Wind power blade orientation angle identification method and system based on rotating frame detection

Abstract

The invention discloses a wind power blade orientation angle identification method and system based on rotating frame detection. The method comprises the following steps: acquiring a wind power blade image; detecting the cabin and the blade in the image by using a rotating frame detection algorithm to obtain rotating frames of the cabin and the blade and angles of the rotating frames; the rotating frame detection algorithm adopts an LSKnet model fused with a PSC technology; determining a cabin center and a blade center according to the position relationship between the cabin and the blade rotating frame; taking the center of the cabin as an original point and the center of the blade as an end point to obtain an orientation angle and judge the quadrant of the orientation angle; and calculating the orientation angle of the blade according to the quadrant where the judged orientation angle is located and the angle of the rotating frame. According to the invention, accurate and efficient identification of the orientation angle of the wind power blade can be realized.

Description

Technical Field

[0001] The present application relates to the technical field of image processing and deep learning, and specifically to a method and system for identifying the orientation angle of a wind turbine blade based on rotating frame detection. Background Art

[0002] With the continuous development of renewable energy, wind turbine blades have attracted widespread attention as a clean and efficient form of energy. In the operation and maintenance of wind turbine blades, it is of great significance to accurately identify and monitor the orientation angle of wind turbine blades. Traditional recognition methods are usually based on manual feature extraction, but this method has limitations when dealing with complex backgrounds and rotating targets.

[0003] In recent years, the rapid development of deep learning technology has provided a new solution for rotating target detection. Compared with traditional methods, deep learning transfers the detection method of general targets to rotating target detection, solves the limitations of horizontal box marking targets, introduces angle information into the neural network, adds branch angle prediction, rotates the original horizontal box according to the predicted angle to fit the actual target, and improves the detection ability and speed of rotating targets. However, when applying deep learning to the recognition of the orientation angle of wind turbine blades, there are still some challenges, such as the discontinuity of the loss function caused by the periodicity of the angle, the 180-degree interval of the angle range of the rectangular symmetry of the detection box, the inability to recognize the orientation of the wind turbine blade target within a 360-degree range, and the impact of image quality on the detection results.

[0004] Therefore, how to combine deep learning and rotating target detection technology to achieve accurate and efficient identification of the orientation angle of wind turbine blades has become a problem that needs to be solved urgently. Summary of the invention

[0005] In order to achieve accurate and efficient recognition of the orientation angle of a wind turbine blade, the present application provides a method and system for identifying the orientation angle of a wind turbine blade based on rotating frame detection.

[0006] In a first aspect, the present application provides a method for identifying the orientation angle of a wind turbine blade based on rotating frame detection, comprising: Collect wind turbine blade images; The nacelle and blades in the image are detected by using a rotation frame detection algorithm to obtain the rotation frames and rotation frame angles of the nacelle and blades; the rotation frame detection algorithm adopts an LSKnet model integrating PSC technology; According to the positional relationship between the nacelle and the blade rotation frame, the nacelle center and the blade center are determined; with the nacelle center as the origin and the blade center as the end point, the orientation angle is obtained and the quadrant in which the orientation angle is located is determined; according to the quadrant in which the determined orientation angle is located, combined with the rotation frame angle, the blade orientation angle is calculated.

[0007] By adopting the above scheme, a deep learning model is used, preferably the LSKnet model, to achieve fast and accurate recognition of the orientation angle of wind turbine blades. At the same time, a phase-shifting coder (PSC) is used to perform phase shift encoding and decoding on the feature information in the image data to solve the discontinuity problem of the loss function caused by the periodicity of the angle, thereby improving the recognition effect of the orientation angle of the wind turbine blades. The wind turbine blade target 360-degree orientation recognition is achieved by relying on the judgment of the center distance relationship of the nacelle blade rotation frame, thereby further improving the accuracy of angle recognition.

