Ship voyage load reverse method, electronic device and storage medium

Abstract

Embodiments of the present disclosure provide a ship voyage load inversion method, an electronic device and a storage medium. The ship voyage load inversion method comprises the following steps: constructing a voyage trajectory and a draft sequence set according to AIS data, archive data and berthing data; determining a draft threshold corresponding to a target ship for each voyage section; determining a loading state of the target ship according to a comparison result of a target draft observation value and the draft threshold; determining a load inversion model according to the loading state for each voyage section; extracting size parameters of the target ship from the archive data; inputting the target draft observation value and the size parameters into the load inversion model to obtain a first predicted load interval of the target ship in the corresponding voyage section; and determining a second predicted load interval of the target ship in the voyage according to the first predicted load intervals of the plurality of voyage sections. Thus, the ship voyage load inversion is realized by splitting the voyage section prediction, and the accuracy and the fineness of the load inversion are improved.

Description

Technical Field

[0001] This disclosure relates to the field of marine technology, and in particular to a method for inverting the load of a ship voyage, electronic equipment, and storage medium. Background Technology

[0002] With the continued growth in demand for global shipping carbon emission accounting, port throughput statistics, and international trade flow monitoring, obtaining accurate ship cargo volume has become a crucial foundation for shipping digitalization and maritime regulation. However, existing cargo volume data mainly relies on port declarations, shipping company reports, or commercial databases. This type of data suffers from problems such as commercial sensitivity, coarse time scale, misalignment across ports, and inconsistency with the actual floating state of ships, making it difficult to support voyage-level carbon emission accounting and trade analysis.

[0003] With the global deployment of the AIS system, it has become possible to infer the physical state of ships from ship behavior data. However, existing methods for estimating ship cargo mass still suffer from poor accuracy and low precision, making it difficult to generate stable and reliable voyage-level cargo volumes. Summary of the Invention

[0004] The main objective of this disclosure is to provide a method for inverting ship voyage load, an electronic device, and a storage medium to solve the aforementioned problems.

[0005] To achieve the above objectives, a first aspect of this disclosure provides a method for inverting ship voyage load, comprising: Acquire the AIS data, file data, and call-to-call data of the target vessel, and construct a set of voyage trajectory and draft sequence based on the AIS data, the file data, and the call-to-call data. The set of voyage trajectory and draft sequence includes the trajectory corresponding to the voyage and multiple original draft observation values ​​in the trajectory. Through data cleaning, the target draft observation values ​​corresponding to each segment of the voyage are extracted from the set of voyage trajectories and draft sequences. For each segment, the draft threshold corresponding to the target vessel is determined. Based on the comparison between the observed target draft and the draft threshold, the loading status of the target vessel is determined. The loading status is used to indicate whether the target vessel is in a ballast state, a partially loaded state, or a fully loaded state in the corresponding segment. For each voyage segment, a load inversion model is determined based on the loading status. The size parameters of the target vessel are extracted from the archive data. The observed draft of the target vessel and the size parameters are input into the load inversion model to obtain the first predicted load range of the target vessel in the corresponding voyage segment. Based on the first predicted load range of multiple voyage segments, the second predicted load range of the target vessel in the voyage segment is determined. The draft threshold includes a ballast draft threshold and a loading draft threshold. Determining the draft threshold corresponding to the target vessel includes: Collect a set of draft samples corresponding to the ship category to which the target ship belongs, and use a Gaussian kernel function to estimate the kernel density of the draft sample set to obtain a draft threshold prediction model based on kernel density estimation. The original draft observation values ​​corresponding to the voyage segment are input into the draft threshold prediction model for prediction, so as to obtain the ballast draft threshold and the loading draft threshold corresponding to the target vessel.

[0006] In some embodiments, inputting the target draft observation value and the size parameters into the load inversion model to obtain the first predicted load range of the target vessel in the corresponding segment includes: The target draft observation value and the size parameters are input into the load inversion model to obtain the initial load range of the target ship in the corresponding segment; Physical and logical constraints are applied to the initial load range to obtain the first predicted load range of the target vessel in the corresponding voyage segment. The physical constraints are used to ensure that the extreme value of the initial load range does not exceed a preset deadweight tonnage and is not less than zero. The logical constraints are used to mark the initial load range as zero or invalid when the loading state is ballast state.

[0007] In some embodiments, the load inversion model is a displacement model, and the step of inputting the target draft observation value and the size parameters into the load inversion model to obtain the initial load range of the target vessel in the corresponding segment includes: The target draft observation value and the size parameters are input into the load inversion model to obtain the displacement of the target ship in the corresponding section; Extract the rated power of the main engine and the fuel consumption rate of the target vessel from the archive data, calculate the sailing time of the target vessel in the segment, estimate the average speed of the target vessel in the segment, determine the actual power of the main engine based on the rated power of the main engine and the average speed, calculate the fuel consumption based on the product of the actual power of the main engine, the fuel consumption rate and the sailing time, and obtain the remaining fuel mass of the segment based on the difference between the initial fuel mass of the target vessel at departure and the fuel consumption. By subtracting the light load mass and the remaining fuel mass from the displacement, the initial load range of the target vessel in the corresponding voyage segment is obtained.

[0008] In some embodiments, constructing a set of voyage trajectory and draft sequence based on the AIS data, the archive data, and the docking data includes: The AIS data is sorted by timestamp, and data points in reverse chronological order, duplicate data points, and drifting data points are removed. The maximum design speed of the target vessel is extracted from the archive data, and unreasonable speed data points are removed based on the maximum design speed to obtain cleaned data points. The cleaned data points are then corrected for temporal continuity to form the ship's navigation trajectory; The vessel's navigation trajectory is segmented based on the berthing data to obtain the trajectory corresponding to each voyage. A set of voyage trajectory and draft sequence is constructed based on the trajectory corresponding to each voyage and the original draft observation values ​​in the AIS data.

