Mooring cable monitoring method and device

Abstract

The invention relates to a mooring rope monitoring method and device, and belongs to the technical field of ship safety, and the method comprises the steps: obtaining the image data and stress data of a mooring rope, and the data of the environment where the mooring rope is located; performing feature extraction on the image data, the stress data and the environment data to obtain image features, stress features and environment features; carrying out feature fusion on the image features, the stress features and the environment features to obtain fusion features; and inputting the fusion features into a preset mooring rope state classifier to obtain a state classification result of the mooring rope. The state of the mooring rope is monitored by fusing the image data, the stress data and the environment data, and the monitoring accuracy can be improved.

Description

Technical Field

[0001] The present invention relates to the field of ship safety technology, and in particular to a mooring cable monitoring method and device. Background Art

[0002] Ship mooring cables are important components for ships to dock safely at the pier. Their operating status directly affects the stability and safety of ships during mooring. Current cable safety monitoring mostly relies on a single type of sensor, such as a stress sensor. Although this monitoring method has been significantly improved compared to manual inspection and can provide continuous monitoring data, in complex working environments, it is difficult for a single sensor to accurately capture safety hazards caused by non-stress factors such as surface damage and wear of mooring cables, which limits the comprehensiveness and accuracy of the monitoring system. Summary of the invention

[0003] In view of this, it is necessary to provide a mooring cable monitoring method and device to solve the problem that the prior art uses a single sensor and is difficult to accurately monitor the status of the mooring cable.

[0004] In order to solve the above problems, in a first aspect, the present invention provides a mooring line monitoring method, comprising: Acquire image data, stress data and environmental data of mooring cables; Extracting features from the image data, the stress data and the environment data respectively to obtain image features, stress features and environment features; Performing feature fusion on the image features, stress features and environmental features to obtain fusion features; The fused features are input into a preset mooring cable state classifier to obtain a state classification result of the mooring cable.

[0005] Optionally, the extracting features from the image data to obtain image features includes: Perform mooring cable detection on the image data using a trained YOLO model to obtain a mooring cable image region; The trained VGG model is used to extract features of the mooring cable image region to obtain image features.

[0006] Optionally, extracting features from the stress data to obtain stress features includes: Performing a fast Fourier transform on the stress data to obtain first stress data; Performing wavelet transformation on the stress data to obtain second stress data at different time scales; A stress signature is generated based on the first stress data and the second stress data.

[0007] Optionally, the obtaining of environmental data of the mooring cable includes: Acquire first environmental data of the mooring rope at a first preset time interval; wherein the first environmental data includes: one of wind speed, wind direction and wave height; Acquire second environmental data of the mooring rope at a second preset time interval; wherein the second environmental data includes at least one of air pressure, humidity and seawater temperature; and the second preset time interval is greater than the first preset time interval.

[0008] Optionally, the performing feature fusion on the image features, stress features and environmental features to obtain fused features includes: Inputting the image features, stress features and environmental features into a multi-layer perceptron to obtain attention weights of the image features, stress features and environmental features; According to the attention weights of the image features, stress features and environmental features, the image features, stress features and environmental features are weightedly fused to obtain fused features.

[0009] Optionally, the method further includes: determining the wear condition of the mooring line according to the image features; Determining whether the attention weight of the image feature exceeds an image attention weight threshold; When the attention weight of the image feature exceeds the image attention weight threshold, determining a stress alarm threshold according to the wear condition; When the stress data exceeds the stress alarm threshold, an alarm message is generated.

[0010] Optionally, the method further includes: Determine whether the attention weight of the environmental data exceeds an environmental attention weight threshold; When the attention weight of the environmental data exceeds the environmental attention weight threshold, determining a stress alarm threshold according to the environmental data; When the stress data exceeds the stress alarm threshold, an alarm message is generated.

[0011] Optionally, the state classification result includes a health state and a probability corresponding to the health state; different health states correspond to different probability thresholds; and the method further includes: Determining a target probability threshold corresponding to the health state according to the health state output by the mooring line state classifier; When the probability output by the mooring line state classifier is greater than the target probability threshold, an alarm message is generated.

[0012] Optionally, the method further includes: According to the material properties of the mooring lines and the opinions of domain experts, the probability thresholds corresponding to each health state are set.