[0008] Preferably, it also includes: Real-time viewing of the wind turbine blades' orientation angle changing from 0 to 360 degrees at a certain rate under the control of the wind turbine blade control system, and counting the time required for the blade orientation angle to change from 0 to 360 degrees at a certain rate, recorded as the standard time; When the orientation angle of the wind turbine blade changes cyclically from 0 to 360 degrees at a certain rate under the control of the wind turbine blade control system, an image of the wind turbine blade at a certain moment is collected at random, and the orientation angle of the blade is detected and calculated using a rotating frame detection algorithm, which is recorded as a first blade orientation angle; an image of the wind turbine blade after a standard time interval from the certain moment is collected, and the orientation angle of the blade is detected and calculated using a rotating frame detection algorithm, which is recorded as a second blade orientation angle; Compare whether the error between the first blade orientation angle and the second blade orientation angle is greater than the preset error. If greater, it is determined that there is a deviation in the blade orientation angle result, and the LSKnet model integrated with PSC technology is optimized and trained until the error between the first blade orientation angle and the second blade orientation angle detected and calculated by the LSKnet model integrated with PSC technology after optimization training is no greater than the preset error, and stop the optimization training.

[0009] By adopting the above scheme, the time required for the orientation angle of the blade to change from 0 degrees to 360 degrees is used to verify whether the orientation angle of the blade detected in the same time period is correct, and then the optimization of the LSKnet model integrating PSC technology is completed to obtain more accurate angle recognition results.

[0010] Preferably, it also includes: The historical wind turbine blade images under different wind speed, wind direction and illumination conditions are used to train the LSKnet model integrated with PSC technology.

[0011] By adopting the above solution, considering the influence of different environmental conditions on the image acquisition of wind turbine blades, training data under different natural conditions are collected to effectively improve the performance of the rotation frame detection algorithm in various practical scenarios.

[0012] Preferably, it also includes: The wind turbine blade control system automatically adjusts the orientation angle of the blade according to the real-time wind direction data, wind speed data and light data, and calculates the orientation angle change of the blade orientation angle in the time series, which is recorded as the first blade orientation angle change sequence; The wind turbine blade image is collected in real time, and the real-time orientation angle of the blade is detected and calculated using a rotating frame detection algorithm, and the orientation angle change of the blade orientation angle in a time series is obtained, which is recorded as a second orientation angle change sequence; Compare the first direction angle change sequence with the second direction angle change sequence. If the difference between the first direction angle change value and the second direction angle change value within a continuous preset time is greater than the preset direction angle change value, add wind turbine blade images collected under the same conditions as the corresponding wind direction data, wind speed data and light data within the continuous preset time to optimize the training of the LSKnet model integrating PSC technology.

[0013] By adopting the above scheme, the control system can obtain the control of the orientation angle of the wind turbine blades in the real-time environment, determine the dynamic change of the orientation angle of the wind turbine blades in the time series, compare it with the orientation angle of the wind turbine blades obtained by actual detection, complete the continuous tracking and identification of the orientation angle of the wind turbine blades, determine that the detection accuracy of the LSKnet model integrating PSC technology is low in some environments, and increase the training of related environments.

[0014] Preferably, it also includes: Collecting environmental data, and comparing the collected environmental data with a corresponding preset threshold of environmental data; the environmental data includes: light intensity, wind speed and wind force; If the number of types of collected environmental data is greater than the preset threshold of the environmental data set for the corresponding type, it is judged that the current environmental data is in a complex environment, and based on the rotating box detection algorithm, a two-stage rotating target detection method is used to detect the cabin and blades in the image; otherwise, it is judged that the current environment is not a complex environment, and based on the rotating box detection algorithm, a single-stage rotating target detection method is used to detect the cabin and blades in the image.

[0015] By adopting the above scheme, considering the increased difficulty of detecting the orientation angle image of the wind turbine blades under complex environmental conditions, a dual-stage rotating target detection method with greater accuracy is adopted for detection to achieve more efficient and accurate recognition.

[0016] Preferably, it also includes: Image processing technology is used to detect whether the blades in the wind turbine blade image are complete. If not, the wind turbine blade images collected at different angles are replaced until the blades in the wind turbine blade image are detected to be complete.

[0017] By adopting the above solution, wind turbine blade images at multiple angles are collected in real time to realize the identification of the orientation angle of a wind turbine blade with a complete blade.