[0009] In some embodiments, the method further includes: When the maximum design speed is missing from the archive data, the ship type and deadweight tonnage of the target ship are extracted from the archive data; The maximum design speed is interpolated based on the ship type and the deadweight tonnage.

[0010] In some embodiments, the step of extracting the target draft observation values ​​corresponding to each segment of the voyage from the voyage trajectory and draft sequence set through data cleaning includes: For each segment of the voyage, the corresponding draft time series is extracted from the voyage trajectory and draft sequence set, and the units of the draft time series are unified. The draft time series includes multiple original draft observations arranged according to the collection time. In the time series of drafts after unit unification, the original draft observations with abnormal magnitudes are removed; In the draft time series after removing the abnormal original draft observations, for data points with missing time or short-term abnormal data points, at least one of forward imputation, backward imputation or interpolation is used to eliminate data point jumps and data point breaks, resulting in a cleaned draft time series. In the post-cleaning draft time series, the original draft observation value that appears most frequently is determined as the target draft observation value corresponding to the current flight segment.

[0011] In some embodiments, determining the load inversion model based on the loading state includes: When the loading state is a partially loaded state, the displacement model is determined as the load inversion model; When the loading state is full load, the linear model is determined as the load inversion model.

[0012] To achieve the above objectives, a second aspect of this disclosure provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the ship voyage load inversion method described in the first aspect embodiment.

[0013] To achieve the above objectives, a third aspect of this disclosure provides a storage medium, which is a computer-readable storage medium storing a computer program that, when executed by a processor, implements the ship voyage load inversion method described in the first aspect embodiment.

[0014] The beneficial effects of the embodiments disclosed herein include: This embodiment acquires the AIS data, archive data, and call-in data of the target vessel. Based on this data, a voyage trajectory and draft sequence set is constructed. This set includes the trajectory corresponding to the voyage and multiple original draft observations within the trajectory. Data cleaning is then performed to extract the target draft observations for each segment of the voyage from the voyage trajectory and draft sequence set, enhancing the robustness of the results. Next, for each segment, a draft threshold corresponding to the target vessel is determined. Based on the comparison between the target draft observations and the draft threshold, the loading status of the target vessel is determined. The loading status indicates whether the target vessel is in a ballast, partially loaded, or fully loaded state in the corresponding segment. Determining the loading status avoids introducing systematic errors. Finally, for each segment, a load inversion model is determined based on the loading status. The target vessel's dimensional parameters are extracted from the archive data. The target draft observations and dimensional parameters are input into the load inversion model to obtain the first predicted load range for the target vessel in the corresponding segment. Based on multiple... The first predicted load range for each voyage segment is determined to identify the second predicted load range for the target vessel within the voyage. This segment-based prediction enables voyage-level load inversion, improving the accuracy and precision of the load inversion. Furthermore, by collecting a set of draft samples corresponding to the vessel category of the target vessel and using a Gaussian kernel function to estimate the kernel density of the draft samples, a draft threshold prediction model based on kernel density estimation is obtained. Multiple original draft observations corresponding to each voyage segment are input into the draft threshold prediction model for prediction, yielding the ballast draft threshold and loading draft threshold for the target vessel, thus improving the accuracy of loading status identification. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the overall process of the ship voyage load inversion method provided in this embodiment of the disclosure; Figure 2 This is a flowchart further included in step S101; Figure 3 This is a flowchart further included in step S102; Figure 4 This is a flowchart further included in step S103; Figure 5 This is a flowchart further included in step S104; Figure 6 This is a schematic diagram of the structure of the ship voyage load inversion device provided in this embodiment of the disclosure; Figure 7 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of this disclosure. Detailed Implementation

[0016] The accompanying drawings in the embodiments clearly and completely describe the technical solutions in the embodiments of this disclosure. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments in this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0017] It is understood that when the above embodiments of this disclosure are applied to specific products or technologies, permission or consent from the subject is required, and the collection, use and processing of related data must comply with relevant laws, regulations and standards.

[0018] Furthermore, when this embodiment of the disclosure needs to retrieve relevant data, it will obtain separate permission or separate consent for the relevant data through pop-up windows or redirection to a confirmation page. After clearly obtaining separate permission or separate consent for the relevant data, it will then obtain the necessary relevant data for enabling this embodiment of the disclosure to operate normally.

[0019] In this disclosure, the terms "module" or "unit" refer to a computer program or part of a computer program that has a predetermined function and works with other related parts to achieve a predetermined goal, and can be implemented wholly or partially using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that includes the functionality of that module or unit.

[0020] Reference Figure 1 , Figure 1 This is a schematic diagram of the overall process of the ship voyage load inversion method provided in this embodiment. The ship voyage load inversion method can be executed by an electronic device, which can be a terminal or a server. The ship voyage load inversion method provided in this embodiment includes steps S101 to S104: Step S101: Obtain the AIS data, file data, and call-to-call data of the target vessel, and construct a set of voyage trajectory and draft sequence based on the AIS data, file data, and call-to-call data; The target vessel can be any transport vessel requiring load inversion, including but not limited to bulk carriers, container ships, tankers, general cargo ships, LNG carriers, and LNG carriers. The method in this disclosure embodiment has no special restrictions on ship type; it is applicable as long as the vessel is equipped with AIS equipment and broadcasts data normally.

