Marine photovoltaic sea ice identification method and system, program product and electronic equipment
Through comprehensive analysis of sea surface images and signal autocorrelation functions, the problem of low sea ice identification accuracy was solved, the comprehensiveness and accuracy of sea ice identification were improved, and the normal operation of offshore photovoltaics was ensured.
Patent Information
- Application Number
- CN202510790532.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-13
- Publication Date
- 2025-09-23
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
In existing technologies, sea ice recognition has blind spots in some areas, resulting in low accuracy of sea ice recognition.
By acquiring sea surface images, direct signals and reflected signals of offshore photovoltaic sites, feature extraction and coherence analysis are performed. By combining the image and signal autocorrelation functions, the sea ice identification results are comprehensively determined, and fine identification is performed to determine the area where the sea ice is located.
The comprehensiveness and accuracy of sea ice identification have been improved, misidentification of single-dimensional detection has been avoided, and the efficiency of offshore photovoltaic production operations has been improved.
Smart Images
Figure CN120689673A_ABST
Abstract
Description
Technical Field
[0001] The embodiments of the present disclosure relate to the field of offshore photovoltaic technology, and in particular, to a method for identifying sea ice for offshore photovoltaics, a system for identifying sea ice for offshore photovoltaics, a computer program product, and an electronic device. Background Art
[0002] Sea ice on the sea surface can threaten the normal operation of offshore photovoltaics, so the identification of sea ice is particularly important.
[0003] Sea ice can be monitored manually or through microwave remote sensing. However, these methods have blind spots in some areas and cannot fully identify sea ice. Furthermore, they only identify sea ice based on one method, resulting in low accuracy. Summary of the Invention
[0004] The purpose of the present disclosure is to provide a method for identifying sea ice for offshore photovoltaics, a system for identifying sea ice for offshore photovoltaics, a computer program product, and an electronic device, thereby at least to a certain extent overcoming the problems of low accuracy and inaccurate sea ice identification caused by the limitations and defects of related technologies.
[0005] According to one aspect of the present disclosure, a sea ice identification method for offshore photovoltaics is provided, comprising: obtaining a sea surface image of a target area where an offshore photovoltaic site is located, and performing feature extraction on the sea surface image to determine image features of the sea surface image; performing feature fitting on the image features to obtain a first sea ice identification result, and performing coherence analysis based on direct signals and reflected signals corresponding to the target area to obtain a signal autocorrelation function to determine a second sea ice identification result; when the first sea ice identification result and the second sea ice identification result are different, determining a preliminary sea ice identification result in combination with the sea surface image and the signal autocorrelation function; when the preliminary sea ice identification result is sea ice, performing fine identification on the sea surface image to determine the area where the sea ice is located.
[0006] In an exemplary embodiment of the present disclosure, image features include one or more of color features, texture features, reflectivity features, transparency features, edge features, and dynamic behavior features; the image features are subjected to feature fitting to obtain a first sea ice recognition result, including: determining corresponding weight parameters according to the importance of the color features, texture features, reflectivity features, transparency features, edge features, and dynamic behavior features; performing weighted fusion of the color features, texture features, reflectivity features, transparency features, edge features, and dynamic behavior features according to the weight parameters to obtain fused features; and inputting the fused features into a classification model to determine the first sea ice recognition result.
[0007] In an exemplary embodiment of the present disclosure, a coherence analysis is performed on the direct signal and the reflected signal corresponding to the target area to obtain a signal autocorrelation function to determine a second sea ice identification result, including: determining a time series of a maximum direct complex correlation value of the direct signal and a time series of a maximum reflected complex correlation value of the reflected signal, and calculating an interference correlation function based on the time series of the maximum direct complex correlation value and the time series of the maximum reflected complex correlation value; determining a signal autocorrelation function of the interference correlation function, taking the absolute value of the signal autocorrelation function and normalizing it; taking the time for the normalized signal autocorrelation function to drop from a peak point to a preset value as the autocorrelation time; and determining the second sea ice identification result based on a comparison result of the autocorrelation time and a time threshold.
[0008] In an exemplary embodiment of the present disclosure, a preliminary sea ice recognition result is determined in combination with a sea surface image and a signal autocorrelation function, including: extracting image features of the sea surface image and autocorrelation features of the signal autocorrelation function, and splicing the image features and the autocorrelation features to obtain comprehensive features; performing average pooling on the comprehensive features, and performing full connection processing based on a fully connected layer to determine the probability that the sea surface image belongs to sea ice, so as to determine the preliminary sea ice recognition result.
[0009] In an exemplary embodiment of the present disclosure, a sea surface image is finely identified to determine the area where the sea ice is located, including: determining the category of each pixel point in the sea surface image and merging the pixel points classified as sea ice; performing edge detection on the merged sea surface image to determine the area where the sea ice is located.
[0010] In an exemplary embodiment of the present disclosure, the category of each pixel point in a sea surface image is determined, including: performing multiple feature extractions on the sea surface image through a multi-layer downsampling network to obtain a feature vector of each layer of the downsampling network; obtaining a corresponding reference feature vector based on the feature vector of each layer of the downsampling network, upsampling the feature vector through a multi-layer upsampling network to obtain an upsampling result, and performing feature fusion on the reference feature vector corresponding to the feature vector output by the same layer and the upsampling result to obtain the category of each pixel point.