[0018] Preferably, it also includes: The collected image is preprocessed, and the preprocessing includes: adjusting the brightness and contrast of the image.

[0019] By adopting the above solution, image acquisition is optimized to improve image quality, thereby improving the recognition effect of the orientation angle of the wind turbine blade.

[0020] In a second aspect, the present application provides a wind turbine blade orientation angle recognition system based on rotating frame detection, comprising: An image acquisition module, used for acquiring images of wind turbine blades; An image detection module is used to detect the cabin and blades in the image using a rotation frame detection algorithm to obtain the rotation frames and rotation frame angles of the cabin and blades; the rotation frame detection algorithm adopts an LSKnet model that integrates PSC technology; The heading angle acquisition module is used to determine the center of the cabin and the center of the blade according to the positional relationship between the cabin and the blade rotation frame; with the cabin center as the origin and the blade center as the end point, the heading angle is acquired and the quadrant in which the heading angle is located is determined; based on the quadrant in which the determined heading angle is located and the angle of the rotation frame, the heading angle of the blade is calculated.

[0021] By adopting the above scheme, a method based on rotating frame detection is used in combination with deep learning technology to automatically and accurately identify the orientation angle of wind turbine blades, thereby improving the recognition accuracy and efficiency.

[0022] In a third aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the method as described above.

[0023] In a fourth aspect, the present application provides a computer device, comprising a memory, a processor, and a program stored and executable on the memory, wherein the program implements the steps of the above method when executed by the processor.

[0024] In summary, this application has the following beneficial effects: 1. A method based on rotating frame detection is adopted in combination with deep learning technology, and the LSKnet model is preferably used to achieve fast and accurate recognition of the orientation angle of wind turbine blades. At the same time, the phase-shifting coder (PSC) is used to solve the discontinuity problem of the loss function caused by the periodicity of the angle, further improving the accuracy of wind turbine blade orientation angle recognition.

[0025] 2. Comprehensively consider the actual environment conditions, collect wind turbine blade images under various environmental conditions for training, and improve the performance of model detection; and select the dual-stage rotating target detection method or the single-stage rotating target detection method for image detection according to the actual environment conditions, by balancing the detection speed and accuracy, and optimizing the image quality of image acquisition, thereby improving the recognition effect of the wind turbine blade orientation angle; 3. Based on the actual real-time adjustment of the orientation angle of the wind turbine blades or the orientation angle of the wind turbine blades collected by the rotation angle sensor, the orientation angle of the wind turbine blades detected and calculated by the model is verified and tracked in real time; based on the verification results and real-time tracking results, the model is further optimized and trained to further improve the detection accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a flow chart of a method for identifying the orientation angle of a wind turbine blade based on rotating frame detection in a specific embodiment; Figure 2 It is a diagram showing the result of identification using the wind turbine blade orientation angle identification method based on rotating frame detection in a specific embodiment; Figure 2 (a) Figure 2 (b) and Figure 2 (c) is a diagram of the rotating frame detection of different wind turbine nacelle blades; Figure 2 (d) is the optimized model detection and Figure 2 (a) Illustration of the rotating frame detection of the blades of the same wind turbine nacelle; Figure 3 It is a structural schematic diagram of a wind turbine blade orientation angle recognition system based on rotating frame detection described in a specific embodiment. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0028] like Figure 1 As shown, the embodiment of the present application discloses a method for identifying the orientation angle of a wind turbine blade based on rotating frame detection, and the specific steps are as follows: S1. Collect wind turbine blade images and perform preprocessing.

[0029] Specifically, a drone is used to collect images of multiple wind turbine blades.

[0030] In order to ensure that the collected wind turbine blade images have complete blades as much as possible and to ensure accurate identification of the blade orientation angle, multiple drones are selected to collect wind turbine blade images at multiple angles. Image processing technology is used to detect whether the blades in the wind turbine blade images are complete. If not, the wind turbine blade images collected at different angles are replaced until the blades in the wind turbine blade images are detected to be complete and the replacement is stopped.

