A port ship unloading area analysis method based on AIS data

By preprocessing AIS data and combining Geohash grid with destination text matching rules and bidirectional retrieval methods, the problem of identifying unloading areas and calculating cargo volume in multi-port clustered environments was solved, achieving high-precision unloading area extraction and cargo volume calculation.

CN120541161BActive Publication Date: 2026-05-05NANJING UNIV
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202510662885.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-05-22
Publication Date
2026-05-05
Estimated Expiration
2045-05-22

AI Technical Summary

Technical Problem

Existing technologies struggle to accurately identify and delineate unloading areas in densely populated port regions, resulting in low accuracy in calculating cargo volumes for energy goods such as crude oil. This is particularly problematic in multi-port clusters where spatial delineation is often ambiguous and erroneous.

Method used

By preprocessing AIS data, the ship's stopping segments are extracted and spatial features are summarized using Geohash grids. Combined with destination text matching rules and bidirectional retrieval methods, the unloading areas are integrated and the areas are divided according to the principles of proximity and obstacle avoidance.

Benefits of technology

It improved the accuracy and stability of unloading area identification, reduced the impact of errors from non-standard destination records, ensured the accuracy of target port identification, and enabled real-time dynamic calculation of the volume of marine energy cargoes such as crude oil.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120541161B_ABST
    Figure CN120541161B_ABST
Patent Text Reader

Abstract

The present application relates to a kind of port ship unloading area analysis method based on AIS data, comprising the following steps: first, AIS data preprocessing-washing and screening of original data and attribute supplementing;Second, ship stay voyage section extraction-define stay voyage section using motion characteristics and express using Geohash grid;Third, port destination text matching-irregular destination is matched with standard port name;Fourth, destination bidirectional search-accurately destination is analyzed using ship trajectory, and the attribution port of stay grid is determined;Fifth, grid integration and route cutting-integrate grid of different port attribution to obtain unloading area and cut ship full cycle route.The present application solves the problem that unloading mode is complex around port, and route is complex, utilizes the complementary advantage of space-time information and destination text in AIS data, and provides reliable technical support for the trade monitoring of bulk energy goods such as crude oil.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of port logistics and shipping data analysis, and in particular to a method for analyzing unloading areas of energy-related goods around ports based on Automatic Identification System (AIS) data, integrating Geohash algorithm, text matching rules and bidirectional port destination retrieval. The resulting unloading areas facilitate decision-making and supervision of port cargo transportation by managers, significantly improving the timeliness and accuracy of port-level trade data analysis. Background Technology

[0002] Energy trade, represented by crude oil, is a crucial component of international trade worldwide, playing a significant role in ensuring national energy security and promoting socio-economic development. Due to its bulk commodity nature and the advantages of low cost and large transport volume via sea freight, crude oil is currently transported primarily by ocean in the global trade network. Organizations such as the United Petroleum Data Project (UPD) aim to provide transparent and reliable crude oil import data; however, these projects only provide the total import and export volume of crude oil for different countries within a time series, failing to differentiate between sea freight and other modes of transport, and also lacking information on the import status of specific ports. With the rapid development of satellite positioning and communication technologies, high-timeliness AIS data can be received by nearby ships, land-based receivers, and satellite-based AIS receivers, containing key fields such as ship position, speed, heading, draft, and destination. Therefore, researching a method for calculating the volume of energy goods such as crude oil at the port level based on AIS data is crucial for improving the spatial granularity of port traffic analysis and energy transport management.

[0003] While AIS data-based analysis methods can provide real-time dynamic information, the automatic identification and delineation of unloading areas at each port in densely populated coastal trade areas presents a critical challenge. This is due to factors such as overlapping ship trajectories, inconsistent manually entered AIS target port information, and complex unloading methods including berth unloading, ship-to-ship unloading, and unloading via offshore oil pipelines. Current technologies often employ traditional spatial clustering algorithms (such as DBSCAN) to spatially cluster and delineate unloading areas at the national level using AIS data. However, when faced with the uncertainty of unloading behavior and target information in multi-port clustered environments at a smaller spatial scale, these algorithms often struggle to balance spatial location and destination semantics, easily leading to ambiguity and errors in the spatial delineation of unloading areas belonging to different ports. This, in turn, affects the accuracy of calculating the volume of marine energy goods such as crude oil. Therefore, there is an urgent need to develop an intelligent method that can fully utilize ship berthing characteristics, unloading behavior information, and destination text information from AIS data to achieve high-precision extraction of unloading areas around ports and, consequently, accurate calculation of the volume of marine energy goods such as crude oil. Summary of the Invention

[0004] The technical problem this invention aims to solve is to overcome the aforementioned shortcomings of existing technologies and propose a port vessel unloading area analysis method based on AIS data. Compared with traditional methods, this invention fully considers the destination text attributes, draft error factors, and actual vessel unloading scenarios in AIS big data, achieving accurate identification and extraction of unloading areas in different ports. This provides a scientific implementation method for port-level maritime energy trade cargo transportation supervision and volume calculation.

