Ship Intelligent Navigation Analysis Method and System Based on Situation Awareness

Abstract

This application relates to the field of intelligent navigation analysis technology. Specifically, it discloses a ship intelligent navigation analysis method and system based on situation awareness. It uses a neural network model based on deep learning to extract multi-level features of navigation environment monitoring information, and conducts context correlation analysis on the multi-level features of the navigation environment monitoring information to achieve the significance distribution modeling of different-level features, thereby identifying key features in the navigation environment monitoring information and filtering out redundant information, so as to facilitate the situation awareness of the navigation environment and the intelligent evaluation of the channel traffic status on this basis. In this way, the processing efficiency and accuracy of the navigation environment monitoring information can be effectively improved, and further the reliability and safety of ship intelligent navigation analysis can be enhanced.

Description

Technical Field

[0001] This application relates to the technical field of intelligent navigation analysis, and more specifically, to a ship intelligent navigation analysis method and system based on situation awareness. Background Art

[0002] With the growth of global trade and the increasing busyness of maritime transportation, the safety and efficiency of ship navigation have become the focus of attention. To improve the autonomous navigation ability of ships in complex and changeable environments, intelligent navigation analysis methods and technologies based on situation awareness have developed rapidly. For example, the invention patent with the publication number CN118245756B proposes a ship intelligent navigation analysis method and system based on situation awareness. It perceives the navigation environment according to the navigation environment monitoring information, performs land-water segmentation on the navigation environment monitoring image, marks non-navigating ships and navigation obstacles at the same time, and then segments the navigable area, and uses digital twin technology to build a channel digital model to realize the real-time monitoring of the channel situation and the assessment of the passing status, so as to provide intelligent navigation decision support for ships.

[0003] In the prior art, the processing of navigation environment monitoring information mainly realizes the perception of the navigation environment based on the extraction and fusion of features from low level to high level of the navigation environment monitoring information. However, this processing method ignores the control of feature quality. Specifically, in the complex ship navigation environment, the sources of monitoring information are extensive and complex, including a large amount of noise and redundant information. For example, affected by factors such as weather changes (such as heavy rain, thick fog), lighting conditions (night, strong light reflection), and water area environment (turbid water body, complex underwater terrain interference), background information and error signals irrelevant to ship navigation decisions may be mixed into the monitoring information. Directly fusing these unfiltered features will result in a feature map after fusion containing a large amount of interference information, reducing the recognition rate of key features. This not only increases the computational burden of subsequent analysis, but also may mislead the judgment of the navigation environment, reducing the accuracy and reliability of subsequent environmental situation awareness and channel passing status assessment.

[0004] Therefore, an optimized ship intelligent navigation analysis method and system based on situation awareness is expected. Summary of the Invention

[0005] To solve the above technical problems, the present application is proposed. An embodiment of the present application provides a method and system for intelligent navigation analysis of ships based on situation awareness. It uses a neural network model based on deep learning to extract multi-level features of navigation environment monitoring information, and performs context correlation analysis on the multi-level features of navigation environment monitoring information to achieve significance distribution modeling of different-level features, thereby identifying key features in the navigation environment monitoring information and filtering out redundant information, so as to facilitate situation awareness of the navigation environment and intelligent evaluation of the channel passage state on this basis. In this way, the processing efficiency and accuracy of navigation environment monitoring information can be effectively improved, and the reliability and safety of ship intelligent navigation analysis can be further improved.

[0006] According to one aspect of the present application, there is provided a method for intelligent navigation analysis of ships based on situation awareness, which includes:

[0007] Obtain navigation environment monitoring information;

[0008] Extract features from low level to high level of the navigation environment monitoring information to obtain a set of navigation environment image monitoring features;

[0009] Perform feature filtering processing based on significance distribution modeling on the set of navigation environment image monitoring features to obtain a set of filtered navigation environment image monitoring features;

[0010] Calculate the attention scores of each feature in the set of filtered navigation environment image monitoring features to generate an attention feature map, and build a feature pyramid network to fuse feature maps of different levels to obtain a multi-level fusion feature map of the navigation environment image;

[0011] Perform object detection and water-land segmentation on the multi-level fusion feature map of the navigation environment image to obtain environment perception information;

[0012] Evaluate the channel situation based on the environment perception information to obtain channel situation evaluation information.

[0013] According to another aspect of the present application, there is provided a system for intelligent navigation analysis of ships based on situation awareness, which includes:

[0014] An environmental monitoring information acquisition module for acquiring navigation environment monitoring information;

[0015] An environmental monitoring feature extraction module for extracting features from low level to high level of the navigation environment monitoring information to obtain a set of navigation environment image monitoring features;

[0016] A feature filtering processing module, configured to perform feature filtering processing on the set of navigation environment image monitoring features based on saliency distribution modeling to obtain a set of filtered navigation environment image monitoring features;

[0017] A feature fusion module, configured to calculate the attention scores of each feature in the set of filtered navigation environment image monitoring features to generate an attention feature map, and build a feature pyramid network to fuse feature maps of different levels to obtain a multi-level fusion feature map of the navigation environment image;

[0018] An environment perception module, configured to perform object detection and water-land segmentation on the multi-level fusion feature map of the navigation environment image to obtain environment perception information;

[0019] A channel situation assessment module, configured to evaluate the channel situation based on the environment perception information to obtain channel situation assessment information.

