Encounter danger identification method and device based on incremental hierarchy density space clustering

Abstract

The invention discloses an encounter danger identification method and device based on incremental hierarchical density space clustering, and relates to the technical field of ship traffic safety and risk assessment. The method comprises the steps of obtaining ship state information in a target water area; processing the state information of the ships in the target water area through a hierarchical density adaptive spatial clustering algorithm, thereby clustering the ships in the target water area into a certain number of clusters; calculating danger encounter identification parameters between every two ships in the clustering cluster, wherein the danger encounter identification parameters comprise the nearest encounter distance, the minimum encounter time and the ship domain overlap index; and calculating the risk identification degree in the clustering cluster according to the risk encounter identification parameters between the ships in the clustering cluster, and carrying out collision risk encounter identification according to the risk degree numerical value. According to the method, the related data of the ship navigation state is fully utilized, the ship entity is abstracted into the area object with the space range by establishing the buffer area model, and the danger encounter identification accuracy is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of ship traffic safety and risk assessment, in particular to a dangerous encounter identification method and device based on incremental hierarchical density spatial clustering. BACKGROUND

[0002] More than 90% of the total global trade transportation is completed by sea transportation. As a key component of global trade, sea transportation carries large-scale goods and has unparalleled advantages in transportation cost and benefit. In recent years, with the expansion of international trade scale and the growth of logistics demand, the amount of ship transportation is increasing, the degree of water traffic is intensifying, and the risk of ship collision is also increasing.

[0003] According to the data statistics of the industry development bulletin of the Ministry of Transport of the People's Republic of China, the incidence of water traffic accidents has shown a significant downward trend in recent years. However, the complexity and uncertainty of the water environment may still pose a threat to safe navigation of ships. With the popularization of AIS equipment, the relevant data available for reference and research is becoming more and more massive, and traditional experience and intuition are not enough to deal with complex decision-making problems.

[0004] Most of the current research on the danger degree of ship collision uses clustering identification methods such as DBSCAN (Density-Based Spatial Clustering of Applications with Noise) and OPTICS (Ordering Points to Identify the Clustering Structure). These methods need to set clustering parameters in advance, and the applicability of AIS data is limited. At the same time, the clustering input is mostly the position of the ship, ignoring the basic information and motion state of the ship. In addition, a large amount of redundant AIS data has a huge impact on processing efficiency. Therefore, in the face of the complexity of ship traffic and the significant increase in the number of ships, it is urgent to design a dangerous encounter identification method that can fully and efficiently adapt to water ship traffic data and consider more comprehensive and various factors. SUMMARY

[0005] In view of the deficiencies of the prior art, the present application provides a dangerous encounter identification method and device based on incremental hierarchical density spatial clustering. The present application makes full use of the relevant data of the ship navigation state, abstracts the ship entity as a regional object with a spatial range by establishing a buffer zone model, and improves the accuracy of dangerous encounter identification.

[0006] The technical means adopted by the present application are as follows: A dangerous encounter identification method based on incremental hierarchical density spatial clustering, comprising the following steps: acquire ship state information in the target water area, the ship state information including ship latitude and longitude, ship speed, ship heading and ship length; process the ship state information in the target water area by a hierarchical density adaptive spatial clustering algorithm, so as to cluster the ships in the target water area into a certain number of clustering clusters; calculate a dangerous encounter identification parameter between each two ships in the clustering cluster, the dangerous encounter identification parameter including a closest encounter distance, a minimum encounter time and a ship domain overlap index; calculate a dangerous degree in the clustering cluster according to the dangerous encounter identification parameter between the ships in the clustering cluster, and identify a collision danger encounter according to the dangerous degree value.

[0007] Further, the ship state information further includes a ring or circular buffer zone constructed with the ship as a center, and the dangerous encounter identification parameter is calculated according to the buffer zone.

