Intelligent sensing and identifying method and system for offshore buoy

Abstract

The invention relates to the technical field of intelligent monitoring of marine equipment, in particular to an intelligent sensing and recognizing method and system for a marine buoy, and the method comprises the following steps: constructing a multi-modal data collection array of the buoy, and obtaining a multi-source heterogeneous sensing data set of the buoy based on the multi-modal data collection array; establishing a spatial-temporal feature fusion model, and performing comprehensive analysis on the multi-source heterogeneous sensing data set to obtain a state feature vector of the buoy; according to a deep residual network architecture, analyzing the state feature vector to obtain a health degree score and anomaly type probability distribution of the buoy; on the basis of a self-adaptive threshold algorithm, the health degree score and the abnormal type probability distribution are combined to obtain early warning information of the buoy; and according to the early warning information, carrying out abnormity verification on the buoy by using unmanned inspection equipment so as to realize intelligent sensing and identification of the offshore buoy. According to the method, innovative technical support is provided for buoy inspection, and the method has important engineering application value.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent monitoring of marine equipment, and in particular to an intelligent sensing and identification method and system for marine buoys. Background Art

[0002] Traditional buoy inspections mostly rely on regular manual inspections and a single sensor's threshold alarm mechanism, which has significant defects such as narrow perception dimensions, delayed response, high false alarm rate, and high maintenance cost. Manual inspections require regular dispatch of ships or divers to check for buoy anchor chain wear, biological attachment, or equipment failure. The cost of a single inspection in remote waters is high and the inspection cycle is long, resulting in an extremely high lag rate for detecting equipment failure. The monitoring system based on fixed sensors can only obtain surface water quality parameters and cannot identify problems such as underwater structure damage, anchor chain breakage, or illegal ship collisions. The mechanical threshold is easily affected by wind and wave noise, and the false alarm rate is high.

[0003] In recent years, breakthroughs in the Internet of Things and artificial intelligence technologies have provided a new path for the intelligentization of buoys: multimodal sensor arrays can simultaneously collect the structural deformation of the buoy body, the status of the anchor chain, underwater corrosion images and dynamic data of the surrounding waters, and build a holographic perception network of "body status-environmental disturbance-abnormal events"; lightweight AI models realize real-time data processing through edge computing modules, increase the speed of anchor chain breakage identification to seconds, and use transfer learning technology to adapt to the environmental characteristics of different sea areas.

[0004] However, the existing technology still faces bottlenecks such as the difficulty in spatiotemporal alignment of multi-source heterogeneous data and insufficient model robustness in complex environments, resulting in a high rate of misjudgment of equipment failures. The present invention focuses on constructing a state feature vector through the multi-source heterogeneous perception data set of the buoy, further obtaining a health score and anomaly type probability distribution, and then generating early warning information, which is verified using unmanned inspection equipment, thereby realizing intelligent perception and recognition of the buoy, and providing real-time and high-precision technical support for intelligent recognition of the buoy. Summary of the invention

[0005] In view of the defects in the prior art, the present invention provides a method and system for intelligent perception and recognition of marine buoys.

[0006] In order to achieve the above-mentioned objectives, in a first aspect, the present invention provides a method for intelligent perception and identification of offshore buoys, the method comprising the following steps: constructing a multimodal data acquisition array for the buoy, and acquiring a multi-source heterogeneous perception data set of the buoy based on the multimodal data acquisition array; establishing a spatiotemporal feature fusion model, and comprehensively analyzing the multi-source heterogeneous perception data set to obtain a state feature vector of the buoy; based on a deep residual network architecture, analyzing the state feature vector to obtain a health score and anomaly type probability distribution of the buoy; based on an adaptive threshold algorithm, combining the health score and the anomaly type probability distribution to obtain early warning information of the buoy; based on the early warning information, using unmanned inspection equipment to perform abnormality verification on the buoy to achieve intelligent perception and identification of offshore buoys. The present invention significantly improves the efficiency and reliability of ocean buoy inspections through multimodal data fusion and deep learning technology; the multimodal data acquisition array breaks through the limitation of a single data dimension and comprehensively perceives the deformation of the buoy structure, the degree of corrosion and the dynamic changes of the environment; the spatiotemporal feature fusion model effectively extracts key state features by mining the spatiotemporal correlation between data, solving the problem of insufficient feature extraction in traditional methods; the deep residual network combined with the adaptive threshold algorithm can not only accurately quantify the health of the buoy, but also identify multiple types of abnormal patterns. Compared with the manual detection method with a fixed threshold, the false alarm rate is effectively reduced; the closed-loop verification mechanism of the unmanned inspection equipment greatly shortens the abnormal response time, and can still achieve all-weather intelligent monitoring in extreme environments such as typhoons; it provides innovative technical support for the stable operation of ocean buoys and has important engineering application value.

[0007] Optionally, the multimodal data acquisition array of the buoy is constructed, and the multi-source heterogeneous perception data set of the buoy is obtained based on the multimodal data acquisition array, including: establishing the multimodal data acquisition array based on meteorological sensors, hydrological sensors, high-precision positioning devices and image acquisition devices; obtaining the meteorological data, hydrological data, location information and image information of the buoy according to the multimodal data acquisition array to construct the multi-source heterogeneous perception data set. The present invention realizes the digital construction of all elements of the buoy's operating status and environmental dynamics by integrating meteorological, hydrological, positioning and visual four-dimensional perception modules; meteorological sensors capture atmospheric parameters such as wind speed, temperature and humidity in real time, hydrological sensors monitor seawater flow rate, salinity and wave characteristics, high-precision positioning devices track the drift trajectory of the buoy, and image acquisition devices identify visual anomalies such as structural corrosion or biological attachment; multi-source heterogeneous data complementary verification can not only locate physical damage, but also distinguish environmental interference from equipment failure, significantly improving data credibility. Compared with a single sensor system, multi-dimensional perception reduces data blind spots and provides high-precision, strongly correlated underlying data support for subsequent intelligent diagnosis.