[0020] Compared with the prior art, the ship intelligent navigation analysis method and system based on situation awareness provided by this application extract multi-level features of navigation environment monitoring information by using a neural network model based on deep learning, and perform context correlation analysis on the multi-level features of navigation environment monitoring information to realize saliency distribution modeling of different-level features, so as to identify key features in navigation environment monitoring information and filter out redundant information, so as to facilitate situation awareness of the navigation environment and intelligent assessment of the channel passage state on this basis. In this way, the processing efficiency and accuracy of navigation environment monitoring information can be effectively improved, and the reliability and safety of ship intelligent navigation analysis can be further improved. Description of the Drawings

[0021] By describing the embodiments of the present application in more detail in conjunction with the drawings, the above and other objects, features and advantages of the present application will become more obvious. The drawings are used to provide a further understanding of the embodiments of the present application, and constitute a part of the specification, and are used to explain the present application together with the embodiments of the present application, and do not constitute a limitation to the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0022] Figure 1 It is a flowchart of the ship intelligent navigation analysis method based on situation awareness according to the embodiment of the present application.

[0023] Figure 2 It is a data flow diagram of the ship intelligent navigation analysis method based on situation awareness according to the embodiment of the present application.

[0024] Figure 3 It is a flowchart of sub-step S3 of the ship intelligent navigation analysis method based on situation awareness according to the embodiment of the present application.

[0025] Figure 4 It is a flowchart of sub-step S31 of the ship intelligent navigation analysis method based on situation awareness according to an embodiment of the present application.

[0026] Figure 5 It is a flowchart of sub-step S32 of the ship intelligent navigation analysis method based on situation awareness according to an embodiment of the present application.

[0027] Figure 6 It is a block diagram of the ship intelligent navigation analysis system based on situation awareness according to an embodiment of the present application. Detailed implementation manners

[0028] As shown in the present application and the claims, unless the context clearly indicates otherwise, words such as "a", "an", "one" and / or "the" are not specifically singular and may also include plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0029] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, however, any number of different modules can be used and run on the user terminal and / or the server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0030] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the operations before or below are not necessarily executed precisely in order. On the contrary, as needed, various steps can be executed in reverse order or simultaneously. At the same time, other operations can also be added to these processes, or one or several steps can be removed from these processes.

[0031] Next, exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. It should be understood that the present application is not limited by the exemplary embodiments described here.

[0032] It is worth noting that in the present application, all actions of obtaining data are carried out on the premise of complying with the corresponding data protection regulations and policies of the country where it is located and obtaining the authorization given by the owner of the corresponding device.

[0033] As described in the above background art, Patent CN118245756B proposes a ship intelligent navigation analysis method and system based on situation awareness. It perceives the navigation environment according to the navigation environment monitoring information, performs land-water segmentation on the navigation environment monitoring image, marks non-driving ships and navigation obstacles at the same time, and then segments the navigable area. It uses digital twin technology to build a channel digital model to realize real-time monitoring of the channel situation and assessment of the traffic conditions, so as to provide intelligent navigation decision support for ships.

[0034] In the prior art, the processing of navigation environment monitoring information mainly relies on feature extraction and fusion from low level to high level to achieve navigation environment perception. However, this method fails to fully focus on the control of feature quality. Specifically, in a complex ship navigation environment, the sources of monitoring information are diverse and complex, including a large amount of noise and redundant data. For example, affected by factors such as weather changes (such as heavy rain, thick fog), lighting conditions (such as night, strong light reflection), and water area environment (such as turbid water body, complex underwater terrain interference), background information or error signals irrelevant to ship navigation decisions may be mixed into the monitoring information. Directly fusing these unfiltered features will result in the fused feature map containing a large amount of interference information, thus reducing the recognition rate of key features. This not only increases the computational burden of subsequent analysis, but also may mislead the judgment of the navigation environment, and then reduce the accuracy and reliability of environmental situation awareness and channel traffic status assessment. To solve the above technical problems, this application proposes an optimized ship intelligent navigation analysis method based on situation awareness. It uses a neural network model based on deep learning to extract multi-level features of navigation environment monitoring information, and performs context correlation analysis on the multi-level features of navigation environment monitoring information to realize the significance distribution modeling of different-level features, so as to identify the key features in the navigation environment monitoring information and filter out redundant information, so as to perform situation awareness of the navigation environment and intelligent assessment of the channel traffic status on this basis. In this way, the processing efficiency and accuracy of navigation environment monitoring information can be effectively improved, and then the reliability and safety of ship intelligent navigation analysis can be improved.

[0035] Figure 1 It is a flowchart of the ship intelligent navigation analysis method based on situation awareness according to an embodiment of this application. Figure 2 It is a schematic diagram of data flow of the ship intelligent navigation analysis method based on situation awareness according to an embodiment of this application. As Figure 1 and Figure 2As shown, the ship intelligent navigation analysis method based on situation awareness includes the steps of: S1, obtaining navigation environment monitoring information; S2, extracting features from low level to high level of the navigation environment monitoring information to obtain a set of navigation environment image monitoring features; S3, performing feature filtering processing based on saliency distribution modeling on the set of navigation environment image monitoring features to obtain a filtered set of navigation environment image monitoring features; S4, calculating the attention scores of each feature in the filtered set of navigation environment image monitoring features to generate an attention feature map, and building a feature pyramid network to fuse feature maps of different levels to obtain a multi-level fusion feature map of the navigation environment image; S5, performing object detection and water-land segmentation on the multi-level fusion feature map of the navigation environment image to obtain environment perception information; S6, evaluating the channel situation based on the environment perception information to obtain channel situation evaluation information.

[0036] In the above ship intelligent navigation analysis method based on situation awareness, in step S1, navigation environment monitoring information is obtained. It should be understood that the environment of ship navigation is complex and changeable, affected by various factors such as weather conditions, water area topography, and activities of other ships. By accurately and comprehensively obtaining this navigation environment information, a reliable data source can be provided for subsequent intelligent navigation decisions. In actual operation, various technical means such as high-resolution cameras, thermal imaging cameras, and satellite remote sensing can be used to obtain image data of the navigation area. Among them, high-resolution cameras and thermal imaging cameras can directly capture the water surface conditions around the ship, providing visual information such as the position, attitude, quantity of other ships, and floating objects on the water surface; satellite remote sensing can monitor the environmental conditions of large water areas from a macroscopic perspective, such as weather changes and water flow trends.

