Method and system for detecting surface defects of pultrusion plate for wind power blade

Abstract

The invention provides a method and system for detecting surface defects of a pultrusion plate for a wind power blade, and relates to the technical field of image processing, and the method comprises the steps: obtaining a cross stroboscopic image of the surface of the pultrusion plate under a reflection field and a backlight field; splitting the image to obtain a reflection field image and a backlight field image; constructing a reflection field data set based on the reflection field image, and constructing a backlight field data set based on the backlight field image; constructing a visual detection model under the reflection field and the backlight field; training a model under the reflection field by using the reflection field data set, and training a model under the backlight field by using the backlight field data set; acquiring a cross stroboscopic image of the surface of the pultrusion plate to be detected in a reflection field and a backlight field; splitting the image to obtain a reflection field image and a backlight field image to be detected; inputting a to-be-detected reflection field image into the model under the reflection field to obtain a reflection field detection result, and inputting a to-be-detected backlight field image into the model under the backlight field to obtain a backlight field detection result; and obtaining a final detection result according to the reflection field and backlight field detection results.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a method and system for detecting surface defects of a pultruded plate material for a wind turbine blade. Background Art

[0002] Pultruded sheet is a composite sheet made by pultrusion process after impregnating glass reinforced fiber with resin. It has a specific cross-sectional shape and continuous length. Due to its excellent mechanical properties and stable quality, pultruded sheet is widely used in the manufacture of wind turbine blades. However, during the production process, pultruded sheet may have various defects, such as bubbles, raw yarn oxidation, inclusions, cracking, and dry yarn. These defects may change during the operation of the wind system, which will seriously affect the normal operation of the wind system and cause significant economic losses and safety risks. Therefore, it is very important to carry out effective defect detection on pultruded sheet.

[0003] In the existing technical solutions, a variety of technologies for pultruded sheet defect detection have been proposed, including image-based detection methods and non-image-based detection methods. These technologies aim to reduce manual intervention and improve the detection accuracy of small target defects.

[0004] However, existing image-based defect detection methods do not fully consider the translucent properties of pultruded sheets for wind turbine blades, which makes it difficult to effectively extract some internal defects under traditional reflected light fields, thus affecting the accuracy of detection. At the same time, although non-image defect detection methods can provide more comprehensive defect information, they also face problems such as noise interference, high equipment and maintenance costs, high technical requirements for operators, and high technical difficulty. Summary of the invention

[0005] In order to solve the problem that traditional image-based defect detection methods do not fully consider the translucent characteristics of pultruded plates for wind turbine blades, which makes it difficult to effectively extract some internal defects under traditional reflected light fields, thereby affecting the accuracy of detection, and at the same time, although non-image defect detection methods can provide more comprehensive defect information, they also face technical problems such as noise interference, high equipment and maintenance costs, high technical requirements for operators, and high technical difficulty, the present invention provides a method and system for detecting surface defects of pultruded plates for wind turbine blades.

[0006] The technical solution provided by the embodiment of the present invention is as follows:

[0007] First aspect:

[0008] A method for detecting surface defects of a pultruded plate material for a wind turbine blade provided in an embodiment of the present invention comprises:

[0009] S1: Obtain cross-stroboscopic images of the pultruded sheet surface in the reflection field and the backlight field;

[0010] S2: Splitting the cross-stroboscopic image to obtain a reflection field image and a backlight field image;

[0011] S3: constructing a reflection field dataset based on the reflection field image, and constructing a backlight field dataset based on the backlight field image;

[0012] S4: Deep learning-based visual detection models built in reflective field and backlit field respectively;

[0013] S5: using the reflection field data set to train the visual detection model in the reflection field, and using the backlight field data set to train the visual detection model in the backlight field;

[0014] S6: Acquire cross-stroboscopic images of the surface of the pultruded sheet to be tested in the reflection field and the backlight field;

[0015] S7: splitting the cross-stroboscopic image of the surface of the pultruded plate to be detected to obtain a reflection field image to be detected and a backlight field image to be detected;

[0016] S8: inputting the reflection field image to be detected into a trained visual detection model under the reflection field for detection, and obtaining a reflection field detection result; inputting the backlight field image to be detected into a trained visual detection model under the backlight field for detection, and obtaining a backlight field detection result;

[0017] S9: Obtaining a final detection result according to the reflection field detection result and the backlight field detection result.

