A method and system for submarine cable cruise safety early warning based on multi-source perception fusion

Abstract

The application discloses a kind of multi-source perception fusion's sea cable cruise safety early warning method and system, it is related to safety early warning related technical field, including: based on the security feature analysis clustering of sea cable laying position, establish safety cruise area;Obtain multi-source perception data, based on the interaction of multi-source feature extraction of sensing cruise area corresponding relationship, obtain interactive security response feature set;Safety static, safety dynamic identification segmentation is carried out, and trace core security response features and dynamic tracking target;Multi-source perception data is tracked and analyzed in time sequence response, generates time sequence safety early warning information.Solve the technical problems that exist in prior art in the sea cable safety early warning multi-source perception data isolation, space area division static extensive, and lack the time sequence causal tracing ability between dynamic target behavior and response feature, reach the technical effect of improving the global situation awareness accuracy of sea cable cruise early warning, dynamic risk tracing ability and emergency response decision efficiency.

Description

Technical Field

[0001] This invention relates to the field of safety early warning technology, specifically to a method and system for safety early warning of submarine cable patrols based on multi-source sensing fusion. Background Technology

[0002] Submarine optical cables and submarine power cables (“submarine cables”) are core infrastructure for international communications, cross-regional power transmission, and marine resource development. However, the risks faced in laying submarine cables are extensive and dynamic, mainly including external mechanical damage, such as human activities such as ship anchoring, trawling, and seabed engineering operations; marine dynamic environment risks, such as seabed erosion, landslides, and insufficient burial depth caused by ocean currents and eddies; aging and faults, such as insulation degradation and joint overheating; and monitoring methods relying on AIS ship positioning or distributed fiber optic temperature measurement, which can only cover local risk dimensions and are difficult to achieve low false alarm and high reliability of global situational awareness in complex marine environments. Current submarine cable safety early warning systems rely on data-level or feature-level stitching, lacking deep interaction based on the spatial semantics of submarine cables. This makes it impossible to distinguish between "risk events" and "environmental noise," resulting in a high false alarm rate. Existing protection zones are usually based on fixed-distance buffers and fail to dynamically analyze and cluster them in conjunction with the actual laying pattern of submarine cables, leading to a lack of spatial adaptability in safety patrol strategies. Current early warning logic only provides instantaneous location alerts for moving targets, without establishing a temporal causal relationship between target behavior and submarine cable response characteristics. This makes it difficult to trace the core risk source, resulting in delayed or misjudged emergency response decisions.

[0003] Therefore, current technologies for submarine cable safety early warning suffer from technical problems such as isolated multi-source sensing data, static and coarse spatial area division, and lack of time-series causal tracing capabilities between dynamic target behavior and response characteristics. Summary of the Invention

[0004] This application provides a method and system for submarine cable patrol safety early warning based on multi-source sensing fusion, which solves the technical problems in existing technologies such as isolated multi-source sensing data, static and coarse spatial area division, and lack of time-series causal tracing ability between dynamic target behavior and response characteristics in submarine cable safety early warning. It achieves the technical effect of improving the accuracy of global situational awareness, dynamic risk tracing ability, and emergency response decision-making efficiency of submarine cable patrol early warning.

[0005] This application provides a multi-source sensing fusion method for submarine cable patrol safety early warning. The method includes: performing safety feature analysis and clustering based on the submarine cable laying location to establish a safety patrol area, wherein each safety patrol area has response safety features and a response risk mapping relationship; acquiring multi-source sensing data; based on the correspondence of the sensing patrol areas, performing multi-source feature extraction interaction according to the response safety features and the response risk mapping relationship to obtain an interactive safety response feature set; performing safety static and safety dynamic identification and segmentation based on the interactive safety response feature set, and tracing core safety response features and dynamic tracking targets; and performing time-series response tracking analysis on the multi-source sensing data according to the traced core safety response features and dynamic tracking targets to generate time-series safety early warning information.

[0006] In one possible implementation, safety feature analysis and clustering are performed based on the submarine cable laying location to establish a safety patrol area. This includes: acquiring geospatial information of the submarine cable laying path and historical risk records along the path; using the submarine cable laying path as an axis, performing multi-dimensional safety feature clustering according to the submarine cable burial depth, seabed geological type, ship traffic flow, distribution of historical anchor damage events, and environmental perception conditions to divide the submarine cable route into multiple safety patrol areas; and cross-combining the clustering results of each dimension to establish response safety feature labels for each safety patrol area; and establishing the response risk mapping relationship under each response safety feature label based on the historical statistical response relationship between the response safety feature labels and risk events.

[0007] In possible implementations, the cable burial depth includes shallow burial, standard burial depth, and exposed state; the seabed geological type includes hard geology, soft geology, and mixed geology; and the environmental sensing conditions include deep-water sonar effective conditions, shallow-water sonar failure magnetic detection effective conditions, and intertidal optical / magnetic synergy conditions.

[0008] In a possible implementation, based on the correspondence of the sensing patrol area, multi-source feature extraction is performed interactively according to the response safety features and response risk mapping relationship to obtain an interactive safety response feature set. This includes: spatially aligning the multi-source sensing data to determine the sensing spatial location corresponding to each sensing data; matching the sensing spatial location with the spatial range of each of the safety patrol areas to determine the safety patrol area to which the current multi-source sensing data belongs and the corresponding response safety feature label; extracting features from the multi-source sensing data according to the preset risk association feature dimension in the response safety feature label to obtain the sensing feature vector of each data source under the risk association feature dimension; calculating the matching degree between each sensing feature vector and the various risk event association feature templates in the response safety feature label, and marking sensing features with a matching degree exceeding a preset matching degree threshold as valid sensing matching features; querying the response risk mapping relationship according to the data source type and feature type corresponding to the valid sensing matching features to determine the sensing contribution and response sensitivity of each valid sensing matching feature to various risk events in the current area, establishing an association mapping between sensing data, safety features, and risk events, and generating the interactive safety response feature set.

[0009] In a possible implementation, the time-domain rate of change and spatial position offset are calculated for each feature item in the interactive security response feature set; based on the time-domain rate of change and spatial position offset, the interactive security response feature set is divided into static security basic features and dynamic security response features; using the static security basic features as a reference benchmark, the core security response features in the dynamic security response features and the dynamic tracking target to which the core security response features belong are traced.

