A method for calculating navigable water depth of a port based on AIS data

Abstract

The present application relates to a kind of port navigable water depth calculation method based on AIS data, comprising the following steps: first, AIS data preprocessing-washing and correcting the error data in AIS;Second, structured grid subdivision-the AIS data is divided into the structured grid in the port area;Third, navigable water depth reliability evaluation-the reliability of the ship record of AIS point in each grid is evaluated;Fourth, navigable water depth calculation-according to the reliability evaluation result of ship, through draft depth, generates navigable water depth.The present application solves the problem of difficult, high cost, poor timeliness of depth measurement in port area, realizes the automatic calculation of navigable water depth, proposes a reliable port navigable water depth calculation method based on AIS data, and provides a process idea for automatically extracting available information based on other big data.

Description

TECHNICAL FIELD

[0001] The present application relates to a port navigable water depth calculation method based on AIS data, in particular to a method which can realize automatic calculation of navigable water depth and evaluation of ship reliability based on AIS data. BACKGROUND

[0002] Maritime trade grows with the development of economic globalization, and the size of the ship and the traffic density also increase, which increases the risk of ship accidents. Among many ship accidents, ship grounding may cause oil spills and ship capsizing, which threatens the environment and life safety. The port is one of the most important hubs of modern global trade and is a distribution center for all kinds of goods, and plays a key role in international trade. The port is generally built near the coast, and the water depth is lower, and because of the action of waves and tides, there will be a phenomenon of siltation, especially in sandy coast, silt coast, and near the estuary of the port, which will appear more serious siltation phenomenon, and ship grounding accidents occur from time to time. Therefore, it is very important to study a port navigable water depth calculation method based on AIS data.

[0003] Generally, the water depth of the port is realized by sonar measurement. Although this way of measuring water depth has high precision, it has the disadvantages of high cost and low efficiency, and is not suitable for large-scale and frequent measurement, which cannot meet the growing demand for shipping. In the aspect of using AIS data for water depth, some scholars have determined the required depth contour area interpolation according to AIS and marine accident data. In addition, some researchers have explored the method of generating water depth map using AIS data, which predicts water depth through chart and AIS data based on machine learning algorithm, which ignores the in-depth evaluation of data reliability and brings great risk to prevent potential safety problems. Therefore, it is necessary to deeply study the port navigable water depth calculation and reliability evaluation method based on AIS data, and provide a more practical, accurate and comprehensive navigable water depth calculation method. SUMMARY

[0004] The technical problem to be solved by the present application is to overcome the above-mentioned shortcomings of the prior art, and to provide a port navigable water depth calculation method based on AIS data. Compared with the traditional method, the present application fully considers the error and other factors in big data, realizes the reliability evaluation of navigable water depth, solves the problem of navigable water depth calculation, and provides a scientific implementation method for port navigable water depth calculation.

[0005] To solve the above technical problems, the present application provides a port navigable water depth calculation method based on AIS data, comprising the following steps:

[0006] Step 1, AIS data preprocessing - select the target port area to extract AIS data, and extract the AIS trajectory data of the ship within one year according to the identification code MMSI of the ship in the region, and clean the trajectory error and attribute error of the extracted AIS data;

[0007] Step 2, structured grid subdivision - generate a structured grid in the port area, and subdivide the extracted AIS data into the grid according to the spatial position, and count the AIS data in the grid, including counting the number of ships, ship types, and ship draft depth distribution in the grid;

[0008] Step 3, navigable water depth reliability evaluation - analyze the factors related to the reliability of the navigable water depth, establish a navigable water depth reliability index system, and use the analytic hierarchy process, which is divided into single ship AIS record reliability evaluation and ship reliability evaluation in the grid, and is divided into the following 7 points:

[0009] 1) Calculate the annual draft dispersion D of the ship by the following formula:

[0010]

[0011] Where n is the number of trajectories of the ship, di is the draft depth of the i-th position in the trajectory of the ship; when D is 0, it means that the data has no dispersion degree, that is, the data concentration is very high, and the data is almost completely equal; when D is 1, the dispersion degree of the data is the highest, that is, the data concentration is extremely low, and the dispersion degree between the data is very large; i