[0018] Second aspect:

[0019] A system for detecting surface defects of a pultruded plate material for a wind turbine blade provided in an embodiment of the present invention includes: a memory and one or more processors;

[0020] One or more application programs are stored in the memory, and the one or more application programs are suitable for being executed by the one or more processors to implement the above-mentioned method for detecting surface defects of pultruded sheet materials for wind turbine blades.

[0021] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0022] In the present invention, by obtaining cross-stroboscopic images of the surface of the pultruded sheet in the reflection field and the backlight field, and splitting the cross-stroboscopic images to obtain the reflection field image and the backlight field image, the translucent characteristics of the pultruded sheet for wind turbine blades are fully considered, which is helpful for effectively extracting internal defects under the traditional reflection light field. By constructing deep learning-based visual detection models in the reflection field and the backlight field respectively, the defect detection of the pultruded sheet no longer faces the problems of noise interference, high equipment and maintenance costs, high technical requirements for operators and high technical difficulty. By constructing a reflection field data set based on the reflection field image, and a backlight field data set based on the backlight field image, and using the reflection field data set to train the visual detection model in the reflection field, and using the backlight field data set to train the visual detection model in the backlight field, the accuracy of surface defect detection of the pultruded sheet is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0024] Figure 1 A schematic flow chart of a method for detecting surface defects of a pultruded plate material for a wind turbine blade provided by an embodiment of the present invention;

[0025] Figure 2 A schematic diagram of a multi-field imaging device for the surface of a pultruded plate material for a wind turbine blade provided by an embodiment of the present invention

[0026] Figure 3 A schematic structural diagram of a system for detecting surface defects of a pultruded plate material for wind turbine blades provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0027] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0028] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0029] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0030] Reference Manual Attached Figure 1 , showing a schematic flow chart of a method for detecting surface defects of a pultruded plate material for a wind turbine blade provided by an embodiment of the present invention.

[0031] The embodiment of the present invention provides a method for detecting surface defects of a pultruded sheet material for a wind turbine blade. The method can be implemented by a device for detecting surface defects of a pultruded sheet material for a wind turbine blade. The device for detecting surface defects of a pultruded sheet material for a wind turbine blade can be a terminal or a server. The processing flow of the method for detecting surface defects of a pultruded sheet material for a wind turbine blade can include the following steps:

[0032] S1: Obtain cross-stroboscopic images of the pultruded sheet surface in the reflection field and the backlight field.

[0033] Reference Manual Attached Figure 2 , showing a schematic structural diagram of a multi-field imaging device for the surface of a pultruded plate material for a wind turbine blade provided by an embodiment of the present invention.

[0034] In a possible implementation manner, S1 specifically includes:

[0035] A time-sharing stroboscopic strategy based on encoder trigger signals is used for alternate acquisition to obtain cross-stroboscopic images of the pultruded sheet surface in the reflection field and the backlight field.

[0036] Specifically, two light sources at different angles and one linear array camera are installed on the upper and lower surfaces of the pultruded sheet, respectively, for a total of four light sources and two linear array cameras. The installation angles and brightness of one linear array camera and two light sources at different angles corresponding to the upper and lower surfaces of the pultruded sheet are adjusted respectively, so as to form a reflection field and a backlight field on the upper surface of the pultruded sheet and a reflection field and a backlight field on the lower surface of the pultruded sheet, respectively. A time-sharing stroboscopic strategy based on an encoder trigger signal is adopted, and the linear array camera is triggered by the unidirectional movement of the pultruded sheet, so that the linear array camera and the light source on the same surface cooperate to perform a rapid alternating acquisition action of the backlight field image and the reflection field image, thereby obtaining cross-stroboscopic images of the upper and lower surfaces of the pultruded sheet under the reflection field and the backlight field.

[0037] It should be noted that the light sources on the same surface of the pultruded sheet are a reflected light source and a backlight light source. The distances between the reflected light source and the backlight light source and the surface of the pultruded sheet are 12 cm and 3 cm respectively, and the acute angles formed between them and the surface of the pultruded sheet are 70° and 80° respectively.