[0011] In an exemplary embodiment of the present disclosure, after determining the area where the sea ice is located, the method further includes: determining input data based on sea ice parameters corresponding to the sea ice and external environmental parameters corresponding to the area; wherein the external environmental parameters include one or more of atmospheric parameters, ocean parameters, and forcing parameters; performing a convolution operation on the input data to obtain local features, and downsampling and pooling the local features to determine a feature vector; and fully connecting the feature vector to determine the changing state of the sea ice.
[0012] According to one aspect of the present disclosure, a sea ice identification system for offshore photovoltaics is provided, comprising: a feature extraction module for acquiring a sea surface image of a target area where an offshore photovoltaic site is located, performing feature extraction on the sea surface image, and determining image features of the sea surface image; a separate identification module for performing feature fitting on the image features to obtain a first sea ice identification result, and performing coherence analysis based on direct signals and reflected signals corresponding to the target area to obtain a signal autocorrelation function to determine a second sea ice identification result; a combined identification module for determining a preliminary sea ice identification result by combining the sea surface image and the signal autocorrelation function when the first sea ice identification result and the second sea ice identification result are different; and a region determination module for performing fine identification on the sea surface image to determine the region where the sea ice is located when the preliminary sea ice identification result is sea ice.
[0013] According to one aspect of the present disclosure, a computer program product is provided. When the computer program is executed by a processor, the computer program implements any of the above-mentioned sea ice identification methods for offshore photovoltaics.
[0014] According to one aspect of the present disclosure, an electronic device is provided, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to implement any one of the above-mentioned sea ice identification methods for offshore photovoltaics by executing the executable instructions.
[0015] The technical solutions provided in the embodiments of the present disclosure, on the one hand, perform coherence analysis of direct and reflected signals, enabling sea ice detection based on the difference in coherence time between sea ice and seawater. This approach is unaffected by changes in satellite altitude, avoiding limitations, enabling comprehensive detection, and improving the reliability of sea ice identification. Furthermore, when the first and second sea ice identification results differ, the preliminary sea ice identification result can be determined from both image and signal dimensions, combining the sea surface image and the signal autocorrelation function obtained through coherence analysis. This increases the predictive dimension of sea ice identification, avoids the problem of misidentification caused by single-dimensional detection, improves the accuracy and reliability of sea ice identification, and enhances the production and operation efficiency of offshore photovoltaics.
[0016] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] The accompanying drawings are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the specification, are used to explain the principles of the present disclosure. Obviously, the drawings described below are only some embodiments of the present disclosure, and those skilled in the art can derive other drawings based on these drawings without inventive effort.
[0018] Figure 1 A schematic flow chart of a method for identifying sea ice for offshore photovoltaics in an embodiment of the present disclosure is shown schematically.
[0019] Figure 2 A schematic diagram of a process for determining preliminary sea ice identification results in an embodiment of the present disclosure is schematically shown.
[0020] Figure 3 The following is a schematic diagram illustrating a process of determining the categories of pixels of a sea surface image in an embodiment of the present disclosure.
[0021] Figure 4 The block diagram of the sea ice identification system for offshore photovoltaics in an embodiment of the present disclosure is schematically shown.
[0022] Figure 5 The block diagram of the electronic device according to the embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION
[0023] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that the present disclosure will be more comprehensive and complete and will fully convey the concepts of the example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced while omitting one or more of the specific details, or that other methods, components, devices, steps, etc. may be employed. In other cases, well-known technical solutions are not shown or described in detail to avoid obscuring various aspects of the present disclosure.
[0024] In addition, the accompanying drawings are merely schematic illustrations of the present disclosure and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.
[0025] Offshore photovoltaics are located in a complex hydrological and climatic environment, especially in winter when there will be sea ice and floating ice at sea. When encountering such a climate, the working risk of the pile foundation increases and it is easy to be damaged by ice loads. The impact of ice loads in various forms of damage on the structure of photovoltaic pile foundations cannot be ignored and will threaten the normal operation of offshore photovoltaics.
[0026] Based on this, an embodiment of the present disclosure provides a method for identifying sea ice for offshore photovoltaics, which can be applied to application scenarios for identifying sea ice on any sea surface. Figure 1 A schematic diagram of the sea ice identification method for offshore photovoltaics is shown in FIG. Figure 1 As shown in , the method mainly includes the following steps:
[0027] Step S110, obtaining a sea surface image of a target area where the offshore photovoltaic site is located, and performing feature extraction on the sea surface image to determine image features of the sea surface image;
[0028] Step S120: performing feature fitting on the image features to obtain a first sea ice recognition result, and performing coherence analysis on the direct signal and the reflected signal corresponding to the target area to obtain a signal autocorrelation function to determine a second sea ice recognition result;
[0029] Step S130, determining a preliminary sea ice recognition result based on the first sea ice recognition result and the second sea ice recognition result, combined with the sea surface image and the signal autocorrelation function;
[0030] Step S140: When the preliminary sea ice identification result is sea ice, fine identification is performed on the sea surface image to determine the area where the sea ice is located.
[0031] Next, the sea ice identification method for offshore photovoltaics in the embodiment of the present disclosure is described in detail with reference to the embodiments.
[0032] Step S110 , obtaining a sea surface image of a target area where the offshore photovoltaic site is located, and performing feature extraction on the sea surface image to determine image features of the sea surface image.