[0031] In this embodiment, blade edge detection is performed based on a traditional image processing algorithm (such as Houqh transformation, etc.), and whether the wind turbine blade in the image is complete is determined based on the edge contour.

[0032] The preprocessing of the wind turbine blade image includes: image cleaning, image enhancement, etc.; the image enhancement includes: adjusting the brightness and contrast of the image, etc., to improve the quality and clarity of the image.

[0033] S2. Detect the cabin and blades in the preprocessed image using a rotating frame detection algorithm.

[0034] Specifically, on the one hand, considering the diversity of the background where the wind turbine blades are located, in order to reduce additional environmental information as much as possible, the LSKnet model is selected for rotating frame detection; on the other hand, considering the problem of discontinuity of the loss function caused by the angle periodicity in rotating target detection, which leads to inaccurate detection results, in order to improve the accuracy of the detection results in this case, the PSC technology is used and integrated into the LSKnet model to perform phase shift encoding on the model features and decode the model output to obtain the accurate rotating frame angle. Therefore, the rotating frame detection algorithm uses the LSKnet model integrated with the PSC technology.

[0035] The input of the LSKnet model is the image of the preprocessed wind turbine blade, and the output is the rotation frame and the angle of the rotation frame of the nacelle and blade in the wind turbine blade image; the LSKnet model is trained and generated by annotating the images of historical wind turbine blades with rotation frame parameters (including the length, width, center coordinates and rotation angle of the rotation frame) in the nacelle and blade, wherein the rotation angle is expressed in the form of a phase shift cosine value after phase shift encoding of the rotation angle using the PSC technology; the angle of the rotation frame output by the LSKnet model is also expressed in the form of a phase shift cosine value, which is encoded using the PSC technology to finally obtain the angle of the rotation frame.

[0036] The preprocessed image is input into the LSKnet model integrated with PSC technology to obtain the detection result.

[0037] S3. Determine whether the nacelle and blades exist in the preprocessed image.

[0038] Specifically, the detection result of the LSKnet model of the preprocessed image fusion PSC technology is obtained, and it is determined whether the cabin and blades exist in the preprocessed image according to the detection result. If not, other preprocessed images are input into the model for further detection; if so, the rotation frame and the angle of the rotation frame of the cabin and blade are obtained.

[0039] S4. Determine the quadrant in which the orientation angle of the formed vector is located according to the positional relationship between the nacelle and the blade rotation frame.

[0040] Considering that the rotation frame detection relies on the 135-degree definition of the long side, since the angle range of rectangular symmetry is spaced 180 degrees apart, a single rotation frame detection cannot identify the orientation of the wind turbine blade target within a 360-degree range. Therefore, according to the positional relationship between the nacelle and the blade rotation frame, the nacelle center and the blade center are determined to determine the nacelle center as the origin and the blade center as the end point. The current orientation angle is determined based on the vector formed by the line connecting the two points, that is, the angle between the line connecting the two points and the horizontal plane, and the quadrant in which the orientation angle is located is determined. Among them, the judgment of the quadrant can be based on artificial settings, such as the quadrant in the coordinate system with the Y axis positive marked as 0° and 360° clockwise; or the quadrant in the coordinate system with the X axis positive marked as 0° and 360° clockwise.

[0041] S5. Determine the quadrant where the orientation angle of the formed vector is located according to the positional relationship between the nacelle and the blade rotation frame, and calculate the orientation angle of the blade in combination with the rotation frame angle.

[0042] Specifically, the 0°-360° range corresponding to the specific orientation angle of the blade is determined by the quadrant where the orientation angle is located. Figure 2 As shown in (c), the x-axis is marked as 0°, and the quadrant of the coordinate system set clockwise 360° is the first quadrant. The angle of the rotation box is the specific angle of the orientation. Figure 2 As shown in (c), the angle of the rotating frame obtained based on the model is 42°, and the orientation angle of the blade is finally determined to be 42°.