[0012] 2) Calculate the annual trajectory integrity G of the ship according to the following formula:

[0013]

[0014] Where n is the number of trajectories of the ship, Q1 is the first quartile of the number of AIS trajectories of the ship in the region within one year, and Q3 is the third quartile of the number of AIS trajectories of the ship in the region within one year;

[0015] 3) For each type of ship, calculate the frequency T of the draft depth of the ship being 0 according to the following formula:

[0016] T = 1 - R1 / R2

[0017] Where R1 is the number of ships with a draft depth of 0 in the type of ship, and R2 is the total number of ships of the type;

[0018] 4) Calculate the ship voyage data integrity S according to the following formula:

[0019]

[0020] ​Wherein, S is the completeness of the ship voyage data in the ship trajectory, L is the number of key fields of the ship voyage data, and the key fields include: ship state, destination, ETA, S j is the missing rate of the jth key field, u is the number of missing fields in the ship trajectory, and n is the number of ship trajectories;

[0021] 5) The single-ship record reliability X of the AIS record of each ship is evaluated by the four factors of the annual water draft dispersion D of the ship, the annual trajectory completeness G of the ship, the frequency T of the ship water depth not being 0, and the ship voyage data completeness S:

[0022] X = a * D + b * G + g * T + d * S

[0023] Wherein, a, b, g and d are the weights of the factors, and the weights of the factors are determined by combining the expert scoring and the analytic hierarchy process AHP, and the specific method is as follows:

[0024] Firstly, a hierarchical structure is established: the target layer is the single-ship record reliability X; the standard layer includes the annual water draft dispersion D of the ship, the annual trajectory completeness G of the ship, the frequency T of the ship water depth not being 0, and the ship voyage data completeness S;

[0025] Secondly, a judgment matrix is constructed: for each factor in the standard layer, 1-9 scale is used for pairwise comparison to construct the judgment matrix A as follows:

[0026]

[0027] Wherein, a DG represents the importance of the annual trajectory completeness G compared with the annual water draft dispersion D, a DT represents the importance of the frequency T of the ship water depth not being 0 compared with the annual water draft dispersion D; a DS represents the importance of the completeness S of the ship voyage data in the ship trajectory compared with the annual water draft dispersion D; a GS represents the importance of the completeness S of the ship voyage data in the ship trajectory compared with the annual trajectory completeness G; a GT represents the importance of the frequency T of the ship water depth not being 0 compared with the annual trajectory completeness G; a TS represents the importance of the completeness S of the ship voyage data in the ship trajectory compared with the frequency T of the ship water depth not being 0;

[0028] Then, the weight vector is calculated: the eigenvalue method is used to concentrate the maximum eigenvalue l max of the judgment matrix A and the corresponding eigenvector W;

[0029] Finally, the consistency is detected according to the following formula:

[0030]

[0031] where CI is the consistency index, v is the order of the judgment matrix, find the average random consistency index RI≈9.0 corresponding to the four-order matrix, calculate the consistency ratio

[0032] If CR<0.1, it is considered that the judgment matrix has satisfactory consistency, otherwise the judgment matrix needs to be adjusted; after passing the consistency test, the weight vector obtained is the weight of each parameter;

[0033] 6) Calculate the relative dispersion degree RD of the draft according to the following formula:

[0034]

[0035] In the formula, d k is the draft of the ship in the grid k, d me is the median of the draft in the ship's track; the larger the RD, the more dispersed the draft, the lower the reliability, otherwise the reliability is higher;

[0036] 7) Calculate the navigable water depth reliability index Gra k of the grid k according to the following formula:

[0037]

[0038] where H max is the maximum draft in the grid k, H w is the draft of other ships w in the grid k, X max is the single-ship record reliability of the ship with the maximum draft in the grid k, X w is the single-ship record reliability of other ships w in the grid k, RD w is the relative dispersion degree of the draft of the ship with the maximum draft in the grid k, RD w is the relative dispersion degree of the draft of other ships w in the grid k;

[0039] Step 4, navigable water depth calculation: based on the statistical results of the draft distribution of ships in the comprehensive grid and the reliability evaluation results of the navigable water depth, the maximum draft that meets the threshold of reliability in the grid is calculated to obtain the navigable water depth.

