane movement distance and the cargo flipping difficulty coefficient.

[0121] It should be noted that the cargo flipping difficulty coefficient is obtained by querying the cargo type code in the cargo feature dataset;

[0122] S3.3: Perform path building operations for each artificial ant.

[0123] Specifically include:

[0124] S3.3.1: Calculate the transition probability based on the pheromone concentration matrix and the heuristic factor matrix and select the next visited node.

[0125] Preferably, the relevant formula for the transition probability is as follows:

[0126]

[0127] in, is the probability that the kth ant moves from node i to node j, τ ij is the pheromone concentration matrix between node i and node j, η ij is the heuristic information matrix between node i and node j (inversely proportional to the distance), γ is the pheromone importance coefficient, δ is the heuristic factor importance coefficient, λ i is the cargo flipping difficulty coefficient of node i, allowed k is the set of nodes that the kth ant is allowed to visit.

[0128] It should be noted that the transition probability The value range of is [0,1], and for any node i, the sum of its transition probabilities to all feasible nodes j is 1.

[0129] S3.3.2: Dynamically update the taboo table to prohibit repeated visits to nodes that have already loaded or unloaded cargo;

[0130] S3.3.3: When three consecutive nodes fail to meet the stacking constraint rules, the backtracking mechanism is triggered to reselect the path.

[0131] S3.4: After all ants have completed their path traversal, the fitness value of each path is calculated. The specific formula is as follows:

[0132]

[0133] Among them, F is the fitness value, ε is the space utilization weight coefficient, V i is the volume of the i-th cargo, V total is the total available volume of the cargo hold, σ is the penalty coefficient for center of gravity offset, Δx is the offset of the cargo center of gravity in the x-axis direction, Δy is the offset of the cargo center of gravity in the y-axis direction, and n is the total number of cargoes in the current cargo hold.

[0134] It should be noted that the fitness value F ranges from [0,1], where 0 represents the worst solution and 1 represents the best solution. When the fitness value F>0.80, it can be considered a better solution. When 0.6≤fitness value F≤0.80, it is an acceptable solution. When the fitness value F<0.6, re-optimization is required.

[0135] S3.5: Update the pheromone concentration matrix using the elite retention strategy.

[0136] Specifically include:

[0137] S3.5.1: Perform pheromone boosting on the top 10% of fitness paths:

[0138]

[0139] Among them, τ ij is the pheromone concentration matrix between node i and node j, ρ is the pheromone volatility coefficient, Q is the pheromone intensity constant, F is the fitness value of this path, ω p is the cargo priority gain coefficient.

[0140] S3.5.2: Perform pheromone decay on the remaining paths: τ ij =(1-ρ·σ p )τ ij .

[0141] S3.5.3: Introduce cargo handling priority as an adjustment factor for the decay coefficient in decay operations.

[0142] S3.6: When the number of iterations reaches the maximum number of iterations or the optimal path fitness value has a multiple consecutive change rate less than the preset change rate, the algorithm is terminated and the loading and unloading operation instruction sequence corresponding to the global optimal path is output.

[0143] It should be noted that the loading and unloading operation instruction sequence is organized in JSON-LD format and contains the following fields: cargo ID list (arranged in loading and unloading order), assigned crane number (determined according to the equipment distribution in the ship loading drawing information), and three-dimensional path coordinate point set (smoothed by B-spline curve).

[0144] S3.7: Verify the feasibility of the loading and unloading operation instruction sequence through discrete event simulation to detect conflicts with dynamic obstacles in the cargo optimization model.

[0145] S3.7: When the conflict rate exceeds the preset conflict rate, the weight parameters in the heuristic factor matrix η are adjusted and the path optimization solution process is re-executed until a loading and unloading operation instruction sequence that meets all constraints and can be directly executed is obtained.

[0146] S4: The cargo loading and unloading operation instruction sequence is transmitted to the terminal's automated loading and unloading equipment to perform the loading and unloading operation, while simultaneously collecting the operation status data frames during the loading and unloading process in real time.

