Fuzzy image recognition method, device and equipment

Abstract

The application discloses a blurred image recognition method, device and equipment. The blurred image recognition method comprises the following steps: acquiring an image of a wind driven generator photographed by a shooting device; in response to the fact that the image comprises a blade of the wind driven generator, inputting the image into a blurred recognition model to obtain a blurred recognition result, wherein the blurred recognition result is used for indicating whether the image is a blurred image; wherein the blurred recognition model is used for recognizing a blurred factor in the image, and the blurred factor comprises a light-emitting element, a lens additional element and a blade outline pixel point. According to the embodiment of the application, whether the image comprises the blade can be recognized first, if the image comprises the blade of the wind driven generator, the blurred factor in the image is recognized based on the blurred recognition model, and then whether the image is a blurred image is judged according to the blurred recognition result, so that the accuracy of the blade fault detection of the wind driven generator is improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of image processing, and particularly relates to a blurred image recognition method, device and equipment. BACKGROUND

[0002] As a clean renewable energy, wind power plays an important role in energy development strategy. Wind power mainly relies on wind turbines to convert wind energy into electrical energy. When the blades of the wind turbine are running, they may encounter faults such as breakage, cracking, icing, etc. In the related art, a pan-tilt camera is usually used to capture images of the wind turbine to detect whether the blades of the wind turbine have these faults. However, the images captured by the pan-tilt camera may be affected by other factors, resulting in blurred images. If detection is performed based on these blurred images, the accuracy of the detection result is usually low. SUMMARY

[0003] The embodiments of the present application provide a blurred image recognition method, device and equipment to solve the technical problem of low accuracy of blade fault detection of a wind turbine.

[0004] In a first aspect, the embodiments of the present application provide a blurred image recognition method, which comprises:

[0005] obtaining an image of a wind turbine captured by a shooting device;

[0006] in response to the image including blades of the wind turbine, inputting the image into a blurred recognition model to obtain a blurred recognition result, the blurred recognition result being used to indicate whether the image is a blurred image;

[0007] wherein the blurred recognition model is used to identify blurred factors in the image, and the blurred factors include a light-emitting element, a lens additional element and a blade outline pixel point.

[0008] In a second aspect, the embodiments of the present application provide a blurred image recognition device, which comprises:

[0009] an obtaining module configured to obtain an image of a wind turbine captured by a shooting device;

[0010] an identifying module configured to, in response to the image including blades of the wind turbine, input the image into a blurred recognition model to obtain a blurred recognition result, the blurred recognition result being used to indicate whether the image is a blurred image;

[0011] wherein the blurred recognition model is used to identify blurred factors in the image, and the blurred factors include a light-emitting element, a lens additional element and a blade outline pixel point.

[0012] In a third aspect, an electronic device is provided, and the device includes:

[0013] a processor and a memory storing program instructions;

[0014] The processor executes the program instructions to implement the method described above.

[0015] In a fourth aspect, a storage medium is provided, and the storage medium stores program instructions, which are executed by a processor to implement the method described above.

[0016] In a fifth aspect, a computer program product is provided, and the instructions in the computer program product are executed by a processor of an electronic device to cause the electronic device to perform the method described above.

[0017] The blurred image recognition method, device, and equipment provided in the embodiments of the present application can obtain an image of a wind turbine photographed by a photographing device; in response to the image including a blade of the wind turbine, the image is input into a blur recognition model to obtain a blur recognition result, which is used to indicate whether the image is a blurred image; wherein the blur recognition model is used to identify a blur factor in the image, and the blur factor includes a light-emitting element, a lens additional element, and a blade contour pixel point. In this way, it can be determined whether the image includes the blade first, and if the image includes the blade of the wind turbine, the blur factor in the image is identified based on the blur recognition model, and then it is determined whether the image is a blurred image according to the blur recognition result, which effectively avoids using a blurred image to detect the fault condition of the blade of the wind turbine, thereby improving the accuracy of the blade fault detection of the wind turbine. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments of the present application. For those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0019] Figure 1 is a flowchart of a blurred image recognition method provided by an embodiment of the present application;

[0020] Figure 2 is a schematic diagram of the working logic of the blurred image recognition method;

[0021] Figure 3 is an example diagram of the network structure of the blur recognition model;