[0008] Optionally, the establishment of a spatiotemporal feature fusion model and the comprehensive analysis of the multi-source heterogeneous perception data set to obtain the state feature vector of the buoy include: based on spatiotemporal modeling, spatiotemporal feature fusion is performed in combination with a hierarchical fusion strategy to obtain the spatiotemporal feature fusion model; cross-modal feature extraction is performed on the multi-source heterogeneous perception data set according to the spatiotemporal feature fusion model to obtain the buoy visual modality, buoy temporal modality and buoy spatial modality; the buoy visual modality, the buoy temporal modality and the buoy spatial modality are jointly characterized and analyzed to construct the state feature vector. The present invention breaks through the limitations of traditional single-modal analysis through hierarchical spatiotemporal modeling, and effectively mines the nonlinear correlation between multi-source heterogeneous data of buoys; the spatiotemporal fusion strategy synchronously analyzes the spatiotemporal coupling laws of environmental parameter fluctuations, equipment displacement trajectories and visual appearance changes, eliminating the fragmentation problem of single-dimensional feature extraction; cross-modal feature extraction deeply interweaves the spatial characteristics of the image, the dynamic evolution of time series data and the spatial topological relationship, capturing the implicit correlation between micro-displacements caused by the accumulation of anchor chain stress and image deformation; the high-information-density feature vectors generated by the joint characterization improve the accuracy of equipment health assessment, especially the early recognition sensitivity of gradual faults, laying the core data foundation for accurate early warning.

[0009] Optionally, the jointly characterizing and analyzing the buoy visual mode, the buoy temporal mode and the buoy spatial mode to construct the state feature vector includes: ; in, is the state feature vector, Representation layer normalization, is the linear transformation parameter, represents the rectified linear unit, is the total amount of the buoy mode, is the traversal count flag, is the spatiotemporal attention weight parameter, is the modal weight gating vector, is the buoy mode, The present invention dynamically allocates multimodal weights through the spatiotemporal attention mechanism, combines the gating mechanism with nonlinear activation to achieve cross-modal feature adaptive fusion; layer normalization ensures gradient stability, avoids feature drift caused by environmental noise, and provides a high-discrimination feature space for anomaly detection.

[0010] Optionally, the state feature vector is analyzed based on the deep residual network architecture to obtain the health score and abnormal type probability distribution of the buoy, including: obtaining a regression task branch and a classification task branch based on the deep residual network architecture, and performing state analysis on the buoy; performing feature dimension matching and feature enhancement on the state feature vector to obtain the state feature optimization vector of the buoy; according to the regression task branch, taking the state feature optimization vector as input to obtain the health score; according to the classification task branch, combining the state feature optimization vector to obtain the abnormal type probability distribution. The present invention establishes a dual-branch structure through a residual network to achieve multi-task collaborative optimization of health assessment and abnormality recognition. The feature enhancement module eliminates environmental noise interference, the dimension matching makes the visual-temporal features adapt to different task requirements, the residual jump connection ensures the gradient stability of the deep network, the regression branch accurately quantifies the structural integrity of the buoy, and the classification branch uses cross-attention to identify common buoy faults, which is significantly better than the traditional single-task analysis model.

[0011] Optionally, performing feature dimension matching and feature enhancement on the state feature vector to obtain the state feature optimization vector of the buoy includes: performing feature dimension matching on the state feature vector to obtain a state feature matching vector, which satisfies the following relationship: ; in, is the state feature matching vector, is the feature dimension mapping matrix, is the state feature vector, is the feature dimension bias term, Represents the dimension, To match the feature dimensions, is the original feature dimension; the state feature matching vector is feature enhanced to obtain a state feature optimization vector, which satisfies the following relationship: ; in, Optimize the vector for state features, is the state feature matching vector, is Gaussian noise, represents a Gaussian distribution, is the standard deviation of Gaussian noise. The present invention realizes multimodal feature space alignment through feature dimension mapping matrix, solves the problem of dimensional heterogeneity of visual, temporal and spatial modalities, and improves the adaptability of the dual-branch input of the residual network. Gaussian noise injection enhances the robustness of the model to environmental noise and avoids overfitting. After feature enhancement, dimensional matching and noise disturbance work together to improve the generalization ability of the model under complex interferences such as typhoons and biological attachment, providing standardized and interference-resistant feature vector representation for subsequent precise evaluation.

[0012] Optionally, the adaptive threshold algorithm is based on the health score and the abnormal type probability distribution to obtain the warning information of the buoy, including: obtaining a dynamic health threshold and an abnormal probability threshold through the adaptive threshold algorithm, analyzing the buoy in combination with environmental factors to obtain a multi-dimensional joint analysis result; based on the health score and the abnormal type probability distribution, performing a spatiotemporal correlation check on the buoy array according to the environmental factors to obtain a consistency check result; combining the multi-dimensional joint analysis result and the consistency check result to comprehensively evaluate the buoy to obtain the warning information. The present invention breaks through the limitations of traditional static warnings through dynamic threshold adjustment, comprehensively considers environmental interference factors such as typhoon intensity and ocean current mutations, so that the health threshold adaptively floats with environmental factors, compares adjacent buoy data through spatiotemporal correlation checks, identifies single-point abnormal false alarms, and the joint analysis can distinguish between mechanical failures and environmental shocks, ensuring efficient operation and maintenance of the ocean buoy network.

[0013] Optionally, the performing of a spatiotemporal correlation check on the buoy array based on the health score and the abnormal type probability distribution according to the environmental factors to obtain a consistency check result includes: ; ; in, Indicates the consistency check result. is the indicative function, is the spatial weight coefficient of the buoy, is the abnormal state of the buoy, is the buoy array, is the dynamic threshold, is the calibration threshold, is the environmental factor control coefficient, It is a real-time risk assessment value of environmental factors. The present invention realizes the spatiotemporal credibility verification of abnormal events through the synergy of spatial weights and dynamic thresholds. The algorithm is weighted according to the abnormal state of adjacent buoys, and the threshold is dynamically adjusted in combination with the risk of sea conditions, which effectively distinguishes local equipment failures from regional disasters. The spatial correlation mechanism reduces the false alarm rate and improves the accuracy of collaborative verification of key faults such as anchor chain breakage. The environmental adaptive characteristics ensure that a high verification reliability is maintained in complex environments.