[0147] S4.1: The loading and unloading operation instruction sequence is sent to the terminal gantry crane control cabinet through the industrial Ethernet protocol, where the loading and unloading operation instruction sequence is encapsulated into an ASN.1 data packet using the ASN.1 encoding format.

[0148] It should be noted that the data fields of the ASN.1 data packet include the instruction header (including timestamp and CRC checksum), the cargo movement task list (sorted by the optimal loading and unloading path planning) and the equipment control parameter set (including the gripper opening and closing speed and the lifting acceleration threshold).

[0149] S4.2: The control cabinet of the gantry crane parses the received ASN.1 data packet, extracts the three-dimensional coordinate information in the cargo movement task list, and converts it into the target angle sequence of each joint motor through inverse kinematics calculation.

[0150] S4.3: Execute the target angle sequence and synchronize the sensor group to collect the operation status data frame.

[0151] It should be noted that the sensor group includes an absolute encoder installed on the wire rope drum (measuring the lifting height, with an accuracy of ±1cm), a laser rangefinder at the four corners of the sling (detecting the tilt angle of the cargo, with a range of 0-15°) and a torque sensor on the slewing bearing (monitoring the load torque, with a sampling rate of 1kHz).

[0152] S4.4: Build a real-time data acquisition pipeline to convert the analog signals of the sensor group into analog signals, align them with the expected values ​​of the cargo movement task list, and generate a time-stamped operation status data frame.

[0153] It should be noted that time alignment uses the PTPv2 precision clock protocol, and the synchronization error is less than 100μs.

[0154] S4.5: Run the status monitoring process in the portal crane control cabinet and process the operation status data frame in real time; specifically, it includes:

[0155] Calculate the Euclidean distance deviation Δd between the actual grasping position and the command coordinates; detect when the spreader swing angle exceeds the allowable value recorded in the cargo feature dataset; and calculate the execution delay Δt of each task node.

[0156] When a Euclidean distance deviation of Δd > 20 cm or Δt > 5 s is detected, the exception handling procedure is triggered, including: freezing the current motor control instructions; sending an exception code to the ship's intelligent dispatching system via the 5G-U industrial private network; and activating the local emergency braking device.

[0157] In an optional embodiment, under normal operation, the operation status data frame is uploaded to the ship intelligent scheduling system through the OPCUA protocol with a period of 200ms, forming a closed-loop verification pair with the original instruction; each data frame is associated with a corresponding instruction sequence ID.

[0158] S5: Analyze the deviation between the operation status data frame and the optimal loading and unloading path planning. When the deviation exceeds a preset threshold, re-solve the cargo optimization model to generate an updated cargo loading and unloading operation instruction sequence.

[0159] S5.1: Receive the processed operation status data frame and extract the actual operation parameter set, where the actual operation parameter set includes cargo grabbing coordinates, spreader posture angle, and task execution timestamp.

[0160] S5.2: Obtain an expected parameter set from the loading and unloading operation instruction sequence, where the expected parameter set includes a planned grasping coordinate, an allowable posture deviation threshold, and a planned completion time.

[0161] S5.3: Calculate the spatial position deviation and the temporal deviation based on the actual operation parameter set and the expected parameter set.

[0162] S5.4: When the spatial position deviation is greater than the preset spatial position deviation or the time deviation is greater than the preset time deviation, the task is marked as an abnormal task node.

[0163] S5.5: Perform root cause analysis on marked abnormal task nodes and adjust the constraints of the cargo optimization model.

[0164] Preferably, if the spatial position deviation is dominant, then check whether the size of the corresponding cargo in the cargo feature data set matches the gripper parameters; if the time deviation is dominant, check whether the spacing between adjacent cargoes in the ship loading drawing information meets the equipment operation requirements.

[0165] Furthermore, the constraints include: adding gripper selection constraints for cargo with mismatched sizes; and modifying the grid availability status in the cargo hold-cargo association matrix for areas with insufficient spacing.

