A floating wind power platform offshore deviation warning method and system

Abstract

The present invention discloses a floating wind power platform offshore offset warning method and system. The method calculates the inclination angle and rotation angle of the floating wind power platform to be measured by combining an image recognition model with a Hough transform technology, thereby improving the accuracy of calculating the inclination angle and rotation angle of the floating wind power platform to be measured, thereby ensuring high accuracy of offset calculation, providing accurate data support for subsequent determination of the offset warning state, and using the image recognition model and the Hough transform technology in combination to reduce the influence of interference from complex environments on the calculated offset, thereby solving the problem that the prior art method of using sensors to directly measure the platform offset is easily interfered by complex environments, resulting in inaccurate data or failure to transmit in real time, thereby causing the floating wind power platform to be unable to issue an offset warning in time.

Description

Technical Field

[0001] The present invention relates to the field of offshore deviation warning of a floating wind power platform, and in particular to an offshore deviation warning method and system for a floating wind power platform. Background Art

[0002] As the global demand for renewable energy grows, offshore wind power, as an important clean energy, is gradually becoming a key component of the energy structure transformation. As the forefront of offshore wind power technology, floating wind power platforms can break through the dependence of traditional fixed wind power platforms on shallow waters and realize wind power development in deep waters. However, the complex marine environment in deep waters, such as strong winds, huge waves and turbulence, poses severe challenges to the stability of floating wind power platforms. The platform's offset and attitude control have become key factors in ensuring safe operation.

[0003] Existing technologies usually use sensors to directly measure the offset of offshore wind power platforms, such as GPS positioning systems and inertial measurement units (IMUs). Although these methods can provide accurate positioning and attitude information, the sensors are easily interfered by complex environments, resulting in inaccurate data or failure to transmit in real time, which in turn causes the floating wind power platform to be unable to issue offset warnings in a timely manner. Summary of the invention

[0004] The present invention provides a floating wind power platform offshore deviation early warning method and system, which can solve the problem that the prior art method of using sensors to directly measure the platform deviation is easily disturbed by complex environments, resulting in inaccurate data or failure to transmit in real time, thereby causing the floating wind power platform to be unable to issue a deviation early warning in a timely manner.

[0005] In order to solve the above technical problems, an embodiment of the present invention provides an offshore deviation warning method for a floating wind power platform, comprising:

[0006] Obtain the real-time offshore image, real-time offshore position and original installation position of the floating wind power platform to be tested;

[0007] Inputting the real-time offshore image of the floating wind power platform to be tested into the trained platform image recognition model, so that the platform image recognition model performs feature extraction on the real-time offshore image of the floating wind power platform to be tested, and obtains a real-time offshore platform feature map of the floating wind power platform to be tested; wherein the platform image recognition model includes a convolution layer; the convolution layer is used to perform feature extraction on the real-time offshore image of the floating wind power platform to be tested;

[0008] According to the real-time offshore platform characteristic map, the tilt angle and the rotation angle of the floating wind power platform to be tested are calculated by Hough transform;

[0009] Calculate the offset of the floating platform to be tested according to the tilt angle, rotation angle, real-time offshore position and original installation position of the floating wind power platform to be tested;

[0010] The relaxation state of the displacement warning of the floating platform to be tested is determined by comparing the displacement of the floating platform to be tested with the preset displacement threshold.

[0011] Furthermore, the model training of the platform image recognition model includes:

[0012] Collect actual pictures of several offshore platforms of floating wind power platforms under different conditions;

[0013] The data of each actual offshore platform image is preprocessed through the input layer to obtain several processed offshore platform images;

[0014] Through the convolution layer, activation layer and pooling layer, feature extraction, feature activation and down sampling are performed on each processed offshore platform image to obtain several initial offshore platform feature images;

[0015] The features of each initial offshore platform feature map are integrated through the fully connected layer to obtain several integrated offshore platform feature maps, and the features of each integrated offshore platform feature map are enhanced through the self-attention mechanism layer to obtain several final offshore platform feature maps;

[0016] According to the final offshore platform feature map and its corresponding actual offshore platform map, the loss value is calculated using the loss function, and the parameters of the platform image recognition model to be trained are optimized according to the loss value until the loss value converges to obtain a trained platform image recognition model.

