Method, device and equipment for detecting spinning beam, and storage medium

Abstract

The present disclosure provides a method, device and equipment for detecting a spinning beam, and a storage medium. The method comprises: determining whether the spinning beam meets a preset defect detection condition; acquiring a to-be-detected image of the spinning beam when the spinning beam meets the preset defect detection condition; inputting the to-be-detected image into a target detection model to obtain a target detection result of the spinning beam; the target detection model is used to detect whether the spinning beam has defects to obtain the target detection result; and the target detection result comprises at least one of the following: a total number of defects, a defect position, and a defect type.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, and in particular to a method, device, equipment, and storage medium for detecting a spinning manifold. Background Art

[0002] In the chemical fiber industry, during the process of producing chemical fiber yarns, due to the long-term operation of the spinning box, abnormal conditions such as yarn drifting, yarn breakage, yarn misalignment, yarn hooking, nozzle tilting, wire guide hook tilting, and yarn not being in the wire guide hook may easily occur in the spinning box. Therefore, staff are usually required to inspect the spinneret, yarn, wire guide hook, and nozzle in the spinning box. However, this inspection method has a strong dependence on manual experience and is inefficient, which affects the production quality and production efficiency of the yarn. Summary of the Invention

[0003] The present disclosure provides a spinning beam detection method, device, equipment, and storage medium to solve or alleviate one or more technical problems in the prior art.

[0004] In a first aspect, the present disclosure provides a method for detecting a spinning beam, comprising:

[0005] When it is determined that the spinning manifold meets the preset defect detection conditions, obtaining an image of the spinning manifold to be inspected;

[0006] The image to be detected is input into the target detection model to obtain the target detection result of the spinning manifold; the target detection model is used to detect whether there are defects in the spinning manifold to obtain the target detection result; the target detection result includes at least one of the following: the total number of defects, the defect location, and the defect type.

[0007] In a second aspect, the present disclosure provides a detection device for a spinning beam, comprising:

[0008] An acquiring unit, configured to acquire an image to be inspected of the spinning manifold when determining that the spinning manifold meets a preset defect detection condition;

[0009] A detection unit is used to input the image to be detected into a target detection model to obtain a target detection result of the spinning manifold; the target detection model is used to detect whether there are defects in the spinning manifold to obtain a target detection result; the target detection result includes at least one of the following: the total number of defects, the defect location, and the defect type.

[0010] According to a third aspect, an electronic device is provided, including:

[0011] at least one processor; and

[0012] a memory communicatively connected with the at least one processor; wherein

[0013] The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of any of the embodiments of the present disclosure.

[0014] In a fourth aspect, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to perform the method according to any of the embodiments of the present disclosure.

[0015] In a fifth aspect, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the method according to any of the embodiments of the present disclosure.

[0016] The technical solutions provided by the present disclosure have at least the following beneficial effects: the target detection model can be used to detect defects of the spinning beam, and a target detection result is obtained, thereby improving the accuracy of defect detection of the spinning beam. Moreover, compared with the existing manual detection, the present solution does not need to rely on manual experience to detect defects of the spinning beam, and realizes the automation process of defect detection of the spinning beam, thereby saving a large amount of human cost and time cost, and improving the detection efficiency of the spinning beam.

[0017] It should be understood that the content described in this part is not intended to identify key or important features of the embodiments of the present disclosure, nor is it used to limit the scope of the present disclosure. Other features of the present disclosure will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0018] In the drawings, like reference numerals refer to same or similar components throughout the several views. These drawings are not necessarily to scale. It should be understood that these drawings only depict some embodiments in accordance with the present disclosure and should not be considered to be limiting of the scope of the present disclosure.

[0019] Figure 1 is a schematic flow of a detection method of a spinning beam according to an embodiment of the present application Figure 1 ;

[0020] FIG. 2(a) is a structural schematic diagram of a first network layer according to an embodiment of the present application;

[0021] FIG. 2(b) is a structural schematic diagram of the first network layer in an example according to an embodiment of the present application;

[0022] FIG. 3(a) is a structural schematic diagram of a first sub-network layer according to an embodiment of the present application;

[0023] FIG. 3(b) is a schematic diagram of a structure of a second network layer in an example according to an embodiment of the present application;

[0024] FIG. 3(c) is a schematic diagram of a structure of a second sub-network layer in another example according to an embodiment of the present application;

[0025] FIG. 3(d) is a schematic diagram of a structure of a feature weight module according to an embodiment of the present application;