[0008] Further, the buffer zone radius is determined according to the ship speed and a preset ship safe speed, when the ship speed is less than or equal to the preset ship safe speed, the buffer zone is circular, and the circular buffer zone radius is a ship domain radius, and a calculation formula of the ship domain radius is:

[0009] wherein, the ship domain radius, a preset ship length and ship domain multiple relationship, the ship length; when the ship speed is greater than the preset ship safe speed, the buffer zone is ring-shaped, and a calculation formula of a ring-shaped buffer zone outer diameter is:

[0010]

[0011] wherein, the ring-shaped buffer zone outer diameter when the ship speed is greater than the ship safe speed, a current ship speed, a ship safe speed, a calculation formula of a ring-shaped buffer zone inner diameter is:

[0012] wherein, the ring-shaped buffer zone inner diameter, a preset ship length and ship domain multiple relationship, the ship length.

[0013] Further, the ship state information in the target water area is processed by a hierarchical density adaptive spatial clustering algorithm, so as to cluster the ships in the target water area into a certain number of clustering clusters, comprising: The ship field and the buffer area are taken as the ship area, the Euclidean distance between each two ships in the target water area is calculated, and the ships are clustered according to the Euclidean distance between the ships.

[0014] Further, the ship state information in the target water area is processed by a hierarchical density adaptive spatial clustering algorithm, so as to cluster the ships in the target water area into a certain number of clustering clusters, further comprising: in the ship dynamic scene, a ship dynamic correlation graph G=(V, E) is constructed according to the ship state information in the target water area, wherein the node set V is used to represent the real-time state of the ship, and the edge set E is used to represent the space-time correlation between the ships. The motion state of each ship is monitored in real time, and when the motion state of a ship reaches a preset condition, the ship is marked as a node to be updated, thereby triggering a local clustering adjustment mechanism.

[0015] Further, the ship state information in the target water area is processed by a hierarchical density adaptive spatial clustering algorithm, so as to cluster the ships in the target water area into a certain number of clustering clusters, further comprising: In the clustering process, when the core point displacement of the clustering cluster meets a preset condition, a cluster splitting detection mechanism is triggered.

[0016] Further, the ship state information in the target water area is processed by a hierarchical density adaptive spatial clustering algorithm, so as to cluster the ships in the target water area into a certain number of clustering clusters, further comprising: The local clustering result is combined with the cluster state cache of the incremental mechanism, and when the core point displacement of the cache cluster meets a preset condition, a local hierarchical density adaptive spatial clustering is triggered to re-cluster.

[0017] Further, the calculation formula of the closest encounter distance is as follows

[0018] Among them, represents the closest encounter distance, represents the distance between the ship and another target ship in the clustering cluster, represents the relative heading angle between the ship and the target ship, represents the relative bearing angle between the ship and the target ship; The calculation formula of the minimum encounter time is as follows

[0019] Among them, represents the relative speed between the ship and the target ship; The formula for calculating the ship field overlap index is as follows:

[0020] Wherein, (C1) represents the virtual center of the ship field of the own ship, (C2) represents the virtual center of the ship field of the target ship, (C3) represents the radius of the ship field of the own ship and the target ship.

[0021] Further, the dangerousness degree in the cluster is calculated according to the dangerous encounter identification parameters between the ships in the cluster, including: The spatial urgency index of the collision danger degree is calculated according to the closest encounter distance, the time urgency index of the collision danger degree is calculated according to the closest encounter time, and the distance and relative azimuth index of the ship collision danger is calculated according to the ship field overlap index; The spatial urgency index of the collision danger degree, the time urgency index of the collision danger degree and the distance and relative azimuth index of the ship collision danger are weighted and fused, so as to obtain the dangerousness degree between any two ships in the cluster; The dangerousness degrees of all the ships in the cluster are weighted and summed, so as to obtain the collision danger degree of the ship cluster.

[0022] The application further discloses a dangerous encounter identification device based on the incremental hierarchical density space clustering, which is used for realizing the dangerous encounter identification method based on the incremental hierarchical density space clustering, and comprises: A ship information acquisition unit is used for acquiring ship state information in a target water area, and the ship state information includes ship longitude and latitude, ship speed, ship heading and ship length. A clustering unit is used for processing the ship state information in the target water area through a hierarchical density adaptive space clustering algorithm, so as to cluster the ships in the target water area into a certain number of cluster groups. A calculation unit is used for calculating dangerous encounter identification parameters between every two ships in the cluster, and the dangerous encounter identification parameters include a closest encounter distance, a minimum encounter time and a ship field overlap index. A collision danger encounter identification unit is used for calculating a dangerousness degree in the cluster according to the dangerous encounter identification parameters between the ships in the cluster, and performing collision danger encounter identification according to the dangerousness degree value.