[0010] In a possible implementation, the interactive security response feature set is divided into static security basic features and dynamic security response features based on the temporal rate of change and spatial position offset. This includes: segmenting feature items whose spatial position offset is less than a preset spatial offset threshold and whose temporal rate of change is less than a first rate of change threshold into the static security basic features; and segmenting feature items whose spatial position offset is greater than or equal to the preset spatial offset threshold, or whose spatial position offset is less than the preset spatial offset threshold but whose temporal rate of change is greater than a second rate of change threshold, into the dynamic security response features, wherein the second rate of change threshold is greater than or equal to the first rate of change threshold.

[0011] In a possible implementation, using the static security basic features as a reference benchmark, tracing the core security response features and the dynamic tracking targets to which the core security response features belong, includes: using the static security basic features as a spatial reference benchmark, spatially associating each dynamic security response feature with the static security basic features to determine the spatial distribution relationship of each dynamic security response feature relative to the static scene; based on the spatial distribution relationship, calculating the degree of deviation of the current feature value of each dynamic security response feature from its respective historical baseline value, wherein the degree of deviation is the absolute value of the difference between the current feature value of each dynamic security response feature and its respective historical baseline value; sorting the dynamic security response features from largest to smallest according to the degree of deviation, selecting the top N dynamic security response features as the core security response features; and marking the dynamic object to which the core security response features belong as the dynamic tracking target.

[0012] In a possible implementation, time-series response tracking analysis is performed on multi-source sensing data based on the core safety response characteristics of the tracing system and the dynamic tracking target to generate time-series safety early warning information. This includes: calculating the rate of change and assessing the signal quality of the core safety response characteristics of the tracing system; determining the risk tracking type of the dynamic tracking target based on the rate of change and signal quality assessment results, wherein the risk tracking type includes one or more of rapidly changing risk targets, weak signal risk targets, and composite risk targets; driving the corresponding multi-source sensors to perform time-series response tracking acquisition on the dynamic tracking target that matches the risk tracking type, thereby obtaining tracking sensing data of the dynamic tracking target within a continuous time window; performing time-series correlation analysis on the tracking sensing data to extract the motion evolution trend and feature change trend of the dynamic tracking target; and generating time-series safety early warning information based on the motion evolution trend and the feature change trend, combined with the spatial positional relationship between the dynamic tracking target and the submarine cable laying path.

[0013] In a possible implementation, based on the risk tracking type, the corresponding multi-source sensor is driven to perform time-series response tracking and acquisition of the dynamically tracked target in a manner matching the risk tracking type. This includes: when the risk tracking type is a rapidly changing risk target, the corresponding multi-source sensor is driven to perform continuous tracking and acquisition of the dynamically tracked target at an acquisition frequency higher than the normal acquisition frequency; when the risk tracking type is a weak signal risk target, the corresponding multi-source sensor is driven to switch to a multi-cycle cumulative acquisition mode, in which the signals acquired within a continuous time window are superimposed in phase to enhance the signal strength; when the risk tracking type is a composite risk target, the corresponding multi-source sensor is driven to simultaneously perform continuous tracking and acquisition at a frequency higher than the normal acquisition frequency and the multi-cycle cumulative acquisition, and the corresponding dynamically tracked target is marked as a high-concern risk target in the generated time-series security warning information.

[0014] This application also provides a multi-source sensing fusion submarine cable patrol safety early warning system, the system comprising: a safety patrol area establishment module, used to perform safety feature analysis and clustering based on the submarine cable laying location to establish a safety patrol area, wherein each safety patrol area has response safety features and a response risk mapping relationship; a safety response feature acquisition module, used to acquire multi-source sensing data, and based on the correspondence of sensing patrol areas, perform multi-source feature extraction interaction according to the response safety features and response risk mapping relationship to obtain an interactive safety response feature set; a feature recognition and segmentation module, used to perform static and dynamic safety recognition and segmentation based on the interactive safety response feature set, and trace core safety response features and dynamic tracking targets; and a safety early warning information generation module, used to perform time-series response tracking analysis on the multi-source sensing data according to the traced core safety response features and dynamic tracking targets, to generate time-series safety early warning information.

[0015] This application proposes a multi-source sensing fusion-based method and system for submarine cable patrol safety early warning. Based on the cable laying location, it performs safety feature analysis and clustering to establish a safety patrol area; acquires multi-source sensing data; and extracts interactive features based on the correspondence between the sensing patrol areas to obtain an interactive safety response feature set. It then performs static and dynamic safety identification and segmentation, traces core safety response features, and dynamically tracks targets. Finally, it analyzes the multi-source sensing data for time-series response tracking to generate time-series safety early warning information. This addresses the technical problems in existing submarine cable safety early warning technologies, such as isolated multi-source sensing data, static and coarse spatial area division, and a lack of time-series causal tracing capabilities between dynamic target behavior and response features. It achieves the technical effect of improving the accuracy of global situational awareness, dynamic risk tracing capabilities, and emergency response decision-making efficiency in submarine cable patrol early warning. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings of the embodiments of this disclosure will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0017] Figure 1 This is a schematic flowchart of a multi-source sensing fusion method for safety early warning of submarine cable patrols, provided as an embodiment of this application.

[0018] Figure 2 This is a schematic diagram of a multi-source sensing fusion submarine cable patrol safety early warning system provided in an embodiment of this application.

[0019] Explanation of reference numerals in the attached diagram: Safe patrol area establishment module 10, safe response feature acquisition module 20, feature recognition and segmentation module 30, and safe early warning information generation module 40. Detailed Implementation

[0020] To further illustrate the technical means and effects adopted by the present invention in order to achieve the intended purpose, the following detailed description is provided in conjunction with the accompanying drawings and preferred embodiments, based on the specific implementation methods, structures, features and effects of the present invention.

[0021] This application provides a multi-source sensing fusion method for submarine cable patrol safety early warning, such as... Figure 1 As shown, the method includes: Step S100: Based on the location of submarine cable laying, perform safety feature analysis and clustering to establish a safety patrol area, wherein each safety patrol area has response safety features and response risk mapping relationship.

[0022] Step S100 further includes acquiring geospatial information of the submarine cable laying path and historical risk records along the path; using the submarine cable laying path as an axis, performing multi-dimensional safety feature clustering according to the submarine cable burial depth, seabed geological type, ship traffic flow, distribution of historical anchor damage events, and environmental perception conditions; dividing the submarine cable path into multiple safety patrol areas; wherein, the clustering results of each dimension are cross-combined to establish response safety feature labels for each of the safety patrol areas; and establishing the response risk mapping relationship under each response safety feature label based on the historical statistical response relationship between the response safety feature labels and risk events.