[0043] like Figure 2 As shown in FIG. 1 , the rotating frame detection diagram of different wind turbine nacelle blades is obtained by using the above method. Figure 2 As shown in (a), the confidence level of wind turbine cabin identification is 99.99%, and the confidence level of wind turbine blade identification is 99.92%. Figure 2 As shown in (b), the confidence level of wind turbine cabin identification is 99.98%, and the confidence level of wind turbine blade identification is 99.88%. Figure 2 As shown in (c), the confidence level of the wind turbine nacelle being identified is 99.99%, and the confidence level of the wind turbine blades being identified is 99.92%.

[0044] In a specific embodiment, in order to further ensure the accuracy of the specific orientation angle obtained by using the model detection and final calculation, the rotation box detection model is further optimized by verifying the accuracy of the model detection result, and the method further includes: By using the display device of the wind turbine blade control system, it is possible to view in real time the cyclic change of the orientation angle of the wind turbine blade from 0 degrees to 360 degrees at a certain rate under the control of the wind turbine blade control system, and to count the time required for the orientation angle of the blade to change from 0 degrees to 360 degrees at a certain rate, which is recorded as the standard time; the certain time is a manually preset time.

[0045] When the orientation angle of the wind turbine blade changes cyclically from 0 to 360 degrees at a certain rate under the control of the wind turbine blade control system, randomly collect the wind turbine blade image at a certain moment, use the rotating frame detection algorithm to detect and calculate the orientation angle of the blade, and record it as the orientation angle of the first blade; collect the wind turbine blade image after a standard time interval from the certain moment, use the rotating frame detection algorithm to detect and calculate the orientation angle of the blade, and record it as the orientation angle of the second blade; Taking into account that the orientation angles of the blades corresponding to the wind turbine blades are cyclically changing, the orientation angles of the blades corresponding to the standard time should be consistent or have only a small error. Therefore, it is determined whether the error between the orientation angle of the first blade and the orientation angle of the second blade is greater than the preset error; if greater, it is determined that there is a deviation in the orientation angle result of the blade, and the LSKnet model integrated with the PSC technology is optimized and trained, and the LSKnet model integrated with the PSC technology after the optimized training is used to re-obtain the error between the orientation angle of the first blade and the orientation angle of the second blade, and compare the calculated error with the preset error, until the error detected and calculated by the LSKnet model integrated with the PSC technology after the optimized training is no greater than the preset error, and the optimization training is stopped.

[0046] Among them, optimization training includes adjusting model parameters or optimizing model structure; for optimization and improvement of LSKnet model, more advanced network structures (such as residual network, attention mechanism, etc.) can be introduced to improve the efficiency and accuracy of feature extraction. For example, in LSKnet model, residual network structure is introduced by replacing the original convolutional layer with residual block, or in LSKnet model, attention mechanism is introduced after convolutional layer in feature extraction stage.

[0047] like Figure 2 (d) shows that compared with Figure 2 For (a), the confidence level of wind turbine cabin identification is 99.99%, and the confidence level of wind turbine blade identification is 99.93%. The accuracy of the detection results using the optimized model can be improved.

[0048] In a specific embodiment, considering that model detection may be affected by environmental factors, in order to avoid the influence of environmental factors, it is possible to select to collect training data in different environmental scenarios for training, so as to improve the detection accuracy of the model in the corresponding environmental scenarios. The method further includes: The historical wind turbine blade images under different wind speed, wind direction and lighting conditions are used to train the LSKnet model integrated with PSC technology, so as to improve the detection performance of the LSKnet model integrated with PSC technology.

[0049] On the basis of the above training, different environmental factors are further considered to have different influences on the accuracy of model detection. For example, complex environments have a greater influence on the accuracy of model detection. The accuracy of model detection in different actual environments can be combined to determine which specific environment training data to add to optimize model performance. The method also includes: The wind turbine blade control system is obtained in real time to automatically adjust the orientation angle of the blade according to the real-time meteorological data, and calculate the orientation angle change of the blade in a time series, which is recorded as the first blade orientation angle change sequence; the meteorological data includes wind direction data, wind speed data, wind force data, light data, wind speed and wind force duration, and rainfall data, etc.