[0037] Specifically, with the help of high-resolution cameras, it is possible to directly capture the situation of the water surface around the ship, including key visual information such as the position, attitude, and quantity of other ships. These cameras are usually installed at key positions on the ship to ensure that as large a field of view as possible can be covered. High-resolution cameras can not only clearly capture the scenes during the day, but also provide sufficient detail support for situations with poor lighting conditions, such as dawn or dusk. Through their advanced optical systems and sensors, such cameras can maintain stable performance under various lighting conditions. Even in low-light environments, they can capture enough information to identify potential obstacles or the presence of other ships. This enables the crew to grasp the changes in the surrounding environment at any time and thus make corresponding operational adjustments in a timely manner. In addition, such cameras can also identify floating objects on the water surface, which is crucial for avoiding collisions and ensuring navigation safety. By analyzing this visual information, potential risks can be detected in a timely manner and corresponding measures can be taken to avoid them. For example, in a busy waterway, by monitoring the surrounding traffic conditions in real time, collisions with other ships can be effectively prevented.

[0038] At the same time, the application of thermal imaging cameras further enhances the perception ability of the navigation environment at night or under other low-light conditions. Since thermal imaging cameras rely on detecting the heat emitted by objects to generate images, they can effectively identify targets in the surrounding environment even in complete darkness or thick fog. This plays an irreplaceable role in maintaining navigation safety in extremely low visibility conditions. For example, when sailing at night, thermal imaging cameras can help the crew detect small ships or other obstacles that may not have their lighting equipment turned on, thus enabling them to react in advance and reduce the risk of collision. Similarly, in thick fog, thermal imaging cameras can penetrate the fog and show the true situation of the water area ahead, greatly enhancing navigation safety. This technology based on imaging by heat difference not only improves the accuracy of detection but also provides an additional layer of safety guarantee for navigation. Especially in some complex water area environments, such as near the coastline or river confluences, there may be a large number of small fishing boats or recreational boats, which may all be potential hazards. Thermal imaging cameras can provide valuable visual assistance in such situations, helping the crew better understand the surrounding environment and make more informed decisions.

[0039] Satellite remote sensing technology provides the ability to monitor the environmental conditions of large water areas from a macroscopic perspective. By receiving data from satellites, important information such as weather change trends, water flow direction and speed can be understood in real time. This helps ship captains and crew prepare in advance for adverse weather, such as adjusting the route to avoid upcoming storms, or optimizing the navigation route according to the water flow direction to save fuel consumption. Satellite remote sensing can also provide valuable data on the marine ecosystem, helping to monitor pollution, track red tide phenomena, etc., thus better protecting the marine environment. For large cargo ships on long voyages, using the information obtained from satellite remote sensing is particularly crucial as it is related to the safety and efficiency of the entire voyage. Satellite remote sensing is not limited to providing meteorological information, but can also be used to monitor sea level height changes, glacier movements, and coastal terrain changes, all of which are important factors affecting ship navigation safety. By comprehensively applying this information, dynamic planning of the navigation route can be achieved to ensure that the ship always moves along the safest and most economical route.

[0040] High-resolution cameras are responsible for fine observations in close-range and local areas; thermal imaging cameras focus on enhancing detection capabilities in low-light or extreme weather conditions; while satellite remote sensing provides a broad perspective, covering the meteorological dynamics of large sea areas and the airspace above. The three complement each other and jointly provide a solid basic data layer for ship navigation, making intelligent navigation analysis based on situation awareness possible. For example, in certain specific navigation scenarios, if it is necessary to cross a sea area where thick fog often appears, then the high-resolution camera and the thermal imaging camera can work together. The former provides detailed water surface structure information, and the latter is responsible for penetrating the thick fog to ensure navigation safety. At the same time, the meteorological forecast information provided by satellite remote sensing can help formulate a detour plan in advance to avoid directly entering dangerous areas. Such combined applications not only increase the safety factor of navigation, but also improve navigation efficiency, reducing unnecessary fuel consumption and time waste.

[0041] In the above-mentioned ship intelligent navigation analysis method based on situation awareness, in step S2, low-level to high-level features of the navigation environment monitoring information are extracted to obtain a set of navigation environment image monitoring features. Specifically, since the navigation environment image usually contains various complex information, such as ships, water surface fluctuations, water area terrain, etc., these information have different scales and feature representation forms in the image. Therefore, in order to comprehensively capture various features in the navigation environment image, this application adopts a multi-scale feature extraction technology to extract low-level to high-level features of the navigation environment monitoring information. In a specific example of this application, a CNN model is used to perform multi-level feature extraction on the navigation environment monitoring information to obtain a set of navigation environment image monitoring features. Those of ordinary skill in the art should know that the CNN model has a powerful feature extraction ability. Through the combination of multiple convolutional layers and pooling layers, it can automatically learn different levels of features from image data. Specifically, the CNN model captures image features of different scales by performing sliding convolution on the navigation environment image using convolutional kernels of different scales, starting from basic simple features such as edges and textures, and gradually deeply extracting more high-level semantic features, such as the navigation state of the ship, details of water surface fluctuations, and complex changes in water area terrain, etc., so as to transform the original navigation environment monitoring information into more representative and analyzable feature data, form a set of navigation environment image monitoring features, and realize a comprehensive description and characterization of the navigation environment monitoring information.