[0017] Furthermore, before calculating the tilt angle and rotation angle of the floating wind power platform to be measured by Hough transform according to the real-time offshore platform characteristic map, it also includes:

[0018] Performing grayscale conversion on the real-time offshore platform characteristic image of the floating wind power platform to be tested to obtain a real-time offshore platform characteristic grayscale image;

[0019] Performing a denoising operation on the real-time offshore platform feature grayscale image to obtain a denoised real-time offshore platform feature grayscale image;

[0020] The edge detection algorithm is used to perform edge detection on the denoised real-time offshore platform feature grayscale image to obtain a real-time offshore platform feature grayscale image with edge points.

[0021] Furthermore, the calculating the tilt angle and rotation angle of the floating wind power platform to be measured by Hough transform according to the real-time offshore platform characteristic map includes:

[0022] Map the edge points in the real-time offshore platform feature grayscale image with edge points to the parameter space of Hough transform to obtain several mapping straight lines;

[0023] The Hough transform accumulator is used to vote and accumulate the parameter space position corresponding to each mapping line to obtain mapping lines of several peak points.

[0024] According to the mapping lines of several peak points, a mapping line corresponding to the peak point with the largest number of cumulative votes is selected, a mapping angle of the mapping line is determined, and the mapping angle of the mapping line is used as the inclination angle of the floating wind power platform to be tested;

[0025] According to the mapping straight lines of several peak points, the mapping angles of the mapping straight lines corresponding to the remaining peak points are determined to obtain several mapping angles, and the average mapping angle value is calculated according to the several mapping angles, and the average mapping angle value is used as the rotation angle of the floating wind power platform to be tested.

[0026] Furthermore, the mathematical expression of the mapping line is:

[0027] ;

[0028] Among them, the Represents the shortest distance from the mapping line to the origin of the image coordinate system; and Respectively represent the horizontal and vertical coordinates of the edge point in the rectangular coordinate system; Represents the mapping angle between the mapping line and the positive direction of the x-axis in the rectangular coordinate system.

[0029] Furthermore, the calculation of the offset of the floating platform to be measured according to the tilt angle, rotation angle, real-time offshore position and original installation position of the floating wind power platform to be measured includes:

[0030] Calculate the deviation distance of the floating wind power platform to be tested according to the real-time offshore position and the original installation position of the floating wind power platform to be tested;

[0031] According to the tilt angle, rotation angle and deviation distance of the floating wind power platform to be tested, the offset of the floating wind power platform to be tested in the east-west direction, north-south direction and vertical direction is calculated respectively by using trigonometric functions.

[0032] Furthermore, the method of calculating the displacement of the floating wind power platform to be measured in the east-west direction, the north-south direction and the vertical direction respectively by using trigonometric functions according to the tilt angle, the rotation angle and the deviation distance of the floating wind power platform to be measured includes:

[0033] The formula for calculating the east-west offset of the floating wind power platform to be tested is:

[0034] ;

[0035] The formula for calculating the north-south offset of the floating wind power platform to be tested is:

[0036] ;

[0037] The formula for calculating the vertical offset of the floating wind power platform to be tested is:

[0038] ;

[0039] Among them, the represents the displacement of the floating wind power platform to be tested in the east-west direction; represents the north-south offset of the floating wind power platform to be tested; Indicates the vertical displacement of the floating wind power platform to be tested; Indicates the deviation distance of the floating wind power platform to be tested; represents the tilt angle of the floating wind power platform to be tested; Indicates the rotation angle of the floating wind turbine platform to be tested.

[0040] Furthermore, the step of comparing the displacement of the floating platform to be tested with a preset displacement threshold to determine the relaxation state of the displacement warning of the floating platform to be tested includes:

[0041] When the offset of the floating wind power platform to be tested in the east-west direction is less than the first offset threshold, and when the offset of the floating wind power platform to be tested in the north-south direction is less than the second offset threshold, and when the offset of the floating wind power platform to be tested in the vertical direction is less than the third offset threshold, the floating platform to be tested does not send an offset warning;

[0042] Otherwise, the floating platform to be tested sends a deviation warning.