[0021] AIS data can be obtained from the International Maritime Organization's global AIS database, public data platforms of national maritime authorities, or third-party commercial data service providers. Archival data can be obtained from classification societies, ship registration agencies, shipyards, or third-party ship databases. Archival data mainly includes static parameters such as ship type, deadweight tonnage, gross tonnage, net tonnage, length, beam, depth, design draft, block coefficient, light load factor, year of construction, main engine rated power, and port of registry. Call-in data can be obtained from port authorities, shipping agencies, liner company websites, or third-party shipping data platforms. Call-in data mainly includes information such as port name, UN port code, berthing time, departure time, berth number, and cargo type.

[0022] The voyage trajectory and draft sequence set includes the trajectory corresponding to the voyage and multiple original draft observations within the trajectory. In this embodiment, the trajectory corresponding to the voyage includes multiple data points collected at different times, each data point corresponding to an original draft observation and other relevant parameters. The voyage trajectory and draft sequence set refers to the sequence of trajectory points and the corresponding draft observation sequence for each voyage after dividing the continuous ship navigation trajectory according to voyages. A voyage refers to the complete navigation process of a ship from leaving one port to arriving at the next port. Limiting the load inversion to the voyage scale avoids drastic fluctuations in load results caused by point-by-point estimation, ensuring semantic consistency and stability of the results.

[0023] Step S102: Through data cleaning, extract the target draft observation values ​​corresponding to each segment of the voyage from the voyage trajectory and draft sequence set; Data cleaning refers to the preprocessing of raw draft observations from the voyage trajectory and draft sequence set to eliminate noise such as outliers, missing values, and jump values ​​in the AIS draft data, thereby improving data quality. AIS draft data is automatically collected by the ship's draft sensors and broadcast through the AIS system. Due to sensor malfunctions, signal interference, data transmission errors, and ship reporting errors, AIS draft data may contain a significant amount of noise. Directly using this noisy draft data for loading status identification and load inversion will lead to inaccurate results. Therefore, data cleaning preprocesses the raw draft observations to extract target draft observations that accurately reflect the ship's buoyancy.

[0024] Meanwhile, in order to improve the level of precision, the voyage is divided into multiple segments. A segment refers to the time period during which the target vessel continuously maintains a certain state, such as when the target vessel is sailing at sea or loading and unloading cargo in a port. Data cleaning is performed on each segment to obtain the target draft observation value corresponding to each segment.

[0025] Step S103: For each voyage segment, determine the draft threshold corresponding to the target vessel, and determine the loading status of the target vessel based on the comparison between the observed draft value and the draft threshold.

[0026] The loading status refers to the cargo-carrying condition of a vessel during navigation, indicating whether the target vessel is in a ballast state, partially loaded state, or fully loaded state in the corresponding segment of the voyage. Ballast state means that the vessel is not carrying cargo or only carrying a small amount of cargo, and ballast water is injected into the ballast tanks to ensure navigational stability; at this time, the vessel's draft is relatively shallow. Partially loaded state means that the vessel is carrying some cargo, and the draft is between the ballast draft and the full load draft. Full load state means that the vessel is carrying the rated maximum cargo capacity, and the draft reaches or is close to the design draft.

[0027] It is important to emphasize that under ballast conditions, the cargo mass is almost zero. If the buoyancy balance formula is used directly to calculate the cargo load under these conditions, it will lead to systematic errors in the results. Therefore, the embodiments of this disclosure first identify the ship's loading state and only perform load inversion on segments that are partially loaded or fully loaded, thereby avoiding miscalculations of ballast voyages.

[0028] Draft thresholds are critical draft values ​​used to distinguish different loading states, including ballast draft thresholds and loading draft thresholds. Specifically, the ballast draft threshold distinguishes between ballast and partially loaded states; when the ship's draft is less than the ballast draft threshold, it is determined to be in ballast state. The loading draft threshold distinguishes between partially loaded and fully loaded states; when the ship's draft is greater than the loading draft threshold, it is determined to be in fully loaded state; when the ship's draft is between the ballast and loading draft thresholds, it is determined to be in partially loaded state. Draft thresholds are related to parameters such as the ship's type, deadweight tonnage, and design draft. Different types and tonnages of ships have different ballast and fully loaded drafts. Therefore, this embodiment of the disclosure determines the corresponding draft threshold based on the ship's type and load range for different ships, thereby improving the accuracy of loading state identification.

[0029] Step S104: For each voyage segment, determine the load inversion model based on the loading status, extract the size parameters of the target vessel from the archive data, input the target draft observation value and size parameters into the load inversion model, obtain the first predicted load range of the target vessel in the corresponding voyage segment, and determine the second predicted load range of the target vessel in the voyage based on the first predicted load range of multiple voyage segments.

[0030] The load inversion model is a mathematical model used to calculate the ship's load based on its draft and dimensional parameters. The relationship between the ship's load and draft differs under different loading conditions, therefore different load inversion models can be used. This disclosure employs a displacement model and a linear model for partially loaded and fully loaded states, respectively, to improve the accuracy of the load inversion.

[0031] The first predicted load range refers to the cargo load range of the vessel within a single voyage segment, while the second predicted load range refers to the cargo load range of the vessel throughout the entire voyage. Because vessels consume fuel and other resources during navigation, or perform loading and unloading operations, their draft and load capacity change. Therefore, the first predicted load range may differ across different voyage segments. This embodiment of the present disclosure determines the second predicted load range for the entire voyage by combining the first predicted load ranges of multiple voyage segments, thereby further improving the stability and reliability of the results.

[0032] Specifically, the maximum value among the right endpoints of the first predicted load interval for all segments can be used as the right endpoint of the second predicted load interval, and the minimum value among the left endpoints of the first predicted load interval for all segments can be used as the left endpoint of the second predicted load interval.