[0033] In the disclosed embodiments, the sea ice identification method for offshore photovoltaic systems can be executed via a terminal, specifically via a mini-program on the terminal, a public account, or an application, without specific limitation. The terminal can be any type of terminal, such as a mobile terminal such as a smartphone or tablet.
[0034] An offshore photovoltaic site refers to the on-site environment used to implement an offshore photovoltaic project. A target area refers to the sea surface area corresponding to the offshore photovoltaic site. It should be noted that the target area can be the entire sea surface of the offshore photovoltaic site, or a portion of the sea surface of the offshore photovoltaic site, depending on actual needs. A sea surface image refers to an image containing the sea surface. This sea surface image can be an image captured by a drone, a satellite, or a camera. The sea surface image can include seawater and, in addition, sea ice, without specific limitations here.
[0035] After acquiring the sea surface image, feature extraction can be performed on the sea surface image to obtain image features of the sea surface image. The image features may include one or more of color features, texture features, reflectivity features, transparency features, edge features, and dynamic behavior features, so as to visually distinguish sea ice from seawater based on the image features.
[0036] Step S120 , performing feature fitting on the image features to obtain a first sea ice recognition result, and performing coherence analysis on the direct signal and the reflected signal corresponding to the target area to obtain a signal autocorrelation function to determine a second sea ice recognition result.
[0037] In the disclosed embodiment, image analysis can be performed on the sea surface image to obtain a first sea ice recognition result of the sea surface image; in addition, coherence analysis can be performed on the direct signal and the reflected signal of the target area to determine the signal autocorrelation function, and then the second sea ice recognition result can be determined based on the signal autocorrelation function.
[0038] In some embodiments, after acquiring the image features of the sea surface image of the target area, the importance of the color features, texture features, reflectivity features, transparency features, edge features, and dynamic behavior features in the image features can be determined, and then the weight parameters corresponding to each feature can be determined according to the importance, and the importance is positively correlated with the weight parameters. Among them, the importance can be determined according to the correlation between each feature and the sea ice. After obtaining the weight parameters of each feature, the corresponding features can be weighted and fused according to the weight parameters to obtain fused features. Based on this, the fused features can be input into the classification model to perform classification and identification of sea ice and seawater to determine the first sea ice identification result. The classification model can be any model that can perform classification processing, for example, it can be a neural network model or a decision tree model, etc., which is not specifically limited here. The first sea ice identification result can be sea ice or seawater.
[0039] In other embodiments, for the sea surface of the target area, the direct signal and the reflected signal of the target area can be obtained. The direct signal and the reflected signal refer to the direct signal and the reflected signal of the navigation satellite. The navigation satellite here can be, for example, a GNSS (Global Navigation Satellite System) navigation satellite.
[0040] Navigation satellites can be fixed to shore-based platforms corresponding to offshore photovoltaic sites. Shore-based platforms are shore-based stations used for collecting and analyzing environmental data. These platforms are located at a fixed altitude above the sea surface. Navigation satellites receive direct signals from the navigation satellites as well as reflected signals from the target area's sea surface after the transmitted direct signals are reflected.
[0041] Furthermore, the maximum direct complex correlation value time series of the direct signal and the maximum reflected complex correlation value time series of the reflected signal can be obtained. The complex correlation value time series is used to characterize the autocorrelation or cross-correlation characteristics of the signal, which can reflect the similarity of the signal in the time domain / frequency domain, and is usually expressed in a complex form containing amplitude and phase information. For example, for the direct signal, the received direct signal can be expressed in a complex form to obtain a complex direct signal. In addition, an ideal direct signal reference copy can be generated as a reference signal based on the parameters of the navigation satellite such as PRN (Pseudo-Random Noise Code) and carrier frequency. On this basis, the complex correlation value time series of the direct signal and the reference signal can be calculated through a sliding time window, and the largest one can be used as the maximum direct complex correlation value time series. The calculation process of the maximum reflected complex correlation value time series is the same as the above process and will not be repeated here.
[0042] On this basis, the interferometric correlation function (ICF) can be calculated based on the ratio of the time series of the maximum direct complex correlation value to the time series of the maximum reflected complex correlation value. For example, the interferometric correlation function can be an ICF (Interferometric Complex Field). The ICF function is used to represent the complex cross-correlation result of two coherent signals. The real and imaginary parts of the ICF function correspond to the in-phase and quadrature components of the signal, respectively.
[0043] Next, the interference correlation function can be processed to obtain the second sea ice identification result. For example, the DC component carried by the interference correlation function can be removed, and the signal autocorrelation function of the interference correlation function can be determined based on the interference correlation function after the DC component is removed, and the absolute value of the signal autocorrelation function is taken and normalized. The time for the normalized signal autocorrelation function to drop from the peak point to the preset value is determined, and this time is taken as the ICF autocorrelation time; the obtained ICF autocorrelation time is compared with the preset time threshold. If the autocorrelation time is greater than the time threshold, the second sea ice identification result is determined to be sea ice; if the autocorrelation time is less than the time threshold, the second sea ice identification result is determined to be sea water. Among them, the signal autocorrelation function of ICF (Interferometric Complex Field) is used to analyze its temporal or spatial coherence, reflecting the similarity of the signal under different time delays or frequency offsets. The signal autocorrelation function can be obtained by inverse Fourier transform of the power spectral density, specifically referring to formula (1):
[0044] R ICF (Δτ)=F -1 {|F{ICF(τ)}| 2} Formula (1)
[0045] The specific steps include: performing Fourier transform on ICF(τ) to obtain the spectrum F(ICF(τ)), calculating the power spectrum |F{ICF(τ)}| 2 , and perform inverse Fourier transform on the power spectrum to obtain the signal autocorrelation function. The preset value can be 1 / e.