[0040] The application firstly preprocesses AIS data, and cleans errors in trajectories and attributes. Secondly, a grid is generated, and ships and draft depths in the grid are analyzed. By identifying main influencing factors of navigable water depth reliability, such as annual draft dispersion degree, annual trajectory integrity degree, ship type, and ship voyage data, single-ship record reliability analysis and quantitative research are performed. Then, taking single-ship record reliability as a starting point, the characteristics of ships in the grid and the mutual relationship with other ships in the grid are comprehensively considered to evaluate the reliability of the port navigable water depth extracted by AIS. Finally, according to the results, the navigable water depth products under different reliabilities are calculated and generated.

[0041] The port navigable water depth calculation method provided by the application has the following advantages:

[0042] (1) The application realizes automatic cleaning of AIS data from two dimensions of trajectories and attributes, and especially considers the correlation between ship draft and length.

[0043] (2) The application establishes a navigable water depth reliability evaluation system based on annual draft dispersion degree, annual trajectory integrity degree, ship type, ship voyage data, draft depth relative dispersion degree, and reliability interaction between similar ships, which overcomes the problem of unreliable big data.

[0044] (3) The application uses reliability evaluation and ship draft statistical calculation to calculate the navigable water depth, which reduces the cost of measuring water depth.

[0045] The application solves the problem of calculating the navigable water depth by AIS data, realizes the evaluation of the reliability of the navigable water depth, proposes a reliable port navigable water depth calculation method based on AIS data, and provides a process idea for calculation and reliability evaluation based on other big data. BRIEF DESCRIPTION OF DRAWINGS

[0046] The application will be further described below with reference to the accompanying drawings.

[0047] Figure 1 It is a general flowchart of the application example.

[0048] Figure 2 It is a port area and AIS data of the application example.

[0049] Figure 3 It is a correlation diagram of ship length and draft depth of the application example.

[0050] Figure 4 It is a comparison of grid number and water depth conditions under different navigable water depth reliability thresholds of the application example.

[0051] Figure 5 It is a navigable water depth reliability result diagram of the application example in the port area.

[0052] Figure 6 A navigable water depth result map of a port area for an example of the present application. DETAILED DESCRIPTION

[0053] The technical route and operation steps of the present application will be clearer from the following detailed description of the present application with reference to the accompanying drawings.

[0054] The present application is applicable to a port with AIS data coverage, and the more AIS data, the more accurate the navigable water depth assessment, such as a port in Southeast Asia. The port in Southeast Asia is one of the busiest ports in the world, with an average of 140,000 ships docking every year, and the amount of AIS data is large.

[0055] And as one of the busiest sea ports in the world, the port in Southeast Asia has the characteristics of narrow space, turbulent water flow, and uneven seabed topography, which makes it difficult to measure the water depth. Secondly, the complexity of the port is reflected in its high accident rate, with about 2000 accidents occurring every month in the strait where the port in Southeast Asia is located.

[0056] Therefore, the present example takes the study area as an example to illustrate a port navigable water depth calculation method based on AIS data, such as Figure 1 The flow chart shows that the present application specifically includes the following steps:

[0057] Step 1, AIS data preprocessing - select the target port area to extract AIS data, and extract the AIS trajectory data of the ships in the area within a year according to the identification code MMSI of the ships. In order to improve the data quality, the extracted AIS data is cleaned in terms of trajectory and attribute, mainly including position error, common MMSI trajectory cleaning and ship type, length, width, speed attribute cleaning, and water depth attribute cleaning according to the relationship between ship draft and length. The specific conditions are as follows:

[0058]

[0059]

[0060] Through the analysis of a large amount of data related to ship length and draft, it is observed that the length of the ship is positively correlated with the draft. Figure 3 The correlation between the draft and length of different types of ships is shown.