[0013] In a second aspect, the present invention further provides a mooring line monitoring device, comprising: A data acquisition module, used to acquire image data, stress data and environmental data of the mooring cable; A feature extraction module, used to extract features from the image data, the stress data and the environmental data respectively, to obtain image features, stress features and environmental features; A feature fusion module, used for fusing the image features, stress features and environmental features to obtain fused features; The state determination module is used to input the fusion feature into a preset mooring cable state classifier to obtain a state classification result of the mooring cable.

[0014] The beneficial effects of the present invention are: The present invention improves the acquisition of image data, stress data and environmental data of mooring cables; extracts features from image data, stress data and environmental data respectively to obtain image features, stress features and environmental features; fuses image features, stress features and environmental features to obtain fusion features; inputs the fusion features into a preset mooring cable state classifier to obtain a state classification result of the mooring cable. By fusing image data, stress data and environmental data to monitor the state of the mooring cable, the safety hazards caused by non-stress factors such as surface damage and wear of the mooring cable can be accurately captured, the mooring cable can be monitored in an all-round way, and the accuracy of monitoring can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A schematic diagram of a flow chart of an embodiment of a mooring line monitoring method provided by the present invention; Figure 2 A data collection flow chart provided by the present invention; Figure 3 A mooring cable monitoring framework diagram provided by the present invention; Figure 4 A data fusion flow chart provided by the present invention; Figure 5 Another mooring line monitoring framework diagram provided by the present invention; Figure 6 A schematic structural diagram of an embodiment of a mooring line monitoring device provided by the present invention. DETAILED DESCRIPTION

[0016] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0017] In the description of the embodiments of the present invention, unless otherwise specified, the meaning of "plurality" is two or more. The "first", "second", etc. involved in the embodiments of the present invention are used to distinguish similar objects, and are not used to describe a specific order or sequence, nor are they used to indicate or imply their relative importance or implicitly indicate the number of technical features indicated. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application can be implemented in an order other than those illustrated or described here, and the objects distinguished by "first", "second", etc. are generally of one type, and the number of objects is not limited. For example, the first object can be one or more.

[0018] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0019] Reference Figure 1 , shows a schematic flow chart of an embodiment of a mooring line monitoring method provided by the present invention, the method comprising: S101, acquiring image data, stress data and environmental data of the mooring cable.

[0020] Regarding the acquisition of image data, cameras can be installed at multiple key locations of the ship, such as at the bow, stern, both sides of the ship, and above the mooring points of the cables, to ensure 360-degree coverage without blind spots and to obtain image data of the mooring cables in real time in all directions. Specifically, first of all, cameras with night vision and infrared functions can be selected to ensure that the cable images can be clearly captured in low-visibility environments. In addition, the camera can also be waterproof, dustproof and corrosion-resistant, with a shell protection level of at least IP67 to adapt to the salt spray erosion and frequent humidity changes in the marine environment. Then, use a special bracket to fix the camera to ensure that the installation is stable and avoid the impact of ship shaking on image quality. Adjust the camera angle to ensure that the mooring cable is centered in the camera image for subsequent image analysis. Then, connect the camera to the internal network of the ship to ensure that the image data can be transmitted in real time via wired or wireless means. For cameras far away from the main network, wireless transmission technology such as Wi-Fi or Lora Wan can be used to ensure a stable signal. Finally, the camera is controlled to capture at least 30 frames per second to capture the dynamic changes of the cable and ensure clear images when the cable moves quickly or the ship shakes.

[0021] After acquiring the image data, part of it can be stored locally to prevent network interruption. At the same time, the image data is also transmitted to the data processing center in real time through the ship's internal network or satellite communication to ensure the real-time and continuity of the data.

[0022] In addition, the online status, storage space and network connection of the camera can be monitored in real time to ensure the stability and continuity of the image data acquisition process. The salt and dirt on the lens should be cleaned regularly, and the sealing of the camera housing should be checked to ensure the long-term stable operation of the equipment.

[0023] Regarding the acquisition of stress data, stress sensors can be installed at multiple key locations of the mooring cable, for example, at the fixed end of the mooring cable, the contact point between the mooring cable and the mooring facility, and the possible wear area on the mooring cable to obtain stress data in real time. Specifically, first of all, a strain gauge or pressure sensor with high precision, good temperature stability, anti-electromagnetic interference ability and long-term stability can be selected as a stress sensor. Then, the stress sensor is calibrated to ensure that its measurement accuracy meets the requirements. The sensitivity and zero drift of the stress sensor are recorded during the calibration process for compensation during the later data analysis. Then, a professional fixing fixture and a high-viscosity, weather-resistant adhesive are used to firmly attach the stress sensor to the surface of the mooring cable to avoid displacement or damage of the stress sensor due to vibration or impact caused by the movement of the ship. Finally, the stress sensor is subjected to preliminary performance tests, including static and dynamic response tests, to verify the stability and accuracy of the stress sensor under actual working conditions. When the performance test passes, the pressure sensor can be put into formal use.