[0023] Preferably, the latitude and longitude coordinate sequence of the actual cable laying route, the measured burial depth values ​​corresponding to each coordinate point, and the seabed geological sampling classification results corresponding to each coordinate point are obtained to determine the geospatial information of the cable laying path. Simultaneously, ship track data, anchoring event data, cable fault repair records caused by anchor damage, and sensor availability data at each coordinate point are obtained along the route within a historical time window to determine historical risk records along the route. The continuous coordinate sequence of the cable route is used as a spatial reference baseline, and the spatial range is defined with this baseline as the center. That is, the spatial coverage of the patrol area extends to both sides along this axis. For each spatial location point along the cable, the cable burial depth status, seabed geological type, ship traffic flow, historical anchor damage event distribution, and environmental perception conditions are simultaneously obtained. The cable burial depth status is obtained by classifying and mapping the burial depth values. The data is categorized into five types: seabed geological type (obtained from geological sampling analysis), vessel traffic flow (statistical value of the number of vessels passing through a defined distance range around the location within a unit time window), historical anchor damage event distribution (statistical value of the frequency of historical anchor damage events within a defined distance range around the location), and environmental perception conditions (combination of sensor types that can operate normally at the location). Then, clustering operations are performed on these five types of data in the spatial dimension. Spatial locations with similar burial depth, similar geological type, similar traffic flow level, similar anchor damage frequency level, and similar perception condition combinations are grouped into the same cluster. The continuous spatial range along the submarine cable is then segmented according to the clustering results, ensuring that the five-dimensional feature combinations within each segment are homogeneous, while the five-dimensional feature combinations between different segments are heterogeneous, thus identifying multiple safe cruising areas.

[0024] Preferably, for each defined safety patrol area, its category in the burial depth dimension, its category in the geological dimension, its level in the traffic flow dimension, its level in the anchor damage frequency dimension, and its combination type in the sensing condition dimension are merged to form a response safety feature label for that area. The labels for different areas are distinguished by different combinations of the five-dimensional categories. Then, for each safety patrol area, the frequency and severity of various risk events such as external mechanical damage events and submarine cable failure events that have occurred in the area in the past are statistically analyzed, as well as the characteristic signal patterns recorded by each sensing sensor when the events occur, to form a statistical response relationship between the response safety feature label of the area and the types and probabilities of risk events. Then, using the response safety feature label as a query index, and using the historical occurrence probability, typical signal patterns, key influencing factors, and other information of various risk events in the area corresponding to the label as query results, a response risk mapping relationship is constructed. Once the location matching relationship between multi-source sensing data and a certain area is determined, the risk types that need to be monitored in the area and the reference value weights of various sensing data for the risk events can be quickly retrieved from the mapping relationship based on the response safety feature label corresponding to the area.

[0025] Furthermore, step S100 also includes the following: the burial depth of the submarine cable includes shallow burial, standard burial depth, and exposed state; the seabed geological type includes hard geology, soft geology, and mixed geology; and the environmental perception conditions include deep-water sonar effective conditions, shallow-water sonar failure magnetic detection effective conditions, and intertidal optical / magnetic synergy conditions.

[0026] Preferably, the burial depth of a submarine cable refers to the vertical distance between the cable surface and the seabed surface. This distance is obtained by measuring point by point along the route using burial depth detection equipment. It includes shallow burial, standard burial depth, and exposed state. Shallow burial means the vertical distance from the cable surface to the seabed surface is less than the minimum burial depth specified in the design standard, for example, less than 1 meter. The cable's protection against external mechanical forces is lower than expected. Standard burial depth means the vertical distance from the cable surface to the seabed surface is within the normal range specified in the design standard, for example, between 1 and 3 meters. Exposed state means the vertical distance from the cable surface to the seabed surface is zero or close to zero, meaning the cable is directly exposed to the seabed surface without any soil cover, and the cable itself directly bears external contact and impact. These three burial depth states correspond to different external risk tolerance capabilities. The probability of direct damage from mechanical forces such as ship anchoring and fishing net dragging is significantly higher in the exposed state than in the standard burial depth state; the shallow burial state is the next lowest probability.

[0027] Preferably, the seabed geological type refers to the classification of the physical properties of the seabed surface sediments at the cable laying location. This classification is obtained through geological sampling, side-scan sonar echo characteristics, or cone penetration resistance measurement. It includes hard geology, soft geology, and mixed geology. Hard geology is characterized by seabed surface sediments dominated by rocks, gravel, calcareous crusts, or dense sand, with high intergranular cementation or large particle size, resulting in high resistance to penetration. Soft geology is characterized by seabed surface sediments dominated by silt, silt, or loose clay, with fine particles and high water content, resulting in low bearing capacity. Mixed geology is characterized by alternating or mixed spatial distribution of hard and soft components in the seabed surface sediments, such as sand-mud interlayers or a mixture of gravel and clay. Different geological types affect the operability of cable laying and the long-term retention of burial depth. They also affect the penetration depth and towing behavior of the anchor after it falls into the seabed, thus determining the magnitude of anchor damage risk and the propagation attenuation characteristics of the sensing signal.

[0028] Preferably, environmental sensing conditions refer to the combination of working states under different water depths and marine environmental conditions along the submarine cable, enabling various sensors to normally acquire effective signals. This includes effective deep-water sonar conditions, effective shallow-water sonar failure magnetic sounding conditions, and intertidal optical / magnetic synergy conditions. Specifically, effective deep-water sonar conditions refer to conditions where the water depth is relatively large, the concentration of suspended matter in the water is low, and the attenuation of sound wave propagation is within an acceptable range, allowing acoustic sensors such as side-scan sonar, multibeam echo sounders, and shallow seismic profilers to acquire clear sonar echo images and seabed target echo signals. Effective shallow-water sonar failure magnetic sounding conditions refer to conditions where the water depth is relatively shallow and the content of air bubbles in the water is high. In some cases, strong absorption of sound waves by geological features can severely interfere with or attenuate the propagation of sonar signals, preventing acoustic sensors from acquiring effective detection data. However, if the water depth allows magnetometers or magnetic gradient meters to function properly, they can detect underwater metal targets by detecting the disturbance of the Earth's magnetic field by ferromagnetic objects. In intertidal optical / magnetic synergistic conditions, where the water depth is extremely shallow, sonar equipment cannot function properly due to interface reverberation interference between sound wave transmission and reception. Under these conditions, optical sensors are used to acquire seabed surface images for visual identification, while magnetometers are used to acquire ferromagnetic target signals. The two sensing methods work together to perform target detection.

[0029] Step S200: Obtain multi-source perception data; based on the correspondence of perception cruise areas, perform multi-source feature extraction interaction according to the response safety features and response risk mapping relationship to obtain an interactive safety response feature set.