[0042] In the above-mentioned ship intelligent navigation analysis method based on situation awareness, in step S3, the set of navigation environment image monitoring features is subjected to feature filtering processing based on saliency distribution modeling to obtain a set of filtered navigation environment image monitoring features. It should be understood that considering the different importance of navigation environment image monitoring features at different levels in a complex navigation environment. For example, some of this information may have a key impact on the judgment of the navigation environment, while another part of the information may be relatively redundant and even interfere with the correct judgment, such as features generated by background noise, features of occasional irrelevant small objects, etc. Therefore, in order to ensure data quality, it is necessary to further perform feature filtering processing on the set of navigation environment image monitoring features. In particular, considering that in different navigation scenarios, the features that play a key role in ship navigation decisions are dynamically changing. For example, when navigating in a narrow channel, the position features of the channel boundary and obstacles are crucial; in a multi-ship intersection scenario, the speed, heading, and relative position features of other ships are more critical. Therefore, in order to achieve targeted screening of features at different levels, the present application proposes a saliency distribution modeling method guided by context features. By using the context information of the set of navigation environment image monitoring features, the saliency description of each navigation environment image monitoring feature is modeled, and thus the set of navigation environment image monitoring features is filtered to retain the key features that have an important impact on the judgment of the navigation environment, filter out redundant and interfering features, so as to improve the quality and effectiveness of the features, and further enhance the reliability of subsequent channel situation analysis. Among them, Figure 3 is a flowchart of sub-step S3 of the ship intelligent navigation analysis method based on situation awareness according to an embodiment of the present application. As Figure 3 shown, step S3 includes the steps of: S31, performing sequence rearrangement on the set of navigation environment image monitoring features based on saliency rough description to obtain a sequence of rearranged navigation environment image monitoring features; S32, performing importance evaluation on each rearranged navigation environment image monitoring feature in the sequence of rearranged navigation environment image monitoring features based on the guidance of pre-order context features to obtain a sequence of closed-loop strength factors for weaving navigation environment image monitoring feature clues; S33, based on the sequence of closed-loop strength factors for weaving navigation environment image monitoring feature clues, performing feature filtering on the set of navigation environment image monitoring features to obtain the set of filtered navigation environment image monitoring features.

[0043] Figure 4 is a flowchart of sub-step S31 of the ship intelligent navigation analysis method based on situation awareness according to an embodiment of the present application. As Figure 4As shown, step S31 includes steps: S311, extracting the maximum eigenvalue of each navigation environment image monitoring feature in the set of navigation environment image monitoring features as the saliency rough description factor to obtain a sequence of saliency rough description factors of navigation environment image monitoring features; S312, based on the ascending order of the sequence of saliency rough description factors of navigation environment image monitoring features, rearranging the set of navigation environment image monitoring features to obtain a sequence of rearranged navigation environment image monitoring features.

[0044] More specifically, step S311 is expressed by the formula:

[0045]

[0046] Where, represents the set of navigation environment image monitoring features, , , and respectively represent the 1st, 2nd, th, and th navigation environment image monitoring features in the set of navigation environment image monitoring features, represents the sequence of saliency rough description factors of navigation environment image monitoring features, represents the number of saliency rough description factors of navigation environment image monitoring features, represents the maximum value function, represents corresponding saliency rough description factor of navigation environment image monitoring features.

[0047] That is, in the image feature space, by extracting the maximum eigenvalue of each navigation environment image monitoring feature and using it as the saliency rough description factor of navigation environment image monitoring features, the high-dimensional image features are projected into a low-dimensional space, completing the initial dimensionality reduction of feature expression and significantly reducing the complexity of subsequent data processing. Here, the extraction of the maximum eigenvalue provides a strongly discriminative sorting criterion for subsequent feature processing, which can effectively characterize the relative saliency degree of each navigation environment image monitoring feature in the overall data distribution, and then assist in completing the preliminary feature screening based on importance, providing a solid data preprocessing foundation for subsequent in-depth analysis and decision-making.

[0048] More specifically, step S312 is expressed by the formula:

[0049]

[0050] Where, represents the sorting function, , , and respectively represent the 1st, 2nd, th, and th rearranged navigation environment image monitoring features in the sequence of the rearranged navigation environment image monitoring features, represents the sequence of the rearranged navigation environment image monitoring features.

[0051] That is, the sequence of the significance rough description factors of the navigation environment image monitoring features is arranged in ascending order. After the sorting is completed, according to this sorting rule, a synchronous rearrangement operation is performed on the sequence of the navigation environment image monitoring features. Here, the ascending order arrangement can make each image monitoring feature arranged in ascending order according to the degree of significance, and naturally introduces a good information flow structure for the subsequent feature filtering operation.

[0052] Figure 5 is a flowchart of sub-step S32 of the ship intelligent navigation analysis method based on situation awareness according to an embodiment of the present application. As Figure 5 shown, the step S32 includes steps: S321, extracting the i-th rearranged navigation environment image monitoring feature from the sequence of the rearranged navigation environment image monitoring features as the navigation environment image monitoring feature to be analyzed; S322, inputting the first i-1 rearranged navigation environment image monitoring features in the sequence of the rearranged navigation environment image monitoring features into the key clue capture network to obtain the key clue guiding feature of the pre-order distribution of the navigation environment image monitoring features; S323, inputting the key clue guiding feature of the pre-order distribution of the navigation environment image monitoring features and the navigation environment image monitoring feature to be analyzed into the clue weaving gating network to obtain the closed-loop strength factor of the navigation environment image monitoring feature clue corresponding to the navigation environment image monitoring feature to be analyzed.

[0053] More specifically, the step S322 includes: calculating the feature distribution significance factor based on the statistical feature values of each rearranged navigation environment image monitoring feature in the first i-1 rearranged navigation environment image monitoring features to obtain a sequence of the navigation environment image monitoring feature distribution significance factors, where the statistical feature values include the feature variance, feature mean, and maximum feature value of the rearranged navigation environment image monitoring feature; performing normalization processing based on the Sigmoid function on the sequence of the navigation environment image monitoring feature distribution significance factors to obtain a sequence of the navigation environment image monitoring feature distribution significance weights; based on the sequence of the navigation environment image monitoring feature distribution significance weights, performing weighted aggregation on the first i-1 rearranged navigation environment image monitoring features to obtain the key clue guiding feature of the pre-order distribution of the navigation environment image monitoring features.