[0022] Figure 4 is a principle diagram of the semantic feature extraction of the blur recognition model;

[0023] Figure 5 is a processing flow diagram for acquiring a blade area based on a fuzzy recognition model;

[0024] Figure 6 is a structural schematic diagram of a fuzzy image recognition device provided by another embodiment of the present application;

[0025] Figure 7 is a structural schematic diagram of an electronic device provided by yet another embodiment of the present application. DETAILED DESCRIPTION

[0026] The features and exemplary embodiments of various aspects of the present application will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are intended to explain the present application, but not to limit the present application. The present application can be implemented without some of the specific details described below. The following description of the embodiments is merely intended to provide a better understanding of the present application by showing examples of the present application.

[0027] It should be noted that, in this document, relational terms such as first and second, and the like, are used solely to distinguish one entity or action from another entity or action, without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without more limitations, an element defined by the phrase "comprising a... " does not exclude the existence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0028] To solve the problems in the prior art, the embodiments of the present application provide a fuzzy image recognition method, device, equipment and computer storage medium. First, the fuzzy image recognition method provided by the embodiments of the present application will be introduced.

[0029] Figure 1 A flowchart of a fuzzy image recognition method provided by an embodiment of the present application is shown. As shown in Figure 1 The fuzzy image recognition method can include the following steps:

[0030] Step 101, acquiring an image of a wind turbine photographed by a photographing device;

[0031] In response to the image including the blade of the wind turbine, the image is input to a blur identification model to obtain a blur identification result, the blur identification result being used to indicate whether the image is a blur image.

[0032] The blur identification model is used to identify a blur factor in the image, and the blur factor includes a light-emitting element, a lens additional element, and a blade outline pixel point.

[0033] The specific implementation of each step will be described in detail below.

[0034] In the embodiment of the present application, the blur image identification method can obtain an image of a wind turbine captured by a shooting device. In response to the image including the blade of the wind turbine, the image is input to a blur identification model to obtain a blur identification result, the blur identification result being used to indicate whether the image is a blur image. The blur identification model is used to identify a blur factor in the image, and the blur factor includes a light-emitting element, a lens additional element, and a blade outline pixel point. In this way, it can be determined whether the image includes the blade first. If the image includes the blade of the wind turbine, the blur factor in the image is identified based on the blur identification model, and then it is determined whether the image is a blur image according to the blur identification result. This effectively avoids using a blur image to detect the fault condition of the blade of the wind turbine, thereby improving the accuracy of the blade fault detection of the wind turbine.

[0035] The specific implementation of each step will be described in detail below.

[0036] In step 101, the image can be obtained by a shooting device such as a gimbal camera or a camera shooting the wind turbine.

[0037] For example, during the operation of the wind turbine, the gimbal camera can continuously or at a preset time interval record a video of the wind turbine, and the image can be a video frame extracted from the video. Alternatively, the image can be obtained by the gimbal camera shooting the wind turbine at a preset frequency.

[0038] In step 102, after obtaining the image, it can be determined whether the image includes the blade of the wind turbine. For example, the image can be input to a pre-trained blade detection model to obtain a blade detection result, which can indicate whether the image includes the blade of the wind turbine. The blade detection model can be a target detection model or an image classification model, which can be trained using sample images including blades and sample images not including blades.

[0039] In response to the image including a blade of the wind turbine, the image can be input into a blur recognition model to obtain a blur recognition result, which can be used to indicate whether the image is a blur image. The blur recognition model can also be a target detection model or an image classification model, etc. The blur recognition model can be used to identify blur factors in the image.

[0040] In the image, the blur factors can include a light-emitting element, a lens attachment element, and a blade outline pixel point. The light-emitting element can be a light source, the sun, or the like. In other words, the blur recognition model can be used to identify whether there is a light source, the sun, or the like in the image. In response to the presence of a light source, the sun, or the like in the image, it can be considered that the image is a backlit image taken in a backlit environment, that is, the blur recognition result can indicate that the image is a blur image. The lens attachment element can be a lens attachment. In other words, the blur recognition model can be used to identify whether there is a water droplet, a stain, or the like in the image. In response to the presence of a lens attachment in the image, it can be considered that there is a water droplet or a stain or the like on the lens when the gimbal camera takes the image, that is, the lens is not clear when taking the image. At this time, the blur recognition result can indicate that the image is a blur image. The blade outline pixel point can be extracted according to an existing edge extraction algorithm. The blur recognition model can be used to identify whether the blade outline composed of the blade outline pixel point in the image is obvious. If the number of blade outline pixel points is small, the blade outline is not obvious, and it can be considered that the image can be caused by poor resolution of the image due to focusing errors or the like. At this time, the blur recognition result can also indicate that the image is a blur image.