[0014] Optionally, according to the warning information, the unmanned inspection equipment is used to verify the abnormality of the buoy to achieve intelligent perception and identification of offshore buoys, including: according to the warning information, the unmanned inspection equipment is collaboratively planned to obtain task configuration and dynamic obstacle avoidance path; the unmanned inspection equipment is used to verify the abnormality of the buoy according to the task configuration and the dynamic obstacle avoidance path to obtain the buoy verification and identification result. The present invention shortens the abnormality verification response time by combining the real-time sea conditions to generate wind and wave resistant equipment obstacle avoidance path through multi-device collaborative scheduling driven by warning levels. The dynamic task configuration automatically switches the sensor module for abnormal types such as corrosion and anchor chain breakage, improves the verification accuracy, and realizes the full intelligent autonomy of the "warning-verification" closed loop of ocean buoys.

[0015] In the second aspect, the present invention provides an intelligent perception and identification system for offshore buoys, the system executes the intelligent perception and identification method for offshore buoys provided by the present invention, the system includes an input device, an output device, a processor and a memory, and its gain lies in: the hardware facilities integrated by the present invention have excellent performance, the input device, the output device, the processor and the memory are interconnected, the information transmission between the various components is smooth, and an efficient information processing system is constructed through the interaction of multiple hardware facilities. The present invention constructs an efficient and collaborative working platform, realizes the intelligence and automation of system functions, improves the processing speed and response capability of the system, ensures the accuracy and timeliness of buoy identification work, and provides stable and reliable hardware equipment support for the intelligent perception and identification of buoys. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a flow chart of a method for intelligently sensing and identifying marine buoys according to an embodiment of the present invention;

[0017] Figure 2 This is a framework diagram of the marine buoy intelligent perception and recognition system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The specific embodiments of the present invention will be described in detail below. It should be noted that the embodiments described herein are only for illustration and are not intended to limit the present invention. In the following description, a large number of specific details are set forth in order to provide a thorough understanding of the present invention. However, it is obvious to those of ordinary skill in the art that these specific details do not need to be adopted to implement the present invention. In other examples, in order to avoid confusing the present invention, known circuits, software or methods are not specifically described.

[0019] Throughout the specification, references to "one embodiment," "an embodiment," "an example," or "an example" mean that a particular feature, structure, or characteristic described in conjunction with the embodiment or example is included in at least one embodiment of the present invention. Therefore, the phrases "in one embodiment," "in an embodiment," "an example," or "an example" appearing in various places throughout the specification do not necessarily all refer to the same embodiment or example. In addition, particular features, structures, or characteristics may be combined in one or more embodiments or examples in any suitable combination and / or subcombination. In addition, it should be understood by those of ordinary skill in the art that the figures provided herein are for illustrative purposes and that the figures are not necessarily drawn to scale.

[0020] See also Figure 1 An embodiment of the present invention provides a method for intelligently sensing and identifying marine buoys, the method comprising the following steps:

[0021] S1. Construct a multimodal data acquisition array for the buoy, and acquire a multi-source heterogeneous perception data set of the buoy based on the multimodal data acquisition array.

[0022] Among them, S1 specifically includes the following steps:

[0023] S11. Establishing the multimodal data acquisition array based on meteorological sensors, hydrological sensors, high-precision positioning devices and image acquisition devices.

[0024] In this embodiment, a multi-modal data acquisition array is constructed through the collaborative deployment of multiple sensors; first, an anti-wind erosion meteorological sensor group is installed on the top of the buoy, including a three-cup anemometer, a temperature and humidity sensor, and a piezoelectric barometer; secondly, according to the actual inspection task requirements, a hydrological sensor unit is deployed at an appropriate position of the buoy to ensure the effective collection of the required data, and the wave motion trajectory is captured by integrating the CTD probe, the acoustic Doppler current profiler and the six-axis inertial measurement unit; the high-precision positioning device uses a dual-frequency GNSS receiver to couple with the inertial navigation system (INS), and eliminates the positioning jitter caused by surges through Kalman filtering; the image acquisition device consists of a panoramic camera, a side-scan sonar and an infrared thermal imager, and a corrosion-resistant image acquisition unit is deployed on the underwater part of the buoy through dynamic waterproof packaging technology, and a time synchronizer is used to ensure the timing alignment of the visual data and the sensor data.

[0025] Specifically, the image acquisition device includes a drone equipped with an intelligent sensing camera and an underwater robot; the drone image acquisition device uses a six-rotor drone equipped with a three-axis gimbal high-resolution camera, which automatically matches the buoy GNSS coordinates through RTK differential positioning, generates a drone surround inspection path based on task planning, and simultaneously loads real-time marine environmental information to avoid surge interference. The camera turns on the multi-spectral imaging mode and scans the buoy surface in different areas at an inclination angle of 30°, and cooperates with the laser rangefinder to maintain the best focusing distance. The image data is processed by the onboard AI chip in real time to perform corrosion identification and structural crack detection, and abnormal areas automatically trigger detailed shooting; the underwater robot information acquisition device uses an electric propulsion type equipped with side-scan sonar, multi-spectral underwater camera and sacrificial anode protection potential sensor, which is linked to the beacon at the bottom of the buoy through an ultra-short baseline positioning system, and performs three-dimensional scanning along the preset spiral descent path. The sonar constructs a three-dimensional point cloud of the anchor chain in real time, and the camera uses a mechanical arm to assist in macro shooting of riveted point corrosion, and simultaneously collects potential data to evaluate the risk of failure of the anti-corrosion layer.