[0023] Compared with the prior art, the application has the following advantages: ​1. The buffer collision danger encounter identification method based on the incremental hierarchical density adaptive spatial clustering provided by the application can fully utilize the relevant data of the ship navigation state, abstract the ship entity into a regional object with a spatial range by establishing a buffer model, instead of simplifying it into a point element, thereby effectively preserving the basic spatial characteristics of the ship. The spatial information loss problem caused by the low input feature dimension in the traditional point element clustering analysis is avoided, and the accuracy and reliability of the clustering analysis are improved.

[0024] 2. The hierarchical density adaptive spatial clustering algorithm used in the application adaptively determines the optimal clustering parameters (including the minimum sample number, distance threshold and compression degree of the number of layers) through a hierarchical clustering mechanism, avoiding the parameter sensitivity problem of traditional density clustering algorithms. Based on the clustering principle of core points and reachability distance, the algorithm accurately identifies the potential collision risk area in the ship encounter process. Thanks to its density-based clustering characteristics, the algorithm can effectively distinguish noise points from effective clustering clusters, reducing the influence of useless calculation and data on the final structure. By referring to three-dimensional parameters, DCPA, TCPA and SDOI, the threshold is set to further update the clustering results, improving the identification accuracy of dangerous encounters.

[0025] 3. The incremental hierarchical density adaptive spatial clustering marks the nodes that need to be updated by introducing an incremental clustering mechanism and a sliding dynamic window, thereby triggering a local clustering adjustment mechanism to realize real-time processing of dynamic data streams and reduce the computational complexity to meet the actual needs of ship dynamic monitoring and analysis. DETAILED DESCRIPTION

[0026] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0027] Figure 1 An encounter danger identification method based on incremental hierarchical density spatial clustering in an embodiment of the application is shown in the flowchart.

[0028] Figure 2 A buffer diagram in an embodiment of the application is shown in the flowchart.

[0029] Figure 3 An improved hierarchical density adaptive spatial clustering diagram in an embodiment of the application is shown in the flowchart.

[0030] Figure 4 A re-screening hierarchical density adaptive spatial clustering diagram in an embodiment of the application is shown in the flowchart.

[0031] Figure 5 A ship motion parameter diagram in an embodiment of the present application.

[0032] Figure 6 A dangerous encounter quantification and ordering in an embodiment of the present application.

[0033] Figure 7 A collision dangerous encounter identification method effect diagram based on incremental hierarchical density adaptive spatial clustering in an embodiment of the present application.

[0034] Figure 8 A ship clustering cluster diagram in an embodiment of the present application. DETAILED DESCRIPTION

[0035] In order to make the personnel in the art better understand the present application scheme, the technical scheme in the embodiments of the present application will be described clearly and completely below in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by the person of ordinary skill in the art without creative labor should belong to the scope of protection of the present application.

[0036] It should be noted that the terms "first", "second", and the like in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0037] Embodiment 1 The embodiment discloses a method for identifying dangerous encounter based on incremental hierarchical density spatial clustering. Firstly, basic information such as ship position, speed and heading is extracted, and after the buffer zone is calculated, a certain number of clustering clusters of ships in the water area are formed. At the same time, the incremental clustering method causes local clustering when the ship movement exceeds the defined threshold, realizes real-time monitoring, and reduces the calculation degree to improve the efficiency. Secondly, from the perspective of geometric encounter, the collision danger with the surrounding ships is described through the collision avoidance parameters, the closest encounter distance DCPA, the minimum encounter time TCPA and the ship domain overlap index SDOI between each two ships in the clustering cluster are calculated, the threshold is set according to the water traffic condition, and the ships that do not meet the conditions are screened out as noise points. Finally, the mapping relationship between the three-dimensional parameters and the collision danger is established, the collision danger result is sorted to realize the danger identification. The identification result can be further graded and highlighted. The invention can mine valuable information from a large amount of data, quickly identify potential risk ship encounters, help maritime supervision personnel better understand and evaluate the maritime traffic situation, improve the monitoring efficiency of maritime traffic safety, effectively reduce the error of human judgment, and provide a scientific basis for maritime traffic management decision-making. Figure 1 The method comprises the following steps.