[0050] The wind turbine blade image is collected in real time, and the real-time orientation angle of the blade is detected and calculated using a rotating frame detection algorithm, and the orientation angle change of the blade orientation angle in a time series is obtained, which is recorded as a second orientation angle change sequence; Compare the first direction angle change sequence with the second direction angle change sequence. If the difference between the first direction angle change value and the second direction angle change value within a continuous preset time is greater than the preset direction angle change value, add wind turbine blade images collected under the same conditions as the corresponding meteorological data within the continuous preset time to optimize the LSKnet model integrating PSC technology.

[0051] In a specific embodiment, considering that the more complex the environmental conditions, the lower the clarity of the captured image, which in turn increases the difficulty of detecting the rotating frame of the captured image, it is necessary to provide a more accurate detection method; the method further includes: Collect environmental data, and compare the collected environmental data with the corresponding preset thresholds of environmental data; the environmental data include: meteorological data such as light intensity, wind speed and wind force, and wind speed and wind force duration; the corresponding preset thresholds of environmental data include: light intensity preset threshold, wind speed preset threshold, wind force preset threshold, a certain wind speed duration preset threshold, and a certain wind force duration preset threshold.

[0052] If the number of types of collected environmental data is greater than the preset threshold of the corresponding type of environmental data, it is judged that the current environmental data is in a complex environment. Based on the rotating frame detection algorithm, a two-stage rotating target detection method is used to detect the cabin and blades in the image. In the first stage, the feature extraction and target recognition capabilities of the LSKnet model are used to generate a series of candidate areas in the image; in the second stage, the wind turbine cabin and blades are identified in these candidate areas, and the rotating frame position information and rotation angle are output. Among them, the preset number is considered to be set, such as more than 3 types of environmental data (light, wind force, wind speed) are greater than the corresponding set environmental data preset threshold.

[0053] Otherwise, based on the rotating frame detection algorithm, a single-stage rotating target detection method is used to detect the nacelle and blades in the image, that is, the wind turbine nacelle and blades are directly identified, and the rotating frame position information and rotation angle are output.

[0054] like Figure 3 As shown, the embodiment of the present application discloses a wind turbine blade orientation angle recognition system based on rotating frame detection, which specifically includes: An image acquisition module 101 is used to acquire wind turbine blade images; An image detection module 102 is used to detect the nacelle and blades in the image using a rotation frame detection algorithm to obtain the rotation frames and rotation frame angles of the nacelle and blades; the rotation frame detection algorithm uses an LSKnet model that integrates the PSC technology; The heading angle acquisition module 103 is used to determine the center of the cabin and the center of the blade according to the positional relationship between the cabin and the blade rotation frame; with the cabin center as the origin and the blade center as the end point, the heading angle is acquired and the quadrant in which the heading angle is located is determined; based on the quadrant in which the determined heading angle is located and the angle of the rotation frame, the heading angle of the blade is calculated.

[0055] The system further comprises: The image detection optimization module 104 is used to check in real time that the orientation angle of the wind turbine blade changes cyclically from 0 to 360 degrees at a certain rate under the control of the wind turbine blade control system, and to count the time required for the orientation angle of the blade to change from 0 to 360 degrees at a certain rate, which is recorded as the standard time; when the orientation angle of the wind turbine blade changes cyclically from 0 to 360 degrees at a certain rate under the control of the wind turbine blade control system, collect the wind turbine blade image at a certain moment at random, use the rotating frame detection algorithm to detect and calculate the orientation angle of the blade, which is recorded as the first blade orientation angle; collect the wind turbine blade image after the standard time interval from the certain moment, use the rotating frame detection algorithm to detect and calculate the orientation angle of the blade, which is recorded as the second blade orientation angle; judge whether the error between the first blade orientation angle and the second blade orientation angle is greater than a preset error, if so, it is determined that there is a deviation in the orientation angle result of the blade, and optimize the LSKnet model integrated with the PSC technology until the error between the first blade orientation angle and the second blade orientation angle detected and calculated by the LSKnet model integrated with the PSC technology after the optimization training is not greater than the preset error, and stop the optimization training.