[0041] For example, the image is input into the blur recognition model, which can identify blur factors such as the light-emitting element, the lens attachment element, and the blade outline pixel point in the image. In response to the presence of blur factors in the image, it can be considered that the image is a blur image. In response to the absence of blur factors in the image, it can be considered that the image is not a blur image, that is, the image is a clear image, which can be used for subsequent blade fault detection.

[0042] In some examples, the related algorithm of the blur recognition model can use an R-CNN system algorithm based on a Region Proposal network, such as R-CNN, Fast R-CNN, or Faster R-CNN. These algorithms can all be referred to as two-stage target detection algorithms, which have the characteristic that a heuristic method (selective search) or a CNN network (RPN) is used to generate a candidate frame (Region Proposal) first, and then classification and regression are performed on the Region Proposal.

[0043] The relevant algorithms for fuzzy recognition models can also adopt the one-stage object detection algorithms of YOLO or SSD, which use only a CNN network to directly predict the category and location of different objects.

[0044] Two-stage algorithms offer relatively high accuracy, while one-stage algorithms are faster and require less computational resources. In this embodiment, since wind turbines are typically located in remote areas where network access to computational resources is often difficult, one-stage algorithms such as YOLO can quickly detect image blurriness.

[0045] In some feasible implementations, when computing resources are sufficient, the above-mentioned fuzzy recognition model can also use a two-stage algorithm to improve the accuracy of fuzzy image recognition results, thereby improving the accuracy of wind turbine blade fault detection.

[0046] like Figure 2 As shown, Figure 2 This is a schematic diagram illustrating the working logic of a fuzzy image recognition method, which can be summarized as follows:

[0047] Step 201: Acquire an image. For example, the image may be obtained by taking a picture of the wind turbine using a gimbal camera or other imaging device.

[0048] Step 202: After obtaining the image, detect whether the image includes leaves. If yes, proceed to step 203; otherwise, proceed to step 207.

[0049] Step 203: Identify the blurring factors of luminous elements and lens-added elements in the image.

[0050] Step 204: Detect whether the image includes at least one of the luminous element and the lens-attached element. If yes, proceed to step 207; otherwise, proceed to step 205.

[0051] Step 205: Identify blurring factors in the leaf outline pixels of the image.

[0052] Step 206: Determine if the relationship between the leaf area and the leaf outline pixels in the image meets the preset outline conditions. If yes, proceed to step 207; otherwise, proceed to step 208.

[0053] Step 207: Determine that the image is a blurred image.

[0054] Step 208: Determine that the image is not a blurry image.

[0055] For example, combined Figure 2In some embodiments, the step 102 can be implemented as follows:

[0056] In response to the image including the blade of the wind turbine, inputting the image to a blur identification model to identify whether the image includes at least one of a light-emitting element and a lens attachment element;

[0057] In response to the image including at least one of the light-emitting element and the lens attachment element, determining the image as a blur image;

[0058] In response to the image not including both the light-emitting element and the lens attachment element, obtaining a blade area and a blade contour pixel point in the image;

[0059] In a case where a relationship between the blade area and the blade contour pixel point meets a preset contour condition, determining the image as the blur image.

[0060] In the embodiments of the present application, in response to the image including the blade of the wind turbine, after inputting the image to the blur identification model, it can be first identified whether the image includes at least one of the light-emitting element and the lens attachment element. If the image includes at least one of the light-emitting element and the lens attachment element, it can be considered that the image is taken in a backlight environment, or there is an attachment such as a water droplet or a stain on the lens when the image is taken. At this time, the image can be directly determined as the blur image.

