Search

CN-122024524-A - Ship arrival state determining method, device and readable storage medium

CN122024524ACN 122024524 ACN122024524 ACN 122024524ACN-122024524-A

Abstract

The application provides a ship arrival state determining method, a ship arrival state determining device and a readable storage medium. The ship arrival state determining method comprises the steps of determining dynamic approach degree according to navigation dynamic data and ship static attribute data, determining a tag sample set according to the dynamic approach degree and actual berthing state of a historical ship and a target port, obtaining historical characteristic data corresponding to the tag sample set according to the tag sample set, space characteristic data of the target port and ship static attribute data, determining a model training set and a model verification set according to the historical characteristic data and a plurality of state tags, performing modeling and verification through the model training set and the model verification set based on a random forest algorithm, determining an arrival judging model, obtaining target characteristic data of the target ship, determining arrival confidence according to the target characteristic data and the arrival judging model, and determining arrival state of the target ship based on the arrival confidence. The method and the device utilize the arrival judgment model to judge the arrival state, thereby improving the prediction precision.

Inventors

  • SHENG ZUNKUO

Assignees

  • 亿海蓝（北京）数据技术股份公司

Dates

Publication Date
20260512
Application Date
20260205

Claims (10)

  1. 1. A method for determining the arrival status of a ship, comprising: Acquiring navigation dynamic data and static attribute data of a historical ship, and determining dynamic approach degree according to the navigation dynamic data and the static attribute data of the ship, wherein the dynamic approach degree is used for representing the approaching target port degree of the ship; Determining a tag sample set according to the dynamic approach degree and the actual berthing state of the historical ship and the target port, wherein the tag sample set comprises a plurality of state tags used for representing whether the historical ship is in a arrived state or a non-arrived state; acquiring historical characteristic data corresponding to the tag sample set according to the tag sample set, the spatial characteristic data of the target port and the static attribute data of the ship; Determining a model training set and a model verification set according to the historical characteristic data and a plurality of state labels; Modeling and verifying through the model training set and the model verification set based on a random forest algorithm to determine a arrival judgment model; Acquiring target characteristic data of a target ship, and determining arrival confidence according to the target characteristic data and an arrival judgment model; and determining the arrival state of the target ship based on the arrival confidence.
  2. 2. The ship arrival status determination method according to claim 1, wherein the acquiring historical characteristic data corresponding to the tag sample set according to the tag sample set, the spatial characteristic data of the target port, and the ship static attribute data comprises: Extracting time sequence statistical characteristics of the dynamic approach according to a preset time window; And determining historical characteristic data corresponding to the tag sample set one by one according to the time sequence statistical characteristic, the space characteristic data of the target port and the static attribute data of the ship.
  3. 3. The ship arrival status determination method according to claim 1, wherein said determining a model training set and a model verification set based on said historical characteristic data and a plurality of said status labels comprises: and dividing the historical characteristic data and the state label corresponding to the historical characteristic data according to a preset proportion to obtain a model training set and a model verification set.
  4. 4. The ship arrival status determination method according to claim 1, wherein the determining the arrival judgment model by modeling and verifying through the model training set and the model verification set based on a random forest algorithm comprises: determining a plurality of sub-samples from the model training set by a self-help sampling method; constructing decision trees based on each sub-sample, each decision tree being trained on a randomly selected feature subset; Forming an initial random forest model through voting results of all decision trees; and carrying out cross validation on the initial random forest model through the model validation set to determine a arrival judgment model.
  5. 5. The ship arrival status determination method according to claim 1, wherein the target characteristic data is consistent with the dimensions and data type of the history characteristic data.
  6. 6. The ship arrival status determination method according to claim 1, wherein the determining the arrival status of the target ship based on the arrival confidence comprises: When the arrival confidence coefficient of the continuous multiple target characteristic data samples is greater than or equal to a first threshold value, determining that the target ship is in an arrived state; when the arrival confidence is larger than or equal to the second threshold value and smaller than the first threshold value and the continuous rising trend exists, determining that the target ship is in the estimated arrival state; when the arrival confidence is smaller than a second threshold, determining that the target ship is in an unoccupied state; wherein the first threshold is greater than the second threshold.
  7. 7. The ship arrival status determination method according to claim 1, further comprising a model updating step of: Acquiring updated historical ship navigation dynamic data, ship static attribute data, space characteristic data of a target port and corresponding actual berthing states; Determining updated dynamic proximity and a state label corresponding to the dynamic proximity based on the updated data to form an updated label sample set and updated historical characteristic data; And adding the updated historical characteristic data and a state label corresponding to the historical characteristic data into an original data set, re-dividing a model training set and a model verification set, performing incremental training on the arrival judgment model, and updating model parameters.
  8. 8. A ship arrival status determining apparatus, comprising: the system comprises a first acquisition module, a second acquisition module and a third acquisition module, wherein the first acquisition module is used for acquiring navigation dynamic data and ship static attribute data of a historical ship, and determining dynamic approach degree according to the navigation dynamic data and the ship static attribute data, wherein the dynamic approach degree is used for representing the approaching target port degree of the ship; The first determining module is used for determining a tag sample set according to the dynamic approach degree and the actual berthing state of the historical ship and the target port, wherein the tag sample set comprises a plurality of state tags used for representing that the historical ship is in a arrived state or in a non-arrived state; The second acquisition module is used for acquiring historical characteristic data corresponding to the tag sample set according to the tag sample set, the space characteristic data of the target port and the static attribute data of the ship; the second determining module is used for determining a model training set and a model verification set according to the historical characteristic data and the plurality of state labels; The third determining module is used for modeling and verifying through the model training set and the model verification set based on a random forest algorithm to determine a arrival judgment model; The fourth determining module is used for acquiring target characteristic data of the target ship and determining arrival confidence according to the target characteristic data and the arrival judgment model; And a fifth determining module, configured to determine an arrival status of the target ship based on the arrival confidence.
  9. 9. A ship arrival status determining apparatus, comprising: A processor; A memory having stored therein a program or instructions which when executed by the processor implement the steps of the ship arrival status determination method according to any one of claims 1 to 7.
  10. 10. A readable storage medium, characterized in that the readable storage medium has stored thereon a program or instructions which, when executed by a processor, implement the steps of the ship arrival status determination method according to any one of claims 1 to 7.