[0030] Step S200 further includes: spatially aligning the multi-source sensing data to determine the sensing spatial location corresponding to each sensing data; matching the sensing spatial location with the spatial range of each of the safety patrol areas to determine the safety patrol area to which the current multi-source sensing data belongs and the corresponding response safety feature label; extracting features from the multi-source sensing data according to the preset risk association feature dimension in the response safety feature label to obtain the sensing feature vector of each data source under the risk association feature dimension; calculating the matching degree between each sensing feature vector and the various risk event association feature templates in the response safety feature label, and marking sensing features with matching degrees exceeding a preset matching degree threshold as valid sensing matching features; querying the response risk mapping relationship according to the data source type and feature type corresponding to the valid sensing matching features, determining the sensing contribution and response sensitivity of each valid sensing matching feature to various risk events in the current area, establishing the association mapping between sensing data-safety features-risk events, and generating the interactive safety response feature set.

[0031] Preferably, raw observation data within the current time or time window is simultaneously acquired from multiple heterogeneous sensors. This includes, but is not limited to, vibration or temperature signal sequences output by a distributed fiber optic sensing system, ship dynamic messages acquired by an AIS receiver, magnetic field strength vector sequences output by a geomagnetic sensor, echo intensity images output by a sonar system, and video frame sequences output by an optical camera. The multi-source sensing data is then determined and spatially aligned. This involves transforming the raw data output by each sensor from its respective local coordinate system or sensor installation coordinate system to a unified spatial reference coordinate system. For example, for a distributed fiber optic sensing system, whose sensing location is known along the coastal cable route, its measurement channel number is directly mapped to the corresponding latitude and longitude coordinates. For ship position reports received by shipborne sonar or AIS, the spatial coordinates of each detection point or target are calculated based on the sensor's own positioning information and beam pointing angle. Then, each sensing data entry is assigned its actual corresponding sensing spatial position. For example, the spatial position of the i-th channel in a distributed acoustic sensing system is XX degrees XX minutes east longitude and XX degrees XX minutes north latitude. The target ship position in the AIS message is directly the unified coordinate system value after the latitude and longitude coordinates of the ship in the message are transformed. The coordinates attached to each sensing data entry are used to determine the spatial topological relationship with the spatial boundary coordinates of each safety patrol area. This involves geometric calculation of the inclusion relationship between a point and a polygon, determining the boundary range of the safety patrol area where the coordinate point is located, determining the safety patrol area to which the current sensing data belongs based on the location matching result, and binding the response safety feature label of that area.

[0032] Preferably, each response safety feature tag predefines a set of risk-related feature dimensions. For example, for the tag of the combination of "shallow burial + soft geology", the preset risk-related feature dimensions include vibration signal amplitude, vibration signal frequency spectrum envelope, temperature change rate, etc. For the tag of the combination of "exposed + hard geology + effective magnetic exploration", the preset risk-related feature dimensions include magnetic anomaly gradient, magnetic anomaly duration, magnetic anomaly movement speed, etc. The risk-related feature dimensions corresponding to different tags are not completely the same. Then, according to the preset risk-related feature dimensions, features are extracted from multi-source sensing data. For example, specific quantitative values ​​such as peak amplitude, zero-crossing rate, and frequency domain dominant frequency are extracted from the vibration time-domain signal of distributed acoustic sensing, and specific quantitative values ​​such as horizontal gradient magnitude and vertical component change are extracted from the time-series magnetic field data of geomagnetic sensors. These are then arranged in a fixed order to form a sensing feature vector.

[0033] Preferably, under the current response safety feature label, the system stores associated feature templates for various risk events (such as anchor dragging, fishing net dragging, and anchor impact). Each associated feature template is a standardized feature vector, representing the numerical range or distribution pattern of each feature dimension when the risk event occurs. The perceived feature vector is compared one by one with the associated feature templates of various risk events in the response safety feature label. The similarity between them is calculated using Euclidean distance or the reciprocal of cosine similarity to determine the matching degree. The higher the matching degree, the closer the current perceived feature is to the historical feature pattern of the risk event. Perceived features with a matching degree exceeding a preset matching degree threshold are marked as valid perceived matching features, indicating that the perceived data does reflect valid information related to the risk event. Perceived features with a matching degree not exceeding the preset matching degree threshold are considered noise or irrelevant information that is not significantly related to the known risk patterns in the current area and are discarded.

[0034] Preferably, for each valid sensing feature, its data source type and feature type are determined, such as distributed acoustic sensing, geomagnetism, AIS, or amplitude, frequency, magnetic gradient, etc. Using the current response safety feature label of the area as an index, the response risk mapping relationship is queried. Each valid sensing matching feature is assigned a sensing contribution and a response sensitivity. The sensing contribution refers to the weight coefficient of the feature in determining whether a certain type of risk event has occurred. For example, when judging anchor damage events in a "shallowly buried + soft geological" area, the vibration amplitude contribution weight of distributed acoustic sensing is 0.7, and the ship's distance from the submarine cable in the AIS data... The distance contribution weight is 0.3; response sensitivity refers to the sensitivity of this feature to changes in the intensity of risk events, that is, the magnitude of change in the feature value when the intensity of a risk event changes by a unit. The higher the response sensitivity, the more suitable the feature is for fine-grained tracking of risk events; then all effective perception matching features are organized into associated records to determine the association mapping between perception data, security features, and risk events; finally, all effective perception matching features and their associated mapping records within the current safe cruise area are summarized to form an interactive safety response feature set, thereby realizing the transformation from raw perception data to safety semantic information.

[0035] Step S300: Based on the interactive security response feature set, perform security static and security dynamic identification and segmentation, and trace the core security response features and dynamic tracking targets.

[0036] Step S300 further includes calculating the time-domain rate of change and spatial position offset of each feature item in the interactive security response feature set; dividing the interactive security response feature set into static security basic features and dynamic security response features based on the time-domain rate of change and spatial position offset; and using the static security basic features as a reference benchmark, tracing the core security response features in the dynamic security response features and the dynamic tracking target to which the core security response features belong.

[0037] Preferably, for each feature in the interactive safety response feature set, the value of the feature at the current moment and its historical values ​​at one or more previous sampling moments are obtained. The difference between the current value and the historical value is divided by the time interval to obtain the rate of change of the feature over time, the magnitude of which reflects the drastic change of the feature in the time dimension. Then, the spatial coordinates of each feature are obtained, such as the location of sonar echo anomalies and the location of AIS target vessels. The spatial displacement of the coordinates between the current moment and the previous moment, i.e., the Euclidean distance, is calculated to determine the spatial offset. Then, based on the temporal rate of change and the spatial offset, the interactive safety response feature set is divided into static safety basic features, which are fixed in space and stable in time, representing the basic background state along the submarine cable at the current moment without external disturbance, and dynamic safety response features, which are additional instantaneous responses that are different from the background baseline caused by significant spatial movement or sudden events.