[0046] Alternatively, the interference features of the interference correlation function can be directly extracted, and the interference features can be convolved through a machine learning model to determine the second sea ice identification result.
[0047] In the disclosed embodiment, by identifying sea ice from an image dimension and a signal dimension respectively, the comprehensiveness and accuracy of sea ice identification can be improved.
[0048] Step S130 , when the first sea ice identification result and the second sea ice identification result are different, a preliminary sea ice identification result is determined by combining the sea surface image and the signal autocorrelation function.
[0049] In the disclosed embodiments, the first sea ice identification result may be the same as or different from the second sea ice identification result, depending on the actual situation. To improve accuracy, when the first sea ice identification result and the second sea ice identification result are the same, the preliminary sea ice identification result may be determined directly based on the first sea ice identification result or the second sea ice identification result. For example, when both the first sea ice identification result and the second sea ice identification result are sea ice, the preliminary sea ice identification result may be determined directly as sea ice; when both the first sea ice identification result and the second sea ice identification result are sea water, the preliminary sea ice identification result may be determined directly as sea water.
[0050] When the first sea ice identification result differs from the second sea ice identification result, a comprehensive determination can be made based on the sea surface image and the signal autocorrelation function of the interferometric correlation function to obtain a preliminary sea ice identification result. For example, the preliminary sea ice identification result can be comprehensively determined based on the sea surface image, the image autocorrelation function of the sea surface image, and the signal autocorrelation function of the interferometric correlation function. The image autocorrelation function of the sea surface image can be a global autocorrelation function. The global autocorrelation function of the sea surface image can be extracted using the convolution theorem of Fourier transform. Alternatively, a sliding window can be used to extract local autocorrelation functions of the sea surface image, and the local autocorrelation functions can be further integrated to determine the global autocorrelation function. For example, the global autocorrelation function can be determined using an averaging method or a median filtering method.
[0051] Next, the image autocorrelation function of the sea surface image and the signal autocorrelation function of the interference correlation function can be fused to obtain an autocorrelation function. Furthermore, the sea surface image and autocorrelation function can be input into a convolutional neural network for feature fitting to obtain preliminary sea ice recognition results. The convolutional neural network may include an input layer, a feature extraction module, a feature fusion module, a pooling layer, a fully connected layer, and an output layer. The feature extraction module may include an image feature extraction branch and an autocorrelation feature extraction branch. The image feature extraction branch may include multiple convolutional layers for extracting image features of different dimensions. The autocorrelation feature extraction branch may include multiple convolutional layers for learning autocorrelation features of different dimensions. The feature fusion module is used to concatenate the image features obtained by the image feature extraction branch and the autocorrelation features obtained by the autocorrelation feature extraction branch in the channel dimension to obtain a comprehensive feature. The comprehensive feature is average pooled using the pooling layer and fully connected using the fully connected layer. The output layer outputs the probability that the sea surface image belongs to seawater and the probability that it belongs to sea ice, thereby determining the preliminary sea ice recognition result based on the maximum probability. Among them, if the probability of belonging to sea ice is the highest, it can be determined that the preliminary sea ice identification result is sea ice. Here, the preliminary sea ice identification result is sea ice, which means that the sea surface image contains sea ice.
[0052] Figure 2 The flowchart for determining the preliminary sea ice identification results is shown schematically in Figure 2 As shown in , it mainly includes the following steps:
[0053] In step S202, the image features of the sea surface image are input into a classification model to determine a first sea ice recognition result.
[0054] In step S204, a signal autocorrelation function is determined based on the maximum complex correlation value time series of the direct signal and the reflected signal of the target area, an autocorrelation time is determined based on the signal autocorrelation function, and a second sea ice identification result is determined based on a comparison result between the autocorrelation time and the time threshold.
[0055] In step S206, it is determined whether the first sea ice identification result is the same as the second sea ice identification result; if they are the same, step S208 is executed; if they are different, step S210 is executed.
[0056] In step S208, a preliminary sea ice identification result is determined based on the first sea ice identification result or the second sea ice identification result. If both are sea water, the preliminary sea ice identification result is sea water. If both are sea ice, the preliminary sea ice identification result is sea ice.
[0057] In step S210 , a preliminary sea ice recognition result is determined based on the sea surface image, the image autocorrelation function of the sea surface image, and the signal autocorrelation function of the interference correlation function.
[0058] The image autocorrelation function of the sea surface image and the signal autocorrelation function of the interference correlation function are fused to obtain the autocorrelation function. The sea surface image and autocorrelation function are input into a convolutional neural network for feature fitting to obtain preliminary sea ice recognition results. Specifically, the features of the sea surface image and autocorrelation function are extracted respectively to obtain image features and autocorrelation features. The obtained image features and autocorrelation features are spliced in the channel dimension to obtain comprehensive features. The comprehensive features are average pooled and fully connected based on the fully connected layer. The probability of the sea surface image belonging to seawater and the probability of belonging to sea ice are output, and the preliminary sea ice recognition result is determined based on the category corresponding to the maximum probability.