[0061] If the image does not include both the light-emitting element and the lens attachment element, it can be considered that the image is not taken in the backlight environment, and the lens is clean when the image is taken, without an attachment such as a water droplet or a stain. At this time, the blade area and the blade contour pixel point in the image can be obtained to determine whether the image has a poor resolution. The blade area in the image can be obtained based on the blur identification model or based on an existing target object area calculation algorithm. The blade contour pixel point can be obtained based on an existing edge extraction algorithm.

[0062] After obtaining the blade area and the blade contour pixel point in the image, it can be determined whether a relationship between the blade area and the blade contour pixel point meets a preset contour condition. For example, a contour constructed according to the blade contour pixel point can be matched with the blade area. If the contour does not match the blade area, it can be considered that the relationship between the blade area and the blade contour pixel point meets the preset contour condition. For another example, the number of the blade contour pixel points can be compared with the blade area. If a ratio of the number of the blade contour pixel points to the blade area meets a certain threshold, it can be considered that the relationship between the blade area and the blade contour pixel point meets the preset contour condition. It can be understood that the specific preset contour condition can be set according to actual conditions, which is not limited here.

[0063] If the relationship between the blade area and the blade contour pixel points meets the preset contour condition, it can be considered that the image may be poor in resolution due to focusing errors and the like, and the image can be determined as a blurred image. If the relationship between the blade area and the blade contour pixel points does not meet the preset contour condition, it can be considered that the resolution of the image is good, and the image can be determined as a clear image.

[0064] In the embodiments of the present application, the blur factors of the light-emitting element and the lens additional element in the image can be identified first. If the image includes at least one of the light-emitting element and the lens additional element, the image can be directly determined as a blurred image. If the image does not include both the light-emitting element and the lens additional element, the blur factors of the blade contour pixel points are identified again. In this way, the calculation amount can be effectively reduced, the computing resource is saved, and the rate of blurred image identification is improved.

[0065] For example, in combination with Figure 2 In some embodiments, after the step 101, the blurred image identification method can further include the following steps:

[0066] In response to the image not including the blades of the wind turbine, the image is determined as a blurred image.

[0067] In the embodiments of the present application, after the image is obtained, it can be detected whether the image includes the blades of the wind turbine. Under normal circumstances, since the image is obtained by shooting the blades of the wind turbine based on a pan-tilt camera or the like, the image includes the blades of the wind turbine. However, if the shooting is performed in the night, snowstorm, heavy fog or the like, the image may not have obvious blade features. Based on this, if the image does not include the blades of the wind turbine, it can be considered that the image may be obtained in the night, snowstorm, heavy fog or the like, and the image can be determined as a blurred image and cannot be used for subsequent blade fault detection.

[0068] In some embodiments, the preset contour condition can be that the relationship between the blade area and the square of the number of blade contour pixel points meets a preset threshold.

[0069] It can be understood that for the image with focusing errors, since the resolution of the image is poor, the contour of the blades of the wind turbine is less obvious than that of a clear image. Therefore, when the edge extraction algorithm in the image processing field is used to extract the blade contour pixel points, the number of blade contour pixel points obtained will be less, and even the contour will be interrupted.

[0070] It can also be understood that, in the case that the clarity of the image does not change, the square of the length of the blade contour is proportional to the blade area, and based on this, in order to simplify the blurred image recognition process, thereby saving computing resources and improving the recognition rate, the blurred degree of the image can be determined according to the relationship between the blade area and the square of the number of contour pixel points. In other words, the preset contour condition can be that the relationship between the blade area and the square of the number of blade contour pixel points satisfies a preset threshold.

[0071] For example, if the ratio of the blade area to the square of the number of blade contour pixel points is greater than or equal to a preset threshold, it can be considered that the number of blade contour pixel points extracted is small, that is, the resolution of the image is poor, and the image can be determined as a blurred image. The number of blade contour pixel points can refer to the number of contour pixel points in the length direction of the blade.

[0072] It can be understood that the preset threshold can be set according to empirical values in combination with actual conditions, or can be calculated based on a plurality of blurred image samples and a plurality of clear image samples. For example, the preset threshold can be 15.3, that is, if the ratio of the blade area to the square of the number of blade contour pixel points in the image is greater than or equal to 15.3, it can be considered that the image is a blurred image, and if the ratio of the blade area to the square of the number of blade contour pixel points in the image is less than 15.3, it can be considered that the image is a clear image.