[0056] The image detection optimization module 104 is also used to use historical wind turbine blade images under different wind speed, different wind direction and different lighting conditions, and use the collected historical wind turbine blade images under different combination scenes to train the LSKnet model integrating the PSC technology; it is also used to obtain in real time that the wind turbine blade control system automatically adjusts the orientation angle of the blade according to the real-time wind direction data, wind speed data and lighting data, and calculates the orientation angle change of the blade orientation angle in the time series, which is recorded as the first blade orientation angle change sequence; collect wind turbine blade images in real time, use the rotating frame detection algorithm to detect and calculate the real-time orientation angle of the blade, and obtain the orientation angle change of the blade orientation angle in the time series, which is recorded as the second orientation angle change sequence; compare the first orientation angle change sequence with the second orientation angle change sequence, if there is a difference between the first orientation angle change value and the second orientation angle change value within a continuous preset time that is greater than the preset orientation angle change value, then add wind turbine blade images collected under the same conditions as the corresponding wind direction data, wind speed data and lighting data within the continuous preset time to optimize the LSKnet model integrating the PSC technology.

[0057] The image detection module 102 is also used to collect environmental data and compare the collected environmental data with the corresponding preset threshold of environmental data; the environmental data includes: light intensity, wind speed and wind force, and wind speed and wind force duration; if the number of types of various environmental data collected is greater than the number of types of environmental data preset thresholds set for corresponding types, it is judged that the current environmental data is in a complex environment, and based on the rotating frame detection algorithm, a two-stage rotating target detection method is used to detect the cabin and blades in the image; otherwise, it is judged that the current environment is not a complex environment, and based on the rotating frame detection algorithm, a single-stage rotating target detection method is used to detect the cabin and blades in the image.

[0058] The image acquisition module 101 is also used to use image processing technology to detect whether the blades in the wind turbine blade image are complete. If not, the wind turbine blade images acquired at different angles are replaced until the blades in the wind turbine blade image are detected to be complete; the acquired images are preprocessed, and the preprocessing includes: adjusting the brightness and contrast of the image.

[0059] The embodiment of the present application also discloses a computer-readable storage medium.

[0060] Specifically, the computer-readable storage medium stores a computer program that can be loaded by a processor and executes the above-mentioned wind turbine blade orientation angle identification method based on rotating frame detection. The computer-readable storage medium includes, for example: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0061] The embodiment of the present application also discloses a computer device.

[0062] Specifically, the computer device includes a memory and a processor, and the memory stores a computer program that can be loaded by the processor and execute the above-mentioned method for identifying the orientation angle of a wind turbine blade based on rotating frame detection.

[0063] The above are all preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Any feature disclosed in this specification (including the abstract and drawings), unless otherwise stated, can be replaced by other equivalent or alternative features with similar purposes. That is, unless otherwise stated, each feature is only an example of a series of equivalent or similar features.

Claims

1. A method for identifying the orientation angle of a wind turbine blade based on rotating frame detection, characterized in that: include: Collect wind turbine blade images; The nacelle and blades in the image are detected using a rotation frame detection algorithm to obtain the rotation frames and rotation frame angles of the nacelle and blades; The rotation frame detection algorithm adopts the LSKnet model integrating PSC technology; According to the positional relationship between the nacelle and the blade rotation frame, the nacelle center and the blade center are determined; with the nacelle center as the origin and the blade center as the end point, the orientation angle is obtained and the quadrant in which the orientation angle is located is determined; according to the quadrant in which the determined orientation angle is located, combined with the rotation frame angle, the blade orientation angle is calculated.