[0061] Small water will affect the stability of the ship when empty, not conducive to the safety of navigation. Therefore, the ship must have a reasonable ballast. Generally considered, the ballast ship sailing in summer, the water depth should be at least 50% of the full load water depth, and in winter, due to the larger wind and wave, the water depth should be greater than 55% of the full load water depth in summer. Therefore, according to the length and type of the ship, find the maximum design water depth of the ship of this length, and calculate the water depth of the empty ship according to this. The water depth limit is obtained, and the water depth of the ship should be within this range. For example, a 200-meter-long cargo ship has a design water depth of 11.40 to 11.52 meters. Then its water depth should not be less than 5.70 meters. Therefore, for a 200-meter-long cargo ship, its water depth should be 5.70-11.52 meters. But according to the actual situation, there may be overloading or recording errors, etc. For example, in February 2017, the mmsi of the cargo ship is 477845200, the length of the ship is 201m, and the recorded water depth is 13.60m, so the maximum design water depth should be increased by 20%. For a 200m cargo ship, the water depth should be between 5.70 and 13.82m. The water depth of a 200m cargo ship should be between 5.70 and 13.82m. If the water depth of the ship is not within this range, it is considered that the water depth is incorrect and needs to be cleaned. Therefore, according to the length and water depth of the ship, the water depth limit is determined. Based on this, the data is cleaned;

[0062] Step 2, structured grid partitioning - generate structured grid in port area, take 50x50m cell as an example, actual cell size can be adjusted according to needs, AIS data extracted is partitioned into grid according to spatial position, compared with single value, grid water depth can better reflect the dimensional information of spatial area. Considering that the navigable water depth is usually defined as the minimum water depth to prevent the ship from touching the bottom, in each cell, record all ship water depth values and ship key information corresponding to different ship trajectory sets. Statistics of AIS data in the grid, including the number of ships in the grid, ship type, and ship water depth distribution;

[0063] Step 3, reliability evaluation of navigable water depth - analyze the reliability of navigable water depth related factors, establish the reliability index system of navigable water depth, adopt AHP method, divided into single ship AIS record reliability evaluation and ship reliability evaluation in grid, divided into the following 7 points:

[0064] 1) Calculate the annual water dispersion D of the ship. The annual water dispersion of the ship is the change range of the recorded water depth of the ship in a year, the larger the dispersion degree, the greater the change of the water depth of the ship, and then reflects the instability of the record. The method of reflecting the dispersion degree by coefficient of variation is used to calculate;

[0065] The annual water depth dispersion D of the ship is calculated by the following formula, the smaller the coefficient of variation, the higher the reliability, and vice versa:

[0066]

[0067] Wherein, n is the number of ship trajectories, d i is the water depth of the i-th position in the trajectory of the ship; when D is 0, it means that the data has no dispersion degree, that is, the data is very concentrated, and the data is almost equal; when D is 1, the dispersion degree of the data is the highest, that is, the data concentration is very low, and the dispersion degree between the data is very large.

[0068] 2) Calculate the annual trajectory integrity G of the ship. The annual trajectory integrity of the ship refers to whether the ship's trajectory in a year is complete, and the higher the integrity, the higher the reliability of the record. The number of AIS trajectories of the ship in a year is calculated.

[0069] If the ship is far away from the base station, the ship trajectory may not be complete, resulting in no record. In addition, the sailing time of the ship will also affect the number of AIS records. If the trajectory of the ship is not complete, the reliability of the recorded water depth will be very low. Statistical method is used to investigate the completeness of AIS records. In the port area, the number of AIS records of each ship in a year is recorded. The first quartile (Q1) is taken as the minimum threshold, which means that the ship with less records may have an incomplete trajectory; the third quartile (Q3) is taken as the maximum threshold, which means that the ship with more records may have a complete trajectory.