[0054] The above step S322 is expressed by the formula:

[0055]

[0056] in, represents the feature variance of the rearranged navigation environment image monitoring feature, represents the feature mean of the rearranged navigation environment image monitoring feature, The first one in the sequence of the rearranged navigation environment image monitoring features is represented Rearrange navigation environment image monitoring features, express The corresponding navigation environment image monitoring feature distribution significance factor, represents an exponential function with a natural constant as base, express The corresponding navigation environment image monitoring feature distribution significance weight, represents the key clue capture network, It represents the key clue guidance feature of the preceding distribution of the navigation environment image monitoring feature.

[0057] That is, after the rearrangement is completed, the first The first rearranged navigation environment image monitoring features are used as the navigation environment image monitoring features to be analyzed. The rearrangement of the navigation environment image monitoring feature is only an exemplary method and does not refer to a specific feature. The rearranged navigation environment image monitoring features are used to carry out feature extraction and analysis of contextual patterns, and the key information hidden therein is excavated, thereby obtaining the key clue guiding features of the preceding distribution of the navigation environment image monitoring features, so as to provide sufficient contextual information support for the navigation environment image monitoring features to be analyzed, making the analysis process of the rearranged navigation environment image monitoring features more comprehensive and accurate, thereby improving the accuracy of feature filtering.

[0058] In particular, in a preferred example of the present application, based on the foregoing The statistical characteristic value of each rearranged navigation environment image monitoring feature in the rearranged navigation environment image monitoring features is used to calculate its characteristic distribution significance factor to obtain a sequence of navigation environment image monitoring feature distribution significance factors, including: on the basis of the statistical characteristic value, an edge correlation factor is also introduced, and the edge correlation factor is used to enhance the edge-global characteristic distribution synergy of the rearranged navigation environment image monitoring feature. Specifically, in this preferred example, the calculation process of the navigation environment image monitoring feature distribution significance factor can be expressed by the formula:

[0059]

[0060] in, is the edge correlation factor.

[0061] Specifically, based on the statistical features of the rearranged navigation environment image monitoring features, the key clue capture network uses the edge significance factor of the extreme value of the feature relative to the overall self-consistent trend, the edge-global consistency factor of the volatility of the feature and the overall self-consistent level, and the edge correlation factor of the fluctuation range of the feature based on the overall self-consistent relationship to measure the global significance self-consistent edge synergy of the rearranged navigation environment image monitoring features, that is, to perform the synergy of the global self-consistent relationship of the rearranged navigation environment image monitoring features based on edge significance, so as to model the consistency and synergy within the global range of the edge distribution under the significance distribution pattern of the rearranged navigation environment image monitoring features. In this way, while ensuring that the extreme edge data distribution is consistent with the overall global self-consistent trend, it is also ensured that the data fluctuation based on the edge distribution is self-consistent with the global whole. Thus, the key clue guiding feature of the pre-order distribution of the navigation environment image monitoring features can substantially improve its contribution to the overall self-consistent mode by re-inferring through the edge-global feature distribution synergy for the rearranged navigation environment image monitoring features with context significance and certain edge fluctuation characteristics, so as to provide a reference index for global self-consistent fusion based on edge significance synergy for the potential pattern of the rearranged navigation environment image monitoring feature distribution.

[0062] More specifically, the step S323 is expressed by the formula:

[0063]

[0064] where represents the inverse matrix, represents and the covariance matrix between represents the transpose of the matrix, represents the closed-loop strength factor of the navigation environment image monitoring feature clue weaving corresponding to

[0065] That is, a gating network is woven using clues to perform feature comparison and analysis on the key clue guiding features of the previous distribution of the navigation environment image monitoring features and the navigation environment image monitoring features to be analyzed, so as to accurately measure the similarity and difference degree between the two at the feature level, and finally obtain the clue weaving closed-loop strength factor of the navigation environment image monitoring features. The clue weaving closed-loop strength factor of the navigation environment image monitoring features can intuitively reflect the tightness of the association between the navigation environment image monitoring features to be analyzed and the key clue guiding features of the previous distribution of the navigation environment image monitoring features. When the degree of tight association is higher, it means that the navigation environment image monitoring features to be analyzed have a higher degree of fit with the distribution law of the previous image monitoring features. Based on this, it is possible to further screen and optimize each navigation environment image monitoring feature, eliminate those features with a low degree of association with the key clue guiding features of the previous distribution and not conforming to the overall self-consistent trend, so as to retain those key and significant features to enhance the internal logic between the features.

[0066] Specifically, in a specific example of the present application, the step S33 includes: determining whether to filter the navigation environment image monitoring features corresponding to the clue weaving closed-loop strength factor of the navigation environment image monitoring features based on the comparison between the clue weaving closed-loop strength factor of the navigation environment image monitoring features and a preset threshold, which is expressed by the formula:

[0067]

[0068] Wherein, represents the preset threshold, represents the selection function.