[0038] It should be noted that the time-sharing strobe strategy based on the encoder trigger signal is:

[0039] A roller-encoder structure is installed to trigger the light source strobe signal and the camera acquisition signal when the pultruded sheet moves in one direction.

[0040] Specifically, the pultruded sheet enters from the equipment entrance and extends out through the equipment exit, and begins to move unidirectionally inside the equipment. The movement rotates the roller, which in turn triggers the encoder to send a trigger signal, causing the four light sources to cross-strobe. Through two linear array cameras, cross-strobe images of the upper and lower surfaces of the pultruded sheet in the reflection field and the backlight field are obtained respectively.

[0041] In the present invention, by installing light sources at different angles on the upper and lower surfaces of the pultruded sheet, images under different lighting conditions can be obtained, which helps to reveal surface defects more comprehensively. This setting can capture different defect characteristics under the reflection field and the backlight field, thereby improving the accuracy and comprehensiveness of defect detection. The time-sharing stroboscopic strategy allows the reflection field and the backlight field images to be quickly and alternately collected on the same surface, thereby improving the efficiency of image acquisition. This method can obtain the required images under multiple lighting conditions in a relatively short time, and is suitable for large-scale production and testing.

[0042] S2: Split the cross-stroboscopic image to obtain a reflection field image and a backlight field image.

[0043] In a possible implementation manner, S2 is specifically:

[0044] The cross-stroboscopic image is split using an image segmentation method based on grayscale threshold to obtain a reflection field image and a backlight field image.

[0045] In a possible implementation, a grayscale threshold-based image segmentation method is used to segment the cross-stroboscopic image to obtain a reflection field image and a backlight field image, specifically including:

[0046] The cross-stroboscopic image is gray-scaled to obtain a gray-scale image.

[0047] Get the global pixel grayscale value information of the grayscale image and determine the grayscale threshold.

[0048] It should be noted that those skilled in the art can set the grayscale threshold value according to actual needs, and the present invention does not limit this.

[0049] Use image rows as the splitting unit to split the grayscale image into multiple image rows.

[0050] Calculate the mean grayscale value of each image row.

[0051] The image rows whose row pixel grayscale mean is greater than the grayscale threshold are selected as the reflection field image rows, and the image rows whose row pixel grayscale mean is less than or equal to the grayscale threshold are selected as the backlight field image rows.

[0052] The screened reflection field image rows and backlight field image rows are spliced ​​according to the arrangement order of the image rows of the cross-stroboscopic image to obtain the reflection field image and the backlight field image respectively.

[0053] In the present invention, by setting the grayscale threshold, the image areas under the reflection field and the backlight field can be accurately distinguished. The flexible adjustment of the grayscale threshold enables the method to adapt to different lighting conditions and material surface characteristics, so as to more accurately extract the required image information. By effectively separating the images of the reflection field and the backlight field, the defect features under different lighting conditions can be extracted respectively. This method helps to discover and identify defects that may be overlooked under single lighting conditions, thereby improving the accuracy of overall detection. By calculating the grayscale mean of each image line and screening it, the image noise caused by ambient light or surface reflection can be effectively reduced. This helps to generate clearer reflection field and backlight field images, and enhances the reliability of subsequent defect detection.

[0054] In a possible implementation manner, after S2 and before S3, the process further includes:

[0055] The reflection field image and the backlight field image are cropped by using a region extraction method based on image science to obtain a reflection field detection region image and a backlight field detection region image.

[0056] In a possible implementation, a region extraction method based on image science is used to crop the reflection field image and the backlight field image to obtain a reflection field detection region image and a backlight field detection region image, specifically including:

[0057] The reflection field image is binarized to obtain a binary image.

[0058] Perform opening and dilation operations on the binary image to extract the black and white pixel areas in the binary image.

[0059] The cropping boundary coordinates are determined according to the coordinates of the dividing line between the black and white pixel areas.

[0060] According to the cropping boundary coordinates, the reflection field image and the backlight field image are cropped respectively to obtain the reflection field detection area image and the backlight field detection area image.

[0061] In the present invention, by cropping the image, retaining and extracting the truly valuable detection area, irrelevant information or interference factors in the image can be effectively reduced. This allows subsequent defect detection to focus more on key parts and improve the accuracy of detection. Through binarization processing and opening and dilation operations, noise or irregular areas in the image can be effectively removed. This process can make the image clearer, help accurately locate the defect area, and improve the overall detection efficiency.