[0026] FIG. 3(e) is a schematic diagram of a structure of a multi-view attention module according to an embodiment of the present application;

[0027] Figure 4 FIG. 2 is a schematic flow chart of a method for detecting a spinning beam according to an embodiment of the present application;

[0028] Figure 5 FIG. 3 is a schematic diagram of a structure of a detection device for a spinning beam according to an embodiment of the present application;

[0029] Figure 6 FIG. 4 is a block diagram of an electronic device for implementing the method for detecting a spinning beam according to an embodiment of the present application. DETAILED DESCRIPTION

[0030] The present disclosure will be further described below with reference to the drawings. The same reference numbers in the drawings represent the same or similar elements or components. Although various aspects of the embodiments are shown in the drawings, the drawings are not necessarily drawn to scale unless specifically indicated.

[0031] In addition, in order to better illustrate the present disclosure, numerous specific details are given in the following detailed description. It should be understood by those skilled in the art that the present disclosure can also be implemented without some specific details. In some examples, methods, means, elements and circuits that are well known to those skilled in the art are not described in detail in order to highlight the main ideas of the present disclosure.

[0032] In the process of spinning beam spinning, due to the long-time operation of the spinning beam, abnormal conditions such as yarn floating, yarn breaking, yarn misalignment, yarn hooking, oil nozzle tilting, guide hook tilting, and yarn not in the guide hook may occur in the spinning beam. Therefore, it is usually necessary for the staff to detect the spinning beam, the spinning beam, the guide hook, and the oil nozzle. However, this detection method requires a large amount of labor cost and time cost, and has a strong dependence on manual experience. Moreover, the defects such as yarn floating and yarn hooking in the spinning beam are relatively subtle and difficult to identify, and the use of manual detection method may have problems such as low efficiency and easy to make mistakes, thereby affecting the production quality and production efficiency of the yarn. Therefore, an automatic processing process is needed to efficiently detect defects in the spinning beam.

[0033] Based on this, the present scheme provides a detection method of a spinning beam to solve the above problems.

[0034] Specifically, Figure 1 is a schematic flow of a detection method of a spinning beam according to an embodiment of the present application Figure 1 . The method can be optionally applied in electronic devices, such as personal computers, servers, server clusters, and the like.

[0035] Further, the method at least includes at least part of the following contents. As Figure 1 indicated, it includes:

[0036] Step S101: In the case where it is determined that the spinning beam meets the preset defect detection condition, acquiring a to-be-detected image of the spinning beam.

[0037] Here, in a specific example, the to-be-detected image of the spinning beam is obtained by image acquisition of the spinning beam by an image acquisition device; further, the image acquisition device can specifically include a silk road inspection robot (with an image acquisition device, such as a camera, built-in), such as taking a picture of the spinning beam by the silk road inspection robot to obtain the to-be-detected image, or acquiring a video of the spinning beam for a preset time length to obtain multiple continuous video frames, and selecting an image from the continuous video frames as the to-be-detected image, so as to facilitate subsequent defect detection of the spinning beam.

[0038] Further, in a specific example, the to-be-detected image at least includes image content of the spinneret, the silk thread, the guide hook, and the oil nozzle.

[0039] Further, in a specific example, the to-be-detected image can be obtained in the following manner, specifically, the above-mentioned acquiring the to-be-detected image of the spinning beam in the case where it is determined that the spinning beam meets the preset defect detection condition (such as the above-mentioned step S101) specifically includes:

[0040] acquiring the to-be-detected image of the spinning beam in the case where it is determined that one of the following conditions is met:

[0041] Condition 1: the detection time of the spinning beam reaches a preset detection opportunity;

[0042] Condition 2: it is determined that there is a degraded silk spool among multiple silk spools obtained based on the spinning beam;

[0043] Condition 3: it is determined that the number of degraded silk spools among the multiple silk spools obtained based on the spinning beam is greater than a first threshold.

[0044] Here, the first threshold value is an empirical value, which can be set according to actual needs in actual application, and the present scheme does not make specific limitations on the first threshold value.

[0045] In this way, the present scheme can start the defect detection of the spinning beam (such as triggering the image acquisition device to acquire images of the spinning beam, and then acquire the to-be-detected images of the spinning beam) when at least one of the above conditions is met, to obtain the detection result of the defect detection, so that the defect condition of the current spinning beam can be quickly known, laying a foundation for subsequent improvement of the quality of spinning in the spinning beam.