[0005] To address the aforementioned technical problems, this invention first preprocesses AIS data, cleaning up useless data and searching for and supplementing detailed ship design parameters. Secondly, it extracts the stopping segments of vessels such as crude oil tankers near ports, uses Geohash grids to summarize spatial characteristics, and performs adjacent merging. With the help of mapping rules between irregular destination information and standard port names in AIS records, a bidirectional destination retrieval method is used to statistically determine port affiliation for each merged grid. Following the principles of proximity and obstacle avoidance, the unloading areas of each port are integrated, preserving routes within and around the unloading areas.

[0006] The advantages of the port vessel unloading area analysis method based on AIS data proposed in this invention are as follows:

[0007] (1) Improved the identification accuracy and stability of unloading areas, effectively avoiding port attribution confusion and instability caused by relying solely on AIS spatial features, while taking into account various unloading methods of energy cargo ships such as crude oil tankers, and improving the completeness of unloading area extraction;

[0008] (2) Combining the text matching rules and destination bidirectional retrieval strategy proposed in this invention, the error impact of non-standard destination records in AIS data is effectively reduced, ensuring the accuracy of target port identification.

[0009] (3) By utilizing AIS statistics and readily available ship design parameters, real-time and dynamic calculations of the volume of marine energy cargoes such as crude oil can be easily performed.

[0010] (4) This method is applicable to various port topology feature areas, has high versatility and scalability, and can be promoted in practical applications in different regions.

[0011] This invention solves the problem of extracting unloading areas for energy goods around ports using AIS data, realizes the spatial extraction and attribution of unloading areas, and proposes a reliable method for transport management and monitoring of maritime energy trade goods around ports based on AIS data at the port level, providing a process approach for the promotion and application in other port scenarios. Attached Figure Description

[0012] The invention will now be further described with reference to the accompanying drawings.

[0013] Figure 1 This is an overall flowchart of an example of the present invention.

[0014] Figure 2 This refers to the AIS data coverage and research port of this invention.

[0015] Figure 3 This is a text matching rule for AIS destination information and standard port names in an example of the present invention.

[0016] Figure 4 This is a schematic diagram illustrating the principle of bidirectional port destination retrieval in an example of the present invention.

[0017] Figure 5 This invention provides an example of a Geohash grid merging integration process that follows the principles of proximity and obstacle avoidance.

[0018] Figure 6 This invention is an example of merging the Geohash grid and unloading area, which belong to different ports.

[0019] Figure 7 This is a schematic diagram illustrating the principle of cutting ship routes based on unloading areas, as described in this invention.

[0020] Figure 8 This invention provides a feasibility verification of using the cut ship route to calculate crude oil unloading volume in an example of the present invention. Detailed Implementation

[0021] The present invention will now be described in detail with reference to the accompanying drawings, which will make the technical approach and operation steps of the present invention clearer.

[0022] This invention is applicable to various port topological scenarios, including multi-port clusters and independent single-port distributions, all covered by AIS data. The greater the straight-line distance between adjacent ports, the more accurate the results. Examples include Chiba, Yokohama, and Kawasaki ports in Tokyo Bay, Yokkaichi and Nagoya ports in Ise Bay, and Tomakomai port in Hokkaido. These six ports are major crude oil import ports for Japan, accounting for over 55% of Japan's total crude oil imports annually from 2014 to 2021. Their geographical relationships exhibit three distinct spatial relationships: adjacent, near, and far.

[0023] Because Kawasaki Port is adjacent to Yokohama Port, and their shipping channels are directly connected, different shipping routes overlap and intersect, which will lead to the following problems: (1) multiple crude oil tankers heading to two different destinations will be anchored in the same small area; (2) crude oil tankers destined for one port will stop in the shipping channels of other ports. If the unloading areas of the two ports are forcibly separated, the highly coupled spatial relationship will bring a large error to the cargo volume calculation. In the end, the two ports were merged to deal with this phenomenon. Although the two adjacent ports were merged into one, this whole was accurately separated from the unloading areas of neighboring ports such as Yokosuka Port and Chiba Port.