[0069] That is, a preset threshold is set. Based on this, according to the size of the clue weaving closed-loop strength factor of the navigation environment image monitoring features, it is determined whether the navigation environment image monitoring features to be analyzed should be retained or filtered. If the clue weaving closed-loop strength factor of the navigation environment image monitoring features is higher than the preset threshold, it indicates that the navigation environment image monitoring features to be analyzed have a high degree of association with the key clue guiding features of the previous distribution of the navigation environment image monitoring features and conform to the overall self-consistent trend, so they are retained; on the contrary, if the clue weaving closed-loop strength factor of the navigation environment image monitoring features is lower than the preset threshold, it indicates that the navigation environment image monitoring features to be analyzed have a weak degree of association with the key clue guiding features of the previous distribution of the navigation environment image monitoring features and may not conform to the distribution law of the overall data, so they are filtered out. In this way, the set of navigation environment image monitoring features can be effectively streamlined, the quality of the set of navigation environment image monitoring features can be improved, the processing speed can be accelerated, and at the same time, the calculation amount and complexity of subsequent data processing can be reduced.

[0070] In the above-mentioned ship intelligent navigation analysis method based on situation awareness, in step S4, calculate the attention scores of each feature in the set of filtered navigation environment image monitoring features to generate an attention feature map, and build a feature pyramid network to fuse feature maps of different levels to obtain a multi-level fused feature map of the navigation environment image. It should be understood that since the navigation environment image monitoring features of different levels carry different amounts of information, in order to understand the navigation environment more accurately, it is necessary to perform further aggregation analysis on the set of filtered navigation environment image monitoring features. For this purpose, in order to effectively focus on key features and make full use of the complementarity between feature maps of different levels, this application uses an attention mechanism and a feature pyramid network for feature fusion. Specifically, the attention mechanism can dynamically emphasize key features that have an important impact on the judgment of the navigation environment by calculating the attention scores of each feature, while relatively weakening the influence of redundant and interfering features, which helps to improve the sensitivity and understanding ability of the model to the navigation environment. The feature pyramid network realizes cross-scale feature fusion by building connections between feature maps of different levels, so that it can capture richer context information and finer feature details, and generate a multi-level fused feature map of the navigation environment image, which not only enhances the expressiveness of the feature map, but also provides a more comprehensive and reliable information basis for subsequent channel situation analysis and intelligent navigation decision-making.

[0071] In the above-mentioned ship intelligent navigation analysis method based on situation awareness, in step S5, perform object detection and water-land segmentation on the multi-level fused feature map of the navigation environment image to obtain environment perception information. Here, object detection aims to identify various target objects in the navigation environment, such as other ships and navigation obstacles on the waterway; water-land segmentation aims to distinguish water areas and land areas and determine the navigable area of the ship. It should be understood that object detection and water-land segmentation of navigation environment image features are existing technologies. For example, the principle disclosed in Chinese Patent CN118245756B can be used for object detection and water-land segmentation of navigation environment image features. Of course, other principles can also be used for object detection and water-land segmentation. This is not limited by this application.

[0072] In the above-mentioned ship intelligent navigation analysis method based on situation awareness, in step S6, the channel situation is evaluated based on the environment perception information to obtain channel situation evaluation information. Specifically, the environment perception information includes the position, type, status of various target objects in the navigation environment, as well as the distribution of water areas and land, etc. Through comprehensive analysis and processing of this information, for example, using digital twin technology to construct a digital model of the navigation environment, and real-time simulating and predicting the changes in the channel situation, the situation information such as the congestion degree of the channel, the distribution of obstacles, and the driving difficulty of the current driving channel can be evaluated, thereby providing a key basis for intelligent navigation decision-making. For example, in the case of a congested channel, the ship can choose to avoid or adjust its speed to avoid potential collision risks; in the case of obstacles in the channel, the ship can plan a detour route in advance to ensure the smooth progress of navigation. It is worth mentioning that the method for evaluating the channel situation using the environment perception information involved in this application can adopt the method disclosed in Chinese Patent CN118245756B, and of course, other equivalent evaluation methods can also be used. This is not limited to this application.

[0073] Specifically, digital twin technology realizes the precise simulation of the real world by mapping the entities and their behaviors in the physical world into the virtual space. For ship navigation, this means being able to create a digital model highly consistent with the actual navigation environment, which not only includes the basic geographical information of water areas and land, but also covers all dynamic and static target objects in the navigation area. For example, the data obtained through high-resolution cameras, thermal imaging cameras, and satellite remote sensing are input into the system, and after a series of processing and conversion, a basic data set for calculation and analysis is formed. In this process, advanced algorithms are used to analyze the images, extract key information such as the position, type, speed of other ships, and the presence or absence of floating objects on the water surface, and integrate them into the digital model. This real-time updated digital model can reflect the actual situation of the current channel, enabling the crew to master the most accurate information at any time point.

[0074] The specific content of the waterway situation assessment includes situation information such as the congestion degree of the waterway, the distribution of obstacles, and the navigation difficulty of the current sailing waterway. For the assessment of the waterway congestion degree, it mainly relies on the real-time tracking and data analysis of all vessels in the navigation area. By comparing the relative distances and speeds between different vessels, it is possible to determine which areas have a higher risk of collision or the possibility of traffic jams. In addition, considering the impacts of weather conditions and water flow changes on vessel navigation, the system will also comprehensively consider these factors to give a more comprehensive risk assessment report. When facing the situation of waterway congestion, vessels can choose appropriate avoidance measures or adjust their speeds according to these assessment results to avoid potential collision risks. For example, when it is found that there is a traffic peak in the front waterway, the probability of meeting other vessels can be reduced by reducing the speed or changing the route to ensure navigation safety.

[0075] As for the assessment of the navigation difficulty, it is mainly based on the analysis of the conditions of the waterway itself, such as the impacts of natural factors like water depth, water flow speed, and tidal changes. At the same time, the complexity of the surrounding environment also needs to be considered, such as whether there are a large number of passing vessels, whether it is close to areas with frequent fishing activities and other human factors. By comprehensively considering these variables, detailed navigation suggestions can be provided for vessels, including the selection of the best route, the appropriate navigation speed, and the arrangement of necessary stopping points, etc. Such an assessment mechanism can not only help the crew better understand the current navigation environment but also provide an important reference for formulating long-term navigation plans.