[0046] Step S102: input the to-be-detected image into the target detection model to obtain the target detection result of the spinning beam.

[0047] Here, the target detection model is used to detect whether the spinning beam has defects to obtain the target detection result; the target detection result includes at least one of the following: total defect quantity, defect position, and defect type. In this way, compared with the existing manual detection, the defect data in the spinning beam can be accurately obtained, and the efficiency of defect detection is further improved.

[0048] Further, in a specific example, the defect detection of the spinning beam can specifically refer to the detection of the following defects: floating yarn, broken yarn, misaligned yarn, misaligned guide yarn hook, misaligned oil nozzle, and foreign matter in the yarn path, etc.

[0049] Further, in a specific example, the main types of yarns involved in the present scheme can include one or more of Partially Oriented Yarns (POY), Fully Drawn Yarns (FDY), etc. For example, the types of yarns can specifically include Polyester Partially Oriented Yarns, Polyester Fully Drawn Yarns, Polyester Drawn Yarns, Polyester Draw Textured Yarns, etc.

[0050] In this way, the present scheme can use the target detection model to detect defects of the spinning beam and obtain the target detection result, so as to improve the accuracy of defect detection of the spinning beam. Moreover, compared with the existing manual detection, the present scheme does not need to rely on manual experience to detect defects of the spinning beam, realizes the automatic process of defect detection of the spinning beam, saves a lot of human cost and time cost, and thus improves the detection efficiency of the spinning beam.

[0051] Further, since the spinning beam can be automatically and accurately detected, subsequent measures can be taken in time according to the obtained detection results (such as target detection results), such as generating prompt information or warning information, to prompt the staff or the robot to repair the spinning beam, thereby providing support for further improving the production quality and production efficiency of spinning.

[0052] In an example, the target detection model can be a dynamic weights-based wavelet attention neural network (DWWA-Net), or can be other models with target detection capability, which is not limited in the present disclosure.

[0053] Further, in an example, the target detection model can be a dynamic weights-based wavelet attention neural network, or can be other models improved based on a dynamic wavelet convolution network (DWCNet), which is not limited in the present disclosure.

[0054] Further, in a specific example, the target detection model at least includes a first network layer, a second network layer, and a third network layer.

[0055] Specifically, the first network layer is used to filter out background noise of the input image to be detected to extract a low-level feature map, for example, the first network layer is a dynamic wavelet convolution network. Further, as shown in FIG. 2(a), the first network layer at least includes an initial convolution module, a weight distribution module, a wavelet convolution module, and a feature fusion module; the initial convolution module is used to at least perform wavelet transform on the image to be detected to obtain a first low-frequency feature map; the weight distribution module is used to obtain a target weight factor of the feature map of the input image; the wavelet convolution module is used to perform wavelet transform on the image to be detected to obtain a global feature map representing low-frequency components and high-frequency components; and the feature fusion module is used to fuse the first low-frequency feature map and the global feature map based on the target weight factor to filter out background noise and obtain a low-level feature map.

[0056] Further, the second network layer is used to perform feature enhancement processing on the low-level feature map to obtain a high-level feature map; for example, in an example, the second network layer can be specifically used to enhance the key feature information in the low-level feature map while suppressing irrelevant feature information in the low-level feature map, such as suppressing feature information irrelevant to the spindles, to obtain a high-level feature map.

[0057] Further, the third network layer is configured to perform defect recognition based on the high-level feature map to obtain a target detection result.

[0058] In this way, the disclosed scheme provides a specific example of a model for quickly detecting defects of a spinning beam, thereby achieving efficient detection of defects in the spinning beam, especially for weak defects (such as floating yarn and hooked yarn) that are not easy to be detected in the spinning beam, and further improving the accuracy of defect detection of the spinning beam, thereby laying a foundation for subsequent improvement of production quality and production efficiency of the spinning.

[0059] Further, in a specific example, the initial convolution module is specifically configured to perform convolution processing on the to-be-detected image to obtain a first convolution feature map, and perform wavelet transform on the obtained first convolution feature map to obtain a first low-frequency feature map; for example, in an example, the to-be-detected image is subjected to convolution processing to obtain a feature map (such as a first convolution feature map) after convolution processing; secondly, the feature map after convolution processing is subjected to wavelet decomposition processing to obtain a plurality of feature maps; finally, the obtained plurality of feature maps are input to a low-pass filter for low-pass filter convolution processing to obtain a first low-frequency feature map. In this way, a low-noise feature map can be effectively extracted, thereby laying a foundation for subsequent improvement of the accuracy of defect detection.