[0024] Therefore, this example uses offshore crude oil monitoring in the study area as a case study to illustrate the method for analyzing port vessel unloading areas based on AIS data. Figure 1 As shown in the flowchart, the process includes the following steps:

[0025] Step 1: AIS Data Preprocessing – Select the corresponding port cluster area for AIS data extraction. Use MMSI and IMO as identification identifiers to extract the yearly AIS trajectory data for each vessel. Preliminary extraction of tanker data is based on vessel type (vessels starting with 8 in AIS attributes are tankers). To improve data quality, the extracted AIS data attributes are cleaned, mainly including three categories: errors, duplicates, and incompleteness. Specific cleaning conditions are shown in the table below:

[0026] Cleaning categories Cleaning conditions Error data MMSI is not a 9-bit integer or IMO is not a 7-bit integer. Duplicate data Same vessel, same recording time Incomplete data The AIS attributes for MMSI, IMO, or draft are empty.

[0027] The extracted ship tracks do not include detailed ship design parameters related to load calculations. This invention uses IMO numbers as the primary retrieval criterion, obtaining information such as deadweight tonnage (DWT), design draft, and tonnage per centimeter (TPC) from ship archives as supplementary data to the ship tracks. Subsequently, detailed ship usage types were retrieved from Marine Traffic, My Ship Tracking, and Vessel Finder websites, and filtered using crude oil tankers as the standard. Draft plays a crucial role in ship deadweight calculations, and crude oil tankers of different sizes often exhibit different draft variation characteristics. Therefore, based on the deadweight tonnage attribute of crude oil tankers, all crude oil tankers are divided into seven categories, from Small Medium to ULCC. Taking 2016 AIS data as an example, the specific classification criteria and the number of ships in each category are shown in the table below:

[0028]

[0029] Step 2, Ship Stop Segment Extraction – Based on three ship motion characteristics, namely ship speed, time interval between records of adjacent points, and straight-line distance between adjacent points, the threshold method is used to extract AIS stop points within the port buffer zone, and stop segments are constructed based on the number of stop points clustered within a certain time period.

[0030] The motion characteristics used to identify ship stopping points are as follows:

[0031]

[0032] Among them, v i Indicates the ship at t i Ship speed at any given time, distance i,i+1 Indicates the ship at t i and t i+1 The straight-line distance between moments, v th t th and distance th These represent the ship speed threshold, the time interval threshold between adjacent points, and the straight-line distance threshold used to identify stopping points, respectively.

[0033] In this embodiment, the ship speed threshold v for stopping point screening th For section 1, the time interval t between records of two adjacent points is a threshold. th The threshold for the straight-line distance between two adjacent points is 1 hour. th The distance is 2 kilometers. For the stop points obtained from the initial screening, based on the statistical characteristics of the time interval between AIS record points in different years, stop points within 30km of the port land boundary with a cluster size of 12 (2014-2017) and 120 (2018-2022) within 12 hours are used to construct stop segments.

[0034] Then, a structured grid is generated based on the range of the stopping segment. The stopping segment is divided into grids according to its spatial location, and the grids are merged according to their adjacency relationships. In this embodiment of the invention, the Geohash algorithm is used to generate the structured grid. The Geohash grid used to express the spatial location of the ship's stopping segment uses Geohash 7 levels, with an error of approximately ±76 meters. The actual cell size can be adjusted according to actual needs. By expressing the extracted spatial location of the ship's stopping segment in grid form, compared to a single ship route, using a grid can not only speed up the calculation of cargo volume, but also reflect the actual unloading scenario of the ship "entering port-stopping-departing port", summarizing the spatial distribution characteristics of crude oil tanker unloading.