[0026] Furthermore, the multimodal data acquisition array adopts a multi-dimensional auxiliary inspection mode to construct a three-dimensional monitoring system. The multi-dimensional auxiliary inspection mode includes manual inspection, which responds to special inspection needs after sudden failures and extreme weather through regular on-site verification and calibration of equipment parameters, and fills in the technical monitoring blind spots; the telemetry and remote control system based on the communication network realizes the real-time collection and remote control of the buoy position, attitude, and hydrological parameters; the AIS system carried by passing ships can automatically receive the status information broadcast by the buoy, and transmit it back to the management platform through the ship-shore data link, forming a mobile monitoring network, which is especially suitable for information blind spots in remote waters and emergency situations; the three-dimensional monitoring system significantly improves the monitoring effect of buoys.

[0027] In an optional embodiment, the navigation light is integrated and installed on the upper part of the buoy light frame, and intelligent switching between day and night is achieved through a photosensitive sensor; it is in standby mode during the day to save energy, and automatically turns on and maintains the strobe function at night to provide ships with clear channel markings; its control unit is interconnected with the buoy central processor, and when the water quality sensor detects that the oil pollution exceeds the standard, the navigation light switches to a red warning strobe and synchronously triggers an alarm signal to the shore-based monitoring platform; in addition, the navigation light is equipped with an AIS receiver that can identify information about ships in close range; the built-in self-diagnostic chip regularly transmits voltage and temperature parameters to the cloud, supports remote adjustment of the flash mode, realizes intelligent coordination with the buoy system, and significantly improves the efficiency of water safety monitoring.

[0028] S12. Acquire the meteorological data, hydrological data, location information and image information of the buoy according to the multimodal data acquisition array to construct the multi-source heterogeneous perception data set.

[0029] In this embodiment, meteorological data of the buoy environment, including wind speed, wind direction, temperature, humidity, air pressure and visibility, are obtained through meteorological sensors; hydrological data of the buoy position, including water temperature, salinity, water flow velocity, pH value, dissolved oxygen, chlorophyll, nutrients, turbidity, oil pollution, wave data, current data, ship wastewater pollution and ship exhaust emissions, are obtained through hydrological sensors; the position information of the buoy, including longitude and latitude, altitude, inclination, movement trajectory and relative positions of adjacent buoys, is obtained through a high-precision positioning device; image information of the buoy, including buoy appearance information, anchor chain status, ship traffic flow information and environmental information, the buoy appearance includes buoy coloring, buoy corrosion and buoy fouling, and the environmental information includes ship search and rescue signals.

[0030] Specifically, the multi-spectral sensor system (integrated with fluorescence, ultraviolet and infrared technologies) carried by the buoy operates around the clock and is deployed close to the water surface. The fluorescence sensor identifies oil by detecting the specific wavelength fluorescence emitted by the oil after being excited by ultraviolet light. The ultraviolet sensor non-contactly scans the ultraviolet band reflected by the water surface to locate the oil film, and the infrared sensor assists in quantifying the thickness of the oil film. At the same time, the buoy combines the positioning device to track the diffusion path of the oil pollution and transmit it to the shore-based monitoring platform to provide precise support for emergency response.

[0031] It should be noted that the relevant influencing factors of visibility in meteorological data include rain, fog, snow and haze.

[0032] Furthermore, firstly, the clocks of various sensors are aligned through the time synchronization controller to ensure that meteorological data, hydrological data, dual-frequency GNSS / INS combined positioning data and panoramic images form a millisecond-level aligned spatiotemporal sequence; secondly, edge computing nodes are deployed, and a sliding window mechanism is used for data cleaning, including three-point calibration of temperature, salinity and depth data, Kalman filter denoising of positioning data, and adaptive histogram equalization of image data; finally, a spatiotemporal index database is constructed to store multi-source data in four dimensions (time + longitude + latitude + depth) according to UTC timestamps and WGS84 coordinates, and the sensor metadata (range, accuracy, unit) is semantically annotated through the JSON-LD framework to form a multi-source heterogeneous perception data set containing physical quantity time series, spatial trajectories, structural images and environmental contexts, providing standardized data input for subsequent feature fusion.

[0033] It should be noted that the temperature, salinity and depth data refer to the temperature data, salinity data and depth data of seawater.

[0034] In an optional embodiment, the multimodal data acquisition array builds an intelligent monitoring network covering the waters around the buoy by integrating multiple environmental perception devices; the high-definition video camera is combined with the millimeter-wave radar to capture the changes in water surface ripples in real time. When a person accidentally falls into the water, the unique "water splash-human posture" coupling recognition algorithm can give an early warning in seconds, complete the judgment of the falling into the water event, and simultaneously trigger the positioning mark; the infrared thermal imager and the multi-spectral water quality analyzer work together to detect people who fall into the water at night through temperature anomalies, and can accurately identify the characteristic absorption peaks of the oil film on the water surface. When hydrocarbon pollutants are detected, the spectral matching database will automatically compare the oil sample fingerprint library to generate pollution, etc. The sonar sensor of the multimodal data acquisition array can penetrate turbid waters, draw underwater three-dimensional terrain maps, and combine with AI models to predict the oil diffusion path to provide support for emergency decision-making; the sensor data is pre-processed by the edge computing node and transmitted to the shore-based monitoring platform through a low-latency network; at the same time, by deploying high-precision tension sensors and vibration monitors, the anchor chain status data is collected in real time, combined with edge computing for millisecond-level analysis. Once the precursor of anchor chain breakage is found, the sound and light alarm is immediately triggered and pushed to the shore-based monitoring platform simultaneously, achieving a response in seconds, winning golden time for emergency disposal, and ensuring a response in seconds in complex environments.

[0035] In addition, in terms of data encryption and physical protection, a layered protection strategy is adopted: the hardware layer equips the sensor nodes with explosion-proof and waterproof casings, built-in three-axis vibration sensors and tracking modules, and abnormal displacement triggers the self-destruct circuit to fuse the data storage chip; the transmission layer deploys data encryption algorithms to encrypt multi-source heterogeneous perception data sets, and synchronizes the data to the shore-based monitoring platform through dual-channel redundant transmission, and rotates the session keys regularly through the dynamic key distribution system; and performs integrity verification on the cleaned data, and automatically isolates abnormal data blocks; the storage layer uses blockchain technology to generate a Merkle tree structure for the encrypted multimodal data according to the UTC timestamp, and realizes tamper-proof verification of distributed storage nodes through the PBFT consensus mechanism; the access layer sets up multi-factor authentication, and implements minimum privilege access control based on the zero-trust architecture, and the audit log is encrypted with random numbers for data synchronization.