[0024] Regarding the acquisition of environmental data, the real-time environmental data of the port where the ship is located can be obtained through ship positioning. Alternatively, the environmental data of the mooring cable can be obtained through temperature and humidity sensors, wind direction monitors and other equipment installed on the ship. Environmental data may include: wind speed and direction (affecting the wind pressure on the cable), air pressure (affecting the atmospheric environment), humidity (affecting the performance of the cable material), seawater temperature (affecting the seawater density and the physical properties of the cable), wave height and period (directly affecting the dynamic force of the cable), etc.

[0025] In addition, the updating frequency of environmental data can be reasonably set according to the monitoring accuracy requirements of the mooring line and the data processing capacity of the mooring line health device. For example, the first environmental data of the mooring line can be obtained at a first preset time interval; the first environmental data can include: one of wind speed, wind direction and wave height; then the second environmental data of the mooring line can be obtained at a second preset time interval; the second environmental data can include: at least one of air pressure, humidity and seawater temperature; the second preset time interval is greater than the first preset time interval.

[0026] Reference Figure 2 , shows a data acquisition flow chart provided by the present invention. Image data is collected through image acquisition equipment, stress data is collected through stress sensors, and a connection is established with a meteorological service provider through an API to automatically request and receive real-time meteorological environment data of the port where the ship is located.

[0027] S102, extracting features from the image data, stress data and environmental data respectively to obtain image features, stress features and environmental features.

[0028] S103, performing feature fusion on the image features, stress features and environmental features to obtain fused features.

[0029] Based on the attention mechanism, image features, stress features and environmental features can be fused to obtain fused features.

[0030] S104, inputting the fused features into a preset mooring cable state classifier to obtain a mooring cable state classification result.

[0031] A classifier is a model that can classify data according to input data features. The input of the mooring rope state classifier can be the above-mentioned fusion features, and the output can be the health state of the mooring rope and the probability of the health state. The health state can include healthy, slightly worn, severely worn, dangerous, etc.

[0032] This embodiment monitors the status of mooring cables by fusing image data, stress data and environmental data, and can accurately capture safety hazards caused by non-stress factors such as surface damage and wear of mooring cables, monitor mooring cables in an all-round manner, and improve monitoring accuracy.

[0033] Reference Figure 3 , showing a mooring cable monitoring framework diagram provided by the present invention. After acquiring image data, stress data and environmental data, the three data are subjected to feature extraction and fusion to obtain fusion features, and the health status of the mooring cable is identified based on the fusion features.

[0034] In one embodiment, after the image data, stress data and environmental data are acquired, the three types of data may be preprocessed respectively, and after the preprocessing is completed, feature extraction may be performed on the three types of data.

[0035] Specifically, image data preprocessing includes: first, denoising the collected image, for example, using Gaussian filtering and other methods. Then, the image is uniformly adjusted to a specific size, such as 256×256 pixels, and normalized so that the image pixel values ​​are distributed in a suitable range to reduce differences caused by factors such as lighting.

[0036] Stress data preprocessing includes: firstly, denoising the stress data, such as using a low-pass filter to remove high-frequency noise. Then, standardizing the stress data to make it conform to the standard normal distribution for subsequent analysis.

[0037] Environmental data preprocessing includes: cleaning environmental data and removing outliers.

[0038] In one embodiment, the step of extracting features from the image data to obtain image features may include: performing mooring cable detection on the image data using a trained YOLO model to obtain a mooring cable image region; and performing feature extraction on the mooring cable image region using a trained VGG model to obtain image features.