[0035] Step 3: Port Destination Text Matching – Using the full port name and IMO number as the matching criteria, calculate the similarity between the AIS destination record and the standard port name. If the two constitute a text matching relationship or the similarity is greater than a certain threshold, then the current record point is mapped to the standard port name; otherwise, the current record point is marked as matching other ports, realizing the mapping between irregular AIS destination records and standard port names. AIS destination attributes are manually recorded and sent, so some problems inevitably occur, such as recording errors and non-standardization. Most destination records are marked with a country's ISO code, followed by an abbreviation or full name of a port, with spaces in the middle or on both sides, and missing or redundant parts of the text, such as >JP YKK. By analyzing a large number of different matching cases between AIS destination information and standard port names, this invention proposes a new method: using the full port name and IMO number as the matching criteria, calculate the similarity between the AIS destination and the standard port name. If the two constitute a matching relationship or the similarity is greater than 80%, then it is inferred that the current record point is related to the standard port name, and the current record point is mapped to the standard port name; otherwise, the current record point is marked as matching other ports. The matching rule is as follows: Figure 3 As shown, the process includes the following steps: First, the AIS destination information to be matched is standardized by removing redundant information, including removing redundant spaces on both sides of the text, capitalizing English characters, and omitting country codes. Second, text slices of the AIS destination information are constructed. Based on the matching relationship of mutual inclusion or equality, matching is only performed on information where the number of port IMO codes in the destination text is less than or equal to 2. The matching rule focusing on the full name of the port divides the mapping relationship into one simple case (the text slice is directly equal to the port IMO code) and three complex cases: if the number of port IMO codes in the destination text is equal to 2, the latter IMO code is used; if the slice contains only 1 port IMO code, it is divided into two cases based on whether it contains or is equal to the full name of other ports. If the slice contains or is equal to the full name of other ports, the full name of the port is used; otherwise, the port IMO code is used. For texts that do not meet the matching relationship, Difflib is used to calculate the matching similarity. If the similarity between the destination text and all the full names of the ports to be matched is less than 80%, the ship destination is set to "Other".

[0036] Some examples are shown below:

[0037] Case AIS Information Relevant matching criteria IMO code interference (full name) Similarity result A JPYKKXXJPCHB YKK, CHB YKK(Yokkaichi) / Chiba B JPYokkaichi YOK, Yokkaichi YOK (Yokohama) / Yokkaichi C >JPYKK YKK / / Yokkaichi D Singapore / / 53% Other E JP YKK YKK / / Yokkaichi

[0038] During the matching process, several AIS destination information entries not in the IMO format formed inclusion relationships with the IMO numbers or full names of multiple ports. In Case A, the destination setting format represents a vessel sailing from Yokkaichi Port (YKK) to Chiba Port (CHB), retaining only the standard port name corresponding to the last item. In Case B, the destination information includes port IMO code interference items (YOK), using the full port name as a highly reliable matching basis. When the destination information matches the IMO number of a port, the full names of other ports are used to check if the voyage belongs to another port. Cases C and E represent simpler mapping relationships, which can be determined solely by simple inclusion or equality relationships of port IMO codes. Case D indicates a situation where no matching relationship is satisfied, and Difflib is used to calculate the similarity of the destination information.

[0039] Step 4, Two-way Destination Search—Apply the two-way destination search method to all AIS data passing through each Geohash merged grid, assigning a port of ownership to each merged grid based on the statistical characteristics of the destination information. The destination information of an AIS stop point may include information about the current port of call and the next voyage's destination port; determining the true port of ownership for a merged grid solely based on the stop point is unreliable. Assume the records of vessels about to enter the port (the port information corresponding to the record point about to enter the Geohash merged grid, denoted as Point). prev (This indicates that) it has been carefully inspected and is generally matched with the port the ship is scheduled to call.

[0040] This invention relates to a two-way port destination retrieval method that considers both the AIS destinations of ships before and after entering a merged grid. In summary, it extracts AIS route data before and after a ship's stopover (i.e., the ship's trajectory before and after entering the merged grid obtained in step 2), maps the destination information from the individual ship's AIS data in step 3, and determines the port affiliation of each merged grid by integrating statistical information on all ship destinations for the current year. The principle of the two-way port destination retrieval method and different retrieval scenarios are as follows: Figure 4 As shown, the implementation will be carried out in three steps:

[0041] (1) By matching all docking routes with the complete trajectory, the start and end time indices are obtained, and then the data interval Δ between each entry of the crude oil tanker into the merged grid and its previous departure is calculated. interval ;

[0042] (2) For each stop route, based on the port destination text matching rules, using the standard port name corresponding to the first record point before entering the grid or within the grid as the benchmark, the maximum distance for reverse retrieval is set to Δ. interval During the reverse retrieval process, if a record point that does not match the baseline is encountered, or if a record point where the vessel previously left the merged grid (i.e., reaching the maximum distance Δ) is encountered...interval If a record point that does not match the baseline is encountered during the forward search, or if a ship leaves the merged grid, the search will exit. The grid is used as a spatial constraint to simulate the ship's entry and departure from the port. Through the above process, a bidirectional destination sequence is obtained, the values ​​of which are all equal to the port baseline.

[0043] (3) After retrieving all crude oil tanker trajectories, merge the two-way destination sequences and select the standard port name with the highest frequency as the port to which the current merged grid belongs.