[0036] S2. Establish a spatiotemporal feature fusion model, and comprehensively analyze the multi-source heterogeneous perception data set to obtain the state feature vector of the buoy.

[0037] Among them, S2 specifically includes the following steps:

[0038] S21. Based on spatiotemporal modeling, spatiotemporal feature fusion is performed in combination with a hierarchical fusion strategy to obtain the spatiotemporal feature fusion model.

[0039] In this embodiment, the construction of the spatiotemporal feature fusion model adopts a three-level hierarchical architecture: the first layer encodes the unimodal spatiotemporal features, the visual modality extracts the spatial-temporal local features through 3D-CNN, the temporal modality uses a bidirectional LSTM to model the dynamic evolution, and the spatial modality uses a graph convolutional network (GCN) to model the topological relationship of the buoy array; the second layer designs a spatiotemporal cross-attention mechanism to achieve cross-modal feature interaction; the third layer generates a joint representation through gated fusion; the spatiotemporal fusion model satisfies the following relationship:

[0040]

[0041] in, is the space-time fusion vector, Indicates the visual mode, represents the time series mode, represents the spatial mode, is the gating coefficient, is the linear transformation parameter of the mode, is the mode vector, is the control coefficient.

[0042] S22. Perform cross-modal feature extraction on the multi-source heterogeneous perception data set based on the spatiotemporal feature fusion model to obtain buoy visual modality, buoy temporal modality and buoy spatial modality.

[0043] In this embodiment, the visual modality uses a pre-trained 3D-CNN network to extract the spatiotemporal features of structural deformation from continuous 10-second image slices (300 frames) (output 512-dimensional vectors), the temporal modality uses a bidirectional LSTM to model dynamic evolution and capture the dynamic evolution patterns of parameters such as wind speed and salinity (output 256-dimensional vectors), and the spatial modality uses a graph convolutional network (GCN) to model the topological relationship of the buoy array (output 128-dimensional vectors). The multi-source heterogeneous perception dataset is used as the model input of the spatiotemporal feature fusion model to obtain the buoy visual modality, buoy temporal modality and buoy spatial modality.

[0044] Specifically, the buoy visual mode satisfies the following relationship:

[0045]

[0046] in, is the buoy visual modality vector, represents a three-dimensional convolutional neural network, is the buoy image sequence within the time window.

[0047] The buoy timing mode satisfies the following relationship:

[0048]

[0049] in, is the buoy time series mode vector, represents a bidirectional long short-term memory network, It is the time series sensor data.

[0050] The buoy space mode satisfies the following relationship:

[0051]

[0052] in, is the buoy space mode vector, represents a graph convolutional network, is the adjacency matrix of the float array, are the geographic coordinates of the buoy.

[0053] S23, jointly characterize and analyze the buoy visual mode, the buoy temporal mode and the buoy spatial mode to construct the state feature vector.

[0054] In this embodiment, multi-source feature fusion is achieved through cross-modal interaction and attention mechanism dynamic weighting strategy, the buoy visual modal vector, buoy temporal modal vector and buoy spatial modal vector are dimensionally aligned, and the state feature vector is obtained by layer normalization and rectified linear unit, which satisfies the following relationship:

[0055]

[0056] in, is the state feature vector, Representation layer normalization, is the linear transformation parameter, represents the rectified linear unit, is the total amount of the buoy mode, is the traversal count flag, is the spatiotemporal attention weight parameter, is the modal weight gating vector, is the buoy mode, is the bias vector.

[0057] S3. According to the deep residual network architecture, the state feature vector is analyzed to obtain the health score and abnormal type probability distribution of the buoy.

[0058] Among them, S3 specifically includes the following steps:

[0059] S31. Obtain a regression task branch and a classification task branch based on the deep residual network architecture, and perform a state analysis on the buoy.

[0060] In this embodiment, a dual-task branch construction based on a deep residual network realizes collaborative analysis through shared feature extraction and task-specific optimization; a shared backbone network is designed, and multiple improved residual blocks are adopted. The first residual block introduces a channel attention mechanism to strengthen key features, and then a dual-branch structure is constructed, including a regression task branch and a classification task branch; the dual tasks are jointly trained through a dynamic weighted loss function, and a gradient normalization algorithm is introduced to balance the learning speed to prevent gradient conflicts caused by sensor noise.

[0061] Specifically, the regression task branch consists of three fully connected layers and is embedded with a spatiotemporal attention module, which focuses on capturing continuous degradation features such as buoy corrosion and offset. At the end, the Tanh activation function is used to constrain the output to the [0,1] interval as the health score.

[0062] Specifically, the classification task branch adopts a channel grouping strategy and splits it into multiple groups of multidimensional sub-features according to the anomaly type. Each group extracts specific patterns through a lightweight module, and outputs the Softmax-normalized anomaly probability distribution after global maximum pooling.

[0063] S32. Perform feature dimension matching and feature enhancement on the state feature vector to obtain the state feature optimization vector of the buoy.

[0064] In this embodiment, the state feature vector is matched with the feature dimension through the feature dimension mapping matrix to obtain the state feature matching vector, which satisfies the following relationship:

[0065]

[0066] in, is the state feature matching vector, is the feature dimension mapping matrix, is the state feature vector, is the feature dimension bias term, Represents the dimension, To match the feature dimensions, is the original feature dimension.

[0067] Furthermore, the state feature matching vector is enhanced using adaptive Gaussian noise to obtain the state feature optimization vector, which satisfies the following relationship:

[0068]

[0069] in, Optimize the vector for state features, is the state feature matching vector, is Gaussian noise, represents a Gaussian distribution, is the standard deviation of Gaussian noise.