[0038] S1, obtaining ship state information in a target water area, the ship state information comprising ship latitude and longitude, ship speed, ship heading and ship length.

[0039] Specifically, this step is used to extract ship latitude and longitude, ship speed, ship heading, ship length and other information data capable of representing the working and motion state of the ship in the target water area through various navigation instruments and equipment, so as to establish a buffer zone incremental clustering model of the ship in the target water area.

[0040] As a preferred embodiment of the present application, a buffer zone is established around the ship in the embodiment, as shown in Figure 2 The buffer zone refers to a virtual area set around the ship domain, which is a circular area with the ship as the center. The size of this area can be adjusted according to various factors, and the present application selects the current speed of the ship and the safe speed of the ship to adjust the size of the ship buffer zone. The ship domain radius and the buffer zone radius are calculated according to the following formula.

[0041] When the ship speed is less than or equal to the preset safe speed of the ship, the buffer zone is a circular buffer zone, and the circular buffer zone radius is the ship domain radius, and the calculation formula of the ship domain radius is:

[0042] wherein, For the ship field radius, For the preset ship length and ship field multiple relationship, For the ship length; The multiple A can be obtained by referring to Table 1.

[0043] Table 1 Ship length and ship field multiple relationship

[0044] When the ship speed is greater than the preset ship safe speed, the buffer zone is an annular buffer zone, and the calculation formula of the outer diameter of the annular buffer zone is:

[0045]

[0046] Wherein, The outer diameter of the annular buffer zone when the ship speed is greater than the ship safe speed, Is the current ship speed, Is the ship safe speed.

[0047] The calculation formula of the inner diameter of the annular buffer zone is:

[0048] Wherein, The inner diameter of the annular buffer zone, For the preset ship length and ship field multiple relationship, For the ship length.

[0049] The calculation formula of the safe speed is as follows:

[0050] Wherein, Indicates the ship safe speed, m / s; Indicates the initial stability height of the ship, m; Indicates the ship length, m; Indicates the height of freeboard, m; Indicates the ship width, m; Indicates the height of the center of gravity of the ship from the baseline, m; Indicates the ship draft, m.

[0051] S2, the ship state information in the target water area is processed by the hierarchical density adaptive spatial clustering algorithm, so as to cluster the ships in the target water area into a certain number of clustering clusters. The number of clustering clusters is automatically generated by clustering analysis.

[0052] This step is used for automatically grouping ship data by incremental hierarchical density adaptive spatial clustering algorithm, revealing the inherent structure and pattern of the data, and the application in the ship clustering model can include the ships with close distance and similar motion trajectory in the same cluster, avoiding the calculation of each two ships in the water area by analytical method, effectively reducing the calculation amount. The ship clustering cluster with possible collision danger is identified, avoiding the calculation between ships without collision danger or with low collision danger, reducing the dilution degree of the final result. As shown in Figure 8 , a ship clustering cluster is shown.

[0053] As a further preferred embodiment of the present application, in the process of ship clustering, the ship is not regarded as a point, but as a region, that is, the buffer zone is considered in the clustering process. The distance between each two ships in the target sea area is calculated according to the latitude and longitude of the ship, which is the minimum distance between the buffer zones of the two ships, and preferably the Euclidean distance is used in the model used in the present application. The Euclidean distance is used to measure the straight line distance between two points in a multi-dimensional space, and the concept is relatively intuitive, and is also applicable to higher dimensional space. The hierarchical density adaptive spatial clustering algorithm used in this embodiment is shown in Table 2.

[0054] Table 2 Pseudo code of hierarchical density adaptive spatial clustering algorithm

[0055] As a further preferred embodiment of the present application, in the ship dynamic scene, real-time clustering update faces the challenge of high computational complexity. In view of this problem, the application introduces an incremental clustering mechanism and a sliding dynamic window, focusing on a local area update strategy, and only performs incremental calculation operation on the area where the data changes due to ship movement, effectively avoiding the re-clustering process in the global range, and then significantly improving the real-time processing efficiency to meet the actual needs of ship dynamic monitoring and analysis.