[0044] In this embodiment, the bidirectional destination retrieval divides the search scenario into four cases based on the spatial relationship between the ship's route and the Geohash merged grid: the ship enters and leaves the grid, the ship is initially inside the grid and leaves, the ship enters the grid but does not leave, and the ship remains inside the grid throughout its journey. The destination information of the point before the ship enters the merged grid (the ship enters from outside the grid) or the first point within the grid (the ship is initially within the grid) is used as the benchmark. Using the text matching rules from step 3, during the forward search according to the chronological order of records, if a point that does not match the benchmark port is encountered or the ship leaves the grid, the search is terminated; during the reverse search, if a point that does not match the benchmark port is encountered or the point where the ship previously left the grid is encountered, the search is terminated. Finally, the mode of the standard port names retrieved around the merged grid for all energy cargo ships such as crude oil tankers is used as the port affiliation.

[0045] Step 5, Grid Integration and Route Cutting—Based on the spatial location and semantic features of the grid, discrete grids are integrated to form unloading areas, and routes within and around the unloading areas are cut and preserved. In multi-port clustered areas, simply merging all grids based on spatial location may include unloading areas of other ports or omit unloading areas of the current port. Therefore, according to the region growth algorithm, unloading areas are extracted among multiple merged grids following the integration principles of proximity and obstacle avoidance. The integration process is as follows: Figure 5 As shown. Specifically, it includes the following steps:

[0046] (a) Sort the importance of Geohash merged grids according to their area size, and select the main merged grids in order of size.

[0047] (b) Based on the diameter estimation of the single-point mooring operation area, and following the principles of proximity and obstacle avoidance, only the nearest grid cells with a shortest distance smaller than the operating diameter of the mooring area are integrated. The formula for calculating the diameter of the single-point mooring operation area is as follows:

[0048] D = 2 × (3 × L) v +L h +L0)

[0049] Where D represents the operating diameter of the mooring area, and L vL is the maximum length of the ship. h L0 represents the length of the mooring cable and the maximum displacement of the buoy under operating conditions. If the current consolidation entity contains or intersects with grids belonging to other ports after merging the latest grid, the consolidation result is rolled back and the consolidation entity is saved as the new unloading area. In this example, the diameter of the single-point mooring operation area is set to 3 kilometers. In this embodiment, the proximity principle means gradually merging merged grids within 3 kilometers of the nearest consolidation entity, iteratively updating the consolidation entity. The obstacle avoidance principle means that if the updated entity contains grids belonging to other ports during the consolidation process, the current entity is rolled back, and a new entity is selected for re-merging based on the area.

[0050] (c) Select a new integration entity based on the grid area and carry out the integration operation.

[0051] (d) If there is no grid that can be merged within the operating diameter (3 km) of the mooring area around the main body, the current main body will be saved as the unloading area and a new main body will be selected to continue merging.

[0052] Repeat the above steps until all merged grids are merged. The final result is represented as multiple merged regions that do not include grids pointing to other ports.

[0053] Taking 2022 data as an example, the final merged grid and unloading area extraction results belonging to different ports are as follows: Figure 6 As shown.

[0054] Since the unloading of crude oil tankers occurs in a near-static state, only the amount of crude oil unloaded needs to be calculated based on the ship's stationary trajectory. AIS records may lag behind the ship's real-time activity; that is, the unloading may have already been completed before the draft is recorded, at which point the ship may have already left the unloading area. Similar to different retrieval scenarios in step 4, such as... Figure 7 As shown, this invention considers four types of boundary cutoff conditions for ship routes in the unloading area:

[0055] (a) Vessel enters and leaves the grid: The vessel enters the grid from outside the unloading area, stops to unload, and then leaves the current area. This type of situation accounts for the majority of route cutting scenarios. (b) Vessel is initially inside the grid and leaves: The vessel is inside the grid at the initial moment of the start of the statistics (i.e., there is no record point outside the unloading area before the start of the statistics) and eventually leaves the unloading area. (c) Vessel enters the grid but does not leave: The vessel enters the grid from outside the unloading area but does not eventually leave the record point of the unloading area. (d) Vessel stays inside the grid the entire time: The vessel stays inside the unloading area the entire time, which mainly applies to vessels transshipping within the port over short distances.