[0060] Further, in a specific example, the weight distribution module is configured to perform convolution processing on the to-be-detected image to obtain a second convolution feature map, perform wavelet transform on the obtained second convolution feature map to obtain a second low-frequency feature map, and perform nonlinear transformation on the obtained second low-frequency feature map to obtain a target weight factor. In this way, the disclosed scheme can make full use of the importance of the low-noise low-frequency feature map in the original image (such as the to-be-detected image) to adjust the background noise in the to-be-detected image, especially for relatively subtle defects (such as floating yarn and hooked yarn), effective feature information can be conveniently identified, thereby laying a foundation for subsequent improvement of the accuracy of defect detection.

[0061] Further, in a specific example, the feature fusion module is configured to perform element-level processing (such as element multiplication) on the first low-frequency feature map and the target weight factor, and perform element-level processing on the global feature map and the target weight factor, and perform element addition on the results obtained after element-level processing to filter out the background noise to obtain a low-level feature map, thereby extracting effective feature information to improve the detection accuracy of defects of the yarn in the spinning beam.

[0062] In this way, the feature map fusion module can effectively combine the extracted local feature information (such as the first low-frequency feature map) and the global feature information (such as the global feature map), obtain a higher-quality feature map, and thus better capture the effective feature information of the defect, thereby improving the accuracy of defect detection and laying a foundation for subsequent improvement of the production quality and production efficiency of the spinning.

[0063] For example, as shown in FIG. 2(b), the first network layer is DWCNet, which includes three branches. The image to be detected is input to the convolution layer via branch 1 (i.e., ① in the figure) for convolution processing, and the obtained result is input to the wavelet neural network to obtain a first low-frequency feature map. The image to be detected is input to the convolution layer via branch 2 (i.e., ② in the figure) for convolution processing, and the obtained result is input to the wavelet neural network to obtain a second low-frequency feature map, which is then input to the S-shaped bending function (Sigmoid) to obtain a target weight factor (such as a target weight factor w or a target weight factor 1-w). The image to be detected is input to the wavelet neural network via branch 3 (i.e., ③ in the figure) to obtain a global feature map representing low-frequency information and high-frequency information. The target weight factor, such as w, is multiplied by the first low-frequency feature map, and 1-w is multiplied by the global feature map, and the results of the element processing are added to filter out background noise and obtain a low-level feature map. In this way, by combining the extracted local feature information (such as the first low-frequency feature map) and the global feature information (such as the global feature map), a higher-quality feature map is obtained, which facilitates better capture of the feature information of the image, thereby improving the accuracy of defect detection and laying a foundation for subsequent improvement of the production quality and production efficiency of the spinning.

[0064] In another example of the present disclosure, the second network layer includes at least a first sub-network layer, a second sub-network layer, and a third sub-network layer.

[0065] Here, the first sub-network layer is configured to extract features from the low-level feature map and fuse the extracted feature maps to obtain M initial fusion feature maps, where M is an integer greater than or equal to 2; the second sub-network layer is configured to extract key features from each of the M initial fusion feature maps and perform feature enhancement processing on the extracted key feature information of each initial fusion feature map to obtain M target enhanced feature maps; and the third sub-network layer is configured to fuse the M target enhanced feature maps to obtain a high-level feature map. In this way, clear and rich feature information (such as a high-level feature map) can be obtained, thereby improving the accuracy of defect detection, and in particular, the detection accuracy of weak defects (such as stray filaments and hooked filaments) in the spinning box, thereby laying a foundation for subsequent improvement of the production quality and production efficiency of the spinning.

[0066] Further, in an example, as shown in FIG. 3(a), the jth initial fusion feature map in the M initial fusion feature maps is obtained based on the following manner:

[0067] performing convolution processing on the jth level feature map to extract features to obtain a (j+1)th level feature map; here, when j is 1, the 1st level feature map is obtained by performing convolution processing on the low-level feature map, and in this way, M level feature maps can be obtained;

[0068] when j takes the value of M, the Mth level feature map is directly taken as the 1st initial fusion feature map;

[0069] when j takes the value of an integer greater than or equal to 1 and less than or equal to M-1, the jth level feature map and the (j-1)th initial fusion feature map are fused to obtain the jth initial fusion feature map.