[0056] According to the space-time constraint condition, a ship dynamic correlation graph is constructed . Among them, the node set is used to represent the real-time state of the ship, specifically covering the position , speed and heading and other key parameters; the edge set is used to represent the space-time correlation between ships. If the space-time distance between two ships and ( is a pre-set neighborhood radius) satisfies the condition of being less than or equal to , and the edge is marked to indicate whether the two ships it connects belong to the same cluster. The spatiotemporal distance is obtained by extracting key information about the ship's position, speed, and heading in the sliding time window, and calculating the distance between the two ships' buffer zones at the same instant as the spatiotemporal distance.

[0057] Real-time monitoring of the ship's movement status. The displacement change Meet or exceed , or its speed A sudden change occurs (i.e. the speed change per unit time exceeds the preset speed change threshold, ) and heading Sudden change (i.e. the change in heading per unit time exceeds the preset heading change threshold) ) when the ship Nodes are marked as requiring updates, triggering a local clustering adjustment mechanism to ensure that clustering results can timely and accurately reflect the dynamic changes of ships. The local clustering adjustment mechanism is configured to monitor ship movements and mark them as update nodes when their distance, navigation, or speed exceed a set threshold. Full clustering is no longer performed, and existing clusters are treated as a whole. Only update nodes are clustered, meaning other clustering results remain unchanged. Update nodes are added to existing clusters, new clusters, or noise points.

[0058] In the clustering process, maintain the core point set C of the cluster _core and the boundary point set C _boundary The hierarchical density adaptive spatial clustering algorithm automatically identifies the core points and boundary points of clusters.

[0059] A cluster has a core point and the others are boundary points. Only when the core point c∈C _core The displacement change Exceed When the cluster split detection mechanism is triggered, unnecessary redundant calculations are avoided, the allocation of computing resources is optimized, the efficiency and performance of incremental clustering are further improved, and the efficient and stable operation of the entire cluster update process is ensured. The cluster split detection mechanism is set as follows: when the core point of the cluster changes more than the set threshold, all clusters are marked as update nodes and participate in clustering again. The local clustering results generated by the hierarchical density adaptive spatial clustering are combined with the cluster state cache of the incremental mechanism to achieve dynamic adjustment. When the core point of the cached cluster shifts, the cluster state cache is automatically generated. When , the local hierarchical density adaptive spatial clustering is triggered to re-cluster and generate new subclusters through agglomerative tree split detection. Since the information is stored in a sliding dynamic window, the storage time of the data can be set independently to obtain cache clusters.

[0060] It should be noted that the concept of buffer zone is mainly used for the calculation of distance in clustering, and there is no buffer zone in the parameters of subsequent collision risk calculation.

[0061] S3, calculate the dangerous encounter identification parameters between each two ships in the clustering cluster, the dangerous encounter identification parameters including the closest encounter distance, the minimum encounter time and the ship domain overlap index.

[0062] The parameter information of the target ship in the clustering cluster is obtained by radar and ship AIS equipment, and the position of the ship is taken as the origin of the coordinate axis, i.e. , the position of the target ship is , the speed of the ship is , the speed of the target ship is , the heading of the ship is , and the heading of the target ship is , as shown in Figure 5 .

[0063] The speed of the ship on the X, Y coordinate axes is and :

[0064] The speed of the target ship on the X, Y coordinate axes is and

[0065]

[0066] The relative speed of the ship and the target ship is :

[0067] The relative bearing of the two ships is :

[0068] The distance between the two ships is :

[0069] The relative heading of the ship and the target ship is :

[0070] The closest encounter distance between the ship and the target ship is The calculation formula is

[0071] wherein, ​​represents the closest encounter distance, represents the distance between the target ship and the cluster, represents the relative bearing angle between the target ship and the own ship, represents the relative bearing angle between the target ship and the own ship; the minimum encounter time The calculation formula is as follows:

[0072] wherein, represents the relative speed between the target ship and the own ship.