[0070] The annual trajectory integrity G of the ship is calculated according to the following formula:

[0071]

[0072] Wherein, G is the trajectory integrity, n is the number of ship trajectories, Q1 is the first quartile of the number of AIS annual trajectories of the ship in the region: 7955, and Q3 is the third quartile of the number of AIS annual trajectories of the ship in the region: 83909.

[0073] 3) Calculate the frequency T of the water depth of each type of ship being 0. Different types of ships have different requirements for their work, and the number of records and the accuracy of the records of their water depth will also be different, which is very important for the reliability analysis of the navigable water depth;

[0074] There are many records of 0 in the ship water depth record, and the water depth of 0 of different types of ships is counted. If the proportion of the water depth of 0 of the ship type is higher, the reliability of the record of the ship type is lower. If more than half of the water depth of a ship is 0, the water depth of the ship is 0.

[0075] For each type of ship, the frequency T of the type of ship draft depth is not 0 is calculated according to the following formula:

[0076] T = 1 - R1 / R2

[0077] Wherein, R1 is the number of ships in the type of ship draft depth is 0, R2 is the total number of the type of ship.

[0078] Taking the proportion of each type of ship in 2017 as an example, see the following table:

[0079]

[0080]

[0081] 4), calculate the ship voyage data integrity S. The AIS voyage data contains draft, so the integrity of the voyage data reflects the reliability of the draft record to some extent;

[0082] AIS provides three types of data. The static data of the ship includes the ship name, call sign, MMSI, IMO, ship type, length, width, etc.; the dynamic data of the ship includes longitude, latitude, ground course (COG), ground speed (SOG), etc.; the voyage data of the ship includes ship status, draft, destination, estimated time of arrival (ETA), etc. The more complete the voyage data is, the higher the reliability of the draft record is. For each key information field such as ship status, destination, ETA, the missing rate is calculated. The missing rate is the ratio of the number of missing data records to the total number of records. The missing rate of each field is calculated by a unified weight to obtain a comprehensive score.

[0083] The ship voyage data integrity S is calculated according to the following formula:

[0084]

[0085] Wherein, S is the completeness of the ship voyage data in the ship trajectory, L is the number of key fields of the ship voyage data (including: ship status, destination, ETA), S j is the missing rate of the jth key field, u is the number of missing fields in the ship trajectory, and n is the number of ship trajectories.

[0086] 5), single ship record reliability evaluation model, through the four factors of annual draft dispersion D, annual trajectory integrity G, ship draft depth frequency T and ship voyage data integrity S, the single ship record reliability X of each ship AIS record is evaluated:

[0087] X = a*D + β*G + γ*T + δ*S

[0088] Wherein, a, β, γ and δ are the weights of each factor.

[0089] To determine the weight of each parameter, the method of combining expert scoring and analytic hierarchy process (AHP) is adopted. AHP is a method combining qualitative and quantitative, which decomposes complex problems into several levels and factors, compares and calculates the order weight reflecting the relative importance of a factor between factors, and calculates the weight reflecting the order of relative importance of all factors through the total order between levels, so as to analyze complex factors. The specific method is as follows:

[0090] Firstly, a hierarchical structure is established: the target layer is the reliability of single ship record (X); the standard layer includes annual water dispersion D, annual trajectory integrity G, frequency of ship draft depth not being 0 T, and the completeness of ship voyage data in ship trajectory S.

[0091] Secondly, a judgment matrix is constructed: for each factor in the standard layer, 1-9 scale is used for pairwise comparison. The constructed judgment matrix A is as follows:

[0092]

[0093] Wherein, a DG represents the importance of annual trajectory integrity G compared with annual water dispersion D, a DT represents the importance of frequency of ship draft depth not being 0 T compared with annual water dispersion D; a DS represents the importance of the completeness of ship voyage data in ship trajectory S compared with annual water dispersion D; a GS represents the importance of the completeness of ship voyage data in ship trajectory S compared with annual trajectory integrity G; a GT represents the importance of frequency of ship draft depth not being 0 T compared with annual trajectory integrity G; a TS represents the importance of the completeness of ship voyage data in ship trajectory S compared with frequency of ship draft depth not being 0 T. For each factor in the standard layer, 1-9 scale is used for pairwise comparison, and scoring is performed according to importance from 1-9 (expert scoring). For example, 1 represents equal importance, 3 represents slight importance, 5 represents very important, 7 represents very important, 9 represents extremely important, and 2, 4, 6 and 8 are intermediate values, representing the importance between two adjacent levels; the inverse represents the opposite importance.