[0070] For example, the first sub-network layer is a Feature Pyramid Network (FPN) layer, the second sub-network layer is a feature enhancement network layer, and the third sub-network layer is an Addition (Add) processing layer; as shown in FIG. 3(b), first, the low-level feature map is input into the FPN layer to obtain M initial fusion feature maps; second, each of the obtained initial fusion feature maps is input into the feature enhancement network layer, such as performing key feature extraction on each initial fusion feature map through the feature enhancement network layer, and performing feature enhancement processing on the extracted feature maps to obtain M target enhanced feature maps; and finally, the M target enhanced feature maps are input into the linear addition processing layer to obtain a high-level feature map. In this way, the accuracy of defect and weak defect detection can be improved, thereby laying a foundation for subsequent improvement of the production quality and production efficiency of the spinning.

[0071] Here, in an example, the i-th target enhanced feature map in the M target enhanced feature maps is obtained based on the following manner:

[0072] The i-th initial fusion feature map is subjected to convolution processing to obtain an i-th weight factor based on a convolution processing result; based on the i-th weight factor, the i+1-th initial fusion feature map is fused with the i-th initial fusion feature map to obtain an i-th target fusion feature map; the obtained i-th target fusion feature map is subjected to key feature extraction, and the extracted multiple feature maps are respectively subjected to enhancement processing to obtain multiple i-th initial enhanced feature maps, and the multiple i-th initial enhanced feature maps are fused to obtain an i-th target enhanced feature map. Here, i is an integer greater than or equal to 1 and less than or equal to M-1.

[0073] Further, in the case where i is M, at this time, the M+1-th initial fusion feature map can be a preset value, at this time, the M-th target fusion feature map can be obtained, and then the M-th target enhanced feature map can be obtained. Alternatively, the M-th target enhanced feature map can also be obtained in the following manner:

[0074] The M-th initial fusion feature map is subjected to key feature extraction, and the extracted multiple feature maps are respectively subjected to enhancement processing to obtain multiple M-th initial enhanced feature maps, and the multiple M-th initial enhanced feature maps are fused to obtain an M-th target enhanced feature map.

[0075] For example, continuing to take the second sub-network layer as an example, at this time, the feature enhancement network layer can further include a feature weight module and a multi-view attention module; as shown in FIG. 3(c), the i-th initial fusion feature map and the i+1-th initial fusion feature map in the M initial fusion feature maps are input into the feature weight module to obtain an i-th target fusion feature map, and the obtained i-th target fusion feature map is input into the multi-view attention module for key feature extraction and feature enhancement processing to obtain an i-th target enhanced feature map; in addition, for the M-th initial fusion feature map, the M-th initial fusion feature map can be directly taken as an M-th target fusion feature map, and the obtained M-th target fusion feature map is input into the multi-view attention module for key feature extraction and feature enhancement processing to obtain an M-th target enhanced feature map.

[0076] Further, as shown in FIG. 3(d), the feature weight module is used to: firstly, the i+1-th initial fusion feature map is subjected to up-sampling processing to obtain a processed i+1-th initial fusion feature map; the i-th initial fusion feature map is subjected to convolution processing, and the obtained result is subjected to S-shaped bending function (Sigmod) processing to obtain an i-th weight factor (such as λ i , or 1-λ i ); secondly, the i-th weight factor, such as 1-λi element-wise multiplication with the i-th initial fusion feature map, and i element-wise multiplication with the i-th initial fusion feature map, and

[0077] Further, as shown in FIG. 3(e), the multi-view attention module is used to, first, perform deformable convolution processing on the i-th target fusion feature map with a first preset convolution kernel (such as 1x1) to perform key feature extraction, to obtain a feature map 1 of key features, and perform deformable convolution processing on the i-th target fusion feature map with a second preset convolution kernel (such as 3x3) to obtain a feature map 2 of key features, and perform deformable convolution processing on the i-th target fusion feature map with a third preset convolution kernel (such as 5x5) to obtain a feature map 3 of key features; second, input the obtained feature maps 1, 2 and 3 into an attention sub-module A to perform enhancement processing on the feature map 1 to obtain an i-th initial enhanced feature map 1, and input the feature map 2 into an attention sub-module B to obtain an i-th initial enhanced feature map 2, and input the feature map 3 into an attention sub-module C to obtain an i-th initial enhanced feature map 3; and finally, perform fusion processing (such as element-wise addition) on the above three i-th initial enhanced feature maps to obtain an i-th target enhanced feature map. Figure 1 Figure 1 Figure 1 Figure 1

[0078] Further, on the basis of FIG. 3(e), a neural network can be added, such as a wavelet neural network, to directly perform feature extraction on the i-th target fusion feature map by using the wavelet neural network to obtain an i-th supplementary feature map, and perform fusion processing with the above three i-th initial enhanced feature maps. Figure 4

[0079] It should be noted that the structure in the attention sub-module (such as the attention sub-module A, or the attention sub-module B, or the attention sub-module C) included in the multi-view attention module can be a Squeeze and Excitation (SE) network, or can be other network structures based on the SE network, and the present disclosure does not limit this.