[0043] Based on the above method embodiment, the present invention provides a corresponding system embodiment;

[0044] An embodiment of the present invention provides an offshore displacement warning system for a floating wind power platform, comprising: a data acquisition module, a characteristic map generation module, a first calculation module, a second calculation module and a displacement warning module;

[0045] The data acquisition module is used to obtain the real-time offshore image, real-time offshore position and original installation position of the floating wind power platform to be tested;

[0046] The feature map generation module is used to input the real-time offshore image of the floating wind power platform to be tested into the trained platform image recognition model, so that the platform image recognition model performs feature extraction on the real-time offshore image of the floating wind power platform to be tested, and obtains the real-time offshore platform feature map of the floating wind power platform to be tested; wherein the platform image recognition model includes a convolution layer; the convolution layer is used to perform feature extraction on the real-time offshore image of the floating wind power platform to be tested;

[0047] The first calculation module is used to calculate the tilt angle and rotation angle of the floating wind power platform to be tested through Hough transformation according to the real-time offshore platform characteristic map;

[0048] The second calculation module is used to calculate the offset of the floating platform to be tested according to the inclination angle, rotation angle, real-time offshore position and original installation position of the floating wind power platform to be tested;

[0049] The deviation warning module is used to compare and judge the deviation of the floating platform to be tested with a preset deviation threshold value to determine the relaxation state of the deviation warning of the floating platform to be tested.

[0050] Furthermore, the floating wind power platform offshore deviation warning system further includes: a model training module;

[0051] The model training module is used to collect several actual offshore platform images of floating wind power platforms under different conditions, perform data preprocessing on each actual offshore platform image through the input layer to obtain several processed offshore platform images, perform feature extraction, feature activation and downsampling on each processed offshore platform image through the convolution layer, the activation layer and the pooling layer to obtain several initial offshore platform feature images, perform feature integration on each initial offshore platform feature image through the fully connected layer to obtain several integrated offshore platform feature images, and perform feature enhancement on each integrated offshore platform feature image through the self-attention mechanism layer to obtain several final offshore platform feature images. According to the final offshore platform feature image and its corresponding actual offshore platform image, a loss value is calculated using a loss function, and parameters of the platform image recognition model to be trained are optimized according to the loss value until the loss value converges to obtain a trained platform image recognition model.

[0052] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: by using a trained platform image recognition model to extract features from the offshore real-time image of the floating wind power platform to be tested, a real-time offshore platform feature map of the floating wind power platform to be tested is generated, and the inclination angle and rotation angle of the floating wind power platform to be tested are calculated according to the real-time offshore platform feature map by using the Hough transform technology, and then the offset of the floating wind power platform to be tested is calculated, and finally the offset warning state of the floating wind power platform to be tested is determined according to the offset, that is, the present invention improves the accuracy of the calculation of the inclination angle and rotation angle of the floating wind power platform to be tested by combining the image recognition model with the Hough transform technology, thereby ensuring the high accuracy of the offset calculation, providing accurate data support for the subsequent determination of the offset warning state, and using the combination of the image recognition model and the Hough transform technology to reduce the influence of the interference of the complex environment on the calculated offset, thereby solving the problem that the method of directly measuring the platform offset using a sensor in the prior art is easily interfered by the complex environment, resulting in inaccurate data or failure to transmit in real time, and then causing the floating wind power platform to fail to issue an offset warning in time. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 : A flowchart of the steps of a method for early warning of offshore deviation of a floating wind power platform provided by an embodiment of the present invention;

[0054] Figure 2 : A flow chart of the model training steps of the platform image recognition model provided by an embodiment of the present invention;

[0055] Figure 3 : A model structure diagram of a platform image recognition model provided by an embodiment of the present invention;

[0056] Figure 4 : A system structure diagram of an offshore deviation warning system for a floating wind power platform provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0057] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0058] In the description of the present invention, it should be understood that the terms "first", "second" and "third" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features.