[0166] S5.6: Use the incremental recalculation method to solve the adjusted cargo optimization model, generate the updated loading and unloading operation instruction sequence, add a version identifier to the modified part, and ensure the integrity of the instruction through digital signature, while retaining the original sequence as a rollback backup.

[0167] Preferably, the adjusted cargo optimization model includes: retaining the solutions of completed tasks unchanged; running the improved ant colony algorithm only on the subset of unexecuted tasks; and inheriting 70% of the initial value of the original pheromone concentration matrix during recalculation.

[0168] In summary, the present invention significantly improves loading and unloading efficiency, reduces energy consumption, enhances loading and unloading safety, and improves space utilization through the organic combination of innovative technical means such as deep learning and multi-source data fusion, multi-objective optimization modeling, improved ant colony algorithm solution, real-time monitoring and closed-loop adjustment. At the same time, it has the adaptive ability to cope with complex environmental changes, providing an innovative solution for intelligent port logistics.

[0169] Example 2

[0170] This embodiment also provides an intelligent ship cargo loading and unloading optimization system, including:

[0171] The recognition module is used to process real-time video data collected from the ship's cargo loading and unloading area through a deep learning model to identify the cargo type, size, and current location information, and generate a cargo feature dataset;

[0172] The construction module builds a cargo optimization model based on the cargo feature dataset and the ship loading drawing information;

[0173] A generation module is used to solve the cargo optimization model using an improved ant colony algorithm to generate a cargo loading and unloading operation instruction sequence, where the cargo loading and unloading operation instruction sequence includes the cargo loading and unloading priority, the operation equipment allocation plan and the optimal loading and unloading path planning;

[0174] The execution module is used to transmit the cargo loading and unloading operation instruction sequence to the terminal automated loading and unloading equipment to execute the loading and unloading operation, and at the same time collect the operation status data frame in real time during the loading and unloading process;

[0175] The update module is used to analyze the deviation between the operation status data frame and the optimal loading and unloading path planning. When the deviation exceeds the preset threshold, the cargo optimization model is re-solved to generate an updated cargo loading and unloading operation instruction sequence.

[0176] This embodiment also provides an electronic device, which includes a processor, memory, a communication interface, a display screen, and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner. The wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a multi-task edge computing resource scheduling method is implemented. The display screen of the computer device can be a liquid crystal display or an electronic ink display screen. The input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad provided on the computer device housing, or an external keyboard, touchpad or mouse.

[0177] This embodiment further provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method proposed in the above embodiment is implemented.

[0178] The storage medium proposed in this embodiment and the method proposed in the above embodiment belong to the same inventive concept. For technical details not fully described in this embodiment, please refer to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0179] Through the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented with the help of software and necessary general-purpose hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute the method of the embodiment of the present invention.

[0180] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

[0181] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The scheme in the embodiment of the present application can be implemented in various computer languages.

[0182] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0183] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0184] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0185] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present application.

[0186] Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims

1. A method for optimizing intelligent cargo loading and unloading of ships, characterized by: include, The deep learning model is used to process real-time video data collected from the ship's cargo loading and unloading area to identify cargo type, size, and current location information, generating a cargo feature dataset. Building a cargo optimization model based on the cargo feature dataset and combined with ship loading drawing information; Solving the cargo optimization model using an improved ant colony algorithm to generate a cargo loading and unloading operation instruction sequence, wherein the cargo loading and unloading operation instruction sequence includes cargo loading and unloading priorities, operation equipment allocation plans, and optimal loading and unloading path planning; Transmitting the cargo loading and unloading operation instruction sequence to the terminal automated loading and unloading equipment to perform the loading and unloading operation, while simultaneously collecting operation status data frames during the loading and unloading process in real time; The deviation between the operation status data frame and the optimal loading and unloading path planning is analyzed. When the deviation exceeds a preset threshold, the cargo optimization model is re-solved to generate an updated cargo loading and unloading operation instruction sequence.