[0059] In the disclosed embodiments, if the first and second sea ice identification results differ, the preliminary sea ice identification result is re-determined based on the sea surface image, the image autocorrelation function of the sea surface image, and the signal autocorrelation function of the interference correlation function. This allows for accurate determination of the presence of sea ice by combining features from multiple dimensions. Furthermore, by combining features from the sea surface image itself, as well as those from the transmitted and reflected signals, this avoids the blind spots inherent in conventional sea ice identification methods. Comprehensive sea ice identification can be achieved based on features from multiple dimensions, improving both the accuracy and comprehensiveness of sea ice identification.
[0060] Step S140: When the preliminary sea ice identification result is sea ice, fine identification is performed on the sea surface image to determine the area where the sea ice is located.
[0061] In the embodiment of the present disclosure, when performing fine recognition of the sea surface image, the category of each pixel point in the sea surface image can be determined. After the category of each pixel point is identified, the pixel points classified as sea ice can be merged to determine the area where the sea ice is located.
[0062] Figure 3 The flowchart for determining the category of pixel points is shown schematically. Figure 3 As shown in , it mainly includes the following steps:
[0063] Step S310, performing multiple feature extractions on the sea surface image through an encoder to obtain a feature vector of each downsampling network layer in the encoder;
[0064] In step S320, the corresponding intermediate feature vector is obtained according to the feature vector of each layer of the downsampling network, the feature vector is upsampled through the decoder to obtain the upsampling result, and the intermediate feature vector of the same layer is feature fused with the upsampling result to obtain the category of each pixel point in the sea surface image.
[0065] A classification model can be used to determine the category of each pixel. The classification model may include an encoder and a decoder. The encoder may include a multi-layer network, each of which may include multiple convolutional layers and maximum pooling layers. Therefore, the encoder's multi-layer network may be a multi-layer downsampling network. Based on this, convolution operations can be performed on the sea surface image in sequence according to each downsampling network layer to obtain a convolution result. The convolution result is then downsampled to obtain a downsampled result, and the downsampled result is determined as a feature vector for each downsampling network layer. This process continues until all convolutional layers and maximum pooling layers in the downsampling network layer have completed feature extraction.
[0066] The decoder can include a multi-layer network, each of which includes an upsampling convolutional layer, feature concatenation, and multiple convolutional layers. The number of convolutional layers in the encoder and decoder can be the same. The decoder's multi-layer network can be a multi-layer upsampling network. First, each upsampling network upsamples the feature vectors output by each downsampling network in the encoder to obtain an upsampling result. Furthermore, feature vectors of the same size are first intercepted from the feature vectors corresponding to the output of the same layer in the encoder as intermediate feature vectors for each upsampling network layer in the decoder. These intermediate feature vectors are then fused with the upsampling result obtained by upsampling the feature vectors in the decoder to obtain fused features for each upsampling network layer. The fused features of each upsampling network layer are further convolved to obtain the category of each pixel in the sea surface image. For example, the category of pixel 1 is seawater, the category of pixels 2-100 is sea ice, the category of pixels 100-150 is seawater, and so on.
[0067] After determining the category of each pixel in the sea surface image, the sea surface image can be converted into a binary image based on the category of each pixel. Connected sea ice pixel regions in the binary image are identified, and edges of the sea ice pixel regions are automatically extracted using an edge detection algorithm or a machine learning model. A mask image or boundary map of the sea ice pixel regions is output to determine the location of the sea ice. The edge detection algorithm can be the Canny edge detection algorithm or other edge detection algorithms, which are not specifically limited here.
[0068] In the disclosed embodiments, if the initial sea ice identification result indicates the presence of sea ice, detailed recognition of the sea surface image can be performed to determine the area where the sea ice is located. By determining the presence and location of sea ice, the system can be applied to a wider range of applications, providing a reference for the construction status of offshore photovoltaic sites.
[0069] In addition, information such as the area and location of the sea ice can be output. For example, a geographic information system can be used to convert image coordinates into geographic coordinates to achieve the geographic positioning of the sea ice and determine its location.
[0070] After determining the sea ice region, the changing state of the sea ice in that region can be predicted based on its parameters and external environmental parameters. Sea ice parameters can include one or more of sea ice thickness, sea ice area, and sea ice type. External environmental parameters include one or more of atmospheric parameters, ocean parameters, and forcing parameters. Atmospheric parameters can include ambient temperature, wind speed, and atmospheric pressure; ocean parameters include seawater temperature, salinity, and ocean circulation; and forcing parameters can be parameters that influence sea ice changes, such as solar radiation.
[0071] In some embodiments, the aforementioned sea ice parameters and external environmental parameters can be integrated to obtain input data. This input data is then fed into a convolutional neural network, where it is convolved using multiple convolution kernels to extract local features. These local features are then downsampled and max pooling is performed to retain the maximum value, resulting in a pooled feature vector. This pooled feature vector is then fully connected to output a predicted value for the changing state of sea ice, such as changes in sea ice thickness or sea ice area, based on task requirements.
[0072] In the disclosed embodiment, the sea ice parameters and external environmental parameters of the sea ice can be used to accurately predict the changes in sea ice in the region, and corresponding processing strategies can be executed based on the changes in sea ice to reduce the impact of sea ice on the offshore photovoltaic site, thereby ensuring the progress of the offshore photovoltaic project and improving the production and operation efficiency of offshore photovoltaics.