[0033] By acquiring the AIS data, archive data, and call-to-call data of the target vessel, a set of voyage trajectory and draft sequence is constructed based on the AIS data, archive data, and call-to-call data. The set of voyage trajectory and draft sequence includes the trajectory corresponding to the voyage and multiple original draft observation values ​​in the trajectory. Then, through data cleaning, the target draft observation values ​​corresponding to each segment of the voyage are extracted from the set of voyage trajectory and draft sequence, which can enhance the robustness of the results. Next, for each voyage segment, the corresponding draft threshold for the target vessel is determined. Based on the comparison between the observed target draft and the draft threshold, the loading status of the target vessel is determined. The loading status indicates whether the target vessel is in a ballast state, a partially loaded state, or a fully loaded state in the corresponding voyage segment. By determining the loading status, systematic errors can be avoided. Finally, for each voyage segment, a load inversion model is determined based on the loading status. The dimensional parameters of the target vessel are extracted from the archive data. The observed target draft and dimensional parameters are input into the load inversion model to obtain the first predicted load range of the target vessel in the corresponding voyage segment. Based on the first predicted load ranges of multiple voyage segments, the second predicted load range of the target vessel in the voyage is determined. Thus, by splitting the voyage segment prediction, voyage-level load inversion is achieved, improving the accuracy and precision of the load inversion.

[0034] In some embodiments, refer to Figure 2 , Figure 2 This is a flowchart further included in step S101. In some embodiments, the process of constructing a set of voyage tracks and draft sequences based on AIS data, archive data, and docking data specifically includes steps S201 to S203: Step S201: Sort the AIS data according to timestamp, remove data points in reverse time order, duplicate data points and drifting data points, and extract the maximum design speed of the target ship from the archive data. Remove unreasonable speed data points based on the maximum design speed to obtain cleaned data points.

[0035] Step S202: Perform time continuity correction on the cleaned data points to form the ship's navigation trajectory.

[0036] Step S203: The ship's navigation trajectory is segmented according to the berthing data to obtain the trajectory corresponding to multiple voyages. Based on the trajectory corresponding to multiple voyages and the original draft observation values ​​in the AIS data, a set of voyage trajectory and draft sequence is constructed.

[0037] Regarding step S201 above, during the collection and transmission of AIS data, problems such as reversed time sequence, duplicate data, trajectory drift, and abnormal flight speed may occur. Therefore, it is necessary to perform preliminary cleaning of the AIS data first.

[0038] First, all AIS data points for the same vessel are sorted in ascending order by timestamp, ensuring the data is arranged chronologically. During the sorting process, any data points in reverse chronological order are checked and removed; that is, data points whose timestamps are earlier than those of previous data points are discarded.

[0039] Secondly, duplicate data points, drifting data points, and data points with unreasonable speeds are removed. Duplicate data points refer to multiple AIS data points generated by the same vessel at the same timestamp; for duplicate data points, only one data point is retained. Drifting data points refer to data points where the vessel's position suddenly jumps significantly, clearly inconsistent with the physical laws of ship navigation. For example, the spatial distance and time interval between two adjacent data points can be calculated, and then the vessel's average speed during this period can be calculated based on the spatial distance and time interval. If the calculated average speed exceeds the target vessel's maximum design speed by more than a certain threshold, the latter data point is determined to be a drifting data point. Unreasonable speed data points refer to data points in the AIS data where the reported speed value clearly does not conform to the actual situation.

[0040] It should be noted that while removing the above-mentioned abnormal data points, the original draft observation values ​​corresponding to these data points were also removed simultaneously to reduce significant noise interference.

[0041] In some embodiments, when the maximum design speed is missing from the archive data, it can be interpolated. Specifically, the ship type and deadweight tonnage of the target vessel can be extracted from the archive data, and the maximum design speed can be interpolated based on the ship type and deadweight tonnage.

[0042] For example, when interpolating the maximum design speed, the average maximum design speed of ships of different types and deadweight tons can be calculated, and then the average speed can be used as the interpolation value for the missing data.

[0043] Alternatively, a machine learning model can be used to interpolate the maximum design speed. Specifically, all ship records with complete maximum design speed data can be extracted from a ship archive database as training samples. Each training sample's features include ship type, deadweight tonnage, gross tonnage, net tonnage, length, beam, depth, year of construction, and main engine rated power, labeled as maximum design speed. The features are then standardized, converting features of different dimensions to a uniform scale. Categorical features (such as ship type) are one-hot encoded and converted to numerical features. The training dataset is then input into the machine learning model for training, adjusting the model's hyperparameters, such as the number of decision trees, tree depth, and minimum number of samples required for node splits, to minimize the model's prediction error. The trained model is evaluated using a test dataset, calculating metrics such as mean squared error and mean absolute error to verify the model's prediction accuracy. When interpolating the maximum design speed, the target ship's ship type, deadweight tonnage, and year of construction are input into the trained machine learning model; the model's output prediction is the interpolated maximum design speed.

[0044] By introducing a ship technical feature interpolation mechanism, the usability of load inversion can be improved.

[0045] Regarding step S202 above, after the initial cleaning of AIS data, it is necessary to perform time continuity correction on the cleaned data points to form a continuous and smooth ship navigation trajectory. Since the update frequency of AIS data is not fixed, the time interval between adjacent data points may range from a few seconds to tens of minutes, resulting in a discontinuous trajectory in time. Specifically, the time interval between adjacent data points can be checked. If the time interval is less than a preset minimum time interval, the timestamp of the subsequent data point is considered to have an error, and its timestamp is corrected to the timestamp of the preceding data point plus the minimum time interval. For example, the minimum time interval can be 1 second. Through time continuity correction, timestamp errors in the AIS data can be eliminated, making the trajectory more continuous and reasonable in time.

[0046] Regarding step S203 above, voyage segmentation refers to dividing a continuous ship navigation trajectory into multiple independent voyages according to the process from the port of origin to the port of destination. Specifically, all berthing and departure records of the target vessel are extracted from the berthing data and arranged in chronological order to form a berthing event sequence. Each berthing event includes berthing time and departure time.