[0073] In some embodiments, the above-mentioned identification of whether the image includes at least one of the light-emitting element and the lens additional element can be specifically performed as follows:

[0074] The image is input to a feature extraction network in the blurred recognition model, and the features in the image are extracted through the feature extraction network to obtain edge features and color features in the image;

[0075] The edge features and the color features are combined through a path aggregation network in the blurred recognition model to obtain semantic features in the image;

[0076] Based on the semantic features in the image, whether the image includes at least one of the light-emitting element and the lens additional element is identified.

[0077] As Figure 3 shown, in this specific application example, the blurred recognition model can use a Yolo network, which can specifically include a feature extraction network, a path aggregation network, and an output network. The feature extraction network can be a BackBone, which can be used to extract rough features. The path aggregation network can be a PANet, which can be used to extract multi-scale detailed features and fuse related features. The output network can be denoted as Output, which mainly functions to organize the output format.

[0078] AsFigure 4 As shown, after the image is input into the blur recognition model, the BackBone can extract the underlying features such as edge features, color features, etc. for the image layer by layer, where the edge features can include contours and shapes, and the color features can include colors. For example,

[0079] The PANet can perform further feature extraction based on the edge features, color features, etc. extracted by the BackBone, to obtain features such as “color is gray or black or white”, “contour is obvious or not obvious”, and “shape is large wing shape”. The PANet can also perform further feature combination on these features to obtain semantic features in the image, such as “crack”, “leaf”, “stain”, “cloud”, etc.

[0080] The semantic features in the image generally correspond to the objects included in the image, and whether at least one of the light-emitting element and the lens additional element is included in the image can be identified through the semantic features in the image. For example, the semantic features in the image indicate that the image includes a sun, a light source, or the like, and it can be considered that the image includes a light-emitting element, and at this time, it can be considered that the image is obtained by shooting in a backlight environment, and is a blurred image, and cannot be used for subsequent blade fault detection. For another example, the semantic features in the image indicate that the image includes lens attachments, and at this time, it can be considered that the image includes a lens additional element, and at this time, it can be considered that there are water droplets or stains and the like on the lens when the pan-tilt camera shoots the image, that is, the lens is not clear when shooting, and the image obtained by shooting is a blurred image, and cannot be used for subsequent blade fault detection.

[0081] The output can directly output the objects included in the image based on the semantic features in the image, or directly output the identification result of whether at least one of the light-emitting element and the lens additional element is included in the image.

[0082] In some examples, detecting whether the image includes a blade of a wind turbine can also be based on the Yolo network. For example, the BackBone can extract edge features, color features, etc. in the image. The PANet can perform further feature extraction and feature combination based on the edge features, color features, etc. extracted by the BackBone to obtain semantic features in the image. If the semantic features in the image include “blade”, it can be considered that the image includes a blade of a wind turbine.

[0083] In some embodiments, obtaining the blade area in the image can include the following steps:

[0084] The edge features and the color features are combined by the path aggregation network in the blur recognition model to obtain semantic features and position features corresponding to the blade;

[0085] According to the semantic feature and the position feature corresponding to the leaf, the area of the leaf in the image is obtained.

[0086] Generally, in the Yolo network, there is also a process of upsampling, that is, filling the extracted feature map to the resolution size of the input image. In the embodiment of the present application, after the semantic feature corresponding to the leaf is extracted, the PANet can determine the position feature of the leaf through upsampling and fuse the semantic feature and the position feature, so that the area of the leaf in the image can be obtained. The output can also output the area of the leaf in the image.

[0087] As shown in Figure 5 , the feature extraction network can extract a first feature map from the image. In combination with an example, the pixel size of the image can be 8x8, and the feature extraction network can perform convolution operation on the image to obtain the first feature map, at this time the first feature map can distinguish that there is an object in the image.

[0088] The path aggregation network can further extract and combine features from the first feature map to obtain a second feature map of 4x4, at this time the second feature map can distinguish that there is a large gray object in the image.

[0089] The path aggregation network can further extract features from the second feature map to obtain a third feature map of 2x2, which can include semantic features corresponding to the leaf, that is, can be used to distinguish that there is a leaf in the image.