[0059] Embodiment 1:

[0060] Reference Figure 1 , is a flowchart of a method for early warning of offshore deviation of a floating wind power platform provided by an embodiment of the present invention, the method comprising steps S1 to S5, each step is specifically as follows:

[0061] Step S1: obtaining a real-time offshore image, a real-time offshore position and an original installation position of the floating wind power platform to be tested;

[0062] In this embodiment, a high-definition camera installed on the wind power platform or a drone with a high-definition camera is used to capture real-time offshore images of the floating wind power platform to be tested, and a GPS positioning system is used to obtain the real-time offshore position and original installation position of the floating wind power platform to be tested.

[0063] Step S2: inputting the real-time offshore image of the floating wind power platform to be tested into the trained platform image recognition model, so that the platform image recognition model performs feature extraction on the real-time offshore image of the floating wind power platform to be tested, and obtains a real-time offshore platform feature map of the floating wind power platform to be tested; wherein the platform image recognition model includes a convolution layer; the convolution layer is used to perform feature extraction on the real-time offshore image of the floating wind power platform to be tested;

[0064] In this embodiment, refer to Figure 2 :A flow chart of the model training steps of the platform image recognition model provided by the embodiment of the present invention, refer to Figure 3 : A model structure diagram of a platform image recognition model provided in an embodiment of the present invention, wherein the platform image recognition model includes an input layer, a convolution layer, an activation layer, a pooling layer, a fully connected layer, a self-attention mechanism layer, and an output layer; the model training steps of the platform image recognition model are as follows:

[0065] Step A1: Collecting actual images of several offshore platforms of floating wind power platforms under different conditions;

[0066] In this embodiment, a high-definition camera installed on a wind power platform or a drone with a high-definition camera is used to collect actual images of several offshore platforms of a floating wind power platform under different conditions;

[0067] Step A2: preprocessing the data of each actual offshore platform image through the input layer to obtain a number of processed offshore platform images;

[0068] In this embodiment, the data preprocessing includes but is not limited to image resizing, noise removal and normalization;

[0069] Step A3: Perform feature extraction, feature activation and downsampling on each processed offshore platform image through the convolution layer, activation layer and pooling layer to obtain several initial offshore platform feature images;

[0070] In this embodiment, local features are extracted from each processed offshore platform image through the convolution kernel of the convolution layer, and the local feature map extracted by the output of the convolution layer is nonlinearly activated through the activation function of the activation layer. Finally, the local feature map after the activation of the activation layer output is downsampled through the pooling layer to obtain several initial local feature maps of the offshore platform; wherein the activation function includes but is not limited to the ReLU function, the Sigmoid function or the Tanh function;

[0071] Step A4: integrating the features of each initial offshore platform feature map through the fully connected layer to obtain several integrated offshore platform feature maps, and enhancing the features of each integrated offshore platform feature map through the self-attention mechanism layer to obtain several final offshore platform feature maps;

[0072] In this embodiment, each initial offshore platform local feature map is feature integrated through a fully connected layer to obtain each integrated offshore platform feature map, and each integrated offshore platform feature map is feature enhanced through a self-attention mechanism layer to obtain several final offshore platform feature maps; wherein the fully connected layer includes an activation function; the activation function includes but is not limited to a ReLU function, a Sigmoid function or a Tanh function;

[0073] Step A5: According to the final offshore platform feature map and its corresponding offshore platform actual map, the loss value is calculated using the loss function, and the parameters of the platform image recognition model to be trained are optimized according to the loss value until the loss value converges to obtain a trained platform image recognition model;

[0074] In this embodiment, the loss function includes but is not limited to the mean square error loss function and the mean absolute error loss; the parameter optimization method includes but is not limited to the gradient descent algorithm, momentum optimizer, Adagrad, RMSprop and Adam; the gradient descent algorithm includes but is not limited to the batch gradient descent algorithm, the stochastic gradient descent algorithm and the mini-batch gradient descent algorithm.