[0047] For two adjacent berthing events, the navigation trajectory between the departure time of the first berthing event and the berthing time of the second berthing event constitutes a complete voyage. In cases where berthing data is missing, voyages are segmented using vessel status information from AIS data. When a vessel's AIS status changes from "on board" to "anchored" or "berthed," and the duration exceeds a preset berthing time threshold, it is considered a berthing event; when the vessel's AIS status changes from "anchored" or "berthed" to "on board," it is considered a departure event. The navigation trajectory between adjacent departure and berthing events constitutes a voyage. Then, a unique voyage identifier is assigned to each voyage, which can be generated by combining the vessel's IMO number, departure time, and arrival time. After voyage segmentation, the trajectory point sequence corresponding to each voyage and the corresponding original draft observation sequence are combined to form a voyage trajectory and draft sequence set. Each voyage track and draft sequence can specifically include the following information: voyage identifier, vessel IMO number, port of departure, port of arrival, departure time, arrival time, list of track points, original draft observation value, and timestamp corresponding to the original draft observation value.

[0048] By using the above-mentioned cruise segmentation method, discrete AIS data points can be constructed into a set of cruise trajectories and draft sequences with clear semantics, so that subsequent load inversion can be carried out within a unified cruise scale, avoiding load fluctuations and semantic inconsistencies caused by point-by-point estimation.

[0049] In some embodiments, refer to Figure 3 , Figure 3 This is a flowchart further included in step S102. In some embodiments, the process of extracting the target draft observation values ​​corresponding to each segment of the voyage from the voyage trajectory and draft sequence set through data cleaning specifically includes steps S301 to S304: Step S301: For each segment of the voyage, extract the corresponding draft time series from the voyage trajectory and draft sequence set, and unify the units of the draft time series.

[0050] Step S302: In the unified draft time series, remove the original draft observations with abnormal size.

[0051] Step S303: In the draft time series after removing abnormal original draft observations, for data points with missing time or short-term abnormal data points, at least one of forward imputation, backward imputation or interpolation is used to eliminate data point jumps and data point breaks, so as to obtain the cleaned draft time series.

[0052] Step S304: In the post-cleaning draft time series, the original draft observation value with the highest frequency is determined as the target draft observation value corresponding to the current flight segment.

[0053] Regarding step S301 above, the draft time series includes multiple raw draft observations arranged according to the acquisition time. Since the units of AIS draft data reported by different vessels may not be consistent, it is necessary to standardize the units of the draft time series to avoid affecting the accuracy of load inversion.

[0054] Regarding step S302 above, after unifying the units, it is necessary to remove raw draft observations that are abnormal in size. Abnormal raw draft observations refer to draft values ​​that significantly exceed the ship's reasonable draft range. The reasonable draft range for a ship can be from 0 meters to its molded depth. By removing raw draft observations that are abnormal in size, obvious errors in the AIS draft data can be eliminated, improving data quality.

[0055] Regarding step S303 above, after removing the original draft observations with significant anomalies, there may still be some missing time data points and short-term anomalous data points in the draft time series. Missing time data points refer to data points where there is no corresponding draft observation at a certain time point, forming a data breakpoint; short-term anomalous data points refer to draft values ​​that change drastically within a short period of time, which does not conform to the physical laws of ship buoyancy changes.

[0056] For example, for time-missing data points and short-term outlier data points, the point can be filled with the most recent valid draft value before the missing point or outlier, or with the most recent valid draft value after the missing point or outlier, or with a linear interpolation of the two valid draft values ​​before and after the missing point or outlier.

[0057] It should be noted that, in the draft data repair of this embodiment, motion characteristic information of each segment within the voyage can be further incorporated, including speed level, duration, and whether the vessel is in a state of maritime navigation. By comparing the similarity of dynamic characteristics between adjacent segments, the actual operating conditions under which the abnormal draft segment is more likely to continue can be determined, thereby completing the correction of the abnormal draft segment. For example, if the vessel is in a stable state of maritime navigation for a period of time, and its speed and course do not change significantly, then its draft value should also remain relatively stable. If a short-term draft jump occurs at this time, it can be determined as an abnormal value, and the stable draft values ​​before and after the jump can be used for repair.

[0058] The above data repair method can effectively eliminate jumps and breakpoints in the draft time series, resulting in a continuous and smooth post-cleaning draft time series.

[0059] Regarding step S304 above, the target draft observation value is an observation value that can characterize the overall floating state of the segment. In this embodiment, the mode draft is used as the target draft observation value corresponding to the segment. The mode draft refers to the draft observation value that appears most frequently in the post-cleaning draft time series. The mode draft is not affected by extreme values ​​and can reflect the main floating state of the ship within the segment. Even if there is a small amount of noise that has not been cleaned up in the draft time series, the mode draft can still remain stable. Specifically, the draft values ​​in the post-cleaning draft time series can be discretized according to a preset precision, that is, the draft values ​​are rounded to one decimal place, the frequency of each discretized draft value is counted, and the draft observation value with the highest frequency is selected as the mode draft of the segment, i.e., the target draft observation value. If multiple draft observation values ​​have the same frequency and are all the highest, the average of these draft observation values ​​is taken as the target draft observation value. By using the mode draft as the target draft observation value, random fluctuations in draft data can be effectively suppressed, accurately reflecting the stable floating state of the ship within the voyage.

[0060] In some embodiments, refer to Figure 4 , Figure 4 This is a flowchart further included in step S103. The draft threshold includes the ballast draft threshold and the loading draft threshold. The process of determining the draft threshold corresponding to the target vessel specifically includes steps S401 to S402: Step S401: Extract the target vessel's ship type and deadweight tonnage from the archive data, and determine the draft threshold prediction model based on the ship type and deadweight tonnage.