[0062] S3: Based on the reflection field image, a reflection field dataset is constructed; based on the backlight field image, a backlight field dataset is constructed.

[0063] It should be noted that the number of images in the reflection field data set is greater than or equal to a first preset number, and the number of images in the backlight field data set is greater than or equal to a second preset number.

[0064] It should be noted that those skilled in the art can set the first preset number and the second preset number according to actual needs, and the present invention is not limited thereto.

[0065] Optionally, the number of images in the reflection field data set is greater than or equal to 1000, and the number of images in the backlight field data set is greater than or equal to 1000.

[0066] The reflection field data set is divided into a reflection field training data set and a reflection field test data set according to a first preset ratio, and the backlight field data set is divided into a backlight field training data set and a backlight field test data set according to a second preset ratio.

[0067] It should be noted that those skilled in the art can set the sizes of the first preset ratio and the second preset ratio according to actual needs, and the present invention is not limited thereto.

[0068] Optionally, the reflection field data set is divided into a reflection field training data set and a reflection field test data set at a ratio of 8:2, and the backlight field data set is divided into a backlight field training data set and a backlight field test data set at a ratio of 8:2.

[0069] In the present invention, by constructing a reflection field data set and a backlight field data set respectively, it can be ensured that the characteristics and defects of the pultruded sheet surface under different lighting conditions are fully characterized. This multi-perspective data construction can capture more details and provide more comprehensive training data. By dividing the data set into a training data set and a test data set in proportion, it is ensured that most of the data is used to train the model, so that the model can obtain sufficient feature information during the learning process. This division method improves the generalization ability of the model, making it more stable in practical applications and able to effectively deal with unseen data.

[0070] In a possible implementation manner, after S3 and before S4, the process further includes:

[0071] The random global brightness noise image enhancement method is used to perform data enhancement on the reflection field dataset and the backlight field dataset.

[0072] In a possible implementation, a random global brightness noise image enhancement method is used to perform data enhancement processing on the reflection field data set and the backlight field data set, specifically including:

[0073] Obtain the brightness value distribution of all images in the reflection field dataset and the backlight field dataset respectively.

[0074] According to the distribution of brightness values, the brightness ranges of the reflection field data set and the backlight field data set are determined respectively.

[0075] With the brightness value of the image being within the brightness range as the goal, random brightness transformations are performed on all images in the reflection field dataset and the backlight field dataset, respectively, so as to perform data enhancement processing on the reflection field dataset and the backlight field dataset.

[0076] In the present invention, by performing random brightness transformation on the image, images under various brightness conditions can be generated to simulate different practical application scenarios. This data diversification enables the model to be exposed to more changes during the training process, improving its robustness in the face of uncertain environments. Since the surface of the pultruded sheet may encounter various lighting conditions in actual detection, by introducing brightness noise for enhancement processing, the model can learn the characteristic performance under different lighting conditions, thereby improving the model's adaptability to lighting changes.

[0077] S4: Deep learning-based visual detection models are constructed in the reflective field and backlit field respectively.

[0078] Optionally, a ternary network-based visual detection model is constructed in the reflection field and the backlight field respectively, and SqueezeNet is selected as the backbone network of the ternary network to extract features of the input image.

[0079] In the present invention, a visual inspection model based on a ternary network and using SqueezeNet as the backbone network is selected, which can achieve lightweight and high efficiency of the model while ensuring high detection accuracy. This design not only improves the real-time and generalization capabilities of the model, but also provides support for rapid deployment and efficient operation in actual industrial applications.

[0080] S5: Use the reflection field dataset to train the visual detection model under the reflection field, and use the backlight field dataset to train the visual detection model under the backlight field.

[0081] Specifically, the reflection field dataset and the backlight field dataset are used respectively to iteratively train the deep learning-based visual detection model in the reflection field and the backlight field. During the training process, the ternary loss function (Triplet Loss) is used to calculate the network loss values ​​of the deep learning-based visual detection model in the reflection field and the backlight field respectively. The obtained network loss values ​​are back-propagated by gradients, and the weights of the deep learning-based visual detection model in the reflection field and the backlight field are updated respectively.