Description

Ship arrival state determining method, device and readable storage medium Technical Field The invention relates to the technical field of shipping and maritime monitoring, in particular to a ship arrival state determining method, a ship arrival state determining device and a readable storage medium. Background In the related technology, the existing ship arrival judging method depends on static rules such as a single speed threshold value, a geofence and the like, lacks analysis of dynamic behavior process of approaching a port to a ship, does not combine multidimensional dynamic characteristics of distance, speed and heading to construct a unified quantitative index, does not introduce a self-learning mechanism to adapt to different ship types and port environments, and causes obvious judging delay and insufficient predictability. Disclosure of Invention The present invention aims to solve at least one of the technical problems existing in the prior art or related art. To this end, a first aspect of the present invention proposes a ship arrival status determination method. A second aspect of the present invention proposes a ship arrival status determination apparatus. A third aspect of the present invention proposes another ship arrival status determination apparatus. A fourth aspect of the application proposes a readable storage medium. In view of the above, a first aspect of the present invention provides a ship arrival status determining method, including obtaining historical ship navigation dynamic data and ship static attribute data, determining a dynamic approach degree according to the navigation dynamic data and the ship static attribute data, the dynamic approach degree being used for representing a ship approach to a target port, determining a tag sample set according to the dynamic approach degree, an actual berthing status of the historical ship and the target port, the tag sample set including a plurality of status tags for representing the historical ship in an arrived status or in a non-arrived status, obtaining historical feature data corresponding to the tag sample set according to the tag sample set, the spatial feature data of the target port and the ship static attribute data, determining a model training set and a model verification set according to the historical feature data and the plurality of status tags, modeling and verifying the model training set and the model verification set based on a random forest algorithm, obtaining target feature data of the target ship, determining an arrival confidence level according to the target feature data and the arrival judgment model, and determining an arrival status of the target ship based on the arrival confidence. The navigation dynamic data of the application refers to real-time motion related data collected by a ship through an automatic identification system (Automatic Identification System, AIS), and the data comprise longitude and latitude, navigational speed, heading, time stamp and the like, and are core data reflecting the real-time navigation state of the ship. The static attribute data of the ship refers to inherent attribute information of the ship, which does not change along with the sailing process, and comprises ship types (cargo ships, oil tankers and the like), capturers, BRT, destination port names and the like, and is used for adapting to judging requirements of different ship types. The dynamic approach degree is a normalized index for representing the approach degree of the ship to the target port, and is obtained by calculating by fusing dynamic factors such as the distance change rate, heading deviation, speed change rate and the like of the ship and the target port, wherein the value range is [0,1], and the larger the value is, the more obvious the trend of the ship approaching the port is. The label sample set is a training data set with a status label of 'arrived' or 'not arrived', is generated based on dynamic approach degree and actual berthing status labels of the historical ship and the target port, and provides a supervision learning basis for subsequent machine learning modeling. The space characteristic data of the target port refers to data related to the geographic space of the target port, including coordinates of the target port, boundary ranges, real-time distance between the ship and the target port, and the like, and is used for supplementing the judgment basis of space dimension. The historical characteristic data is a multidimensional characteristic set formed by integrating time sequence statistical characteristics of dynamic approaches, ship static attribute data and target port space characteristic data, and can comprehensively reflect the dynamic process, self attribute and port environment characteristics of the ship approaching a port. The model training set is a data subset which is divided from the historical characteristic data and the corresponding state label and is used for training a