[0070] It should be noted that the standard deviation of the Gaussian noise is dynamically adjusted by the environmental risk assessment value of the area where the buoy is located.

[0071] S33. According to the regression task branch, the state feature optimization vector is used as input to obtain the health score.

[0072] Specifically, in the regression task branch, the generation of health scores is achieved through multi-layer feature refinement and physical constraint optimization. First, the state feature optimization vector is input into a feature extractor composed of multiple residual blocks, and the dilated convolution is used to capture the temporal degradation trend and output high-dimensional features. Then, a dual-path self-attention module is used to focus on key indicators: one path extracts the cumulative effect characteristics of the corrosion rate through a time sliding window, and the other path strengthens the spatial correlation based on the spatial convolution kernel. Finally, it is mapped to the health score through a fully connected layer, satisfying the following relationship:

[0073]

[0074] in, Score your health. represents the hyperbolic tangent activation function, is the learnable parameter matrix, Optimize the vector for state features, is the bias term.

[0075] Furthermore, a loss function is used to train the learnable parameters, and a health derivative constraint is introduced to prevent sudden changes in the health score.

[0076] S34. According to the classification task branch, the abnormal type probability distribution is obtained in combination with the state feature optimization vector.

[0077] Specifically, in the classification task branch, the generation of abnormal type probability distribution is achieved through the collaboration of multi-channel specific modeling and cross-modal attention; the state feature optimization vector is input into the abnormal classification module composed of multiple parallel sub-networks, and each sub-network adopts the following processing flow: first, feature decoupling is performed, and the input features are split into multiple groups through group convolution, and each group is associated with a specific abnormal pattern; secondly, attention is enhanced, and spatiotemporal cross attention is applied to each group of features to strengthen abnormality-related features; finally, the abnormal type probability distribution is generated, and after each group undergoes lightweight convolution and global maximum pooling, the initial probability is output through the Sigmoid activation function, and then the abnormal type probability distribution is obtained through Softmax normalization.

[0078] Furthermore, in the classification task branch, an improved cross-entropy loss function is used to alleviate the imbalance problem of samples, and a cross-task comparison constraint is introduced to ensure the consistency of the classification results with the physical level of the health score.

[0079] S4. Based on an adaptive threshold algorithm, the warning information of the buoy is obtained by combining the health score and the probability distribution of the abnormal type.

[0080] Wherein, S4 specifically includes the following steps:

[0081] S41. A dynamic health threshold and an abnormal probability threshold are obtained through the adaptive threshold algorithm, and the buoy is analyzed in combination with environmental factors to obtain a multi-dimensional joint analysis result.

[0082] In this embodiment, the marine environmental parameters are counted through a sliding window to calculate the real-time risk assessment value of the environmental factor, satisfying the following relationship:

[0083]

[0084] in, Real-time risk assessment value for environmental factors, is the ambient wind speed, is the effective wave height, is the ocean current speed.

[0085] Furthermore, the dynamic health threshold is set based on the real-time risk assessment value of environmental factors to satisfy the following relationship:

[0086]

[0087] in, is the dynamic health threshold, is the buoy health baseline threshold, is the adjustment coefficient, Provide real-time risk assessment values ​​for environmental factors.

[0088] In this embodiment, based on the adaptive threshold algorithm, according to the real-time risk assessment value of environmental factors, combined with the category adaptive strategy, the abnormal probability threshold is obtained, which satisfies the following relationship:

[0089]

[0090] in, For the The anomaly probability threshold of the class anomaly, For the The baseline probability threshold of the class anomaly, For the Environmental sensitivity coefficient of class anomaly, Provides real-time risk assessment values ​​for environmental factors.

[0091] Specifically, during the multidimensional joint analysis, a health-environmental parameter decision plane is constructed, the buoys are divided into normal or abnormal buoys through support vector domain description, and the abnormal probability distribution is integrated to construct a three-dimensional feature space. The isolation forest algorithm is used to detect complex anomalies. Based on the dynamic health threshold and abnormal probability threshold, the potential impact of marine environmental factors on the operating status of the buoy is comprehensively considered, and a multidimensional joint analysis is performed to obtain a comprehensive evaluation result of the buoy status that integrates environmental factors as the result of the multidimensional joint analysis.

[0092] S42. Based on the health score and the probability distribution of the abnormal type, a spatiotemporal correlation check is performed on the buoy array according to the environmental factors to obtain a consistency check result.

[0093] In this embodiment, firstly, current and historical marine environmental factors (such as wind speed, wave height, current speed, etc.) are considered, which may affect the operation status of the buoy; then, a temporal sequence analysis is performed on each buoy in the buoy array to check whether the change trend of its health score and abnormal type probability over time is consistent; at the same time, a spatial correlation analysis is performed to compare whether the health scores and abnormal type probabilities of adjacent buoys have logical coherence or similarity; finally, the analysis results of both time and space are combined to determine whether the overall state of the buoy array remains consistent, thereby obtaining a consistency check result that satisfies the following relationship:

[0094]

[0095]

[0096] in, Indicates the consistency check result. is the indicative function, is the spatial weight coefficient of the buoy, is the abnormal state of the buoy, is the buoy array, is the dynamic threshold, is the calibration threshold, is the environmental factor control coefficient, Provides real-time risk assessment values ​​for environmental factors.

[0097] S43. Combining the multi-dimensional joint analysis result and the consistency check result, a comprehensive evaluation is performed on the buoy to obtain the warning information.