[0080] ​​​​​In this way, the scheme disclosed in the present application can make full use of the second sub-network layer to extract key features in the low-level feature map and perform feature enhancement processing on the extracted results, thereby guiding the attention of the network to the potential target (i.e., the defects in the spinning beam) and facilitating subsequent better defect identification. In particular for the yarns and background plates in the spinning beam that are difficult to distinguish in color, key features can be more comprehensively obtained, thereby improving the detection accuracy of defects and laying a foundation for subsequent improvement of the production quality and production efficiency of spinning.

[0081] In still another example of the scheme disclosed in the present application, the third network layer includes a fourth sub-network layer and a fifth sub-network layer.

[0082] Here, the fourth sub-network layer is configured to frame a candidate region of a suspected defect in the image based on the high-level feature map, and the fifth sub-network layer is configured to identify whether the candidate region has a defect based on the high-level feature map and the framed candidate region. In this way, defects that may exist in the spinning beam can be efficiently identified, thereby laying a foundation for subsequent improvement of the production quality and production efficiency of spinning.

[0083] Figure 1 FIG. 2 is a schematic flowchart of a method for detecting a spinning beam according to an embodiment of the present application. The method can be optionally applied in electronic devices such as personal computers, servers, server clusters, and the like. It can be understood that the related content of the method shown in FIG. 1 can also be applied in this example, and the associated content will not be described again. Figure 4 The method shown in FIG. 1 can also be applied in this example, and the associated content will not be described again.

[0084] Further, the method at least includes at least part of the following content. As shown in FIG. 2, the method includes: Figure 5

[0085] Step S401: When it is determined that the spinning beam meets the preset defect detection condition, acquiring a to-be-detected image of the spinning beam.

[0086] Step S402: Inputting the to-be-detected image into a target detection model to obtain a target detection result of the spinning beam.

[0087] Here, the target detection model is configured to detect whether the spinning beam has defects to obtain the target detection result, and the target detection result includes at least one of the following: the total number of defects, the defect position, and the defect type.

[0088] Step S403: Determining whether the target detection result of the spinning beam meets a preset defect requirement. If yes, proceed to step S405, otherwise proceed to step S404.

[0089] ​Step S404: In a case where the target detection result of the spinning beam does not satisfy the preset defect requirement, generating abnormal prompt information corresponding to the target detection result.

[0090] Here, the abnormal prompt information is used to prompt that the spinning beam has defects.

[0091] Further, in a specific example, the abnormal prompt information can be generated at a time when it is convenient to prompt a worker or a robot to repair the spinning beam, so as to further ensure the production quality and production efficiency of spinning. Specifically, the above-mentioned determination that the target detection result of the spinning beam does not satisfy the preset defect requirement, and the generation of the abnormal prompt information corresponding to the target detection result (such as the above-mentioned step S404) specifically include:

[0092] In a case where at least one of the following conditions is satisfied, the abnormal prompt information corresponding to the target detection result is generated:

[0093] The total number of defects contained in the target detection result is greater than a second threshold value;

[0094] The defect position contained in the target detection result is within a preset defect position range (for example, the total defect size is within a preset defect position range, and exceeds a preset area);

[0095] The defect type contained in the target detection result is within a preset defect type range (for example, if the target defect type contains all the defect types contained in the first detection result, and at least one defect type falls within the preset defect type range, it is considered that the target defect type is within the preset defect type range).

[0096] Here, the second threshold value is an empirical value, which can be set according to actual needs in actual application, and the present disclosure does not make specific limitations on the second threshold value.

[0097] Step S405: End the process.

[0098] In this way, the present disclosure can generate corresponding prompt information in a case where the obtained defect detection result does not satisfy the preset defect requirement, so as to facilitate a worker or a robot to take corresponding measures on the spinning beam, such as repairing the spinning beam, so that the production quality of spinning can meet the expected requirements, and the production efficiency of the workshop is ensured.