[0076] In summary, the ship intelligent navigation analysis method based on situation awareness according to the embodiments of the present application is clarified. It uses a neural network model based on deep learning to extract multi-level features of the navigation environment monitoring information, and conducts context correlation analysis on the multi-level features of the navigation environment monitoring information to realize the significance distribution modeling of different-level features, thereby identifying the key features in the navigation environment monitoring information and filtering out redundant information, so as to facilitate the situation awareness of the navigation environment and the intelligent assessment of the waterway passage state on this basis. In this way, the processing efficiency and accuracy of the navigation environment monitoring information can be effectively improved, and further the reliability and safety of the ship intelligent navigation analysis can be enhanced.

[0077] Furthermore, a ship intelligent navigation analysis system based on situation awareness is also provided.

[0078] Figure 6 It is a block diagram of the ship intelligent navigation analysis system based on situation awareness according to the embodiments of the present application. As Figure 6As shown, the ship intelligent navigation analysis system 100 based on situation awareness according to an embodiment of the present application includes: an environmental monitoring information acquisition module 110 for acquiring navigation environmental monitoring information; an environmental monitoring feature extraction module 120 for extracting low-level to high-level features of the navigation environmental monitoring information to obtain a set of navigation environmental image monitoring features; a feature filtering and processing module 130 for performing feature filtering and processing based on saliency distribution modeling on the set of navigation environmental image monitoring features to obtain a set of filtered navigation environmental image monitoring features; a feature fusion module 140 for calculating the attention scores of each feature in the set of filtered navigation environmental image monitoring features to generate an attention feature map, and building a feature pyramid network to fuse feature maps of different levels to obtain a multi-level fusion feature map of the navigation environmental image; an environmental perception module 150 for performing target detection and water-land segmentation on the multi-level fusion feature map of the navigation environmental image to obtain environmental perception information; and a waterway situation assessment module 160 for assessing the waterway situation based on the environmental perception information to obtain waterway situation assessment information.

[0079] Here, those skilled in the art can understand that the specific operations of each module in the above ship intelligent navigation analysis system based on situation awareness have been introduced in detail in the description of the ship intelligent navigation analysis method based on situation awareness above with reference to Figures 1 to 5 and therefore, the repeated description thereof will be omitted.

[0080] The basic principles of the present invention have been described above in conjunction with specific embodiments. However, it should be noted that the advantages, advantages, effects, etc. mentioned in the present invention are only examples and not limitations, and it cannot be considered that these advantages, advantages, effects, etc. are essential for each embodiment of the present invention. In addition, the specific details of the above embodiments are only for the purpose of illustration and easy understanding, rather than limitations, and the above details do not limit the present invention to necessarily adopt the above specific details to implement.

[0081] In the above embodiments, the descriptions of each embodiment have their own emphases. For parts not detailed or recorded in a certain embodiment, reference may be made to the relevant descriptions of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed system and method can be implemented in other ways. For example, the system embodiments described above are only illustrative. For example, the unit division is only a logical function division, and there may be other division methods in actual implementation. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0082] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above-described exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be embraced by the present invention. Any reference signs in the claims should not be construed as limiting the claims involved.

[0083] In addition, it is obvious that the term "comprising" does not exclude other units or steps, and the singular does not exclude the plural. The multiple units stated in the system claims can also be implemented by one unit through software or hardware.

[0084] Finally, it should be noted that the above description has been given for purposes of illustration and description. In addition, the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for analyzing ship intelligent navigation based on situational awareness, characterized in that: include: Obtain navigation environment monitoring information; Extracting low-level to high-level features of the navigation environment monitoring information to obtain a set of navigation environment image monitoring features; Performing feature filtering processing based on significance distribution modeling on the set of navigation environment image monitoring features to obtain a filtered set of navigation environment image monitoring features; Calculating the attention score of each feature in the set of filtered navigation environment image monitoring features to generate an attention feature map, and building a feature pyramid network to fuse feature maps of different levels to obtain a multi-level fusion feature map of the navigation environment image; Performing target detection and land-water segmentation on the multi-level fusion feature map of the navigation environment image to obtain environmental perception information; Evaluate the waterway situation based on the environmental perception information to obtain waterway situation evaluation information; Performing feature filtering processing based on significance distribution modeling on the set of navigation environment image monitoring features to obtain a filtered set of navigation environment image monitoring features, including: The set of navigation environment image monitoring features is subjected to sequence rearrangement based on the significance rough description to obtain a sequence of rearranged navigation environment image monitoring features, specifically comprising: extracting the maximum eigenvalue of each navigation environment image monitoring feature in the set of navigation environment image monitoring features as a significance rough description factor to obtain a sequence of significance rough description factors of navigation environment image monitoring features; based on the ascending arrangement of the sequence of significance rough description factors of navigation environment image monitoring features, the set of navigation environment image monitoring features is rearranged to obtain the sequence of rearranged navigation environment image monitoring features; Extract the i-th rearranged navigation environment image monitoring feature from the sequence of rearranged navigation environment image monitoring features as the navigation environment image monitoring feature to be analyzed; calculate the feature distribution significance factor based on the statistical feature value of each rearranged navigation environment image monitoring feature in the first i-1 rearranged navigation environment image monitoring features to obtain a sequence of navigation environment image monitoring feature distribution significance factors, wherein the statistical feature value includes the feature variance, feature mean and maximum feature value of the rearranged navigation environment image monitoring feature; perform a sigmoid function based on the sequence of navigation environment image monitoring feature distribution significance factors. Normalization processing is performed to obtain a sequence of navigation environment image monitoring feature distribution significance weights; based on the sequence of navigation environment image monitoring feature distribution significance weights, the first i-1 rearranged navigation environment image monitoring features are weighted aggregated to obtain the navigation environment image monitoring feature pre-order distribution key clue guidance feature; the navigation environment image monitoring feature pre-order distribution key clue guidance feature and the navigation environment image monitoring feature to be analyzed are input into the clue weaving gating network to obtain the navigation environment image monitoring feature clue weaving closed-loop strength factor corresponding to the navigation environment image monitoring feature to be analyzed, which is expressed by the formula: ; in, represents the inverse matrix, express and The covariance matrix between represents the transpose of a matrix, express The corresponding closed-loop strength factor of the navigation environment image monitoring feature clue weaving, represents the key clue guiding feature of the preceding distribution of the navigation environment image monitoring feature, The first one in the sequence of the rearranged navigation environment image monitoring features is represented Rearrange navigation environment image monitoring features; Based on the sequence of closed-loop strength factors woven from the navigation environment image monitoring feature clues, feature filtering is performed on the set of navigation environment image monitoring features to obtain the set of filtered navigation environment image monitoring features.