[0075] Step S3: Calculate the tilt angle and rotation angle of the floating wind power platform to be tested by Hough transform according to the real-time offshore platform characteristic map;

[0076] In this embodiment, before calculating the tilt angle and rotation angle of the floating wind power platform to be measured by Hough transform according to the real-time offshore platform characteristic map, the method further includes:

[0077] Performing grayscale conversion on the real-time offshore platform characteristic image of the floating wind power platform to be tested to obtain a real-time offshore platform characteristic grayscale image;

[0078] Performing a denoising operation on the real-time offshore platform feature grayscale image to obtain a denoised real-time offshore platform feature grayscale image;

[0079] In this embodiment, the denoising operation method includes but is not limited to mean filtering, median filtering, Gaussian filtering, bilateral filtering and non-local mean filtering;

[0080] The edge detection algorithm is used to perform edge detection on the denoised real-time offshore platform feature grayscale image to obtain a real-time offshore platform feature grayscale image with edge points;

[0081] In this embodiment, the edge detection algorithm includes but is not limited to a Canny edge detection algorithm, a Roberts operator algorithm, a Prewitt operator algorithm, a Sobel operator algorithm, a Scharr operator algorithm, and a Laplacian operator algorithm.

[0082] In this embodiment, the method of calculating the tilt angle and rotation angle of the floating wind power platform to be measured by Hough transform according to the real-time offshore platform characteristic map includes:

[0083] Map the edge points in the real-time offshore platform feature grayscale image with edge points to the parameter space of Hough transform to obtain several mapping straight lines;

[0084] The Hough transform accumulator is used to vote and accumulate the parameter space position corresponding to each mapping line to obtain mapping lines of several peak points.

[0085] According to the mapping lines of several peak points, a mapping line corresponding to the peak point with the largest number of cumulative votes is selected, a mapping angle of the mapping line is determined, and the mapping angle of the mapping line is used as the inclination angle of the floating wind power platform to be tested;

[0086] According to the mapping straight lines of several peak points, the mapping angles of the mapping straight lines corresponding to the remaining peak points are determined to obtain several mapping angles, and the average mapping angle value is calculated according to the several mapping angles, and the average mapping angle value is used as the rotation angle of the floating wind power platform to be tested.

[0087] In this embodiment, the mathematical expression of the mapping straight line is:

[0088] ;

[0089] Among them, the represents the shortest distance from the mapping line to the origin of the image coordinate system; x and y represent the horizontal and vertical coordinates of the edge point in the rectangular coordinate system respectively; Represents the mapping angle between the mapping line and the positive direction of the x-axis in the rectangular coordinate system.

[0090] Step S4: calculating the offset of the floating platform to be tested according to the tilt angle, rotation angle, real-time offshore position and original installation position of the floating wind power platform to be tested;

[0091] In this embodiment, the deviation distance of the floating wind power platform to be measured is calculated according to the real-time offshore position and the original installation position of the floating wind power platform to be measured, and the offsets of the floating wind power platform to be measured in the east-west direction, the north-south direction and the vertical direction are calculated respectively by using trigonometric functions according to the inclination angle, the rotation angle and the deviation distance of the floating wind power platform to be measured; wherein the offset of the floating wind power platform to be measured in the east-west direction refers to the displacement of the floating wind power platform to be measured on the sea level along the east-west direction (usually corresponding to the longitude direction in the geographic coordinates) relative to the original installation position; the offset of the floating wind power platform to be measured in the north-south direction refers to the displacement of the floating wind power platform to be measured on the sea level along the north-south direction (usually corresponding to the latitude direction in the geographic coordinates) relative to the original installation position; the offset of the floating wind power platform to be measured in the vertical direction refers to the displacement of the floating wind power platform to be measured in the direction perpendicular to its sea level (i.e., the gravity direction) relative to the original installation position.

[0092] In this embodiment, by calculating the offsets of the floating wind power platform in various directions, the floating wind power platform can be comprehensively analyzed and judged to improve the accuracy of judging whether the floating wind power platform should initiate an offset warning.