[0090] Subsequently, the path aggregation network can perform upsampling based on the second feature map and the third feature map to obtain a fourth feature map that distinguishes the approximate range of the leaf, for example, the fourth feature map can indicate that there is a leaf in the middle and lower area of the image. For example, the path aggregation network can fill the third feature map of 2x2 to obtain a feature map of 4x4, and fuse it with the second feature map of 4x4 to obtain a fourth feature map of 4x4 that distinguishes the approximate range of the leaf.

[0091] The fourth feature map can be considered to include the position feature of the leaf, at this time the accuracy of the position feature is relatively low.

[0092] The path aggregation network further upsamples the fourth feature map to obtain a feature map equal in size to the first feature map, and fuses it with the first feature map, because the pixel size of the first feature map can be 8x8, that is, equal to the pixel size of the original image, at this time the area of the leaf in the image can be determined according to the first feature map and Figure Four The fifth feature map obtained after the fusion of the feature maps distinguishes the area ratio of the leaf in the image. For example, the fifth feature map can indicate that there is a leaf with an area ratio of 20% in the middle and lower area of the image.

[0093] Further, the path aggregation network can fuse the fifth feature map with the original image to obtain a sixth feature map capable of determining the accurate position of the leaf.

[0094] In the embodiment of the present application, the area of the leaf in the image can be obtained based on the blur recognition model. For example, the edge features and color features can be combined by the path aggregation network in the blur recognition model to obtain semantic features corresponding to the leaf, and then the position features of the leaf can be obtained by upsampling fusion, and then the area of the leaf in the image can be obtained according to the semantic features and the position features corresponding to the leaf.

[0095] Figure 6 The structure diagram of the blur image recognition device provided by another embodiment of the present application is shown, and only the parts related to the embodiments of the present application are shown for ease of illustration.

[0096] Reference Figure 6 The blur image recognition device 600 can include:

[0097] The acquisition module 601 is configured to acquire an image of a wind turbine captured by a shooting device.

[0098] The recognition module 602 is configured to, in response to the image including a leaf of the wind turbine, input the image into a blur recognition model to obtain a blur recognition result, the blur recognition result being used to indicate whether the image is a blur image.

[0099] The blur recognition model is used to identify blur factors in the image, and the blur factors include a light-emitting element, a lens additional element, and a leaf outline pixel point.

[0100] In some embodiments, the above-mentioned recognition module 602 can include

[0101] The recognition unit is configured to, in response to the image including the leaf of the wind turbine, input the image into the blur recognition model to identify whether the image includes at least one of the light-emitting element and the lens additional element.

[0102] The first determination unit is configured to, in response to the image including at least one of the light-emitting element and the lens additional element, determine the image as a blur image.

[0103] The acquisition unit is configured to, in response to the image not including both the light-emitting element and the lens additional element, acquire the area of the leaf and the leaf outline pixel point in the image.

[0104] The second determination unit is configured to, in a case where the relationship between the area of the leaf and the leaf outline pixel point satisfies a preset outline condition, determine the image as a blur image.

[0105] In some embodiments, the preset contour condition can be that a relationship between the leaf area and a square of the number of leaf contour pixels meets a preset threshold.

[0106] In some embodiments, the identifying unit can be specifically configured to:

[0107] input the image into a feature extraction network in the fuzzy recognition model, extract features in the image through the feature extraction network, and obtain edge features and color features in the image;

[0108] combine the edge features and the color features through a path aggregation network in the fuzzy recognition model, and obtain semantic features in the image;

[0109] identify whether the image includes at least one of the light-emitting element and the lens additional element based on the semantic features in the image.

[0110] In some embodiments, the obtaining unit can be specifically configured to:

[0111] combine the edge features and the color features through a path aggregation network in the fuzzy recognition model, and obtain semantic features and position features corresponding to the leaf;

[0112] obtain a leaf area in the image according to the semantic features and the position features corresponding to the leaf.

[0113] In some embodiments, the fuzzy image recognition device 600 can further include:

[0114] a determining module configured to determine the image as the fuzzy image in response to the image not including the leaf of the wind turbine.

[0115] In some embodiments, the light-emitting element is a light-emitting body, and the lens additional element is a lens attachment.