[0061] Step S402: Input multiple original draft observation values ​​corresponding to the voyage segment into the draft threshold prediction model for prediction, and obtain the ballast draft threshold and loading draft threshold corresponding to the target vessel.

[0062] Regarding step S401 above, firstly, ships can be classified into different categories based on their ship type and deadweight tonnage. For example, ship types are mainly divided into container ships, bulk carriers, oil tankers, liquefied gas carriers, and general cargo ships. Then, for each ship category, a large amount of historical draft data and corresponding loading status data of ships within that category are collected to construct a training dataset. Each training sample contains the ship's design draft, multiple historical draft observations, and corresponding loading status labels, including ballast, partially loaded, and fully loaded. Finally, based on the training dataset for each category, a corresponding draft threshold prediction model is trained.

[0063] For example, embodiments of this disclosure may employ a draft threshold prediction model based on kernel density estimation. Kernel density estimation is a non-parametric statistical method that can estimate the probability density function of the population based on sample data. It does not require any assumptions about the distribution of the data and has strong adaptability and robustness. The construction process of the draft threshold prediction model based on kernel density estimation is as follows: For a training dataset of a specific ship category, historical draft observations of all ships are collected to form a draft sample set. ,in Let be the i-th historical draft observation, and n be the number of samples.

[0064] A Gaussian kernel function is used to estimate the kernel density of the draft sample set, resulting in the probability density function of the draft. : ; in, Let be the independent variable of the probability density function. For Gaussian kernel function, This is the bandwidth parameter. The expression for the Gaussian kernel function is: , = ; in, It is a natural constant.

[0065] Specifically, the probability density function of draft is obtained. Then, the probability density function is analyzed. The peak distribution of draft. In the scenario of this embodiment, the probability density function of draft can exhibit a trimodal distribution, where the left peak corresponds to the ballast state, the middle peak corresponds to the partially loaded state, and the right peak corresponds to the fully loaded state. Then, to improve accuracy, the draft threshold can be determined based on the valley value of the probability density function. The draft value corresponding to the valley value between two peaks is the threshold for distinguishing between the two states. For example, the draft value corresponding to the valley value between the peak value of the ballast state and the peak value of the partially loaded state is the ballast draft threshold; the draft value corresponding to the valley value between the peak value of the partially loaded state and the peak value of the fully loaded state is the loading draft threshold.

[0066] Regarding step S402 above, after determining the draft threshold prediction model, multiple original draft observation values ​​corresponding to the voyage segment are input into the model to predict the ballast draft threshold and loading draft threshold corresponding to the target vessel.

[0067] Specifically, all original draft observations corresponding to the voyage segment are extracted from the voyage trajectory and draft sequence set to form a draft sample set for that segment. This sample set is then preliminarily cleaned to remove obvious outliers, resulting in a cleaned draft sample set. The cleaned draft sample set is then input into the corresponding draft threshold prediction model. Based on the draft distribution characteristics of the voyage segment and historical data for that vessel type, the model predicts the ballast draft threshold and loading draft threshold for that segment.

[0068] In addition, in some embodiments, since the draft of similar vessels may vary, to further improve the accuracy of the draft threshold, the draft threshold prediction model can extract draft observation data from multiple historical voyages of the target vessel. For each segment of the draft observation data from multiple historical voyages, the mode draft observation value corresponding to that segment is extracted, and a draft distribution histogram of the target vessel is generated based on multiple mode draft observation values. Then, the ballast peak and full load peak in the draft distribution histogram are identified. Based on the ballast peak and full load peak, the general ballast draft threshold and the general loading draft threshold of the target vessel are determined. The first difference between the general ballast draft threshold and the ballast draft threshold output by the draft threshold prediction model is determined. The initial ballast correction coefficient is determined based on the ratio of the first difference to the ballast draft threshold output by the draft threshold prediction model, which can be specifically expressed as: ; This is the initial ballast correction factor. For general ballast draft thresholds, This refers to the ballast draft threshold.

[0069] The weighting adjustment factor is determined based on the number of historical voyages. The target ballast correction factor is obtained by multiplying the weighting adjustment factor by the initial ballast correction factor, which can be expressed as follows: ; in, The target ballast correction factor. The weighting adjustment factor increases with the number of historical voyages.

[0070] Finally, the ballast draft threshold output by the draft threshold prediction model is corrected according to the target ballast correction factor to obtain the final ballast draft threshold, which can be expressed as follows: ; in, This is the final ballast draft threshold.

[0071] Similarly, a second difference is determined between the general loading draft threshold and the loading draft threshold output by the draft threshold prediction model. An initial loading correction coefficient is determined based on the ratio of this second difference to the loading draft threshold output by the draft threshold prediction model. A target loading correction coefficient is obtained by multiplying the weight adjustment factor by the initial loading correction coefficient. The loading draft threshold output by the draft threshold prediction model is then corrected based on the target loading correction coefficient to obtain the final loading draft threshold. The relevant calculation formulas are similar to those used for correcting the ballast draft threshold and will not be repeated here.

[0072] In some embodiments, refer to Figure 5 , Figure 5 This is a flowchart further included in step S104. The process of inputting the target draft observations and dimensional parameters into the load inversion model to obtain the first predicted load range of the target vessel in the corresponding voyage segment specifically includes steps S501 to S502: Step S501: Input the target draft observation value and size parameters into the load inversion model to obtain the initial load range of the target ship in the corresponding voyage segment.

[0073] Step S502: Apply physical and logical constraints to the initial load range to obtain the first predicted load range of the target vessel in the corresponding voyage segment.