[0093] In this embodiment, the offset of the floating wind power platform to be measured in the east-west direction, the north-south direction and the vertical direction is calculated by using trigonometric functions according to the tilt angle, the rotation angle and the deviation distance of the floating wind power platform to be measured, including:

[0094] The formula for calculating the east-west offset of the floating wind power platform to be tested is:

[0095] ;

[0096] The formula for calculating the north-south offset of the floating wind power platform to be tested is:

[0097] ;

[0098] The formula for calculating the vertical offset of the floating wind power platform to be tested is:

[0099] ;

[0100] Among them, the represents the displacement of the floating wind power platform to be tested in the east-west direction; represents the north-south offset of the floating wind power platform to be tested; Indicates the vertical displacement of the floating wind power platform to be tested; Indicates the deviation distance of the floating wind power platform to be tested; represents the tilt angle of the floating wind power platform to be tested; Indicates the rotation angle of the floating wind turbine platform to be tested.

[0101] Step S5: comparing and judging the displacement of the floating platform to be tested with a preset displacement threshold, and determining the relaxation state of the displacement warning of the floating platform to be tested.

[0102] In this embodiment, the calculated offsets of the floating platform to be tested in the east-west direction, the north-south direction and the vertical direction are compared with the preset offset thresholds. When the offset of the floating wind power platform to be tested in the east-west direction is less than the first offset threshold, when the offset of the floating wind power platform to be tested in the north-south direction is less than the second offset threshold, and when the offset of the floating wind power platform to be tested in the vertical direction is less than the third offset threshold, the floating platform to be tested does not send an offset warning; otherwise, the floating platform to be tested sends an offset warning.

[0103] In this embodiment, the accuracy of judging the deviation warning state of the floating platform to be measured is improved by comprehensively judging the calculated deviations of the floating platform to be measured in the east-west direction, the north-south direction and the vertical direction.

[0104] Embodiment 2:

[0105] Reference Figure 4 , is a system structure diagram of an offshore displacement warning system for a floating wind power platform provided by an embodiment of the present invention, the system comprising: a data acquisition module, a characteristic map generation module, a first calculation module, a second calculation module and a displacement warning module;

[0106] The data acquisition module is used to obtain the real-time offshore image, real-time offshore position and original installation position of the floating wind power platform to be tested;

[0107] The feature map generation module is used to input the real-time offshore image of the floating wind power platform to be tested into the trained platform image recognition model, so that the platform image recognition model performs feature extraction on the real-time offshore image of the floating wind power platform to be tested, and obtains the real-time offshore platform feature map of the floating wind power platform to be tested; wherein the platform image recognition model includes a convolution layer; the convolution layer is used to perform feature extraction on the real-time offshore image of the floating wind power platform to be tested;

[0108] The first calculation module is used to calculate the tilt angle and rotation angle of the floating wind power platform to be tested through Hough transformation according to the real-time offshore platform characteristic map;

[0109] The second calculation module is used to calculate the offset of the floating platform to be tested according to the inclination angle, rotation angle, real-time offshore position and original installation position of the floating wind power platform to be tested;

[0110] The deviation warning module is used to compare and judge the deviation of the floating platform to be tested with a preset deviation threshold value to determine the relaxation state of the deviation warning of the floating platform to be tested.

[0111] In this embodiment, the floating wind power platform offshore deviation warning system further includes: a model training module;

[0112] The model training module is used to collect several actual offshore platform images of floating wind power platforms under different conditions, perform data preprocessing on each actual offshore platform image through the input layer to obtain several processed offshore platform images, perform feature extraction, feature activation and downsampling on each processed offshore platform image through the convolution layer, the activation layer and the pooling layer to obtain several initial offshore platform feature images, perform feature integration on each initial offshore platform feature image through the fully connected layer to obtain several integrated offshore platform feature images, and perform feature enhancement on each integrated offshore platform feature image through the self-attention mechanism layer to obtain several final offshore platform feature images. According to the final offshore platform feature image and its corresponding actual offshore platform image, a loss value is calculated using a loss function, and parameters of the platform image recognition model to be trained are optimized according to the loss value until the loss value converges to obtain a trained platform image recognition model.