[0073] To improve the calculation accuracy of the collision risk, the present application adopts SDOI (ship domain overlap index), DCPA and TCPA, three-dimensional parameters to calculate the ship collision risk. The ship domain model used in this paper is a circular ship domain.

[0074]

[0075] wherein, represents the virtual center of the ship domain of the own ship, represents the virtual center of the ship domain of the target ship, represents the radius of the ship domain of the own ship and the target ship. Considering the influence of the 1972 International Regulations on the Prevention of Collisions at Sea, the actual ship domain center needs to be moved to the port side and the stern of the ship to highlight the collision risk of the ship coming from the right front position. The original center of the ship domain is the virtual center. The specific movement can be set according to different water conditions.

[0076] After obtaining the calculation results of the above three parameters, the ships in the ship cluster are reselected according to the threshold limit condition, such as Figure 4 as shown, the ships that do not meet the requirements are excluded from the cluster as noise points, and the dangerous encounter is accurately identified. The threshold limit condition is to limit the threshold value of the three parameters of DCPA, TCPA and SDOI, for example, the DCPA should be less than 2 nautical miles, and more than 2 nautical miles is not qualified.

[0077] S4, according to the dangerous encounter identification parameters between the ships in the cluster, the dangerous identification degree in the cluster is calculated, and the collision danger encounter is identified according to the dangerous degree value.

[0078] With the continuous development of shipping industry, the carrying capacity, tonnage, ship size and speed of the ship are also increasing, and the difficulty and time required for collision avoidance are also increasing. The method is mainly aimed at calculating the potential risk of ship in open waters, so the related parameters in the model are less restrictive. Considering the value range of CRI and its nonlinear negative correlation with the reference factor, statistical data of critical risk values corresponding to different indicators are collected.

[0079] By applying the negative exponential function, the specific influence of DCPA (closest point of approach), TCPA (time to closest point of approach) and SDOI (ship domain overlap index) on collision risk index CRI is quantified respectively. Finally, by setting a critical risk threshold, the specific collision risk index CRI between ships is determined.

[0080]

[0081]

[0082]

[0083] In the above equation, represents the spatial urgency index of DCPA collision risk, represents the time urgency index of TCPA collision risk, represents the distance and relative bearing index of SDOI ship collision risk. The coefficient of negative exponential equation and can be obtained by setting the values of independent variables and dependent variables and calculating, and the reference coefficients used in the invention for open waters are shown in Tables 3-5.

[0084] Table 3 Value function reference coefficient table

[0085] Table 4 Value function reference coefficient table

[0086] Table 5 Value function reference coefficient table

[0087] After considering the three-dimensional collision risk index of DCPA, TCPA and SDOI, the collision risk calculation formula is proposed, which integrates the risk index of the three dimensions to calculate the overall collision risk between ships. The specific calculation method is shown in the formula:

[0088] wherein, , and are the target weights, previous studies have shown that the importance of DCPA and TCPA is higher than 70%, so set =0.4, =0.4, =0.2 CRI is defined as a represents the collision risk of ship a, the collision risk of ship a is closely related to all ships in the cluster, so ship a is paired with each ship in the cluster to calculate. After calculating the collision risk value of each ship, the collision risk of the ship cluster can be calculated by the following formula:

[0089] wherein, n is the number of ships in the cluster, is the collision risk of the ship, is the weight of the ship.

[0090]

[0091] wherein, is the length of the ship, N is the number of ships in the water area, and n is the number of ships in the cluster.

[0092] As a preferred embodiment of the present application, the present embodiment faces a large amount of ship data, and for the convenience of supervisors, a simple and clear real-time collision risk evaluation display method is proposed, which divides the ship collision risk into four levels: safe, unsafe, dangerous, and very dangerous.

[0093] Safe: that is, the collision risk between ships is low, and no collision risk will occur without changing the direction of navigation, that is, no collision risk is identified, and the ship is set as a noise point.

[0094] Unsafe: that is, there is a certain collision risk between ships, and the collision risk is identified in the collision risk encounter identification, and the quantitative mapping collision risk index is 30%≤CRC<60%.