[0074] Regarding step S501 above, when the loading state is partially loaded, the displacement model is determined as the load inversion model; when the loading state is fully loaded, the linear model is determined as the load inversion model.

[0075] The displacement model is based on Archimedes' principle of buoyancy. The displacement of the corresponding segment is determined by the target draft observation and size parameters. Then, the light load mass and the remaining fuel mass are deducted. Since the remaining fuel mass is a range, the initial load range of the target ship in the corresponding segment can be obtained.

[0076] The basic idea of ​​the linear model is that there is a linear relationship between the ship's load and draft. The light-load design draft and full-load design draft of the target ship are determined by the size parameters. Then, the target proportion is calculated by the ratio of the target draft observation value to the light-load design draft and the full-load design draft. The initial load range can be obtained by multiplying the target proportion by the light-load mass and the full-load mass.

[0077] In addition, the displacement model can be used as the main model, applicable to effective loading segments other than ballast conditions; the linear model can be used as a substitute model when the block coefficient, depth or hydrostatic parameters are missing, or for empirical correction near the full load condition.

[0078] Regarding step S502 above, physical constraints are used to ensure that the extreme value of the initial load range does not exceed the preset load tonnage and is not less than zero, while logical constraints are used to mark the initial load range as zero or invalid when the loading state is ballast state.

[0079] If the upper limit of the initial load range is greater than the preset deadweight tonnage, the upper limit is adjusted to the preset deadweight tonnage; if the lower limit of the initial load range is less than 0, the lower limit is adjusted to 0. Furthermore, since the ship is not carrying cargo or only carrying a small amount of cargo under ballast conditions, the cargo mass is almost zero, making load inversion meaningless. Therefore, when the ship is determined to be under ballast conditions, the initial load range for that segment is directly marked as zero or invalid. Invalid means that no specific load value is output, applicable to scenarios where the research objective is cargo transport volume; this segment does not participate in cargo load inversion.

[0080] The remaining fuel mass varies with the ship's voyage, and the displacement deducts the remaining fuel mass at the start and end of the voyage. The remaining fuel mass is related to factors such as the ship's range, speed, and main engine rated power. The embodiments of this disclosure estimate the remaining fuel mass in the following way: The rated power of the main engine and fuel consumption rate of the target vessel are extracted from the archival data. The sailing time of the target vessel in this segment is calculated, i.e., the end time of the segment minus the start time. The average speed of the target vessel in this segment is estimated. The actual power of the main engine is determined based on the rated power of the main engine and the average speed. The fuel consumption is calculated by multiplying the actual power of the main engine, the fuel consumption rate, and the sailing time. The remaining fuel mass of the target vessel at departure is obtained by the difference between the remaining fuel mass and the fuel consumption. The remaining fuel mass at departure can be estimated based on the vessel's fuel tank capacity and empirical values. Generally, the remaining fuel mass at departure is approximately 80% to 90% of the fuel tank capacity. For vessels without fuel tank capacity data, it can be estimated based on the vessel's deadweight tonnage, which is generally approximately 5% to 10% of the deadweight tonnage. Estimating the remaining fuel mass using the above method and subtracting it from the displacement mass can further improve the accuracy of load inversion.

[0081] In some embodiments, refer to Figure 6 , Figure 6 This is a schematic diagram of the structure of the ship voyage load inversion device provided in this embodiment of the disclosure. The ship voyage load inversion device includes: The set construction module 601 is used to acquire the AIS data, file data and berthing data of the target vessel, and construct the voyage trajectory and draft sequence set based on the AIS data, file data and berthing data. The voyage trajectory and draft sequence set includes the trajectory corresponding to the voyage and multiple original draft observation values ​​in the trajectory. The target draft observation extraction module 602 is used to extract the target draft observation values ​​corresponding to each segment of the voyage from the voyage trajectory and draft sequence set through data cleaning. The loading status determination module 603 is used to determine the draft threshold corresponding to the target vessel for each segment, and to determine the loading status of the target vessel based on the comparison result between the observed draft value and the draft threshold. The loading status is used to indicate whether the target vessel is in ballast state, partially loaded state or fully loaded state in the corresponding segment. The load prediction module 604 is used to determine the load inversion model for each voyage segment based on the loading status, extract the size parameters of the target vessel from the archive data, input the target draft observation value and size parameters into the load inversion model, obtain the first predicted load range of the target vessel in the corresponding voyage segment, and determine the second predicted load range of the target vessel in the voyage based on the first predicted load range of multiple voyage segments.

[0082] The aforementioned ship voyage load inversion device and ship voyage load inversion method are based on the same inventive concept and have the same beneficial effects, and will not be elaborated further here.

[0083] This disclosure also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the aforementioned ship voyage load inversion method. This electronic device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0084] Please see Figure 7 , Figure 7 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this disclosure. The electronic device includes: The processor 701 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this disclosure. The memory 702 can be implemented as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 702 can store operating devices and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 702 and is called and executed by the processor 701 to execute the ship voyage load inversion method of the embodiments of this disclosure. The input / output interface 703 is used to implement information input and output; The communication interface 704 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). Bus 705 transmits information between various components of the device (e.g., processor 701, memory 702, input / output interface 703, and communication interface 704); The processor 701, memory 702, input / output interface 703, and communication interface 704 are connected to each other within the device via bus 705.

[0085] This disclosure also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described ship voyage load inversion method.

[0086] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0087] The embodiments described in this disclosure are for the purpose of more clearly illustrating the technical solutions of this disclosure and do not constitute a limitation on the technical solutions provided by this disclosure. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by this disclosure are also applicable to similar technical problems.