[0038] Furthermore, step S300 also includes: segmenting feature items whose spatial position offset is less than a preset spatial offset threshold and whose temporal change rate is less than a first change rate threshold into the static security basic features; and segmenting feature items whose spatial position offset is greater than or equal to the preset spatial offset threshold, or feature items whose spatial position offset is less than the preset spatial offset threshold but whose temporal change rate is higher than a second change rate threshold, into the dynamic security response features, wherein the second change rate threshold is greater than or equal to the first change rate threshold.

[0039] Preferably, a preset spatial offset threshold is set based on the spatial sampling interval or positioning accuracy of the submarine cable laying path; a first rate of change threshold is set based on the statistical value of normal environmental background fluctuations, which is the lower limit threshold for distinguishing between static and dynamic states; and a second rate of change threshold is set based on the statistical value of the characteristic rate of change caused by risk events, which is the threshold for distinguishing between general dynamic and strong dynamic states, and the second rate of change threshold is greater than or equal to the first rate of change threshold. Then, feature items that simultaneously satisfy the condition that the spatial position offset is less than the preset spatial offset threshold and the temporal rate of change is lower than the first rate of change threshold are divided into static safety basic features, representing the background baseline values ​​that each sensing feature should present when the submarine cable is not affected by external disturbances in the current state. Feature items that have a spatial position offset greater than or equal to the preset spatial offset threshold, or feature items that have a spatial position offset less than the preset spatial offset threshold but a temporal rate of change higher than the second rate of change threshold, are divided into dynamic safety response features, representing the additional response amount that is different from the background baseline caused by external moving targets or instantaneous disturbance events.

[0040] Furthermore, step S300 also includes: using the static security basic feature as a spatial reference benchmark, spatially associating each dynamic security response feature with the static security basic feature to determine the spatial distribution relationship of each dynamic security response feature relative to the static scene; based on the spatial distribution relationship, calculating the deviation degree of the current feature value of each dynamic security response feature relative to its respective historical baseline value, wherein the deviation degree is the absolute value of the difference between the current feature value of each dynamic security response feature and its respective historical baseline value; sorting the dynamic security response features from largest to smallest according to the deviation degree, selecting the top N dynamic security response features as the core security response features; and marking the dynamic object to which the core security response features belong as the dynamic tracking target.

[0041] Preferably, all values ​​of the segmented static safety basic features are used as baseline reference values. Each value of the dynamic safety response feature is compared with the corresponding static baseline value to determine the direction and magnitude of the deviation of the current value of the dynamic feature from the normal background value at the current moment. Among all the segmented dynamic safety response features, the absolute value of the deviation of the current value of each dynamic feature from its historical baseline value is calculated. The top N values ​​are sorted from largest to smallest by the degree of deviation, and the items with the most significant deviation from the background state are selected and marked as core safety response features, which correspond to the highest intensity of risk signals. Since each feature item in the interactive safety response feature set originates from a specific sensing target or sensing event, each feature item determined as a core safety response feature can be traced back to the specific dynamic object that generated the response to determine the dynamic tracking target. For example, a certain vibration amplitude abnormality feature belongs to a specific AIS-positioned ship, and a certain magnetic anomaly feature belongs to a certain metal target.

[0042] Step S400: Based on the traceability core security response characteristics and dynamic tracking targets, perform time-series response tracking and analysis on the multi-source sensing data to generate time-series security early warning information.

[0043] Step S400 further includes: calculating the rate of change and assessing the signal quality of the traceability core safety response characteristics; determining the risk tracking type of the dynamic tracking target based on the rate of change and signal quality assessment results; the risk tracking type includes one or more of the following: rapidly changing risk targets, weak signal risk targets, and composite risk targets; driving the corresponding multi-source sensors to perform time-series response tracking and acquisition on the dynamic tracking target in accordance with the risk tracking type, obtaining tracking perception data of the dynamic tracking target within a continuous time window; performing time-series correlation analysis on the tracking perception data to extract the motion evolution trend and feature change trend of the dynamic tracking target; and generating time-series safety early warning information based on the motion evolution trend and the feature change trend, combined with the spatial positional relationship between the dynamic tracking target and the submarine cable laying path.

[0044] Preferably, for the top N core security response features selected, the numerical sequence of the feature at multiple consecutive time points is obtained, the numerical sequence is differentially processed, the change between each adjacent time point is calculated and divided by the time interval to obtain the instantaneous rate of change sequence of the feature on the time axis, and the mean of the rate sequence or the latest rate value at the current time point is calculated. The magnitude of the rate of change reflects the urgency of the current change of the feature, and the higher the value, the more drastic the change in the intensity of the risk response signal. The quality of the original sensing signals corresponding to the core security response characteristics is evaluated, and quantitative indicators of signal quality are calculated. These include the ratio of the energy of the effective signal components to the energy of the background noise components; a higher signal-to-noise ratio indicates better signal quality. The system also analyzes whether there are interruptions, packet loss, or missing data within a continuous time window and calculates the proportion of effective data. Furthermore, it calculates the variance or standard deviation of the signal within a short time window; a smaller variance indicates a more stable signal and higher measurement repeatability. The system detects the presence of stray components from non-target sources, such as environmental electromagnetic interference or false echoes caused by multipath reflections, and calculates the proportion of stray components in the total signal energy. Finally, a weighted synthesis is performed according to preset weights to obtain a signal quality score, reflecting the reliability of the sensing data upon which the current core security response characteristics depend.

[0045] Preferably, the calculated rate of change value is compared with a preset rate of change grading threshold, and the signal quality score is compared with a preset quality pass threshold. The comparison results of the two dimensions are combined to determine the risk tracking type of the dynamically tracked target. Specifically, if the rate of change is higher than the preset fast change threshold, it indicates that the perception response caused by the target is rapidly increasing or moving rapidly, and the risk state is changing drastically, thus it is a fast-changing risk target. If the signal strength or signal-to-noise ratio in the signal quality assessment result is lower than the preset weak signal threshold, it indicates that the perception response amplitude caused by the target is weak and close to the background noise level, indicating that the weak signal still has risk significance, thus it is a weak signal risk target. If the rate of change is higher than the fast change threshold and the signal quality is lower than the weak signal threshold, it indicates that the target is both in a rapidly changing state and the signal is weak, belonging to the target type with high detection difficulty and risk urgency, thus it is a composite risk target.