[0073] In some embodiments, remote sensing satellite data of the target area can also be obtained, and the drift state of the sea ice can be determined in real time based on the remote sensing satellite data. The remote sensing satellite data can be Sentinel-1 synthetic aperture radar data or medium resolution imaging spectrometer data. Exemplarily, the remote sensing satellite data is denoised, corrected and spliced to obtain pre-processed remote sensing satellite data, which can be a sea ice image. The feature points in the sea ice image are extracted by the SIFT (Scale Invariant Feature Transform) algorithm, and the displacement of the feature points in different time intervals is calculated. Based on the displacement and time interval of the feature points, the drift speed and direction are calculated to obtain the drift vector; the drift vector is combined with the geographic information system to generate a sea ice drift dynamics map. By determining the drift vector of the sea ice, it can provide a reference for subsequent offshore photovoltaic sites and assist in the normal execution of offshore photovoltaic projects.
[0074] In the disclosed embodiment, by performing coherence analysis on the direct signal and the reflected signal, sea ice detection can be achieved based on the difference in coherence time between sea ice and seawater. The detection accuracy is high, and it will not be affected by changes in satellite altitude or limitations in other dimensions. It can achieve comprehensive detection and improve the reliability of sea ice identification. In addition, when the first sea ice identification result and the second sea ice identification result are different, the sea surface image and the signal autocorrelation function obtained by the coherence analysis can be combined to comprehensively determine the preliminary sea ice identification result from the two dimensions of image and signal, thereby increasing the prediction dimension of sea ice identification, avoiding limitations, and improving the accuracy and reliability of sea ice identification. When the preliminary sea ice identification result is determined to be sea ice, the sea surface image can also be finely identified to determine the area where the sea ice is located. The accuracy of sea ice detection can be improved through secondary identification.
[0075] In the embodiment of the present disclosure, a sea ice recognition system for offshore photovoltaics is also provided. Figure 4 As shown in FIG, the offshore photovoltaic sea ice identification system 400 includes:
[0076] The feature extraction module 401 is used to obtain a sea surface image of a target area where the offshore photovoltaic site is located, and perform feature extraction on the sea surface image to determine image features of the sea surface image;
[0077] A separate recognition module 402 is configured to perform feature fitting on the image features to obtain a first sea ice recognition result, and perform coherence analysis based on the direct signal and the reflected signal corresponding to the target area to obtain a signal autocorrelation function to determine a second sea ice recognition result;
[0078] A combined identification module 403 is configured to determine a preliminary sea ice identification result by combining the sea surface image and the signal autocorrelation function when the first sea ice identification result and the second sea ice identification result are different;
[0079] The area determination module 404 is used to perform fine identification on the sea surface image to determine the area where the sea ice is located when the preliminary sea ice identification result is sea ice.
[0080] In an exemplary embodiment of the present disclosure, image features include one or more of color features, texture features, reflectivity features, transparency features, edge features, and dynamic behavior features; the image features are subjected to feature fitting to obtain a first sea ice recognition result, including: determining corresponding weight parameters according to the importance of the color features, texture features, reflectivity features, transparency features, edge features, and dynamic behavior features; performing weighted fusion of the color features, texture features, reflectivity features, transparency features, edge features, and dynamic behavior features according to the weight parameters to obtain fused features; and inputting the fused features into a classification model to determine the first sea ice recognition result.
[0081] In an exemplary embodiment of the present disclosure, a coherence analysis is performed on the direct signal and the reflected signal corresponding to the target area to obtain a signal autocorrelation function to determine a second sea ice identification result, including: determining a time series of a maximum direct complex correlation value of the direct signal and a time series of a maximum reflected complex correlation value of the reflected signal, and calculating an interference correlation function based on the time series of the maximum direct complex correlation value and the time series of the maximum reflected complex correlation value; determining a signal autocorrelation function of the interference correlation function, taking the absolute value of the signal autocorrelation function and normalizing it; taking the time for the normalized signal autocorrelation function to drop from a peak point to a preset value as the autocorrelation time; and determining the second sea ice identification result based on a comparison result of the autocorrelation time and a time threshold.
[0082] In an exemplary embodiment of the present disclosure, a preliminary sea ice recognition result is determined in combination with a sea surface image and a signal autocorrelation function, including: extracting image features of the sea surface image and autocorrelation features of the signal autocorrelation function, and splicing the image features and the autocorrelation features to obtain comprehensive features; performing average pooling on the comprehensive features, and performing full connection processing based on a fully connected layer to determine the probability that the sea surface image belongs to sea ice, so as to determine the preliminary sea ice recognition result.
[0083] In an exemplary embodiment of the present disclosure, a sea surface image is finely identified to determine the area where the sea ice is located, including: determining the category of each pixel point in the sea surface image and merging the pixel points classified as sea ice; performing edge detection on the merged sea surface image to determine the area where the sea ice is located.
[0084] In an exemplary embodiment of the present disclosure, the category of each pixel point in a sea surface image is determined, including: performing multiple feature extractions on the sea surface image through a multi-layer downsampling network to obtain a feature vector of each downsampling network layer; obtaining a corresponding reference feature vector based on the feature vector of each downsampling network layer, upsampling the feature vector through a multi-layer upsampling network to obtain an upsampling result, and performing feature fusion of the reference feature vector corresponding to the feature vector output by the same layer and the upsampling result to obtain the category of each pixel point.