[0116] It should be noted that the information interaction, execution process and the like between the above-described devices / units are based on the same concept as the method embodiments of the present application, and are devices corresponding to the above fuzzy image recognition method. All implementation manners in the above method embodiments are applicable to the embodiments of the device, and the specific functions and technical effects brought by the device can be referred to the method embodiments part, and will not be described here.

[0117] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional units and modules is taken as an example, and in actual application, the above functions can be completed by different functional units and modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the above described functions. Each functional unit and module in the embodiment can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of software functional unit. In addition, the specific names of each functional unit and module are only for convenient distinction, and do not limit the protection scope of the present application. The specific working process of the units and modules in the system can refer to the corresponding process in the foregoing method embodiments, which will not be repeated here.

[0118] Figure 7 A hardware structure schematic diagram of an electronic device provided by yet another embodiment of the present application is shown.

[0119] The device can include a processor 701 and a memory 702 storing program instructions.

[0120] The processor 701 executes the program to implement the steps in any of the above method embodiments.

[0121] For example, the program can be divided into one or more modules / units, one or more modules / units are stored in the memory 702 and executed by the processor 701 to complete the present application. One or more modules / units can be a series of program instruction segments capable of completing a specific function, which is used to describe the execution process of the program in the device.

[0122] Specifically, the above processor 701 can include a central processing unit (CPU), or a specific integrated circuit (ASIC), or can be configured as one or more integrated circuits that implement the embodiments of the present application.

[0123] The memory 702 can include mass storage for data or instructions. As an example and not by way of limitation, the memory 702 can include a hard disk drive (HDD), a floppy disk drive, flash memory, an optical disc (e.g., a compact disc (CD) or a digital versatile disc (DVD)), a solid-state drive (SSD), a USB drive, or a combination of two or more of these. Where appropriate, the memory 702 can include removable or non-removable (or fixed) media. Where appropriate, the memory 702 can be internal or external to the integrated gateway disaster recovery appliance. In particular embodiments, the memory 702 is non-volatile, solid-state memory.

[0124] The memory can include read-only memory (ROM), random-access memory (RAM), magnetic disk storage mediums, optical storage mediums, flash memory devices, electrical, optical, or other physical / tangible memory storage devices. Thus, in general, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., a memory device) encoded with software that, when executed (by one or more processors), is operable to access the data and / or instructions as described with reference to the methods according to the aspects of the present disclosure.

[0125] The processor 701 implements any of the above-described methods by reading and executing program instructions stored in the memory 702.

[0126] In one example, the electronic device further includes a communication interface 703 and a bus 710. The processor 701, the memory 702, and the communication interface 703 are connected through the bus 710 and complete communication therebetween.

[0127] The communication interface 703 is mainly used to realize the communication between the modules, devices, units, and / or equipment in the embodiments of the present application.

[0128] Bus 710 includes hardware, software, or both, to couple components of the online data traffic metering device to each other and to couple components to other systems. For example, but not limited to, the bus can include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand (IB) interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association local (VLB) bus, or another suitable bus or a combination of two or more of these. Where suitable, bus 710 can include one or more buses. Although a particular bus is described and illustrated in this embodiment, the application contemplates any suitable bus or interconnect.

[0129] In addition, in combination with the method in the above-mentioned embodiments, the embodiments of the present application can provide a storage medium for implementation. The storage medium has program instructions stored thereon; the program instructions are executed by a processor to implement any of the methods in the above-mentioned embodiments.

[0130] The embodiments of the present application further provide a chip, which includes a processor and a communication interface, the communication interface is coupled to the processor, the processor is used to run programs or instructions, to implement various processes of the above-mentioned method embodiments and achieve the same technical effects. To avoid repetition, details are not described here.

[0131] It should be understood that the chip mentioned in the embodiments of the present application can also be referred to as a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.

[0132] The embodiments of the present application provide a computer program product, which is stored in a storage medium, and the program product is executed by at least one processor to implement various processes of the above-mentioned method embodiments and achieve the same technical effects. To avoid repetition, details are not described here.

[0133] It should be clear that the present application is not limited to the specific configurations and processes described above and shown in the drawings. For the sake of brevity, detailed descriptions of well-known methods are omitted here. In the above-mentioned embodiments, several specific steps are described and shown as examples. However, the method processes of the present application are not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between steps, after understanding the spirit of the present application.