[0046] Furthermore, step S400 also includes: when the risk tracking type is a rapidly changing risk target, driving the corresponding multi-source sensor to perform continuous tracking and acquisition of the dynamic tracking target at an acquisition frequency higher than the normal acquisition frequency; when the risk tracking type is a weak signal risk target, driving the corresponding multi-source sensor to switch to a multi-cycle cumulative acquisition mode, in which the signals acquired within the continuous time window are superimposed in phase to improve the signal strength; when the risk tracking type is a composite risk target, driving the corresponding multi-source sensor to simultaneously perform continuous tracking and acquisition at a frequency higher than the normal acquisition frequency and the multi-cycle cumulative acquisition, and marking the corresponding dynamic tracking target as a high-concern risk target in the generated time-series security warning information.

[0047] Preferably, the risk tracking type is used as the control command input, and a differentiated sensor driving strategy is invoked to perform different modes of acquisition operations for different types of targets. The physical implementation of this operation is to send specific acquisition parameter configuration commands to each sensor system, including sampling frequency adjustment commands, acquisition mode switching commands, and signal processing parameter configuration commands. For rapidly changing risk targets, a command to increase the sampling frequency is sent, so that the sensor continuously samples the spatial area where the target is located at a frequency higher than the conventional set value. For weak signal risk targets, a command to switch to multi-cycle cumulative acquisition mode is sent, so that the sensor repeatedly acquires signals for multiple cycles at the same spatial location and superimposes the acquisition data of each cycle. For complex risk targets, both types of commands are sent simultaneously, so that the sensor both increases the sampling frequency and performs multi-cycle cumulative acquisition. Finally, the tracking perception data of the dynamic tracking target within a continuous time window is obtained, including the spatial position coordinates, signal strength value, and characteristic parameter value of the target at each sampling moment within the time window.

[0048] Preferably, the tracking and sensing data within a continuous time window are arranged in chronological order, and the temporal correlation between data at each moment is calculated. Specifically, the spatial coordinate sequence of the target is temporally differentiald, the target's velocity sequence and direction sequence are calculated, the target's characteristic parameter sequence is correlated with the time axis, the evolution trajectory of the characteristic parameters over time is calculated, and the temporal synchronization relationship between the changes in characteristic parameters and changes in spatial position is detected, thereby determining the causal relationship between the target's behavior and the sensing response. Based on the results of temporal correlation analysis, the spatial position sequence of the target is trend-fitted, such as linear fitting, polynomial fitting, or Kalman filtering prediction, to obtain the predicted trajectory of the target in a short future time window, including the predicted heading angle, speed value, and the predicted intersection point and arrival time of the predicted path and the submarine cable route. At the same time, based on the results of temporal correlation analysis, the characteristic parameter sequence of the target is trend-fitted to obtain the expected value of the direction and amplitude of change of the characteristic parameter as it monotonically increases, monotonically decreases, or periodically fluctuates over time. For example, the vibration amplitude shows an exponential growth trend, and the magnetic anomaly intensity shows a decaying trend, reflecting that the threat intensity posed by the target to the submarine cable is increasing or decreasing. The motion evolution trend and characteristic change trend are used as two input parameters for generating early warning information. The spatial distance sequence between the predicted future position sequence of the target and each coordinate point of the submarine cable laying path is calculated from the motion evolution trend. The minimum distance value and the corresponding coordinates and estimated arrival time of the submarine cable position point are determined to determine whether the target will enter the safe warning distance range of the submarine cable and the estimated entry time. Finally, all calculation results are summarized to output structured time-series safety early warning information, including but not limited to the identification information of the dynamically tracked target, current spatial position coordinates, motion evolution trend, characteristic change trend, closest distance to the submarine cable path and estimated arrival time, the currently determined risk tracking type, and the risk level or suggested action based on the above information.

[0049] In the above text, refer to Figure 1 This paper describes in detail a multi-source sensing fusion method for submarine cable patrol safety early warning according to an embodiment of the present invention. Next, we will refer to... Figure 2 This invention describes a multi-source sensing fusion-based submarine cable patrol safety early warning system according to an embodiment of the present invention.

[0050] According to an embodiment of the present invention, a multi-source sensing fusion submarine cable patrol safety early warning system addresses the technical problems in existing submarine cable safety early warning technologies, such as isolated multi-source sensing data, static and coarse spatial area division, and a lack of temporal causal tracing capabilities between dynamic target behavior and response characteristics. This system achieves the technical effect of improving the accuracy of global situational awareness, dynamic risk tracing capabilities, and emergency response decision-making efficiency in submarine cable patrol early warning. Figure 2As shown, a multi-source sensing fusion submarine cable patrol safety early warning system includes: a safety patrol area establishment module 10, a safety response feature acquisition module 20, a feature recognition and segmentation module 30, and a safety early warning information generation module 40.

[0051] The safety patrol area establishment module 10 is used to perform safety feature analysis and clustering based on the submarine cable laying location to establish a safety patrol area, wherein each safety patrol area has response safety features and response risk mapping relationship; the safety response feature acquisition module 20 is used to acquire multi-source sensing data, and based on the correspondence of sensing patrol areas, perform multi-source feature extraction interaction according to the response safety features and response risk mapping relationship to obtain an interactive safety response feature set; the feature recognition and segmentation module 30 is used to perform safety static and safety dynamic recognition and segmentation based on the interactive safety response feature set, and trace the core safety response features and dynamic tracking targets; the safety early warning information generation module 40 is used to perform time-series response tracking analysis on multi-source sensing data according to the traced core safety response features and dynamic tracking targets to generate time-series safety early warning information.

[0052] The specific configuration of the safety patrol area establishment module 10 will be described in detail below. The safety patrol area establishment module 10 further includes: acquiring geospatial information of the submarine cable laying path and historical risk records along the path; using the submarine cable laying path as an axis, performing multi-dimensional safety feature clustering according to the submarine cable burial depth, seabed geological type, ship traffic flow, distribution of historical anchor damage events, and environmental perception conditions to divide the submarine cable path into multiple safety patrol areas; wherein, the clustering results of each dimension are cross-combined to establish response safety feature labels for each safety patrol area; and establishing the response risk mapping relationship under each response safety feature label based on the historical statistical response relationship between the response safety feature labels and risk events.