[0113] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A floating wind power platform offshore deviation early warning method, characterized in that: include: Obtain the real-time offshore image, real-time offshore position and original installation position of the floating wind power platform to be tested; Inputting the real-time offshore image of the floating wind power platform to be tested into the trained platform image recognition model, so that the platform image recognition model performs feature extraction on the real-time offshore image of the floating wind power platform to be tested, and obtains a real-time offshore platform feature map of the floating wind power platform to be tested; wherein the platform image recognition model includes a convolution layer; the convolution layer is used to perform feature extraction on the real-time offshore image of the floating wind power platform to be tested; According to the real-time offshore platform characteristic map, the tilt angle and the rotation angle of the floating wind power platform to be tested are calculated by Hough transform; Calculate the offset of the floating platform to be tested according to the tilt angle, rotation angle, real-time offshore position and original installation position of the floating wind power platform to be tested; According to the comparison and judgment between the displacement of the floating platform to be tested and the preset displacement threshold, the sending state of the displacement warning of the floating platform to be tested is determined; Before calculating the tilt angle and rotation angle of the floating wind power platform to be measured by Hough transform according to the real-time offshore platform characteristic map, the method further includes: Performing grayscale conversion on the real-time offshore platform characteristic image of the floating wind power platform to be tested to obtain a real-time offshore platform characteristic grayscale image; Performing a denoising operation on the real-time offshore platform feature grayscale image to obtain a denoised real-time offshore platform feature grayscale image; The edge detection algorithm is used to perform edge detection on the denoised real-time offshore platform feature grayscale image to obtain a real-time offshore platform feature grayscale image with edge points; The method of calculating the tilt angle and rotation angle of the floating wind power platform to be tested by Hough transform according to the real-time offshore platform characteristic map includes: Map the edge points in the real-time offshore platform feature grayscale image with edge points to the parameter space of Hough transform to obtain several mapping straight lines; The Hough transform accumulator is used to vote and accumulate the parameter space position corresponding to each mapping line to obtain mapping lines of several peak points. According to the mapping lines of several peak points, a mapping line corresponding to the peak point with the largest number of cumulative votes is selected, a mapping angle of the mapping line is determined, and the mapping angle of the mapping line is used as the inclination angle of the floating wind power platform to be tested; According to the mapping straight lines of the several peak points, the mapping angles of the mapping straight lines corresponding to the remaining peak points are determined to obtain several mapping angles, and the average mapping angle value is calculated according to the several mapping angles, and the average mapping angle value is used as the rotation angle of the floating wind power platform to be tested; The mathematical expression of the mapping straight line is: ； Among them, the represents the shortest distance from the mapping line to the origin of the image coordinate system; x and y represent the horizontal and vertical coordinates of the edge point in the rectangular coordinate system respectively; Indicates the mapping angle between the mapping line and the positive direction of the x-axis in the rectangular coordinate system.

2. A floating wind power platform offshore deviation early warning method according to claim 1, characterized in that: The model training of the platform image recognition model includes: Collect actual pictures of several offshore platforms of floating wind power platforms under different conditions; The data of each actual offshore platform image is preprocessed through the input layer to obtain several processed offshore platform images; Through the convolution layer, activation layer and pooling layer, feature extraction, feature activation and down sampling are performed on each processed offshore platform image to obtain several initial offshore platform feature images; The features of each initial offshore platform feature map are integrated through the fully connected layer to obtain several integrated offshore platform feature maps, and the features of each integrated offshore platform feature map are enhanced through the self-attention mechanism layer to obtain several final offshore platform feature maps; According to the final offshore platform feature map and its corresponding actual offshore platform map, the loss value is calculated using the loss function, and the parameters of the platform image recognition model to be trained are optimized according to the loss value until the loss value converges to obtain a trained platform image recognition model.

3. The method for early warning of offshore deviation of a floating wind power platform according to claim 1, characterized in that: The method of calculating the offset of the floating platform to be measured according to the tilt angle, rotation angle, real-time offshore position and original installation position of the floating wind power platform to be measured comprises: Calculate the deviation distance of the floating wind power platform to be tested according to the real-time offshore position and the original installation position of the floating wind power platform to be tested; According to the tilt angle, rotation angle and deviation distance of the floating wind power platform to be tested, the offset of the floating wind power platform to be tested in the east-west direction, north-south direction and vertical direction is calculated respectively by using trigonometric functions.