[0082] In the present invention, by separately training the visual detection models in the reflection field and the backlight field, it is possible to ensure that each model is specifically adapted to specific lighting conditions, thereby achieving the best detection effect in the corresponding scene. The illumination characteristics of the reflection field and the backlight field are different, so the model needs to extract different features, and this separate training method can better capture these features. By optimizing the network loss value using gradient backpropagation, the model can gradually adjust its parameters to better identify defects. This precise optimization process can effectively reduce false detections and missed detections, and improve the overall detection accuracy of the model.

[0083] S6: Acquire cross-stroboscopic images of the surface of the pultruded sheet to be inspected in the reflection field and the backlight field.

[0084] S7: splitting the cross-stroboscopic image of the surface of the pultruded plate to be detected, and obtaining the reflection field image to be detected and the backlight field image to be detected.

[0085] S8: Input the reflection field image to be detected into the trained visual detection model under the reflection field for detection to obtain the reflection field detection result; input the backlight field image to be detected into the trained visual detection model under the backlight field for detection to obtain the backlight field detection result.

[0086] Specifically, the reflection field image to be detected is input into the trained visual detection model under the reflection field, and the updated visual detection model weights under the reflection field are used to detect the reflection field image to obtain the reflection field detection result; the backlight field image to be detected is input into the trained visual detection model under the backlight field, and the updated visual detection model weights under the backlight field are used to detect the backlight field image to obtain the backlight field detection result.

[0087] In the present invention, the updated weights after training are used for detection, which can ensure that the model can accurately identify and process the image to be detected after sufficient training. The weights optimized during the training process can maximize the detection performance of the model. The reflection field image and the backlight field image to be detected are input into the trained special models respectively, so that each model can optimize the detection effect for specific lighting conditions, thereby improving the accuracy of detection.

[0088] S9: Obtain a final detection result according to the reflection field detection result and the backlight field detection result.

[0089] In a possible implementation, S9 specifically includes:

[0090] Perform an OR operation on the reflection field detection result and the backlight field detection result to obtain the final detection result:

[0091]

[0092] Among them, Out represents the final detection result, OR represents the logical OR operator, Out1 and Out2 represent the reflection field detection result and the backlight field detection result respectively, wherein the values ​​of Out1 and Out2 are both 1 or 0, wherein 1 represents the detection of defects, and 0 represents the non-detection of defects.

[0093] It should be noted that when the reflection field detection result and / or the backlight field detection result is 1, the final detection result is a detected defect; when the reflection field detection result and the backlight field detection result are both 0, the final detection result is an undetected defect.

[0094] In the present invention, the reflection field and the backlight field provide different visual information, which can detect potential defects from different angles and lighting conditions. The logical OR operation can combine the two types of information to ensure that defects under any condition will not be missed. Since the reflection field and the backlight field each have their own characteristics and limitations, relying on one condition alone may lead to missed detection or false detection. By combining the two detection results, the robustness and accuracy of the final detection result are enhanced, and the sensitivity and reliability of the detection system to various defects are improved.

[0095] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0096] In the present invention, by obtaining cross-stroboscopic images of the surface of the pultruded sheet in the reflection field and the backlight field, and splitting the cross-stroboscopic images to obtain the reflection field image and the backlight field image, the translucent characteristics of the pultruded sheet for wind turbine blades are fully considered, which is helpful for effectively extracting internal defects under the traditional reflection light field. By constructing deep learning-based visual detection models in the reflection field and the backlight field respectively, the defect detection of the pultruded sheet no longer faces the problems of noise interference, high equipment and maintenance costs, high technical requirements for operators and high technical difficulty. By constructing a reflection field data set based on the reflection field image, and a backlight field data set based on the backlight field image, and using the reflection field data set to train the visual detection model in the reflection field, and using the backlight field data set to train the visual detection model in the backlight field, the accuracy of surface defect detection of the pultruded sheet is improved.

[0097] Reference Manual Attached Figure 3 , showing a schematic structural diagram of a system for detecting surface defects of a pultruded plate material for a wind turbine blade provided by the present invention.

[0098] The present invention further provides a system 30 for detecting surface defects of a pultruded plate material for a wind turbine blade, comprising: a memory 303 and one or more processors 301 .