[0053] The specific configuration of the safe cruise area establishment module 10 will be described in detail below. The safe cruise area establishment module 10 further includes: the submarine cable burial depth status including shallow burial status, standard burial depth status, and exposed status; the seabed geological type including hard geology, soft geology, and mixed geology; and the environmental perception conditions including deep-water sonar effective conditions, shallow-water sonar failure magnetic detection effective conditions, and intertidal optical / magnetic synergy conditions.

[0054] The specific configuration of the safety response feature acquisition module 20 will be described in detail below. The safety response feature acquisition module 20 further includes: spatially aligning the multi-source sensing data to determine the sensing spatial location corresponding to each sensing data point; matching the sensing spatial location with the spatial range of each of the safety patrol areas to determine the safety patrol area to which the current multi-source sensing data belongs and its corresponding response safety feature label; extracting features from the multi-source sensing data according to the preset risk association feature dimension in the response safety feature label to obtain the sensing feature vector of each data source under the risk association feature dimension; calculating the matching degree between each sensing feature vector and various risk event association feature templates in the response safety feature label, marking sensing features with matching degrees exceeding a preset matching degree threshold as valid sensing matching features; querying the response risk mapping relationship according to the data source type and feature type corresponding to the valid sensing matching features, determining the sensing contribution and response sensitivity of each valid sensing matching feature to various risk events in the current area, establishing an association mapping between sensing data, safety features, and risk events, and generating the interactive safety response feature set.

[0055] The specific configuration of the feature recognition and segmentation module 30 will be described in detail below. The feature recognition and segmentation module 30 further includes: calculating the temporal rate of change and spatial position offset of each feature item in the interactive security response feature set; segmenting the interactive security response feature set into static security basic features and dynamic security response features based on the temporal rate of change and spatial position offset; and tracing the core security response features in the dynamic security response features and the dynamic tracking target to which the core security response features belong, using the static security basic features as a reference.

[0056] The specific configuration of the feature recognition and segmentation module 30 will be described in detail below. The feature recognition and segmentation module 30 further includes: segmenting feature items whose spatial position offset is less than a preset spatial offset threshold and whose temporal change rate is less than a first change rate threshold into the static security basic features; and segmenting feature items whose spatial position offset is greater than or equal to the preset spatial offset threshold, or feature items whose spatial position offset is less than the preset spatial offset threshold but whose temporal change rate is greater than a second change rate threshold, into the dynamic security response features, wherein the second change rate threshold is greater than or equal to the first change rate threshold.

[0057] The specific configuration of the feature recognition and segmentation module 30 will be described in detail below. The feature recognition and segmentation module 30 further includes: using the static security basic features as a spatial reference benchmark, spatially associating each dynamic security response feature with the static security basic features to determine the spatial distribution relationship of each dynamic security response feature relative to the static scene; based on the spatial distribution relationship, calculating the deviation degree of the current feature value of each dynamic security response feature relative to its respective historical baseline value, wherein the deviation degree is the absolute value of the difference between the current feature value of each dynamic security response feature and its respective historical baseline value; sorting the dynamic security response features from largest to smallest according to the deviation degree, selecting the top N dynamic security response features as the core security response features; and marking the dynamic object to which the core security response features belong as the dynamic tracking target.

[0058] The specific configuration of the safety warning information generation module 40 will be described in detail below. The safety warning information generation module 40 further includes: calculating the rate of change and assessing the signal quality of the traceable core safety response characteristics; determining the risk tracking type of the dynamic tracking target based on the rate of change and signal quality assessment results; the risk tracking type includes one or more of rapidly changing risk targets, weak signal risk targets, and composite risk targets; driving corresponding multi-source sensors to perform time-series response tracking and acquisition of the dynamic tracking target matching the risk tracking type, obtaining tracking perception data of the dynamic tracking target within a continuous time window; performing time-series correlation analysis on the tracking perception data to extract the motion evolution trend and feature change trend of the dynamic tracking target; and generating time-series safety warning information based on the motion evolution trend and the feature change trend, combined with the spatial positional relationship between the dynamic tracking target and the submarine cable laying path.

[0059] The specific configuration of the security warning information generation module 40 will be described in detail below. The security warning information generation module 40 further includes: when the risk tracking type is a rapidly changing risk target, driving the corresponding multi-source sensor to perform continuous tracking and acquisition of the dynamically tracked target at a higher acquisition frequency than the conventional acquisition frequency; when the risk tracking type is a weak signal risk target, driving the corresponding multi-source sensor to switch to a multi-cycle cumulative acquisition mode, in which the signals acquired within a continuous time window are superimposed in phase to enhance signal strength; when the risk tracking type is a composite risk target, driving the corresponding multi-source sensor to simultaneously perform continuous tracking and acquisition at a higher acquisition frequency than the conventional acquisition frequency and the multi-cycle cumulative acquisition, and marking the corresponding dynamically tracked target as a high-concern risk target in the generated time-series security warning information.

[0060] The multi-source sensing fusion submarine cable patrol safety early warning system provided in this embodiment of the invention can execute the multi-source sensing fusion submarine cable patrol safety early warning method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0061] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for safety early warning of submarine cable patrols using multi-source sensing fusion, characterized in that, include: Based on the location of submarine cable laying, safety feature analysis and clustering are performed to establish safety patrol areas, where each safety patrol area has response safety features and response risk mapping relationships. Acquire multi-source perception data, and based on the corresponding relationship of perception cruise areas, perform multi-source feature extraction interaction according to the aforementioned response safety features and response risk mapping relationship to obtain an interactive safety response feature set; Based on the interactive security response feature set, perform static and dynamic security identification and segmentation, and trace core security response features and dynamically track targets; Based on the core security response characteristics and dynamic tracking targets, the multi-source sensing data is analyzed for time-series response tracking to generate time-series security early warning information.

2. The multi-source sensing fusion method for submarine cable patrol safety early warning according to claim 1, characterized in that, Based on the location of submarine cables, safety features are analyzed and clustered to establish a safe patrol area, including: The geospatial information of the submarine cable laying route and historical risk records along the route are obtained. Taking the submarine cable laying route as the axis, multi-dimensional safety feature clustering is performed according to the submarine cable burial depth, seabed geological type, ship traffic flow, distribution of historical anchor damage events and environmental perception conditions. The submarine cable route is divided into multiple safety patrol areas. The clustering results of each dimension are cross-combined to establish response safety feature labels for each of the safety patrol areas. Based on the historical statistical response relationship between response security feature labels and risk events, establish the response risk mapping relationship under each response security feature label.