[0056] The segmented shipping routes can quickly respond when calculating the cargo trade volume of a specific port. Taking maritime crude oil trade as an example, after obtaining the unloading areas of important crude oil import ports according to steps 3 and 4, this embodiment uses the crude oil unloading areas of each port to segment the full-cycle shipping routes of crude oil tankers. Taking the method based on load status conversion as an example, the segmented routes are used to calculate the crude oil unloading volume. Of course, there are other methods for calculating crude oil unloading volume. The main purpose of this invention is to realize the segmentation of shipping routes based on unloading areas through AIS data. This embodiment calculates crude oil unloading volume based on the load status conversion method, specifically by distinguishing the empty and full load states of ships according to draft and using deadweight tonnage (DWT) to express cargo loading and unloading volume. This is based on the assumption that the load status of bulk energy cargo ships such as oil and natural gas is mostly one-way full load or one-way empty load. The calculation formula is as follows:

[0057]

[0058] Unload i,t =α×dwt i Full load → No load

[0059] Among them, draft t+1 and draft t It refers to the AIS draft at adjacent times in each shipping route after the crude oil tanker is cut up. d This serves as the basis for determining whether a crude oil tanker has transitioned between empty and full-load states. Since ship draft is closely related to size, the draft changes of each ship across all seven different crude oil tanker sizes throughout the entire voyage can be statistically analyzed for all years. The cumulative frequency of draft changes for each ship size, arranged in ascending order, reaching 10% is used as the threshold for determining the transition between empty and full-load states. d Unload i,t Let represent the amount of crude oil unloaded by ship i at time t, and α represent the conversion rate between unloaded volume and deadweight tonnage (DWT), which is taken as 0.87 here. i This represents the deadweight tonnage (DWT) of vessel i. AIS data statistics for the entire shipping route are performed simultaneously during the data preprocessing stage. When calculating crude oil unloading volume, only the segmented route is used. This focuses on areas of interest such as ports, accelerating the calculation process, and reduces interference from occasional draft record errors during crude oil tanker transport.

[0060] Before calculating the crude oil unloading volume, this embodiment also proposes the following indicators to determine whether crude oil unloading has occurred: minimum draft change, maximum permissible draft change, and maximum speed at adjacent recording points. The determination is based on the following formula:

[0061]

[0062] Among them, draft t+1 and draft t It refers to the AIS draft at adjacent times in each shipping route after the crude oil tanker is cut up. min This represents the minimum draft change, taken as the first percentile of the cumulative frequency of the current crude oil tanker's draft changes across all years, arranged in ascending order (mostly 0.1m, representing random errors during manual recording). Maximum permissible draft change (draft) max This is the difference between the full-load design draft and the empty-load draft. The full-load draft is the design draft in the vessel's records; if empty, it is set to the maximum draft across all years. The empty-load draft is the first percentile of the cumulative frequency of the current crude oil tanker's draft across all years, arranged in ascending order. t and v t+1 This refers to the ship's speed at adjacent times along each route after the crude oil tanker was cut up, with the maximum speed v at adjacent recording points. max The data was set to section 2 to accommodate the lag in AIS records. AIS statistics such as draft changes and maximum and minimum draft of the crude oil tanker across all years were obtained during the ship's full-cycle route extraction process in step 1 of the data preprocessing.

[0063] The correlation between crude oil tanker unloading volumes calculated using the segmented crude oil tanker routes and official data is as follows: Figure 8 As shown, the annual trend and growth rate of crude oil imports are calculated using Chiba Port as an example. The method based on load condition transformation has a high correlation (R... 2 =0.95), showing a direct proportional relationship with official data. The strong correlation and similar trend curves between the two indicate that using AIS data for port trade monitoring is feasible, and also demonstrate that the technical process of delineating unloading areas around ports is reliable.

[0064] Beyond crude oil transport volume calculations, the segmented vessel routes can also be used for safety management purposes such as collision risk assessment and real-time tracking in port-adjacent waters and channels. By analyzing the loading and unloading times of vessels at different anchorages and berths within the port, spatial scheduling and rational allocation of port facilities and transportation resources can be achieved, improving port operational efficiency. Vessel dwell areas and segmented vessel routes can help identify high-frequency congestion areas and abnormal vessel dwelling behavior around the port, providing data support for infrastructure expansion or route adjustments. Combining vessel type (especially energy cargo ships such as oil tankers) and navigation paths can help quickly locate potential pollution sources that may cause cargo leaks, providing real-time decision support for sudden environmental events (such as oil spills). By using segmented (or partial) vessel routes focusing on key areas, rapid decision support can be provided for different needs.

[0065] The port vessel unloading area analysis method based on AIS data of the present invention is not limited to the specific technical solutions and implementation areas described in the above embodiments. All technical solutions formed by equivalent substitutions are within the scope of protection claimed by the present invention.