4. A method for early warning of offshore deviation of a floating wind power platform according to claim 3, characterized in that: The method of calculating the offset of the floating wind power platform to be tested in the east-west direction, the north-south direction and the vertical direction respectively by using trigonometric functions according to the tilt angle, the rotation angle and the deviation distance of the floating wind power platform to be tested includes: The formula for calculating the east-west offset of the floating wind power platform to be tested is: ； The formula for calculating the north-south offset of the floating wind power platform to be tested is: ； The formula for calculating the vertical offset of the floating wind power platform to be tested is: ； Among them, the represents the displacement of the floating wind power platform to be tested in the east-west direction; represents the north-south offset of the floating wind power platform to be tested; Indicates the vertical displacement of the floating wind power platform to be tested; Indicates the deviation distance of the floating wind power platform to be tested; represents the tilt angle of the floating wind power platform to be tested; Indicates the rotation angle of the floating wind turbine platform to be tested.

5. A method for early warning of offshore deviation of a floating wind power platform according to claim 4, characterized in that: The step of comparing and judging the displacement of the floating platform to be tested with a preset displacement threshold value to determine the sending state of the displacement warning of the floating platform to be tested includes: When the offset of the floating wind power platform to be tested in the east-west direction is less than the first offset threshold, and when the offset of the floating wind power platform to be tested in the north-south direction is less than the second offset threshold, and when the offset of the floating wind power platform to be tested in the vertical direction is less than the third offset threshold, the floating platform to be tested does not send an offset warning; Otherwise, the floating platform to be tested sends a deviation warning.

6. A floating wind power platform offshore displacement warning system, the floating wind power platform offshore displacement warning system is used to implement the floating wind power platform offshore displacement warning method according to claim 1, characterized in that: include: A data acquisition module, a feature map generation module, a first calculation module, a second calculation module and a deviation warning module; The data acquisition module is used to obtain the real-time offshore image, real-time offshore position and original installation position of the floating wind power platform to be tested; The feature map generation module is used to input the real-time offshore image of the floating wind power platform to be tested into the trained platform image recognition model, so that the platform image recognition model performs feature extraction on the real-time offshore image of the floating wind power platform to be tested, and obtains the real-time offshore platform feature map of the floating wind power platform to be tested; wherein the platform image recognition model includes a convolution layer; the convolution layer is used to perform feature extraction on the real-time offshore image of the floating wind power platform to be tested; The first calculation module is used to calculate the tilt angle and rotation angle of the floating wind power platform to be tested through Hough transformation according to the real-time offshore platform characteristic map; The second calculation module is used to calculate the offset of the floating platform to be tested according to the inclination angle, rotation angle, real-time offshore position and original installation position of the floating wind power platform to be tested; The deviation warning module is used to compare and judge the deviation of the floating platform to be tested with a preset deviation threshold value to determine the sending state of the deviation warning of the floating platform to be tested.

7. The floating wind power platform offshore deviation warning system according to claim 6, characterized in that: Also includes: Model training module; The model training module is used to collect several actual offshore platform images of floating wind power platforms under different conditions, perform data preprocessing on each actual offshore platform image through the input layer to obtain several processed offshore platform images, perform feature extraction, feature activation and downsampling on each processed offshore platform image through the convolution layer, the activation layer and the pooling layer to obtain several initial offshore platform feature images, perform feature integration on each initial offshore platform feature image through the fully connected layer to obtain several integrated offshore platform feature images, and perform feature enhancement on each integrated offshore platform feature image through the self-attention mechanism layer to obtain several final offshore platform feature images. According to the final offshore platform feature image and its corresponding actual offshore platform image, a loss value is calculated using a loss function, and parameters of the platform image recognition model to be trained are optimized according to the loss value until the loss value converges to obtain a trained platform image recognition model.

Images