[0095] Dangerous: that is, there is a collision risk, which needs to be avoided by reducing speed, changing direction, etc. The quantitative mapping collision risk index is 60%≤CRC<80%.

[0096] Very dangerous: that is, the dangerous encounter needs to be communicated immediately, and the road ship needs to take action immediately to avoid ship collision accidents. The quantitative mapping collision risk index is CRC>80%.

[0097] In visualization, such as Figure 6As shown, the unsafe ship clustering cluster is yellow, the dangerous ship clustering cluster is orange, and the very dangerous clustering cluster is red. When the mouse moves to the clustering cluster, the danger level, DCPA, TCPA, SDOI, and CRC index of the clustering cluster are displayed. At the same time, the clustering clusters in the monitored water area are sorted from top to bottom according to the danger degree, and are highlighted according to different danger degree indexes, so as to improve the supervision efficiency. The overall effect is as shown in Figure 7

[0098] Embodiment 2 The embodiment discloses a device for identifying encounter danger based on incremental hierarchical density spatial clustering, which is used to realize the method for identifying encounter danger based on incremental hierarchical density spatial clustering as disclosed in Embodiment 1, and comprises the following steps: A ship information acquisition unit is configured to acquire ship state information in a target water area, wherein the ship state information comprises ship longitude and latitude, ship speed, ship heading, and ship length.

[0099] A clustering unit is configured to process the ship state information in the target water area by using a hierarchical density adaptive spatial clustering algorithm, so as to cluster the ships in the target water area into a certain number of clustering clusters.

[0100] A calculation unit is configured to calculate a dangerous encounter identification parameter between each two ships in a clustering cluster, wherein the dangerous encounter identification parameter comprises a closest encounter distance, a minimum encounter time, and a ship domain overlap index.

[0101] A collision danger encounter identification unit is configured to calculate a danger identification degree in a clustering cluster according to the dangerous encounter identification parameter between the ships in the clustering cluster, and to identify a collision danger encounter according to the danger degree value.

[0102] For the device for identifying collision danger encounter based on incremental hierarchical density adaptive spatial clustering, since it corresponds to the method for identifying collision danger encounter based on incremental hierarchical density adaptive spatial clustering in the above embodiment, the description is relatively simple, and the relevant similar parts can be referred to the description of the method for identifying collision danger encounter based on incremental hierarchical density adaptive spatial clustering in the above embodiment.

[0103] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, but not to limit the same; although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the above embodiments, or make equivalent replacement for part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.​

Claims

1. A method for identifying hazard encounters based on incremental hierarchical density spatial clustering, characterized in that: The following steps are involved: Obtaining ship status information within the target waters, the ship status information including the ship's latitude and longitude, ship speed, ship heading, and ship length; The ship status information in the target waters is processed by a hierarchical density adaptive spatial clustering algorithm, so that the ships in the target waters are clustered into a certain number of clusters; Calculating dangerous encounter identification parameters between every two ships in the cluster, wherein the dangerous encounter identification parameters include the closest encounter distance, the minimum encounter time, and the ship area overlap index; Calculating a risk recognition degree within a cluster based on risk encounter recognition parameters between ships in the cluster, and performing collision risk encounter recognition based on the risk degree value; The ship status information also includes an annular or circular buffer zone constructed with the ship as the center, and the dangerous encounter identification parameter is calculated based on the buffer zone; the buffer zone radius is determined according to the ship speed and the preset ship safety speed. When the ship speed is less than or equal to the preset ship safety speed, the buffer zone is circular, and the circular buffer zone radius is the ship area radius. The calculation formula of the ship area radius is: in, is the radius of the ship's area, For the preset relationship between the captain and the ship area multiple, is the length of the ship; when the ship speed is greater than the preset safe speed of the ship, the buffer zone is annular, and the calculation formula for the outer diameter of the annular buffer zone is: in, is the outer diameter of the annular buffer zone when the ship speed is greater than the safe speed of the ship, is the current ship speed, is the safe speed of the ship, The calculation formula for the inner diameter of the ring buffer is: in, is the inner diameter of the ring buffer, For the preset relationship between the captain and the ship area multiple, is the length of the ship; Among them, the ship status information in the target waters is processed by the hierarchical density adaptive spatial clustering algorithm, so that the ships in the target waters are clustered into a certain number of clusters, including: The ship area and buffer zone are used as the ship area, the Euclidean distance between every two ships in the target waters is calculated, and the ships are clustered according to the Euclidean distance between the ships.