[0085] In an exemplary embodiment of the present disclosure, after determining the area where the sea ice is located, the device is further configured to: determine input data based on sea ice parameters corresponding to the sea ice and external environmental parameters corresponding to the area; wherein the external environmental parameters include one or more of atmospheric parameters, ocean parameters and forcing parameters; perform convolution operation on the input data to obtain local features, and downsample and pool the local features to determine feature vectors; and perform full connection processing on the feature vectors to determine the changing state of the sea ice.
[0086] It should be noted that the specific details of each module in the above-mentioned offshore photovoltaic sea ice identification system have been described in detail in some implementation methods of the corresponding methods, and will not be repeated here.
[0087] It should be noted that although several modules or units of the device for action execution are mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be concretized in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided into multiple modules or units to be concretized.
[0088] Furthermore, although the steps of the method of the present disclosure are described in a particular order in the accompanying drawings, this does not require or imply that the steps must be performed in this particular order, or that all steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps.
[0089] In an exemplary embodiment of the present disclosure, an electronic device capable of implementing the above method is also provided.
[0090] Those skilled in the art will appreciate that various aspects of the present disclosure may be implemented as systems, methods, or program products. Therefore, various aspects of the present disclosure may be implemented in the following forms: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software implementations, which may be collectively referred to herein as "circuits," "modules," or "systems."
[0091] Refer to the following Figure 5 hereinafter, an electronic device 500 according to this embodiment of the present disclosure is described. Figure 5 The electronic device 500 shown is merely an example and should not limit the functions and scope of use of the embodiments of the present disclosure.
[0092] like Figure 5 As shown, electronic device 500 is implemented as a general-purpose computing device. Components of electronic device 500 may include, but are not limited to, the aforementioned at least one processing unit 510, the aforementioned at least one storage unit 520, a bus 530 connecting various system components (including storage unit 520 and processing unit 510), and a display unit 540.
[0093] The storage unit stores program codes, which can be executed by the processing unit 510, so that the processing unit 510 performs the steps according to various exemplary embodiments of the present disclosure described in the above “Exemplary Method” section of this specification. For example, the processing unit 510 can perform the following steps: Figure 1 Follow the steps shown in .
[0094] The storage unit 520 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 5201 and / or a cache memory unit 5202 , and may further include a read-only memory unit (ROM) 5203 .
[0095] The storage unit 520 may also include a program / utility 5204 having a set (at least one) of program modules 5205, such program modules 5205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.
[0096] Bus 530 may represent one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.
[0097] The electronic device 500 may also communicate with one or more external devices 600 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), one or more devices that enable a user to interact with the electronic device 500, and / or any device that enables the electronic device 500 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed through an input / output (I / O) interface 550. Furthermore, the electronic device 500 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through a network adapter 560. Figure 5 As shown, the network adapter 560 communicates with other modules of the electronic device 500 via the bus 530. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 500, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.
[0098] It should be noted that some embodiments of the present disclosure further provide a computer program product, which includes a computer program, and the computer program implements the above method when executed by a processor.
[0099] In one embodiment, a computer program product may be a tangible product containing a computer program, such as a computer-readable storage medium storing a computer program. The readable storage medium may be a storage medium based on electrical, magnetic, optical, electromagnetic, infrared, or other signals, including but not limited to: random access memory (RAM), read-only memory (ROM), magnetic tape, floppy disk, flash memory (Flash), mechanical hard disk (HDD), solid-state drive (SSD), and the like. Exemplarily, a computer program product may be implemented as a non-volatile storage medium storing a computer program, such as a read-only memory, a NAND flash memory (Nand Flash), and the like. In one embodiment, a computer program product may be an intangible product containing a computer program. Exemplarily, a computer program product may be implemented as a virtual digital product, such as a digital file such as an executable file or installation package storing a computer program.
[0100] The code of the computer program can be written in one or more programming languages. Programming languages include C, Java, C++, etc. The program code can be executed entirely on the user computing device, partially on the user computing device, or as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device via any type of network, such as a local area network (LAN), a wide area network (WAN), etc., or can be connected to an external computing device (e.g., via an Internet connection provided by a carrier).
[0101] Computer programs can be carried or transmitted via electrical, magnetic, optical, electromagnetic, infrared, or other signals. Electronic devices can convert signals carrying computer programs into digital signals to run the computer programs. When the computer program is run on an electronic device, its code causes the electronic device to execute (more specifically, the processor of the electronic device to execute) the method steps of various exemplary embodiments of the present disclosure.
[0102] Through the description of the above embodiments, it is easy for those skilled in the art to understand that the example embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to the embodiments of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the embodiments of the present disclosure.
[0103] Furthermore, the figures above are merely illustrative of the processes included in the methods according to exemplary embodiments of the present disclosure and are not intended to be limiting. It is readily understood that the processes illustrated in the figures above do not indicate or limit the temporal order of these processes. Furthermore, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.
[0104] It should be noted that although several modules or units of the device for action execution are mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be concretized in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided into multiple modules or units to be concretized.
[0105] Other embodiments of the present disclosure will readily occur to those skilled in the art after considering the specification and practicing what is disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered as exemplary only, with the true scope and spirit of the present disclosure being indicated by the claims.