[0094] Then, the weight vector is calculated: using the eigenvalue method, the maximum eigenvalue λ max of the judgment matrix A and the corresponding eigenvector W are calculated.

[0095] Finally, consistency detection is performed according to the following formula:

[0096]

[0097] where CI is the consistency index, v is the order of the judgment matrix, find the average random consistency index RI≈9.0 corresponding to the four-order matrix, and calculate the consistency ratio

[0098] If CR<0.1, it is considered that the judgment matrix has satisfactory consistency, otherwise the judgment matrix needs to be adjusted; after passing the consistency test, the weight vector obtained is the weight of each parameter.

[0099] In this example, the maximum eigenvalue of matrix A is determined as λ_max=4.077, the consistency index CI=0.026, and the consistency ratio CR=0.0029<0.1, so the judgment matrix has satisfactory consistency. The corresponding eigenvector W=[0.948, 0.259, 0.098, 0.157], and the weight vector obtained by normalizing W is [0.648, 0.177, 0.067, 0.107].

[0100] The weights of each factor are set as follows: the weight of annual water dispersion is the highest, which is 0.648; the weight of trajectory integrity is the second, which is 0.177; the weight of ship type is 0.067; and the weight of ship voyage data is 0.107.

[0101] 6), calculate the relative dispersion degree RD of the draft depth. The relative dispersion degree of the draft depth refers to the difference between the draft depth of a ship in a specific region and the median of the draft depth records of the ship within a year. If the dispersion degree is large, it indicates that the reliability of the draft depth records in the region is low.

[0102] In a specific region, the draft depth of a ship and the draft depth records of the ship within a year will have differences. If the difference is too large, the reliability of the record is lower. Therefore, the relative dispersion degree of the median of the draft depth is used to detect the reliability of the draft depth in the grid within the whole year of the ship. The relative dispersion degree of the draft depth is:

[0103]

[0104] where d k is the draft depth of the ship in the grid k, d me is the median of the draft depth in the trajectory of the ship; the larger the RD, the more dispersed the draft depth, and the lower the reliability, otherwise the reliability is higher.

[0105] 7), the reliability interaction between draft depth similar ships, refers to the influence of draft depth similar ships on the reliable depth of water in the same region. If the number of draft depth similar ships is large and the reliability of the ship records is high, it may also affect the reliability of the reliable depth of water;

[0106] The reliability level of the ship record is taken as the quality of the ship, and the difference in draft is taken as the distance between the ships; at this time, the ship record reliability is proportional to the attraction, and the draft difference is inversely proportional to the attraction, which is consistent with cognition.

[0107] The navigable water depth reliability index Gra of the grid k is calculated according to the following formula k :

[0108]

[0109] where H max is the maximum draft in the grid k, H w is the draft of other ships w in the grid k, X max is the single-ship record reliability of the ship with the maximum draft in the grid k, X w is the single-ship record reliability of other ships w in the grid k, RD max is the relative dispersion degree of the draft of the ship with the maximum draft in the grid k, RD w is the relative dispersion degree of the draft of other ships w in the grid k.

[0110] Step 4, navigable water depth calculation - the statistical results of the draft distribution of the ships in the comprehensive grid and the reliability evaluation results of the navigable water depth are compared with the chart depth to form Figure 4 , and the reliability threshold is determined to be 0.53 according to Figure 4 , and the navigable water depth reliability results are formed as shown in Figure 5 , where the lighter the grid color, the higher the reliability, and the maximum draft that meets the threshold of the reliability in the grid is calculated to obtain the navigable water depth as shown in Figure 6 , where the lighter the grid color, the deeper the water depth.