2. The method for identifying hazard encounters based on incremental hierarchical density spatial clustering according to claim 1, characterized in that: The ship status information in the target waters is processed by a hierarchical density adaptive spatial clustering algorithm, thereby clustering the ships in the target waters into a certain number of clusters. The method also includes: in a ship dynamic scene, constructing a ship dynamic association graph G = (V, E) based on the ship status information in the target waters, where the node set V is used to represent the real-time status of the ships, and the edge set E is used to represent the spatiotemporal correlation between the ships; The motion status of each ship is monitored in real time. When the motion status of a ship reaches the preset conditions, the ship is marked as a node that needs to be updated, thereby triggering the local clustering adjustment mechanism.

3. The method for identifying hazard encounters based on incremental hierarchical density spatial clustering according to claim 2, characterized in that: The ship status information in the target waters is processed by a hierarchical density adaptive spatial clustering algorithm, thereby clustering the ships in the target waters into a certain number of clusters, which also includes: During the clustering process, when the core point displacement of the cluster meets the preset conditions, the cluster split detection mechanism is triggered.

4. The method for identifying hazard encounters based on incremental hierarchical density spatial clustering according to claim 2, characterized in that: The ship status information in the target waters is processed by a hierarchical density adaptive spatial clustering algorithm, thereby clustering the ships in the target waters into a certain number of clusters, which also includes: The local clustering results are combined with the cluster state cache of the incremental mechanism. When the displacement of the core point of the cached cluster meets the preset conditions, the local hierarchical density adaptive spatial clustering is triggered to re-cluster.

5. The method for identifying hazard encounters based on incremental hierarchical density spatial clustering according to claim 1, characterized in that: The formula for calculating the closest encounter distance is as follows in, Indicates the closest distance. Indicates the distance between the ship and another target ship in the cluster, Indicates the relative heading angle between own ship and target ship. Indicates the relative bearing between own ship and target ship; The calculation formula for the minimum encounter time is as follows in, Indicates the relative speed between own ship and target ship; The calculation formula for the ship sector overlap index is as follows: in,( ) represents the virtual center of the own ship's ship domain, ( ) represents the virtual center of the target ship’s ship domain, Indicates the radius of the ship area of ​​own ship and target ship.

6. The method for identifying hazard encounters based on incremental hierarchical density spatial clustering according to claim 1, characterized in that: The hazard recognition degree within the cluster is calculated based on the hazard encounter recognition parameters between the ships in the cluster, including: The spatial urgency index of the collision risk is calculated based on the closest approach distance, the temporal urgency index of the collision risk is calculated based on the closest approach time, and the distance and relative bearing index of the ship collision risk is calculated based on the ship area overlap index; The spatial urgency index of collision risk, the temporal urgency index of collision risk, the distance and relative orientation index of ship collision risk are weighted and fused to obtain the risk recognition degree between any two ships in the cluster. The risk recognition degrees of all ships in the cluster are weighted and summed to obtain the collision risk degree of the ship cluster.

7. A device for identifying danger encounters based on incremental hierarchical density spatial clustering, used to implement the method for identifying danger encounters based on incremental hierarchical density spatial clustering as claimed in claim 1, characterized in that: include: A ship information acquisition unit, which is used to acquire ship status information in the target waters, wherein the ship status information includes the ship's latitude and longitude, ship speed, ship heading, and ship length; A clustering unit is used to process the ship status information in the target waters by using a hierarchical density adaptive spatial clustering algorithm, thereby clustering the ships in the target waters into a certain number of clusters; A calculation unit, which is used to calculate dangerous encounter identification parameters between every two ships in the cluster, wherein the dangerous encounter identification parameters include the closest encounter distance, the minimum encounter time and the ship area overlap index; The collision risk encounter identification unit is used to calculate the risk recognition degree within the cluster according to the risk encounter identification parameters between the ships in the cluster, and perform collision risk encounter identification according to the risk degree value.

Images