2. The method for optimizing intelligent cargo loading and unloading of ships according to claim 1, characterized in that: Analyze the deviation between the operation status data frame and the optimal loading and unloading path planning. When the deviation exceeds a preset threshold, re-solve the cargo optimization model to generate an updated cargo loading and unloading operation instruction sequence, including: Receive the processed operation status data frame and extract the actual operation parameter set, where the actual operation parameter set includes cargo grabbing coordinates, spreader posture angle and task execution timestamp; Acquiring an expected parameter set from a loading and unloading operation instruction sequence, wherein the expected parameter set includes a planned grasping coordinate, an allowable posture deviation threshold, and a planned completion time; Calculating a spatial position deviation and a temporal deviation based on the actual operation parameter set and the expected parameter set; When the spatial position deviation is greater than the preset spatial position deviation or the time deviation is greater than the preset time deviation, the task is marked as an abnormal task node; Performing root cause analysis on the marked abnormal task nodes and adjusting the constraints of the cargo optimization model; An incremental recalculation method is used to solve the adjusted cargo optimization model and generate an updated sequence of loading and unloading operation instructions. A version identifier is added to the modified part, and the integrity of the instructions is ensured by a digital signature. At the same time, the original sequence is retained as a rollback backup.

3. The intelligent cargo loading and unloading optimization method for ships according to claim 2, characterized in that: The cargo loading and unloading operation instruction sequence is transmitted to the terminal automated loading and unloading equipment to perform the loading and unloading operation, and the operation status data frame of the loading and unloading process is collected in real time, including: Sending the loading and unloading operation instruction sequence to the control cabinet of the port gantry crane via the industrial Ethernet protocol, wherein the loading and unloading operation instruction sequence is encapsulated into an ASN.1 data packet using the ASN.1 encoding format; The control cabinet of the gantry crane parses the ASN.1 data packet, extracts the three-dimensional coordinate information in the cargo movement task list, and converts it into a target angle sequence of each joint motor through inverse kinematics calculation; Executing the target angle sequence, the synchronous sensor group collects the operation status data frame; Constructing a real-time data acquisition pipeline to convert the analog signals of the sensor group into analog signals, aligning them with the expected values ​​of the cargo movement task list, and generating a time-stamped operation status data frame; A status monitoring process is run in the portal crane control cabinet to process the operation status data frame in real time.

4. The method for optimizing intelligent cargo loading and unloading of ships according to claim 3, characterized in that: The method for generating the cargo loading and unloading operation instruction sequence is as follows: Initializing a parameter set of the improved ant colony algorithm, wherein the parameter set includes a pheromone concentration matrix, a heuristic factor matrix, and the number of ants; Construct a solution space mapping relationship, mapping the effective grid cells of the cargo hold-cargo association matrix to nodes in the topological graph, where the edge weights between nodes include the crane movement distance and the cargo flipping difficulty coefficient; Perform path building operations for each artificial ant; Adopting an elite retention strategy to update the pheromone concentration matrix; When the number of iterations reaches the maximum number of iterations or the optimal path fitness value has a multiple consecutive change rate less than the preset change rate, the algorithm is terminated and the loading and unloading operation instruction sequence corresponding to the global optimal path is output; feasibility verification of the loading and unloading operation instruction sequence through discrete event simulation to detect conflicts with dynamic obstacles in the cargo optimization model; When the conflict rate exceeds the preset conflict rate, the weight parameters in the heuristic factor matrix are adjusted and the path optimization solution process is re-executed until a loading and unloading operation instruction sequence that meets all constraints and can be directly executed is obtained.