[0098] In this embodiment, the generation of warning information is intended to ensure that the health status of the buoy can be reflected and communicated to the engineering personnel in a timely and accurate manner. First, based on the adaptive threshold algorithm and the multi-dimensional joint analysis results, it is determined whether the health score of the buoy is lower than the dynamic health threshold and whether the probability distribution of the abnormal type exceeds the abnormal probability threshold. Once the abnormal condition is detected, the warning mechanism will be triggered. Then, according to the consistency check result, the abnormality in the buoy array is located and classified to determine whether it is an isolated event or a regional problem, and whether the abnormal problem may affect or escalate each other, and identify potential risk points and impact ranges. Then, based on the comprehensive information, the warning information is automatically generated. The warning information includes the identification information of the buoy, the health score, the warning level, the abnormal type and probability, the possible cause of the failure, the recommended countermeasures, etc., and is presented in the form of charts, lists or short text descriptions, so as to quickly grasp the health status and potential threats of the buoy. Finally, the warning information will be sent to relevant personnel in a timely manner through preset communication channels, such as SMS, email, application notification, etc. The warning information can be quickly conveyed to provide strong support for subsequent troubleshooting, maintenance and emergency response.

[0099] S5. Based on the warning information, use unmanned inspection equipment to verify the abnormality of the buoy to achieve intelligent perception and identification of offshore buoys.

[0100] Among them, S5 specifically includes the following steps:

[0101] S51. Coordinate planning of the unmanned inspection equipment is performed based on the warning information to obtain task configuration and dynamic obstacle avoidance path.

[0102] In this embodiment, the unmanned inspection equipment is first collaboratively planned based on the early warning information, and the specific inspection task requirements are determined. Then, the corresponding task configuration is assigned to the unmanned inspection equipment based on the task requirements, such as the inspection priority, the location information of the target buoy, and the required detection equipment. At the same time, the path planning algorithm is used in combination with real-time environmental data to generate a dynamic optimal path for the unmanned inspection equipment to ensure that it can safely and efficiently reach the target buoy for anomaly verification.

[0103] Specifically, the task priority is analyzed by the warning level and the abnormality type to generate a task configuration matrix, including but not limited to the task duration, sensor requirements and target buoy coordinates; based on the real-time sea condition data, the path planning algorithm is combined with the Bezier curve to plan the initial path, and the path smoothness and safety are optimized through dynamic weights; at the same time, a real-time obstacle avoidance module is embedded: sudden obstacles (such as floating objects) are perceived through the onboard LiDAR and millimeter-wave radar, and the path nodes are dynamically adjusted using the reinforcement learning algorithm, and the operating timing of the unmanned inspection equipment is coordinated based on the time synchronization protocol to ensure the efficient execution of the abnormality verification task.

[0104] It should be noted that unmanned inspection equipment includes but is not limited to drones and unmanned inspection ships.

[0105] S52: Utilize the unmanned inspection equipment to perform abnormal verification on the buoy according to the task configuration and the dynamic obstacle avoidance path to obtain a buoy verification and identification result.

[0106] In this embodiment, after obtaining the task configuration and dynamic obstacle avoidance path, the unmanned inspection equipment autonomously navigates to the target buoy position according to the preset instructions and planned path; after arriving, the unmanned inspection equipment will use its own high-precision sensors (such as cameras, infrared sensors, etc.) to perform detailed abnormality detection on the buoy, and can capture key information such as the appearance status of the buoy, changes in the surrounding environment, etc., and through built-in image processing and recognition algorithms, such as target detection and classification models based on deep learning, the collected data is analyzed in real time. By comparing with the preset normal buoy status, the unmanned inspection equipment can intelligently identify whether there are any abnormalities in the buoy, such as structural damage, biological attachment or deviation from the original position, and generate a detailed buoy verification and identification result report to provide a basis for subsequent processing decisions, thereby realizing rapid and accurate verification and evaluation of the status of offshore buoys, and improving the efficiency and intelligence level of offshore buoy monitoring and maintenance.

[0107] In an optional embodiment, the engineering and technical personnel wear AR smart helmets and receive all-round monitoring data of the buoy in real time through high-speed networks; the high-definition display screen built into the helmet superimposes the buoy's three-dimensional model, real-time hydrological parameters, equipment operating status and other information in the form of holographic projection in the field of vision of the engineering and technical personnel, forming an immersive monitoring interface that integrates virtual and real; relying on high-precision positioning data, the buoy position deviation is intuitively presented in a dynamic trajectory diagram, combined with the sensor readings corrected by the compensation algorithm, to accurately judge the structural stability. When early warning information is detected, the engineering and technical personnel use the AR helmet to remotely control the buoy status and assist in decision-making; engineers can call the historical inspection database, retrieve the parameter change curves of the buoys in the same area over the years, and remotely formulate maintenance strategies in combination with the fault probability model deduced by the AI ​​algorithm; significantly improve the operation and maintenance efficiency and emergency response capabilities of offshore facilities.

[0108] See also Figure 2In an optional embodiment, the present invention provides an intelligent sensing and identification system for offshore buoys, the system comprising an input device, an output device, a processor and a memory, the hardware facilities being interconnected, wherein the memory is used to store a computer program, the computer program comprising program instructions, and the processor is configured to call the program instructions to execute the specific steps of the embodiment of the intelligent sensing and identification method for offshore buoys provided by the present invention. The intelligent sensing and identification system for offshore buoys provided by the present invention has a complete structure, objective stability, and improves the overall applicability and practical application capability of the present invention.

[0109] In summary, the method and system for intelligent perception and identification of offshore buoys provided by the method of the present invention construct a buoy perception closed-loop system through multimodal data fusion and deep learning; the multimodal sensor array integrates meteorological, hydrological, positioning and visual data to realize full-dimensional perception of buoy structural deformation, corrosion status and environmental dynamics; the spatiotemporal feature fusion model combines the attention mechanism to mine cross-modal spatiotemporal associations and generate high-discrimination state feature vectors; the deep residual network synchronously outputs the health score and abnormal probability distribution through a dual-branch architecture, and reduces the false alarm rate in conjunction with the dynamic threshold algorithm; the unmanned inspection equipment generates the best path based on the warning level to realize abnormality verification in a short time, reduces the buoy inspection cost, improves the fault response efficiency, and provides an all-weather, reliable intelligent operation and maintenance solution for the buoy observation system, which has significant engineering application value; the method of the present invention is easy to understand, simple to calculate, small workload, and convenient for engineering application, providing a theoretical basis and technical support for the further development of the field of intelligent monitoring technology of marine equipment.