[0111] In order to verify the reliability and stability of the port navigable water depth calculation method based on AIS data, the chart depth data is used for comparison, and the Pearson correlation coefficient (r), root mean square error (RMSE) and average relative error (ARE) are calculated respectively. The larger the r, the better the effect of the navigable water depth. On the contrary, the smaller the RMSE and ARE, the better the effect of the navigable water depth.

[0112] After calculation, the results of each statistic of different areas (such as wharf, inland channel, anchorage and strait main channel) and the whole port are shown in the following table. The results of the wharf and channel are better, and the anchorage and strait main channel are relatively worse, which is due to the fact that the water depth of the anchorage and strait main channel is usually deeper. In contrast, when the water depth is limited to 25 meters or shallower, the statistical results of different areas will be improved, especially in the anchorage and strait main channel. Therefore, when using AIS data to estimate the navigable water depth, the water depth limit of 25 meters or shallower will be more accurate.

[0113]

[0114]

[0115] The port navigable water depth calculation method based on AIS data is not limited to the specific technical solutions and implementation areas described in the above embodiments, and any technical solution formed by equivalent replacement is within the protection scope of the present application.

Claims

1. A method for calculating navigable water depth in ports based on AIS data, comprising the following steps: Step 1: AIS data preprocessing - Select the target port area to extract AIS data, and extract the AIS trajectory data of the ships within one year based on the MMSI identification code of the ships in the area. Perform trajectory error cleaning and attribute error cleaning on the extracted AIS data. Step 2, Structured Grid Subdivision – Generate a structured grid within the port area, subdivide the extracted AIS data into the grid according to spatial location, and perform statistics on the AIS data within the grid, including the number of ships, ship types, and ship draft distribution within the grid. Step 3, Navigable Depth Reliability Assessment – ​​Analyze the relevant factors of navigable depth reliability, establish a navigable depth reliability index system, and use the analytic hierarchy process (AHP) to divide it into two aspects: single-ship AIS record reliability assessment and grid-based ship reliability assessment. Specifically, it includes the following 7 points: 1) Calculate the annual draft dispersion D of the ship using the following formula: in, n is the number of the ship's trajectory, d i Let D be the draft at the i-th position in the ship's trajectory. When D is 0, it means that the data has no dispersion, that is, the data centrality is very high and the data are almost completely equal. When D is 1, the data has the highest dispersion, that is, the data centrality is extremely low and the data are very dispersed. 2) Calculate the annual track integrity G of the ship according to the following formula: Where G is the trajectory completeness, n is the number of ship trajectories, Q1 is the first quartile of the annual number of ship AIS trajectories in the region, and Q3 is the third quartile of the annual number of ship AIS trajectories in the region. 3) For each type of vessel, calculate the frequency T of non-zero draft for that type of vessel using the following formula: T = 1 - R1 / R2 Where R1 is the number of vessels of this type with a draft of 0, and R2 is the total number of vessels of this type; 4) Calculate the ship voyage data completeness S according to the following formula: Where S represents the completeness of the ship's voyage data in the ship's trajectory, and L represents the number of key fields in the ship's voyage data. Key fields include: ship status, destination, and ETA. j is the missing rate of the j-th key field, u is the number of missing values ​​of the j-th field in the ship trajectory, and n is the number of ship trajectories; 5) The reliability X of each ship's AIS record is evaluated using four factors: annual draft dispersion (D), annual trajectory integrity (G), frequency of non-zero ship draft (T), and ship voyage data integrity (S). X = α*D + β*G + γ*T + δ*S Wherein, α, β, γ, and δ are the weights of each factor. The weights of each factor are determined using a combination of expert scoring and the Analytic Hierarchy Process (AHP), as follows: First, a hierarchical structure is established: the target layer is the reliability of single-ship records X; the standard layer includes annual draft dispersion D, annual trajectory integrity G, frequency of ship draft not being zero T, and ship voyage data integrity S. Secondly, a judgment matrix is ​​constructed: for each factor in the standard layer, pairwise comparisons are performed using a ratio of 1-9, and the judgment matrix A is constructed as follows: Among them, a DG This indicates the relative importance of annual trajectory integrity G compared to annual draft dispersion D, a DT This indicates the relative importance of the frequency T (where the ship's draft is not zero) compared to the annual draft dispersion D; a DS This indicates the importance of the completeness (S) of the ship's voyage data in its trajectory compared to the annual draft dispersion (D); a GS This indicates the importance of the completeness S of ship voyage data in a ship trajectory compared to the annual trajectory completeness G; a GT This indicates the relative importance of the frequency T (where the ship's draft is not zero) compared to the annual trajectory integrity G; a TS The degree of completeness (S) of ship voyage data in the ship trajectory is compared to the importance of the frequency (T) of the ship's draft not being zero. Then, the weight vector is calculated: using the eigenvalue method, the largest eigenvalue λ of the judgment matrix A is calculated in one step. max and its corresponding eigenvector W; Finally, a consistency check is performed according to the following formula: Where CI is the consistency index, v is the order of the judgment matrix, and the average random consistency index RI≈9.0 corresponding to the fourth-order matrix is ​​found. The consistency ratio is then calculated. If CR < 0.1, the judgment matrix is ​​considered to have satisfactory consistency; otherwise, the judgment matrix needs to be adjusted. After passing the consistency test, the resulting weight vector is the weight of each parameter. 6) Calculate the relative dispersion of draft RD according to the following formula: In the formula, d k Let d be the draft of the vessel within grid k. me This represents the median draft of the ship's trajectory; a larger RD indicates a more dispersed draft and lower reliability, while a larger RD indicates higher reliability. 7) Calculate the navigable depth reliability index Gra of grid k according to the following formula. k : Where, H max H represents the maximum draft within the grid k. w X represents the draft of other vessels w within the grid k. max X represents the single-ship record reliability for the vessel with the maximum draft within grid k. w For the reliability of single-ship records of other ships w within grid k, RD max RD represents the relative dispersion of the draft of ships with the maximum draft within grid k. w The relative dispersion of the draft depth of other ships w within the grid k; Step 4: Navigable depth calculation – Based on the statistical results of the ship draft distribution within the grid and the reliability assessment results of navigable depth, calculate the maximum draft within the grid that meets the reliability threshold, and obtain the navigable depth.