2. The method for identifying the orientation angle of a wind turbine blade based on rotating frame detection according to claim 1, characterized in that: Also includes: Real-time viewing of the wind turbine blades' orientation angle changing from 0 to 360 degrees at a certain rate under the control of the wind turbine blade control system, and counting the time required for the blade orientation angle to change from 0 to 360 degrees at a certain rate, recorded as the standard time; When the orientation angle of the wind turbine blade changes cyclically from 0 to 360 degrees at a certain rate under the control of the wind turbine blade control system, an image of the wind turbine blade is collected at any time, and the orientation angle of the blade is detected and calculated using a rotating frame detection algorithm, which is recorded as a first blade orientation angle; Collect the wind turbine blade image after a standard time interval from a certain moment, and use the rotating frame detection algorithm to detect and calculate the orientation angle of the blade, which is recorded as the second blade orientation angle; Determine whether the error between the first blade orientation angle and the second blade orientation angle is greater than the preset error. If so, it is determined that there is a deviation in the obtained blade orientation angle result, and the LSKnet model integrating the PSC technology is optimized and trained until the error between the first blade orientation angle and the second blade orientation angle detected and calculated by the LSKnet model integrating the PSC technology after optimization training is no greater than the preset error, and stop the optimization training.

3. The method for identifying the orientation angle of a wind turbine blade based on rotating frame detection according to claim 1, characterized in that: Also includes: The historical wind turbine blade images under different wind speed, wind direction and illumination conditions are used to train the LSKnet model integrated with PSC technology.

4. The method for identifying the orientation angle of a wind turbine blade based on rotating frame detection according to claim 3 is characterized in that: Also includes: The wind turbine blade control system automatically adjusts the orientation angle of the blade according to the real-time wind direction data, wind speed data and light data, and calculates the orientation angle change of the blade orientation angle in the time series, which is recorded as the first blade orientation angle change sequence; The wind turbine blade image is collected in real time, and the real-time orientation angle of the blade is detected and calculated using a rotating frame detection algorithm, and the orientation angle change of the blade orientation angle in a time series is obtained, which is recorded as a second orientation angle change sequence; Compare the first direction angle change sequence with the second direction angle change sequence. If the difference between the first direction angle change value and the second direction angle change value within a continuous preset time is greater than the preset direction angle change value, add wind turbine blade images collected under the same conditions as the corresponding wind direction data, wind speed data and light data within the continuous preset time to optimize the training of the LSKnet model integrating PSC technology.

5. The method for identifying the orientation angle of a wind turbine blade based on rotating frame detection according to claim 1, characterized in that: Also includes: Collect environmental data, and compare the collected environmental data with a corresponding preset threshold of environmental data; The environmental data include: light intensity, wind speed and wind force; If the number of types of collected environmental data is greater than the preset threshold of the environmental data set for the corresponding type, it is judged that the current environmental data is in a complex environment, and based on the rotating box detection algorithm, a two-stage rotating target detection method is used to detect the cabin and blades in the image; otherwise, it is judged that the current environment is not a complex environment, and based on the rotating box detection algorithm, a single-stage rotating target detection method is used to detect the cabin and blades in the image.

6. The method for identifying the orientation angle of a wind turbine blade based on rotating frame detection according to claim 1, characterized in that: Also includes: Image processing technology is used to detect whether the blades in the wind turbine blade image are complete. If not, the wind turbine blade images collected at different angles are replaced until the blades in the wind turbine blade image are detected to be complete.

7. The method for identifying the orientation angle of a wind turbine blade based on rotating frame detection according to claim 1, characterized in that: Also includes: The collected image is preprocessed, and the preprocessing includes: adjusting the brightness and contrast of the image.

8. A wind turbine blade orientation angle recognition system based on rotating frame detection, characterized in that: include: An image acquisition module, used for acquiring images of wind turbine blades; An image detection module is used to detect the cabin and blades in the image using a rotation frame detection algorithm to obtain the rotation frames and rotation frame angles of the cabin and blades; the rotation frame detection algorithm adopts an LSKnet model that integrates PSC technology; The heading angle acquisition module is used to determine the center of the cabin and the center of the blade according to the positional relationship between the cabin and the blade rotation frame; with the cabin center as the origin and the blade center as the end point, the heading angle is acquired and the quadrant in which the heading angle is located is determined; based on the quadrant in which the determined heading angle is located and the angle of the rotation frame, the heading angle of the blade is calculated.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the method according to any one of claims 1 to 7.

10. A computer device, characterized in that: The computer device comprises a memory, a processor and a program stored and executable on the memory, and the program implements the steps of the method according to any one of claims 1 to 7 when executed by the processor.

Images