[0110] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.

Claims

1. A method for intelligently sensing and identifying marine buoys, characterized in that: The steps include: Constructing a multimodal data acquisition array for the buoy, and acquiring a multi-source heterogeneous sensing data set of the buoy based on the multimodal data acquisition array; Establishing a spatiotemporal feature fusion model, and comprehensively analyzing the multi-source heterogeneous perception data set to obtain the state feature vector of the buoy; According to the deep residual network architecture, the state feature vector is analyzed to obtain the health score and abnormal type probability distribution of the buoy; Based on an adaptive threshold algorithm, the warning information of the buoy is obtained by combining the health score and the probability distribution of the abnormal type; According to the warning information, unmanned inspection equipment is used to verify the abnormality of the buoy to achieve intelligent perception and identification of offshore buoys.

2. The method for intelligent sensing and identifying marine buoys according to claim 1, characterized in that: The method of constructing a multimodal data acquisition array for a buoy and acquiring a multi-source heterogeneous sensing data set for the buoy based on the multimodal data acquisition array includes: Establishing the multimodal data acquisition array based on meteorological sensors, hydrological sensors, high-precision positioning devices and image acquisition devices; The meteorological data, hydrological data, location information and image information of the buoy are acquired according to the multimodal data acquisition array to construct the multi-source heterogeneous perception data set.

3. The method for intelligent sensing and identifying marine buoys according to claim 1, characterized in that: The step of establishing a spatiotemporal feature fusion model and comprehensively analyzing the multi-source heterogeneous sensing data set to obtain the state feature vector of the buoy includes: Based on spatiotemporal modeling, spatiotemporal feature fusion is performed in combination with a hierarchical fusion strategy to obtain the spatiotemporal feature fusion model; Performing cross-modal feature extraction on the multi-source heterogeneous perception data set according to the spatiotemporal feature fusion model to obtain buoy visual modality, buoy temporal modality and buoy spatial modality; The state feature vector is constructed by jointly characterizing and analyzing the buoy visual mode, the buoy temporal mode and the buoy spatial mode.

4. The method for intelligent sensing and identifying marine buoys according to claim 3 is characterized in that: The step of jointly characterizing and analyzing the buoy visual mode, the buoy temporal mode, and the buoy spatial mode to construct the state feature vector includes: ； in, is the state feature vector, Representation layer normalization, is the linear transformation parameter, represents the rectified linear unit, is the total amount of the buoy mode, is the traversal count flag, is the spatiotemporal attention weight parameter, is the modal weight gating vector, is the buoy mode, is the bias vector.

5. The method for intelligent sensing and identifying marine buoys according to claim 1, characterized in that: The state feature vector is analyzed based on the deep residual network architecture to obtain the health score and abnormal type probability distribution of the buoy, including: Based on the deep residual network architecture, a regression task branch and a classification task branch are obtained, and a state analysis is performed on the buoy; Performing feature dimension matching and feature enhancement on the state feature vector to obtain a state feature optimization vector of the buoy; According to the regression task branch, the state feature optimization vector is used as input to obtain the health score; According to the classification task branch, the abnormal type probability distribution is obtained in combination with the state feature optimization vector.

6. The method for intelligent sensing and identifying marine buoys according to claim 5, characterized in that: The step of performing feature dimension matching and feature enhancement on the state feature vector to obtain the state feature optimization vector of the buoy includes: The state feature vector is matched with the feature dimension to obtain a state feature matching vector, which satisfies the following relationship: ； in, is the state feature matching vector, is the feature dimension mapping matrix, is the state feature vector, is the feature dimension bias term, Represents the dimension, To match the feature dimensions, is the original feature dimension; The state feature matching vector is enhanced to obtain a state feature optimization vector, which satisfies the following relationship: ； in, Optimize the vector for state features, is the state feature matching vector, is Gaussian noise, represents a Gaussian distribution, is the standard deviation of Gaussian noise.

7. The method for intelligent sensing and identifying marine buoys according to claim 1, characterized in that: The method of obtaining the warning information of the buoy based on the adaptive threshold algorithm in combination with the health score and the probability distribution of the abnormal type includes: The dynamic health threshold and the abnormal probability threshold are obtained by the adaptive threshold algorithm, and the buoy is analyzed in combination with environmental factors to obtain a multi-dimensional joint analysis result; Based on the health score and the probability distribution of the abnormality type, performing a spatiotemporal correlation check on the buoy array according to the environmental factors to obtain a consistency check result; The buoy is comprehensively evaluated in combination with the multi-dimensional joint analysis result and the consistency check result to obtain the warning information.

8. The method for intelligent sensing and identifying marine buoys according to claim 7, characterized in that: The step of performing a spatiotemporal correlation check on the buoy array based on the health score and the abnormal type probability distribution according to the environmental factors to obtain a consistency check result includes: ； ； in, Indicates the consistency check result. is the indicative function, is the spatial weight coefficient of the buoy, is the abnormal state of the buoy, is the buoy array, is the dynamic threshold, is the calibration threshold, is the environmental factor control coefficient, Provide real-time risk assessment values ​​for environmental factors.

9. The method for intelligent sensing and identifying marine buoys according to claim 1, characterized in that: According to the warning information, the buoy is verified to be abnormal by using unmanned inspection equipment to realize intelligent perception and recognition of offshore buoys, including: According to the warning information, the unmanned inspection equipment is collaboratively planned to obtain task configuration and dynamic obstacle avoidance path; The unmanned inspection equipment is used to perform abnormal verification on the buoy according to the task configuration and the dynamic obstacle avoidance path to obtain a buoy verification and identification result.

10. The marine buoy intelligent perception and recognition system is characterized by: The system includes an input device, an output device, a processor and a memory, and the input device, the output device, the processor and the memory are interconnected, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to call the program instructions to execute the offshore buoy intelligent perception and recognition method as described in any one of claims 1-9.

Images