5. The method for optimizing intelligent cargo loading and unloading of ships according to claim 4, characterized in that: Perform path building operations for each artificial ant, including: Calculating the transition probability according to the pheromone concentration matrix and the heuristic factor matrix, and selecting the next visited node; Dynamically update the taboo table to prohibit repeated visits to nodes that have already loaded or unloaded goods; When three consecutive nodes fail to meet the stacking constraint rules, the backtracking mechanism is triggered to reselect the path; After completing the path traversal of all ants, the fitness value of each path is calculated. The specific formula is as follows: Among them, F is the fitness value, ε is the space utilization weight coefficient, V i is the volume of the i-th cargo, V total is the total available volume of the cargo hold, σ is the penalty coefficient for center of gravity offset, Δx is the offset of the cargo center of gravity in the x-axis direction, Δy is the offset of the cargo center of gravity in the y-axis direction, and n is the total number of cargoes in the current cargo hold.

6. The method for optimizing intelligent cargo loading and unloading of ships according to claim 4, characterized in that: The method for constructing the cargo optimization model is: Extracting ship loading plan information from a ship management database, wherein the ship loading plan information includes cargo capacity data of each cargo hold, a deck load distribution diagram, and the coordinates of securing point positions; Spatially matching the cargo position coordinate data with the cabin boundary coordinates of the ship loading plan information to determine the target cargo hold number to which each cargo belongs; According to the cargo type code and weight level, query the cargo attribute knowledge base to obtain the corresponding stacking constraint rules; Using the target cargo hold number as an index, establish a cargo hold-cargo association matrix, where rows of the cargo hold-cargo association matrix correspond to cargo hold grid cells, and columns store cargo IDs, sizes, and stacking constraints; Calculating the current cumulative load of each cargo hold grid cell based on the grid code of the deck load distribution diagram; The cargo hold-cargo association matrix, stacking constraint rules, and securing point position coordinates are integrated to construct a cargo optimization model with cargo hold space utilization and center of gravity balance coefficient as optimization objectives; The cargo optimization model is verified for solvability. When there are cargoes that cannot satisfy the stacking constraints, an adjustment suggestion list including the conflicting cargo IDs and constraint types is generated and fed back to the ship cargo management system.

7. The method for optimizing intelligent cargo loading and unloading of ships according to claim 6, characterized in that: The method for generating the cargo feature dataset is as follows: Install multiple cameras in the cargo loading and unloading area of ​​the ship to collect real-time video data and multi-view image data of the operation area; Inputting the real-time video data into a deep learning model to extract cargo image features, wherein the deep learning model is designed based on an improved ResNet-101 network structure; Matching the cargo image features with a preset cargo type feature library to obtain cargo type information; Based on the multi-view image data, obtaining cargo size data through a three-dimensional reconstruction algorithm; Based on the feature point information extracted from the multi-view image data, a visual positioning algorithm is used to obtain the coordinate data of the cargo position; The cargo type information, the cargo size data and the cargo location coordinate data are integrated into a cargo feature data set, wherein the cargo feature data set includes cargo location coordinate data, cargo type code and weight grade.

8. A ship intelligent cargo loading and unloading optimization system, based on the ship intelligent cargo loading and unloading optimization method according to any one of claims 1 to 7, characterized in that: include, The recognition module is used to process real-time video data collected from the ship's cargo loading and unloading area through a deep learning model to identify the cargo type, size, and current location information, and generate a cargo feature dataset; A construction module, which constructs a cargo optimization model based on the cargo feature dataset and the ship loading drawing information; a generation module, configured to solve the cargo optimization model using an improved ant colony algorithm to generate a cargo loading and unloading operation instruction sequence, wherein the cargo loading and unloading operation instruction sequence includes cargo loading and unloading priorities, operation equipment allocation plans, and optimal loading and unloading path planning; An execution module is used to transmit the cargo loading and unloading operation instruction sequence to the terminal automated loading and unloading equipment to execute the loading and unloading operation, and simultaneously collect the operation status data frames during the loading and unloading process in real time; The updating module is used to analyze the deviation between the operation status data frame and the optimal loading and unloading path planning, and when the deviation exceeds a preset threshold, re-solve the cargo optimization model to generate an updated cargo loading and unloading operation instruction sequence.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the ship intelligent cargo loading and unloading optimization method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the ship intelligent cargo loading and unloading optimization method according to any one of claims 1 to 7 are implemented.

Images