[0039] The YOLO model can be a YOLO v8 model. As an efficient target detection model, the YOLO v8 model can comprehensively analyze the image and quickly and accurately locate the position of the mooring cable in the image. The VGG model can be a VGG16 model. The VGG16 model consists of multiple convolutional layers, pooling layers, and fully connected layers. In the convolution layer, the convolution operation is performed by sliding the convolution kernel on the image area to extract features at different levels layer by layer, from simple low-level features such as edges and textures to more abstract mooring cable structural features (such as diameter, weaving pattern) and damage features (such as wear and cracks). Then, after the downsampling operation of the pooling layer, the feature scale is compressed, and the main features are retained. Finally, the previously extracted features are comprehensively processed in the fully connected layer to obtain a feature vector of a fixed length, which is the image feature vector. (Dimensions are ).

[0040] In one embodiment, the step of extracting features from stress data to obtain stress features may include: performing a fast Fourier transform on the stress data to obtain first stress data; performing a wavelet transform on the stress data to obtain second stress data at different time scales; and generating stress features based on the first stress data and the second stress data.

[0041] The original stress data will be affected by the environment and generate noise during the acquisition process, so denoising can be performed first. The commonly used denoising method is to use a low-pass filter, which can allow low-frequency signals to pass through and effectively suppress high-frequency noise, making the stress data smoother and more accurate. After denoising, the stress data can be subjected to spectral analysis, in which the fast Fourier transform (FFT) is the key operation. FFT converts the denoised stress data from the time domain to the frequency domain to obtain the first stress data. By performing frequency domain analysis on the first stress data, the frequency components in the first stress data can be clearly identified. For example, the frequency peaks in the first stress data can be found. These peaks are often related to the vibration frequency or external excitation frequency of the cable under specific working conditions. In addition, wavelet transform can also be performed. By selecting appropriate wavelet basis functions, the stress data can be decomposed into components of different frequencies and time resolutions, namely, the second stress data. Information reflecting the change characteristics of the second stress data can be extracted from the second stress data, such as the fluctuation of stress at different time scales (the maximum value, minimum value and mean value of stress data at different time scales, etc.) and the stress mutation point, etc. The above-identified frequency components and stress fluctuations at different time scales are sorted out to obtain the stress characteristic vector (Dimensions are ) In one embodiment, an LSTM neural network may also be used to extract features from the denoised stress data to generate stress features.

[0042] In one embodiment, the step of extracting features from environmental data to obtain environmental features may include: first, because the sensor may be faulty or interfered, there may be outliers and missing values ​​in the collected environmental data, so data cleaning operations are required. For example, an outlier detection algorithm based on statistical methods can be used to detect and filter out outliers and fill in missing values ​​through linear interpolation to ensure the integrity and reliability of environmental data. Then, key variables are screened out from the cleaned environmental data based on domain knowledge, such as wind speed and wave height that have a direct impact on the force of mooring cables, and temperature and humidity that affect the material properties of cables (such as corrosion rate). For these key variables, feature engineering operations need to be performed based on physical principles and practical experience, such as calculating the product of wind speed and wave height to represent the combined force of wind and waves on cables, or calculating the ratio of temperature and humidity to reflect the comprehensive impact of the environment on cable corrosion. Finally, the product of wind speed and wave height and the ratio of temperature and humidity obtained after the above processing are used as environmental feature vectors. (Dimensions are ).

[0043] In one embodiment, after obtaining the image features, stress features and environmental features, these three features can be spliced ​​to obtain a spliced ​​feature , the dimension of the concatenated features is . You can then stitch the features Input to the multi-layer perceptron MLP, MLP has two layers, the first layer activation function is ReLU, and the second layer output dimension is 3 (corresponding to the three modalities of image, stress, and environment). Assume that the second layer of MLP calculates , and then the attention weights of image features, stress features and environmental features can be calculated through the softmax function: Attention weights of image features: .

[0044] Attention weights for stress features: .

[0045] Attention weights for environmental features: .

[0046] According to the attention weights of image features, stress features and environmental features, the image features, stress features and environmental features are weighted fused to obtain the fusion feature .

[0047] In one embodiment, the state classification result of the mooring cable output by the classifier includes the health state of the mooring cable and the probability corresponding to the health state, wherein the health state includes healthy, relatively healthy, slightly worn, severely worn, dangerous, etc., and different health states correspond to different probability thresholds; the mooring cable monitoring method further includes: determining a target probability threshold corresponding to the health state according to the health state output by the mooring cable state classifier; when the probability output by the mooring cable state classifier is greater than the target probability threshold, generating an alarm message.