3. The multi-source sensing fusion method for submarine cable patrol safety early warning according to claim 2, characterized in that, The cable burial depth status includes shallow burial, standard burial depth, and exposed state; the seabed geological types include hard geology, soft geology, and mixed geology; the environmental sensing conditions include deep-water sonar effective conditions, shallow-water sonar failure magnetic detection effective conditions, and intertidal optical / magnetic synergy conditions.

4. The multi-source sensing fusion method for submarine cable patrol safety early warning according to claim 2, characterized in that, Based on the correspondence of perceived cruise areas, multi-source feature extraction is performed according to the aforementioned response safety features and response risk mapping relationship to obtain an interactive safety response feature set, including: Spatial alignment is performed on the multi-source sensing data to determine the sensing spatial location corresponding to each sensing data. The sensing spatial location is then matched with the spatial range of each of the safety cruise areas to determine the safety cruise area to which the current multi-source sensing data belongs and the corresponding response safety feature label. Based on the preset risk association feature dimension in the response security feature label, feature extraction is performed on the multi-source perception data to obtain the perception feature vector of each data source under the risk association feature dimension. Calculate the matching degree between each of the perception feature vectors and the risk event associated feature templates in the response security feature label, and mark the perception features with a matching degree exceeding a preset matching degree threshold as valid perception matching features; Based on the data source type and feature type corresponding to the effective perception matching feature, query the response risk mapping relationship, determine the perception contribution and response sensitivity of each effective perception matching feature to various risk events in the current area, establish the association mapping between perception data, security features, and risk events, and generate the interactive security response feature set.

5. The multi-source sensing fusion method for submarine cable patrol safety early warning according to claim 1, characterized in that, include: The temporal rate of change and spatial position offset are calculated for each feature item in the interactive security response feature set; Based on the time-domain rate of change and spatial location offset, the interactive security response feature set is divided into static security basic features and dynamic security response features; Using the static security basic characteristics as a reference, the core security response characteristics in the dynamic security response characteristics and the dynamic tracking targets to which the core security response characteristics belong are traced.

6. The multi-source sensing fusion method for submarine cable patrol safety early warning according to claim 5, characterized in that, Based on the time-domain rate of change and spatial location offset, the interactive security response feature set is divided into static security basic features and dynamic security response features, including: The feature items whose spatial position offset is less than a preset spatial offset threshold and whose temporal change rate is less than a first change rate threshold are segmented into the static security basic features; The feature items whose spatial position offset is greater than or equal to the preset spatial offset threshold, or the feature items whose spatial position offset is less than the preset spatial offset threshold but whose time domain change rate is higher than the second change rate threshold, are segmented into the dynamic safety response features, wherein the second change rate threshold is greater than or equal to the first change rate threshold.

7. The multi-source sensing fusion method for submarine cable patrol safety early warning according to claim 5, characterized in that, Using the aforementioned static security fundamental characteristics as a reference benchmark, the core security response characteristics within the dynamic security response characteristics and the dynamic tracking targets to which the core security response characteristics belong are traced, including: Using the static security basic features as a spatial reference benchmark, the dynamic security response features are spatially correlated with the static security basic features to determine the spatial distribution relationship of each dynamic security response feature relative to the static scene. Based on the spatial distribution relationship, the deviation of the current feature value of each dynamic security response feature from its respective historical baseline value is calculated, wherein the deviation is the absolute value of the difference between the current feature value of each dynamic security response feature and its respective historical baseline value. The dynamic security response features are sorted from largest to smallest according to the degree of deviation, and the top N dynamic security response features are selected as the core security response features. The dynamic object to which the core security response feature belongs is marked as the dynamic tracking target.

8. The multi-source sensing fusion method for submarine cable patrol safety early warning according to claim 1, characterized in that, Based on the aforementioned core security response characteristics and dynamic tracking targets, multi-source sensing data is analyzed for time-series response tracking to generate time-series security early warning information, including: The change rate and signal quality of the core security response characteristics are calculated and evaluated. Based on the change rate and signal quality evaluation results, the risk tracking type of the dynamic tracking target is determined. The risk tracking type includes one or more of the following: rapidly changing risk targets, weak signal risk targets, and composite risk targets. Based on the risk tracking type, drive the corresponding multi-source sensors to perform time-series response tracking and acquisition on the dynamic tracking target in a manner that matches the risk tracking type, and obtain tracking and perception data of the dynamic tracking target within a continuous time window; The tracking and sensing data is analyzed by time-series correlation to extract the motion evolution trend and feature change trend of the dynamically tracked target; Based on the motion evolution trend and the characteristic change trend, combined with the spatial positional relationship between the dynamic tracking target and the submarine cable laying path, a time-series safety early warning information is generated.

9. The multi-source sensing fusion method for submarine cable patrol safety early warning according to claim 8, characterized in that, Based on the risk tracking type, the corresponding multi-source sensors are driven to perform time-series response tracking and acquisition on the dynamically tracked target, matching the risk tracking type, including: When the risk tracking type is a rapidly changing risk target, the corresponding multi-source sensor is driven to perform continuous tracking and acquisition of the dynamic tracking target at a higher acquisition frequency than the conventional acquisition frequency; When the risk tracking type is a weak signal risk target, the corresponding multi-source sensor is driven to switch to the multi-cycle cumulative acquisition mode. In the multi-cycle cumulative acquisition mode, the signals acquired within the continuous time window are superimposed in phase to improve the signal strength. When the risk tracking type is a composite risk target, the corresponding multi-source sensor is driven to simultaneously perform continuous tracking and acquisition at a higher frequency than the conventional acquisition frequency and the multi-cycle cumulative acquisition, and the corresponding dynamic tracking target is marked as a high-concern risk target in the generated time-series safety warning information.

10. A multi-source sensing fusion submarine cable patrol safety early warning system, characterized in that, The system is used to implement the multi-source sensing fusion method for submarine cable patrol safety early warning as described in any one of claims 1 to 9, the system comprising: The safety patrol area establishment module is used to perform safety feature analysis and clustering based on the submarine cable laying location to establish safety patrol areas, where each safety patrol area has response safety features and response risk mapping relationships. The safety response feature acquisition module is used to acquire multi-source perception data, and based on the correspondence of perception cruise areas, perform multi-source feature extraction interaction according to the response safety features and response risk mapping relationship to obtain an interactive safety response feature set. The feature recognition and segmentation module is used to perform static and dynamic security recognition and segmentation based on the interactive security response feature set, and to trace core security response features and dynamically track targets. The safety early warning information generation module is used to perform time-series response tracking and analysis on multi-source sensing data based on the core safety response characteristics of the traceability core and the dynamic tracking target, and generate time-series safety early warning information.

Images