[0099] In summary, compared with the prior art, the present disclosure has the following advantages, specifically including:

[0100] First, it achieves automated processing. Compared with existing manual methods, the disclosed solution can efficiently detect defects in the spinning beam without relying on manual experience. This realizes an automated processing flow, thereby saving a lot of manpower and time costs and further improving the efficiency of defect detection.

[0101] Second, it improves detection accuracy. Traditional manual inspection methods have difficulty detecting subtle defects such as loose threads and snagged threads in the spinning beam, and are prone to missed detection. The disclosed solution uses a neural network model suitable for defect detection, which improves detection accuracy, especially for weak defects that are difficult to detect, thus ensuring the quality of the spindles produced by the spinning beam.

[0102] Third, improve production efficiency. The disclosed solution can automatically and accurately detect defects in the spinning manifold, and then take corresponding measures in a timely manner based on the obtained detection results (such as target detection results), such as generating prompts or warning messages to prompt staff or robots to repair the spinning manifold, thereby further ensuring the normal production of spinning in the workshop.

[0103] The disclosed solution also provides a detection device for a spinning box, such as Figure 6 Shown, including:

[0104] An acquiring unit 501 is configured to acquire an image of the spinning manifold to be inspected when determining that the spinning manifold meets a preset defect detection condition;

[0105] The detection unit 502 is used to input the image to be detected into the target detection model to obtain the target detection result of the spinning manifold; the target detection model is used to detect whether there are defects in the spinning manifold to obtain the target detection result; the target detection result includes at least one of the following: the total number of defects, the defect location, and the defect type.

[0106] In a specific example of the disclosed solution, the acquisition unit is specifically configured to:

[0107] When it is determined that one of the following conditions is met, the image to be detected of the spinning beam is obtained:

[0108] The detection time of the spinning box reaches the preset detection time;

[0109] determining whether there are degraded spindles among the plurality of spindles obtained based on the spinning beam;

[0110] It is determined that the number of the degraded spindles among the plurality of spindles obtained based on the spinning beam is greater than a first threshold.

[0111] In a specific example of the present disclosure, the method further includes: an exception prompt unit; wherein

[0112] The exception prompt unit is configured to generate exception prompt information corresponding to the target detection result when the target detection result of the spinning beam does not meet the preset defect requirement; and the exception prompt information is used to prompt the defect degree of the spinning beam.

[0113] In a specific example of the present disclosure, the exception prompt unit is specifically configured to:

[0114] Generate the exception prompt information corresponding to the target detection result when at least one of the following conditions is met:

[0115] The total number of defects contained in the target detection result is greater than a second threshold value;

[0116] The defect position contained in the target detection result is within a preset defect position range;

[0117] The defect type contained in the target detection result is within a preset defect type range.

[0118] In a specific example of the present disclosure, the target detection model at least includes a first network layer, a second network layer, and a third network layer;

[0119] The first network layer is configured to filter out background noise of an input image to be detected to extract a low-level feature map; wherein the first network layer at least includes an initial convolution module, a weight distribution module, a wavelet convolution module, and a feature fusion module; the initial convolution module is configured to at least perform wavelet transform on the image to be detected to obtain a first low-frequency feature map; the weight distribution module is configured to obtain a target weight factor of a feature map of the input image; the wavelet convolution module is configured to perform wavelet transform on the image to be detected to obtain a global feature map representing low-frequency components and high-frequency components; and the feature fusion module is configured to perform fusion processing on the first low-frequency feature map and the global feature map based on the target weight factor to filter out background noise and obtain a low-level feature map.

[0120] The second network layer is configured to perform feature enhancement processing on the low-level feature map to obtain a high-level feature map.

[0121] The third network layer is configured to perform defect recognition based on the high-level feature map to obtain a target detection result.

[0122] In a specific example of the present disclosure, the initial convolution module is configured to perform convolution processing on the image to be detected to obtain a first convolution feature map, and perform wavelet transform on the obtained first convolution feature map to obtain a first low-frequency feature map.

[0123] In a specific example of the present disclosure, the weight distribution module is configured to perform convolution processing on the to-be-detected image to obtain a second convolution feature map, perform wavelet transform on the obtained second convolution feature map to obtain a second low-frequency feature map, and perform nonlinear transformation on the obtained second low-frequency feature map to obtain a target weight factor.