[0048] For example, when the mooring rope status is classified as severely worn by the classifier, the probability of severe wear exceeds 50%, the alarm mechanism is triggered and an alarm message is generated; when the mooring rope status is classified as dangerous by the classifier, the probability of danger exceeds 30%, the alarm mechanism is triggered and an alarm message is generated.

[0049] The probability thresholds corresponding to each health state can be set based on the material properties of the mooring line and combined with the opinions of domain experts.

[0050] In one embodiment, in addition to determining whether to trigger the alarm mechanism based on the output result of the classifier, it is also possible to determine whether to trigger the alarm mechanism based on the result of the auxiliary judgment. Specifically, it is also possible to: determine the wear of the mooring rope based on the image features; determine whether the attention weight of the image features exceeds the image attention weight threshold; when the attention weight of the image features exceeds the image attention weight threshold, determine the stress alarm threshold based on the wear; when the stress data exceeds the stress alarm threshold, generate an alarm message.

[0051] For example, if the attention weight of the image feature and the attention weights of stress features are higher (e.g., >0.3 and >0.3), and the classifier outputs a high probability of cable wear or abnormality, focusing on the fiber breakage of the mooring cable identified based on image data and whether the stress data is close to or exceeds the rated bearing capacity. For example, if the image analysis shows that more than 3% of the cable has fiber breakage and the stress data exceeds 70% of the rated bearing capacity, the early warning mechanism is triggered.

[0052] In one embodiment, the auxiliary judgment also includes: determining whether the attention weight of the environmental data exceeds the environmental attention weight threshold; when the attention weight of the environmental data exceeds the environmental attention weight threshold, determining the stress alarm threshold based on the environmental data; when the stress data exceeds the stress alarm threshold, generating an alarm message.

[0053] For example, when the attention weight of environmental features is high (e.g. >0.3), use environmental data to assist in judging the health status of mooring cables. Through historical data analysis, a linear or nonlinear regression model is established between the stress of mooring cables and environmental factors (such as wind speed and wave height). For example, if the wind speed exceeds 15 m / s, the regression model predicts that the cable stress is expected to increase by 20%. If the stress data exceeds 80% of the rated bearing capacity at this time, the early warning mechanism is triggered.

[0054] In one embodiment, the auxiliary judgment also includes: setting different stress level thresholds according to the material characteristics and historical stress data of the mooring cable, combined with the opinions of field experts. For example, the stress level threshold for health can be set at 80% of the rated bearing capacity, and the stress level threshold for danger can be set at 40% of the rated bearing capacity.

[0055] Reference Figure 4 , shows a data fusion flow chart provided by the present invention. First, a data fusion rule is defined, which is an auxiliary judgment rule. Then, multi-modal data is fused or associated based on the fusion rule.

[0056] In one embodiment, the training process of the classifier, the YOLO model, and the VGG model may include the following process: ① Dataset division: Randomly divide the collected labeled data (including images, stress and environmental data in normal and abnormal states) to ensure consistent data distribution. Usually the training set accounts for 70% - 80%, and the validation set and test set each account for 10% - 15%.

[0057] ② Initialization parameters: Use weight initialization based on a pre-trained model. When using weight initialization based on a pre-trained model, for example, when processing image data, use a model (VGG16) pre-trained on a large image dataset (such as ImageNet), and load the weights of the pre-trained model into the current model. If the current model structure is not exactly the same as the pre-trained model, for example, the current model may only need some layers in the pre-trained model (such as convolutional layers), you can choose to intercept the corresponding layer weights for initialization. For newly added layers (such as custom fully connected layers), the weights of these layers can be randomly initialized.

[0058] ③ Select appropriate loss functions (such as cross entropy loss) and optimization algorithms (such as Adam), set the initial learning rate and decay strategy. The cross entropy loss function has a natural advantage in dealing with classification problems. It can well measure the difference between the probability distribution predicted by the model and the true category. When the model predicts the correct category with a higher probability, the cross entropy loss value is smaller, and vice versa. This helps guide the model to continuously adjust parameters during training and improve the classification accuracy of different health states. The Adam optimization algorithm combines the advantages of the momentum method and RMSProp and can adaptively adjust the learning rate. When dealing with complex neural network models, the update frequency of different parameters may be different. Adam can automatically adjust the learning rate of each parameter according to the gradient history information of the parameter, making the training process more stable and efficient. In the model training of cable health status assessment, since the model may be more complex (involving multimodal data fusion, multi-layer neural network, etc.), the Adam optimization algorithm can avoid falling into the local optimal solution to a certain extent, accelerate the convergence speed of the model, and thus find better model parameters faster.