[0106] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes can be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.
Claims
1. A method for identifying sea ice in offshore photovoltaics, characterized in that: include: Acquire a sea surface image of a target area where the offshore photovoltaic site is located, and perform feature extraction on the sea surface image to determine image features of the sea surface image; Performing feature fitting on the image features to obtain a first sea ice recognition result, and performing coherence analysis based on the direct signal and the reflected signal corresponding to the target area to obtain a signal autocorrelation function to determine a second sea ice recognition result; When the first sea ice identification result and the second sea ice identification result are different, determining a preliminary sea ice identification result by combining the sea surface image and the signal autocorrelation function; When the preliminary sea ice identification result is sea ice, the sea surface image is finely identified to determine the area where the sea ice is located.
2. The method for identifying sea ice for offshore photovoltaics according to claim 1, characterized in that: The image features include one or more of color features, texture features, reflectivity features, transparency features, edge features, and dynamic behavior features; The performing feature fitting on the image features to obtain a first sea ice recognition result includes: Determining corresponding weight parameters according to the importance of the color feature, the texture feature, the reflectivity feature, the transparency feature, the edge feature, and the dynamic behavior feature; Performing weighted fusion on the color feature, texture feature, reflectivity feature, transparency feature, edge feature and dynamic behavior feature according to the weight parameter to obtain a fusion feature; The fused features are input into a classification model to determine the first sea ice recognition result.
3. The method for identifying sea ice for offshore photovoltaics according to claim 1, characterized in that: The performing coherence analysis on the direct signal and the reflected signal corresponding to the target area to obtain a signal autocorrelation function to determine a second sea ice identification result includes: Determine a time series of maximum direct complex correlation values of a direct signal and a time series of maximum reflected complex correlation values of a reflected signal, and calculate an interference correlation function based on the time series of maximum direct complex correlation values and the time series of maximum reflected complex correlation values; determining a signal autocorrelation function of the interference correlation function, taking an absolute value of the signal autocorrelation function and performing normalization processing; The time it takes for the normalized signal autocorrelation function to drop from the peak point to the preset value is taken as the autocorrelation time; The second sea ice identification result is determined according to a comparison result of the autocorrelation time and a time threshold.
4. The method for identifying sea ice for offshore photovoltaics according to claim 1, characterized in that: Determining a preliminary sea ice recognition result by combining the sea surface image and the signal autocorrelation function includes: extracting image features of the sea surface image and autocorrelation features of the signal autocorrelation function, and concatenating the image features and the autocorrelation features to obtain comprehensive features; The comprehensive features are average pooled and fully connected based on a fully connected layer to determine the probability that the sea surface image belongs to sea ice, so as to determine the preliminary sea ice recognition result.
5. The method for identifying sea ice for offshore photovoltaics according to claim 1, characterized in that: The finely identifying the sea surface image to determine the area where the sea ice is located includes: Determining the category of each pixel in the sea surface image, and merging the pixels classified as sea ice; Edge detection is performed on the merged sea surface image to determine the area where the sea ice is located.
6. The method for identifying sea ice for offshore photovoltaics according to claim 5, characterized in that: Determining the category of each pixel in the sea surface image includes: Performing multiple feature extractions on the sea surface image through a multi-layer downsampling network to obtain a feature vector of each layer of the downsampling network; The corresponding reference feature vector is obtained according to the feature vector of each layer of the downsampling network, and the feature vector is upsampled through a multi-layer upsampling network to obtain an upsampling result. The reference feature vector corresponding to the feature vector output by the same layer is feature fused with the upsampling result to obtain the category of each pixel point.
7. The method for identifying sea ice for offshore photovoltaics according to claim 1, characterized in that: After determining the area where the sea ice is located, the method further includes: Determining input data based on sea ice parameters corresponding to the sea ice and external environmental parameters corresponding to the region; wherein the external environmental parameters include one or more of atmospheric parameters, ocean parameters, and forcing parameters; Performing a convolution operation on the input data to obtain local features, and performing downsampling and pooling processing on the local features to determine a feature vector; The eigenvectors are fully connected to determine the changing state of the sea ice.
8. A sea ice identification system for offshore photovoltaics, characterized in that: include: a feature extraction module, configured to obtain a sea surface image of a target area where the offshore photovoltaic site is located, and perform feature extraction on the sea surface image to determine image features of the sea surface image; a separate recognition module, configured to perform feature fitting on the image features to obtain a first sea ice recognition result, and perform coherence analysis based on the direct signal and the reflected signal corresponding to the target area to obtain a signal autocorrelation function to determine a second sea ice recognition result; a merging and identifying module, configured to determine a preliminary sea ice identification result by combining the sea surface image and the signal autocorrelation function when the first sea ice identification result and the second sea ice identification result are different; The area determination module is used to perform fine identification on the sea surface image to determine the area where the sea ice is located when the preliminary sea ice identification result is sea ice.
9. A computer program product, characterized in that When the computer program is executed by a processor, the sea ice identification method for offshore photovoltaics according to any one of claims 1 to 7 is implemented.
10. An electronic device, characterized in that: include: processor; a memory for storing executable instructions of the processor; The processor is configured to implement the sea ice identification method for offshore photovoltaics according to any one of claims 1 to 7 by executing the executable instructions.