[0134] The functional modules shown in the above-described block diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer grids such as the Internet, intranets, etc.

[0135] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0136] The aspects of this disclosure have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and program products according to embodiments of this disclosure. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to create a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by special-purpose hardware performing the specified functions or actions, or can be implemented by a combination of special-purpose hardware and computer instructions.

[0137] The above description is merely a specific implementation of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the protection scope of this application.

Claims

1. A method for recognizing fuzzy images, characterized in that, include: Acquire images of the wind turbine taken by the imaging equipment; In response to the image including the blades of the wind turbine, the image is input into a fuzzy recognition model to obtain a fuzzy recognition result, which is used to indicate whether the image is a fuzzy image; The fuzzy recognition model is used to identify fuzzy factors in an image, including: luminous elements, lens-added elements, and leaf outline pixels. In response to the image including the blades of the wind turbine, the image is input into a fuzzy recognition model to obtain a fuzzy recognition result, including: In response to the image including the blades of the wind turbine, the image is input into a fuzzy recognition model to identify whether the image includes at least one of a light-emitting element and a lens-attached element; In response to the image including at least one of luminous elements and lens-attached elements, the image is determined to be a blurred image; In response to the absence of both light-emitting elements and lens-attached elements in the image, the leaf area and the leaf outline pixels in the image are obtained; If the relationship between the blade area and the blade outline pixels satisfies a preset outline condition, the image is determined to be a blurred image.

2. The method according to claim 1, characterized in that, The preset contour condition is that the relationship between the area of ​​the blade and the square of the number of pixels in the blade contour satisfies a preset threshold.

3. The method according to claim 1, characterized in that, The identification of whether the image includes at least one of luminous elements and lens-added elements includes: The image is input into the feature extraction network of the fuzzy recognition model, and the feature extraction network extracts the features in the image to obtain the edge features and color features in the image; The semantic features in the image are obtained by combining the edge features and the color features through the path aggregation network in the fuzzy recognition model. Based on the semantic features in the image, identify whether the image includes at least one of luminous elements and lens-attached elements.

4. The method according to claim 3, characterized in that, Obtaining the leaf area in the image includes: The edge features and color features are combined by the path aggregation network in the fuzzy recognition model to obtain the semantic features and positional features corresponding to the leaf. The area of ​​the leaf in the image is obtained based on the semantic and positional features corresponding to the leaf.

5. The method according to claim 1, characterized in that, After acquiring the image of the wind turbine captured by the imaging device, the method further includes: In response to the fact that the image does not include the blades of the wind turbine, the image is determined to be a blurred image.

6. The method according to claim 1, characterized in that, The light-emitting element is a light-emitting body; and The lens attachment is the lens attachment.

7. A fuzzy image recognition device, characterized in that, The device includes: The acquisition module is used to acquire images of the wind turbine taken by the imaging device; The recognition module is used to input the image into a fuzzy recognition model in response to the image including the blades of the wind turbine, and obtain a fuzzy recognition result, the fuzzy recognition result being used to indicate whether the image is a fuzzy image; The fuzzy recognition model is used to identify fuzzy factors in an image, including: luminous elements, lens-added elements, and leaf outline pixels. The identification module includes: The recognition unit is configured to, in response to the image including the blades of the wind turbine, input the image into a fuzzy recognition model to identify whether the image includes at least one of a light-emitting element and a lens-attached element; A first determining unit is configured to determine the image as a blurred image in response to the image including at least one of luminous elements and lens-attached elements. The acquisition unit is configured to acquire the leaf area and the leaf outline pixels in the image in response to the absence of both light-emitting elements and lens-attached elements in the image. The second determining unit is used to determine the image as a blurred image when the relationship between the blade area and the blade outline pixels satisfies a preset outline condition.

8. An electronic device, characterized in that, The device includes: a processor and a memory storing program instructions; When the processor executes the program instructions, it implements the method as described in any one of claims 1-6.

9. A storage medium, characterized in that, The storage medium stores program instructions, which, when executed by a processor, implement the method as described in any one of claims 1-6.

10. A computer program product, characterized in that, When the instructions in the computer program product are executed by the processor of the electronic device, the electronic device causes the electronic device to perform the method as described in any one of claims 1-6.

Images