Claims

1. A method for analyzing port vessel unloading areas based on AIS data, comprising the following steps: Step 1: AIS Data Preprocessing – Select the corresponding port area to extract AIS data, and extract the AIS data for each ship year by year based on the ship's MMSI number and IMO number. Select the AIS data for oil tankers and perform data cleaning. Supplement the necessary parameters for cargo volume calculation based on ship files and query websites, and classify the specifications according to the deadweight tonnage attributes of crude oil tankers. Step 2, Ship Stop Segment Extraction—Based on three ship motion characteristics—ship speed, time interval between records of adjacent points, and straight-line distance between adjacent points—the threshold method is used to extract AIS stop points within the port buffer zone. Stop segments are constructed based on the number of stop points clustered within a certain time period. A structured grid is generated based on the range of the stop segments. The stop segments are then divided into the structured grid according to their spatial location. Finally, the structured grids are merged according to their adjacency relationships to obtain the Geohash merged grid. Step 3: Port Destination Text Matching – Using the full port name and IMO number as the matching criteria, calculate the similarity between the AIS destination record and the standard port name; if the two constitute a text matching relationship or the similarity is greater than a certain threshold, then map the current record point to the standard port name; otherwise, mark the current record point as matching other ports. Step 4, Two-way Destination Search – Apply the port destination two-way search method to all AIS data passing through each Geohash merged grid, and assign a home port to each Geohash merged grid based on the statistical characteristics of the destination information. Step 5, Grid Integration and Route Cutting: Based on the single-point mooring operation area of ​​the vessel, following the principle of proximity and obstacle avoidance, the region growth algorithm integrates the Geohash merged grid with port destination attributes from Step 4 to generate the unloading area range of the target port, and uses the unloading area to cut the vessel's full-cycle route. Step 4 of the port destination bidirectional retrieval method specifically includes the following steps: 4.

1. By matching all stopping routes with the complete trajectory, the start and end time indices are obtained. Then, the data interval ∆ between each entry of the crude oil tanker into the Geohash merged grid and its previous departure is calculated. interval ; 4.2 For each stop route, based on the port destination text matching rules, using the standard port name corresponding to the first record point before entering the grid or within the grid as the benchmark, the maximum distance for reverse retrieval is set to ∆. interval If a record point that does not conform to the baseline is encountered during the reverse retrieval process, or if a record point where the ship previously left the Geohash merged grid is encountered, the retrieval will be terminated. If a record point that does not conform to the baseline is encountered during the forward retrieval process, or if the ship sails out of the Geohash merged grid, the retrieval will be terminated. The grid is used as a spatial constraint to simulate the process of ships entering and leaving the port. Through the above process, a bidirectional destination sequence is obtained, and its value is equal to the port baseline. 4.3 After retrieving all crude oil tanker trajectories, merge the bidirectional destination sequences and select the standard port name with the highest frequency as the port to which the current Geohash merged grid belongs.

2. The port vessel unloading area analysis method based on AIS data according to claim 1, characterized in that: The data preprocessing standards in step 1 based on AIS attribute data and ship archives are as follows: By analyzing the standard format of attribute fields in AIS data, firstly, extract the tanker trajectories whose ship purpose type starts with 8, and then remove the following three types of data points: a) Incomplete data: MMSI, IMO, draft, or speed are empty; b) Erroneous data: MMSI is not a 9-digit integer or IMO is not a 7-digit integer; c) Duplicate data: the same record time for the same ship; Based on ship archive information and ship query website crawler programs, supplement each ship with detailed purpose, design draft, draft in centimeters (TPC), and deadweight tonnage (DWT); Finally, subdivide the ship database according to the detailed ship purpose, retain only crude oil tankers with the detailed ship purpose of "Crude Oil Tanker", and classify them into 7 ship sizes according to deadweight tonnage (DWT).

3. The port vessel unloading area analysis method based on AIS data according to claim 1, characterized in that: In step 2, the motion characteristic conditions used to identify the ship's stopping point are as follows: in, Indicates that the ship is The ship's speed at that moment, Indicates that the ship is and The straight-line distance between moments , and These represent the speed threshold, time interval threshold, and straight-line distance threshold used to identify stopping points, respectively.