[0099] One or more application programs are stored in the memory 303 , and the one or more application programs are suitable for being executed by the one or more processors 301 to implement the method for detecting surface defects of pultruded sheet materials for wind turbine blades described in the method embodiment.

[0100] The detection system 30 for surface defects of pultruded plates for wind turbine blades comprises: a processor 301 and a memory 303. The processor 301 and the memory 303 are connected, for example, via a bus 302.

[0101] The structure of the system 30 for detecting surface defects of pultruded sheet materials for wind turbine blades does not constitute a limitation on the embodiment of the present invention.

[0102] Processor 301 may be a CPU, a general purpose processor, a DSP, an ASIC, an FPGA or other programmable logic device, a transistor logic device, a hardware component or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of the present invention. Processor 301 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0103] The bus 302 may include a path to transmit information between the above components. The bus 302 may be a PCI bus or an EISA bus, etc. The bus 302 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0104] The memory 303 can be a ROM or other type of static storage device that can store static information and instructions, a RAM or other type of dynamic storage device that can store information and instructions, or an EEPROM, a CD-ROM or other optical disk storage, an optical disk storage (including a compressed optical disk, a laser disk, an optical disk, a digital versatile disk, a Blu-ray disk, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited to these.

[0105] It should be noted that the detection system 30 for surface defects of pultruded sheet materials for wind turbine blades can implement the above-mentioned detection method for surface defects of pultruded sheet materials for wind turbine blades, and can achieve the same or similar technical effects. To avoid repetition, the present invention will not go into details.

[0106] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0107] In the present invention, by obtaining cross-stroboscopic images of the surface of the pultruded sheet in the reflection field and the backlight field, and splitting the cross-stroboscopic images to obtain the reflection field image and the backlight field image, the translucent characteristics of the pultruded sheet for wind turbine blades are fully considered, which is helpful for effectively extracting internal defects under the traditional reflection light field. By constructing deep learning-based visual detection models in the reflection field and the backlight field respectively, the defect detection of the pultruded sheet no longer faces the problems of noise interference, high equipment and maintenance costs, high technical requirements for operators and high technical difficulty. By constructing a reflection field data set based on the reflection field image, and a backlight field data set based on the backlight field image, and using the reflection field data set to train the visual detection model in the reflection field, and using the backlight field data set to train the visual detection model in the backlight field, the accuracy of surface defect detection of the pultruded sheet is improved.

[0108] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program can be loaded and executed by a processor to implement the method for detecting surface defects of pultruded sheet materials for wind turbine blades as described in the first aspect.

[0109] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0110] There are a few points to note:

[0111] (1) The drawings of the embodiments of the present invention only involve structures related to the embodiments of the present invention. Other structures may refer to conventional designs.

[0112] (2) For the sake of clarity, in the drawings used to describe the embodiments of the present invention, the thickness of layers or regions is exaggerated or reduced, that is, these drawings are not drawn according to the actual scale. It is understood that when an element such as a layer, film, region or substrate is referred to as being "on" or "under" another element, the element may be "directly" "on" or "under" the other element or there may be intermediate elements.

[0113] (3) In the absence of conflict, the embodiments of the present invention and the features therein may be combined with each other to obtain new embodiments.

[0114] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. The protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for detecting surface defects of pultruded plates for wind turbine blades, characterized in that: include: S1: Obtain cross-stroboscopic images of the pultruded sheet surface in the reflection field and the backlight field; S2: Splitting the cross-stroboscopic image to obtain a reflection field image and a backlight field image; S3: constructing a reflection field data set based on the reflection field image, and constructing a backlight field data set based on the backlight field image; S4: Deep learning-based visual detection models built in reflective field and backlit field respectively; S5: using the reflection field data set to train the visual detection model in the reflection field, and using the backlight field data set to train the visual detection model in the backlight field; S6: Acquire cross-stroboscopic images of the surface of the pultruded sheet to be tested in the reflection field and the backlight field; S7: splitting the cross-stroboscopic image of the surface of the pultruded plate to be detected to obtain a reflection field image to be detected and a backlight field image to be detected; S8: inputting the reflection field image to be detected into a trained visual detection model under the reflection field for detection, and obtaining a reflection field detection result; inputting the backlight field image to be detected into a trained visual detection model under the backlight field for detection, and obtaining a backlight field detection result; S9: Obtaining a final detection result according to the reflection field detection result and the backlight field detection result.