2. The method for calculating navigable water depth of a port based on AIS data according to claim 1, characterized in that: Data cleaning mainly includes cleaning attributes such as ship type, length, width, and speed, as well as cleaning draft attributes based on the relationship between ship draft and length.

3. The method for calculating navigable water depth of a port based on AIS data according to claim 1, characterized in that: In step 1, the criteria for trajectory error cleaning are that the AIS point is not on land and the average speed between any trajectory point of the same MMSI ship and two adjacent trajectory points is less than 120 knots; the criteria for attribute error cleaning are that the ship type, time, and draft are not empty, the length threshold is 460 meters, the width threshold is 80 meters, and the speed threshold is 30 knots.

4. The method for calculating navigable water depth of a port based on AIS data according to claim 1, characterized in that: In step 2, the grid division can be set manually, with the default value being 50 meters.

5. The method for calculating navigable water depth of a port based on AIS data according to claim 1, characterized in that: In step 3, the analytic hierarchy process (AHP) is used, which is divided into two aspects: single-ship AIS record reliability assessment and grid-based ship reliability assessment. The single-ship AIS record reliability assessment includes four factors: annual draft dispersion, annual trajectory integrity, ship type, and ship voyage data. The grid-based ship reliability assessment includes two parts: relative draft dispersion and reliability interaction between ships with similar drafts.

6. The method for calculating navigable water depth of a port based on AIS data according to claim 1, characterized in that: In step 3, the maximum and minimum thresholds for annual trajectory completeness are the third quartile and the first quartile, respectively.

7. The method for calculating navigable port depth based on AIS data according to claim 1, characterized in that: In step 4, the reliability threshold within the grid is 53%.

Images