[0088] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this disclosure, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0089] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0090] Those skilled in the art will understand that all or some of the steps, apparatuses, or functional modules / units in the methods disclosed above can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0091] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in this disclosure and the foregoing drawings are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, apparatus, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0092] It should be understood that in this disclosure, "at least one item" means one or more, and "more than one" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0093] In the several embodiments provided in this disclosure, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0094] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0095] Furthermore, the functional units in the various embodiments of this disclosure can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0096] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this disclosure, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0097] The preferred embodiments of the present disclosure have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present disclosure. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and spirit of the present disclosure shall be within the scope of the claims of the present disclosure.

Claims

1. A method for inverting the load of a ship voyage, characterized in that, include: Acquire the AIS data, file data, and call-to-call data of the target vessel, and construct a set of voyage trajectory and draft sequence based on the AIS data, the file data, and the call-to-call data. The set of voyage trajectory and draft sequence includes the trajectory corresponding to the voyage and multiple original draft observation values ​​in the trajectory. Through data cleaning, the target draft observation values ​​corresponding to each segment of the voyage are extracted from the set of voyage trajectories and draft sequences. For each segment, the draft threshold corresponding to the target vessel is determined. Based on the comparison between the observed target draft and the draft threshold, the loading status of the target vessel is determined. The loading status is used to indicate whether the target vessel is in a ballast state, a partially loaded state, or a fully loaded state in the corresponding segment. For each voyage segment, a load inversion model is determined based on the loading status. The size parameters of the target vessel are extracted from the archive data. The observed draft of the target vessel and the size parameters are input into the load inversion model to obtain the first predicted load range of the target vessel in the corresponding voyage segment. Based on the first predicted load range of multiple voyage segments, the second predicted load range of the target vessel in the voyage segment is determined. The draft threshold includes a ballast draft threshold and a loading draft threshold. Determining the draft threshold corresponding to the target vessel includes: Collect a set of draft samples corresponding to the ship category to which the target ship belongs, and use a Gaussian kernel function to estimate the kernel density of the draft sample set to obtain a draft threshold prediction model based on kernel density estimation. The original draft observation values ​​corresponding to the voyage segment are input into the draft threshold prediction model for prediction, so as to obtain the ballast draft threshold and the loading draft threshold corresponding to the target vessel. The step of inputting the target draft observation value and the size parameters into the load inversion model to obtain the first predicted load range of the target vessel in the corresponding voyage segment includes: The target draft observation value and the size parameters are input into the load inversion model to obtain the initial load range of the target ship in the corresponding segment; Physical and logical constraints are applied to the initial load range to obtain the first predicted load range of the target vessel in the corresponding voyage segment. The physical constraints are used to ensure that the extreme value of the initial load range does not exceed a preset deadweight tonnage and is not less than zero. The logical constraints are used to mark the initial load range as zero or invalid when the loading state is ballast state. The step of constructing a set of voyage trajectory and draft sequence based on the AIS data, the archive data, and the docking data includes: The AIS data is sorted by timestamp, and data points in reverse chronological order, duplicate data points, and drifting data points are removed. The maximum design speed of the target vessel is extracted from the archive data, and unreasonable speed data points are removed based on the maximum design speed to obtain cleaned data points. The cleaned data points are then corrected for temporal continuity to form the ship's navigation trajectory; The ship's navigation trajectory is segmented based on the berthing data to obtain the trajectory corresponding to each voyage. A set of voyage trajectory and draft sequence is constructed based on the trajectory corresponding to each voyage and the original draft observation values ​​in the AIS data. The step of data cleaning, which involves extracting the target draft observations for each segment of the voyage from the set of voyage trajectories and draft sequences, includes: For each segment of the voyage, the corresponding draft time series is extracted from the voyage trajectory and draft sequence set, and the units of the draft time series are unified. The draft time series includes multiple original draft observations arranged according to the collection time. In the time series of drafts after unit unification, the original draft observations with abnormal magnitudes are removed; In the draft time series after removing the abnormal original draft observations, for data points with missing time or short-term abnormal data points, at least one of forward imputation, backward imputation, or interpolation is used to eliminate data point jumps and data point breaks, resulting in a cleaned draft time series. In the cleaned draft time series, the original draft observation with the highest frequency is determined as the target draft observation corresponding to the current flight segment.

2. The method for inverting ship voyage load according to claim 1, characterized in that, The load inversion model is a displacement model. The process of inputting the target draft observations and the dimensional parameters into the load inversion model to obtain the initial load range of the target vessel in the corresponding voyage segment includes: The target draft observation value and the size parameters are input into the load inversion model to obtain the displacement of the target ship in the corresponding section; The rated power of the main engine and the fuel consumption rate of the target vessel are extracted from the archived data. The sailing time of the target vessel in the segment is calculated, and the average speed of the target vessel in the segment is estimated. The actual power of the main engine is determined based on the rated power of the main engine and the average speed. The fuel consumption is calculated based on the product of the actual power of the main engine, the fuel consumption rate, and the sailing time. The remaining fuel mass of the segment is obtained based on the difference between the initial fuel mass of the target vessel at departure and the fuel consumption. The initial load range of the target vessel in the corresponding segment is obtained by deducting the light load mass and the remaining fuel mass from the displacement.

3. The method for inverting ship voyage load according to claim 1, characterized in that, The method further includes: When the maximum design speed is missing from the archive data, the ship type and deadweight tonnage of the target ship are extracted from the archive data; The maximum design speed is interpolated based on the ship type and the deadweight tonnage.

4. The method for inverting ship voyage load according to claim 1, characterized in that, The step of determining the load inversion model based on the loading state includes: When the loading state is a partially loaded state, the displacement model is determined as the load inversion model; When the loading state is full load, the linear model is determined as the load inversion model.

5. An electronic device, characterized in that, The electronic device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the ship voyage load inversion method according to any one of claims 1 to 4.

6. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the ship voyage load inversion method according to any one of claims 1 to 4.

Images