2. The method for detecting surface defects of pultruded sheet materials for wind turbine blades according to claim 1, characterized in that: The S1 is specifically: A time-sharing stroboscopic strategy based on encoder trigger signals is used for alternate acquisition to obtain cross-stroboscopic images of the pultruded sheet surface in the reflection field and the backlight field.

3. The method for detecting surface defects of pultruded sheet materials for wind turbine blades according to claim 1, characterized in that: The S2 is specifically: The cross-stroboscopic image is split using an image splitting method based on grayscale threshold to obtain the reflection field image and the backlight field image.

4. The method for detecting surface defects of pultruded sheet materials for wind turbine blades according to claim 3, characterized in that: The method of using the image segmentation method based on grayscale threshold to segment the cross-stroboscopic image to obtain the reflection field image and the backlight field image specifically includes: grayscale the cross-stroboscopic image to obtain a grayscale image; Obtaining global pixel grayscale value information of the grayscale image and determining a grayscale threshold; Taking the image row as a splitting unit, splitting the grayscale image into a plurality of image rows; Calculate the mean grayscale value of each image row; Filter out image rows whose row pixel grayscale mean is greater than the grayscale threshold as reflection field image rows, and filter out image rows whose row pixel grayscale mean is less than or equal to the grayscale threshold as backlight field image rows; The screened reflection field image rows and the backlight field image rows are spliced ​​according to the arrangement order of the image rows of the cross-stroboscopic image to obtain the reflection field image and the backlight field image respectively.

5. The method for detecting surface defects of pultruded sheet materials for wind turbine blades according to claim 1, characterized in that: After S2 and before S3, the method further includes: The reflection field image and the backlight field image are cropped using an image-based region extraction method to obtain a reflection field detection region image and a backlight field detection region image.

6. The method for detecting surface defects of pultruded sheet materials for wind turbine blades according to claim 5, characterized in that: The method of using an area extraction method based on image science to crop the reflection field image and the backlight field image to obtain a reflection field detection area image and a backlight field detection area image specifically includes: Binarizing the reflection field image to obtain a binary image; Performing an opening operation and a dilation operation on the binary image to extract a black and white pixel area in the binary image; Determining the cutting boundary coordinates according to the coordinates of the dividing line between the black and white pixel areas; The reflection field image and the backlight field image are respectively cropped according to the cropping boundary coordinates to obtain a reflection field detection area image and a backlight field detection area image.

7. The method for detecting surface defects of pultruded sheet materials for wind turbine blades according to claim 6, characterized in that: After S3 and before S4, the method further includes: The random global brightness noise image enhancement method is used to perform data enhancement processing on the reflection field dataset and the backlight field dataset.

8. The method for detecting surface defects of pultruded sheet materials for wind turbine blades according to claim 7, characterized in that: The use of the random global brightness noise image enhancement method to perform data enhancement processing on the reflection field data set and the backlight field data set specifically includes: Respectively obtaining the brightness value distribution of all images in the reflection field data set and the backlight field data set; According to the brightness value distribution, respectively determine the brightness intervals of the reflection field data set and the backlight field data set; With the brightness value of the image being within the brightness range as the goal, brightness transformations of random sizes are performed on all images in the reflection field dataset and the backlight field dataset, respectively, so as to perform data enhancement processing on the reflection field dataset and the backlight field dataset.

9. The method for detecting surface defects of pultruded sheet materials for wind turbine blades according to claim 1, characterized in that: The S9 is specifically: Perform an OR operation on the reflection field detection result and the backlight field detection result to obtain a final detection result: ； Among them, Out represents the final detection result, OR represents the logical OR operator, Out1 and Out2 represent the reflection field detection result and the backlight field detection result respectively, wherein the values ​​of Out1 and Out2 are both 1 or 0, wherein 1 represents the detection of defects, and 0 represents the non-detection of defects.

10. A system for detecting surface defects of pultruded plates for wind turbine blades, characterized in that: include: memory and one or more processors; One or more application programs are stored in the memory, and the one or more application programs are suitable for being executed by the one or more processors to implement the method for detecting surface defects of pultruded sheet materials for wind turbine blades according to any one of claims 1 to 9.

Images