[0059] ④ Hyperparameter tuning: Through Bayesian optimization and other techniques, hyperparameters such as learning rate, batch size, and regularization coefficient are adjusted to find the best configuration. Bayesian optimization is a model-based hyperparameter tuning method that uses a probability model to guide the search for hyperparameters. The advantage of Bayesian optimization is that it can find a better hyperparameter combination with fewer evaluations, which is especially suitable for situations where the cost of hyperparameter evaluation is high (such as long training time).

[0060] ⑤Model evaluation: Use the test set to evaluate the model performance, pay attention to indicators such as accuracy, recall, F1 score, etc., to ensure the generalization ability and stability of the model.

[0061] The trained model can be deployed to the production environment to monitor the health of the mooring line in real time. The model performance can be checked regularly and retrained or optimized as needed.

[0062] In one embodiment, the alarm information can be released to all crew members through the ship's internal radio, intercom or dedicated alarm system. At the same time, satellite communications or the Internet are used to send emails or text messages to the shore-based operation center, maintenance team and management to ensure that relevant parties on shore receive the information simultaneously. Maritime safety applications can also be developed or integrated to push real-time warnings to designated personnel through mobile phones or tablets to ensure the rapid dissemination of information. The alarm information can clearly identify the severity of the alarm for rapid identification and response; provide specific location information of the mooring rope to facilitate on-site inspection and maintenance; and list preliminary response strategies, such as slowing down, changing the route, and starting spare ropes.

[0063] In one embodiment, once a danger warning is received, an alarm command can be automatically issued. In extreme risk situations, the port terminal is automatically contacted to request immediate assistance. Once the alarm information is triggered, the emergency response team is immediately activated to organize professionals to conduct a comprehensive inspection and assessment of the cable. According to the pre-established emergency manual, the corresponding cable replacement, reinforcement or temporary repair work is performed. A command center is established to summarize on-site information and maintain close communication with the decision-making level to ensure that all actions are approved and supported by the top management.

[0064] Reference Figure 5 , showing another mooring cable monitoring framework diagram provided by the present invention. Collect image data, stress data and environmental data. Preprocess the three types of data respectively. After the preprocessing is completed, perform feature extraction and feature fusion on the three types of data. Identify the health status of the mooring cable based on the fused features. Determine whether to trigger the alarm mechanism based on the auxiliary judgment rules and the mooring cable status recognition results.

[0065] In summary, the present invention has the following beneficial effects: ① Significantly improve monitoring efficiency and accuracy. The traditional manual inspection method is limited by manpower and time, and cannot achieve continuous monitoring. The present invention can collect mooring cable status information 24 hours a day by integrating various sensors and video monitoring, ensuring the real-time and integrity of the data. Especially in complex environmental conditions, such as at night, in bad weather or when the cable is under high stress, multimodal data fusion technology can integrate information from all aspects and provide more comprehensive analysis results, avoiding the misjudgment that may be caused by a single data source, and greatly improving the accuracy and reliability of monitoring.

[0066] ② Enhanced the intelligence level of mooring cable status assessment. By applying deep learning and machine learning algorithms, the present invention can automatically identify cable surface damage, wear degree, and stress change trend, and even predict potential failure risks. This intelligent assessment method not only saves labor costs, but also can provide early warning, providing port managers with sufficient time to take preventive measures, effectively avoiding safety accidents caused by cable failures, and greatly improving the safety of port operations.

[0067] ③ It promotes the refinement and predictability of maintenance work. Traditional maintenance strategies are often based on experience or regular inspections, while the present invention can grasp the health of cables in real time through continuous monitoring and intelligent analysis, helping port managers to formulate more scientific and reasonable maintenance plans. For example, when it is detected that the stress of the mooring cable exceeds the safety threshold or the surface damage reaches a certain level, it can automatically trigger a maintenance reminder to guide the staff to repair or replace it in time, avoiding excessive or insufficient maintenance, thereby reducing maintenance costs and extending the service life of the mooring cable.

[0068] ④ Provides strong support for the digital transformation of port operations. By integrating various sensor data, a digital twin model of the mooring rope status is constructed, providing a rich and accurate data foundation for the information management of the port. These data can not only be used for real-time monitoring, but also for historical data analysis, failure mode research and future trend prediction, providing valuable decision-making basis for port operation management, and helping ports develop in the direction of intelligence and efficiency.