4. The port vessel unloading area analysis method based on AIS data according to claim 1, characterized in that: In step 3, firstly, the AIS destination information to be matched is standardized by removing redundant information, including removing redundant spaces on both sides of the text, capitalizing English characters, and omitting country codes. Secondly, text slices of the AIS destination information are constructed. Based on the matching relationship of mutual inclusion or equality, matching is only performed on information where the destination text contains less than or equal to 2 port IMO codes. The matching rules focusing on the full port name divide the mapping relationship into one simple case and three complex cases. The simple case is that the text slice is directly equal to the port IMO code. The three complex cases are: if the destination text contains two port IMO codes, the latter IMO code is used; if the slice contains only one port IMO code, it is divided into two cases based on whether it contains or is equal to the full standard name of another port. If the slice contains or is equal to the full name of another port, the full port name is used; otherwise, the port IMO code is used. For text that does not meet the matching relationship, Difflib is used to calculate the matching similarity. If the similarity between the destination text and all the full names of the ports to be matched is less than 80%, the ship destination is set to "Other".

5. The port vessel unloading area analysis method based on AIS data according to claim 1, characterized in that: In step 4, the bidirectional destination retrieval is divided into four scenarios based on the spatial relationship between the ship's route and the Geohash merged grid: the ship enters and leaves the grid, the ship is initially inside the grid and leaves, the ship enters the grid but does not leave, and the ship stays inside the grid the entire time. The destination information of the point before the ship enters the Geohash merged grid or the first point within the grid is used as the benchmark. Using the text matching rules in step 3, during the forward search according to the chronological order of the records, if a point that does not match the benchmark port is encountered or the ship leaves the grid, the search is terminated. During the reverse search, if a point that does not match the benchmark port is encountered or the point where the ship previously left the grid is encountered, the search is terminated. Finally, the mode of the standard port names retrieved by all energy cargo ships such as crude oil tankers around the Geohash merged grid is used as the port affiliation.

6. The port vessel unloading area analysis method based on AIS data according to claim 1, characterized in that: In step 5, the region growing algorithm includes the following steps: 5.1 Rank the importance of Geohash grid merging according to area size, and select the main merging entities in order of size; 5.

2. Based on the diameter estimation of the single-point mooring operation area, following the principles of proximity and obstacle avoidance, only the nearest grid with a shortest distance less than the operating diameter of the mooring area is integrated. If the current integration entity contains or intersects with grids belonging to other ports after merging the latest grid, the integration result is rolled back, and the integration entity is saved as a new unloading area. Then, step 5.3 is executed. The formula for calculating the diameter of the single-point mooring operation area is as follows: in, Indicates the operating diameter of the mooring area. The maximum length of the ship. The length of the mooring cable. This indicates the maximum displacement of the buoy under operating conditions; 5.

3. Select a new integration entity based on the grid area and perform the integration operation; 5.4 If there are no mergeable grids within the operating diameter of the mooring area surrounding the main body, save the current main body as the unloading area and select a new main body to continue merging; 7. The port vessel unloading area analysis method based on AIS data according to claim 6, characterized in that: The diameter of the single-point mooring operation area for ships is set at 3 kilometers. The principle of proximity means that Geohash merged grids within 3 kilometers of the main body are gradually merged and the main body is iteratively updated. The principle of obstacle avoidance means that if the updated main body contains grids whose ownership information belongs to other ports during the merging process, the current main body is reverted and a new main body is selected for re-merging based on the area.

8. The port vessel unloading area analysis method based on AIS data according to claim 1, characterized in that: Using unloading areas to cut a vessel's entire lifecycle route includes the following four scenarios: (a) Vessel enters and leaves the grid: Vessel enters the grid from outside the unloading area, stops to unload, and then leaves the current area. This type of situation accounts for the majority of route cutting scenarios. (b) Vessel initially inside the grid and leaves: The vessel is inside the grid at the initial moment of the count and eventually leaves the unloading area; (c) Vessel enters the grid but does not leave: Recording points where a vessel enters the grid from outside the unloading area but does not ultimately leave the unloading area; (d) Vessel stays within the grid throughout the entire journey: This mainly applies to vessels that transship within the port over short distances.

9. The port vessel unloading area analysis method based on AIS data according to claim 1, characterized in that: This also includes step 6, calculating the crude oil unloading volume based on the cut-off route—distinguishing between empty and full load states of the vessel based on draft depth, and calculating the cargo loading and unloading volume based on deadweight tonnage (DWT): This represents the amount of crude oil unloaded by ship i at time t. This indicates the conversion rate between unloading volume and deadweight tonnage (DWT). Indicates the deadweight tonnage (DWT) of vessel i; Load condition is determined by the following formula: in, and It refers to the AIS draft at adjacent times in each shipping route after the crude oil tanker was cut up. It is the threshold for determining whether an oil tanker has undergone a transition from empty to full load status.

Citation Information

Patent Citations

  • AIS data-driven crude oil ship destination port prediction method

    CN116805172A

  • Marine traffic network construction method and device, electronic equipment and readable storage medium

    CN119811139A