2. The method for analyzing ship intelligent navigation based on situational awareness according to claim 1 is characterized in that: Extracting low-level to high-level features of the navigation environment monitoring information to obtain a set of navigation environment image monitoring features, including: A CNN model is used to perform multi-level feature extraction on the navigation environment monitoring information to obtain a set of navigation environment image monitoring features.

3. The method for analyzing ship intelligent navigation based on situational awareness according to claim 2 is characterized in that: The method further comprises calculating the feature distribution significance factor of each rearranged navigation environment image monitoring feature in the first i-1 rearranged navigation environment image monitoring features based on the statistical feature value of each rearranged navigation environment image monitoring feature to obtain a sequence of navigation environment image monitoring feature distribution significance factors, including: On the basis of the statistical characteristic value, an edge correlation factor is further introduced, and the edge correlation factor is used to enhance the edge-global feature distribution synergy of the rearranged navigation environment image monitoring feature.

4. The method for analyzing ship intelligent navigation based on situational awareness according to claim 3 is characterized in that: Based on the sequence of closed-loop strength factors of the navigation environment image monitoring feature clues, feature filtering is performed on the set of navigation environment image monitoring features to obtain the set of filtered navigation environment image monitoring features, including: Based on the comparison between the navigation environment image monitoring feature clue weaving closed-loop strength factor and a preset threshold, it is determined whether to filter the navigation environment image monitoring feature corresponding to the navigation environment image monitoring feature clue weaving closed-loop strength factor.

5. A ship intelligent navigation analysis system based on situational awareness, characterized in that: include: Environmental monitoring information acquisition module, used to obtain navigation environment monitoring information; An environment monitoring feature extraction module, used to extract low-level to high-level features of the navigation environment monitoring information to obtain a set of navigation environment image monitoring features; A feature filtering processing module, used for performing feature filtering processing on the set of navigation environment image monitoring features based on significance distribution modeling to obtain a set of filtered navigation environment image monitoring features; A feature fusion module is used to calculate the attention score of each feature in the set of filtered navigation environment image monitoring features to generate an attention feature map, and to build a feature pyramid network to fuse feature maps of different levels to obtain a multi-level fusion feature map of the navigation environment image; An environment perception module, used for performing target detection and land-water segmentation on the multi-level fusion feature map of the navigation environment image to obtain environment perception information; A channel situation assessment module, used to assess the channel situation based on the environmental perception information to obtain channel situation assessment information; The feature filtering processing module is used to: The set of navigation environment image monitoring features is subjected to sequence rearrangement based on the significance rough description to obtain a sequence of rearranged navigation environment image monitoring features, specifically comprising: extracting the maximum eigenvalue of each navigation environment image monitoring feature in the set of navigation environment image monitoring features as a significance rough description factor to obtain a sequence of significance rough description factors of navigation environment image monitoring features; based on the ascending arrangement of the sequence of significance rough description factors of navigation environment image monitoring features, the set of navigation environment image monitoring features is rearranged to obtain the sequence of rearranged navigation environment image monitoring features; Extract the i-th rearranged navigation environment image monitoring feature from the sequence of rearranged navigation environment image monitoring features as the navigation environment image monitoring feature to be analyzed; calculate the feature distribution significance factor based on the statistical feature value of each rearranged navigation environment image monitoring feature in the first i-1 rearranged navigation environment image monitoring features to obtain a sequence of navigation environment image monitoring feature distribution significance factors, wherein the statistical feature value includes the feature variance, feature mean and maximum feature value of the rearranged navigation environment image monitoring feature; perform a sigmoid function based on the sequence of navigation environment image monitoring feature distribution significance factors. Normalization processing is performed to obtain a sequence of navigation environment image monitoring feature distribution significance weights; based on the sequence of navigation environment image monitoring feature distribution significance weights, the first i-1 rearranged navigation environment image monitoring features are weighted aggregated to obtain the navigation environment image monitoring feature pre-order distribution key clue guidance feature; the navigation environment image monitoring feature pre-order distribution key clue guidance feature and the navigation environment image monitoring feature to be analyzed are input into the clue weaving gating network to obtain the navigation environment image monitoring feature clue weaving closed-loop strength factor corresponding to the navigation environment image monitoring feature to be analyzed, which is expressed by the formula: ; in, represents the inverse matrix, express and The covariance matrix between represents the transpose of a matrix, express The corresponding closed-loop strength factor of the navigation environment image monitoring feature clue weaving, represents the key clue guiding feature of the preceding distribution of the navigation environment image monitoring feature, The first one in the sequence of the rearranged navigation environment image monitoring features is represented Rearrange navigation environment image monitoring features; Based on the sequence of closed-loop strength factors woven from the navigation environment image monitoring feature clues, feature filtering is performed on the set of navigation environment image monitoring features to obtain the set of filtered navigation environment image monitoring features.

Images