[0069] Reference Figure 6 , shows a schematic structural diagram of an embodiment of a mooring line monitoring device provided by the present invention, the device 60 comprises: The data acquisition module 601 is used to acquire the image data, stress data and environmental data of the mooring cable; A feature extraction module 602 is used to extract features from the image data, stress data and environmental data to obtain image features, stress features and environmental features; A feature fusion module 603 is used to fuse image features, stress features and environmental features to obtain fused features; The state determination module 604 is used to input the fused features into a preset mooring line state classifier to obtain a mooring line state classification result.

[0070] It should be noted that the implementation principle or implementation process of the above modules can refer to the above-mentioned embodiment of the mooring line monitoring method, and will not be described one by one here.

[0071] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0072] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A mooring line monitoring method, characterized in that: include: Acquire image data, stress data and environmental data of mooring cables; Extracting features from the image data, the stress data and the environment data respectively to obtain image features, stress features and environment features; Performing feature fusion on the image features, stress features and environmental features to obtain fusion features; The fused features are input into a preset mooring cable state classifier to obtain a state classification result of the mooring cable.

2. The mooring line monitoring method according to claim 1, characterized in that: The step of extracting features from the image data to obtain image features includes: Perform mooring cable detection on the image data using a trained YOLO model to obtain a mooring cable image region; The trained VGG model is used to extract features of the mooring cable image region to obtain image features.

3. The mooring line monitoring method according to claim 1, characterized in that: The extracting features of the stress data to obtain stress features includes: Performing a fast Fourier transform on the stress data to obtain first stress data; Performing wavelet transformation on the stress data to obtain second stress data at different time scales; A stress signature is generated based on the first stress data and the second stress data.

4. The mooring line monitoring method according to claim 1, characterized in that: The step of obtaining the environmental data of the mooring rope includes: Acquire first environmental data of the mooring rope at a first preset time interval; wherein the first environmental data includes: one of wind speed, wind direction and wave height; Acquire second environmental data of the mooring rope at a second preset time interval; wherein the second environmental data includes at least one of air pressure, humidity and seawater temperature; and the second preset time interval is greater than the first preset time interval.

5. The mooring line monitoring method according to claim 1, characterized in that: The step of fusing the image features, stress features and environmental features to obtain fused features includes: Inputting the image features, stress features and environmental features into a multi-layer perceptron to obtain attention weights of the image features, stress features and environmental features; According to the attention weights of the image features, stress features and environmental features, the image features, stress features and environmental features are weightedly fused to obtain fused features.

6. The mooring line monitoring method according to claim 5, characterized in that: The method further comprises: determining the wear condition of the mooring line according to the image features; Determining whether the attention weight of the image feature exceeds an image attention weight threshold; When the attention weight of the image feature exceeds the image attention weight threshold, determining a stress alarm threshold according to the wear condition; When the stress data exceeds the stress alarm threshold, an alarm message is generated.

7. The mooring line monitoring method according to claim 5, characterized in that: The method further comprises: Determine whether the attention weight of the environmental data exceeds an environmental attention weight threshold; When the attention weight of the environmental data exceeds the environmental attention weight threshold, determining a stress alarm threshold according to the environmental data; When the stress data exceeds the stress alarm threshold, an alarm message is generated.

8. The mooring line monitoring method according to claim 1, characterized in that: The status classification result includes the health status and the probability corresponding to the health status; Different health states correspond to different probability thresholds; the method further includes: Determining a target probability threshold corresponding to the health state according to the health state output by the mooring line state classifier; When the probability output by the mooring line state classifier is greater than the target probability threshold, an alarm message is generated.

9. The mooring line monitoring method according to claim 7, characterized in that: The method further comprises: According to the material properties of the mooring lines and the opinions of domain experts, the probability thresholds corresponding to each health state are set.

10. A mooring line monitoring device, characterized in that: include: A data acquisition module, used to acquire image data, stress data and environmental data of the mooring cable; A feature extraction module, used to extract features from the image data, the stress data and the environmental data respectively, to obtain image features, stress features and environmental features; A feature fusion module, used for fusing the image features, stress features and environmental features to obtain fused features; The state determination module is used to input the fusion feature into a preset mooring cable state classifier to obtain a state classification result of the mooring cable.

Images