[0124] In a specific example of the present disclosure, the feature fusion module is configured to perform element-level processing on the first low-frequency feature map and the target weight factor, and perform element-level processing on the global feature map and the target weight factor, and perform element addition on the results obtained after the element-level processing to filter out background noise to obtain a low-level feature map.

[0125] The specific functions and examples of the units of the apparatus of the embodiments of the present disclosure are described above in the related description of the corresponding steps in the method embodiments, which will not be described here again.

[0126] In the technical solutions of the present disclosure, the acquisition, storage and application of user personal information comply with relevant laws and regulations and do not violate public order and good customs.

[0127] Figure 6 FIG. 1 is a structural block diagram of an electronic device according to an embodiment of the present disclosure. As shown in FIG. 1, the electronic device includes a memory 610 and a processor 620, and the memory 610 stores a computer program that can run on the processor 620. Figure 6 The number of the memory 610 and the processor 620 can be one or more. The memory 610 can store one or more computer programs, which, when executed by the electronic device, cause the electronic device to perform the method provided in the above method embodiments. The electronic device can further include a communication interface 630 for communicating with external devices and performing data transmission and reception.

[0128] If the memory 610, the processor 620 and the communication interface 630 are independently implemented, the memory 610, the processor 620 and the communication interface 630 can be connected to each other through a bus and complete communication therebetween. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus and a control bus. For the convenience of representation, ​Only one bus or only one type of bus can be present.

[0129] Optionally, if the memory 610, the processor 620 and the communication interface 630 are integrated on a chip, the memory 610, the processor 620 and the communication interface 630 can complete the communication among each other through an internal interface.

[0130] It should be understood that the processor described above can be a central processing unit (CPU), and can also be other general-purpose processors, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc. It should be noted that the processor can be a processor supporting an advanced RISC machine (ARM) architecture.

[0131] Further, the aforementioned memory can include a read-only memory and a random access memory, and can also include a non-volatile random access memory. The memory can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memory. The non-volatile memory can include a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically EPROM (EEPROM), or a flash memory. The volatile memory can include a random access memory (RAM) used as an external cache. By way of example, and not limitation, many forms of RAM can be used. For example, a static random access memory (SRAM), a dynamic random access memory (DRAM), a synchronous dynamic random access memory (SDRAM), a double data rate synchronous dynamic random access memory (DDR SDRAM), an enhanced SDRAM (ESDRAM), a Synchlink DRAM (SLDRAM), and a direct Rambus RAM (DR RAM) can be used.

[0132] In the above embodiments, all or part of the steps can be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the steps can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present disclosure are generated. The computer can be a general purpose computer, a special purpose computer, a computer network or other programmable apparatus. The computer instructions can be stored in a computer readable storage medium or transferred from one computer readable storage medium to another computer readable storage medium, for example, the computer instructions can be transferred from one website, computer, server or data center to another website, computer, server or data center through wired (for example: coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (for example: infrared, Bluetooth, microwave, etc.) mode. The computer readable storage medium can be any available medium accessible by a computer, or a data storage device such as a server, data center, etc. integrated with one or more available media. The available media can be a magnetic medium (for example: floppy disk, hard disk, magnetic tape), an optical medium (for example: digital versatile disc (DVD)) or a semiconductor medium (for example: solid state disk (SSD)) etc. It is worth noting that the computer readable storage medium mentioned in the present disclosure can be a non-volatile storage medium, in other words, it can be a non-transitory storage medium.

[0133] A person of ordinary skill in the art can understand that all or part of the steps of the above embodiments can be completed by hardware, or the program can instruct the related hardware to complete, and the program can be stored in a computer readable storage medium. The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk.

[0134] In the description of the embodiments of the present disclosure, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example" or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present disclosure. Moreover, the specific features, structures, materials or characteristics described can be combined in any appropriate manner in any one or more embodiments or examples. In addition, those skilled in the art can combine and combine the different embodiments or examples described in the specification and the features of the different embodiments or examples without contradiction.

[0135] In the description of the embodiments of the present disclosure, unless otherwise specified, " / " means or, for example, A / B can mean A or B. "And / or" in this document only describes the relationship between associated objects, which means that there can be three relationships, for example, A and / or B, which can mean: A exists alone, A and B exist together, and B exists alone.

[0136] In the description of the embodiments of the present disclosure, the terms "first", "second" are only for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more of the features. In the description of the embodiments of the present disclosure, unless otherwise specified, the meaning of "multiple" is two or more.

[0137] The above only describes exemplary embodiments of the present disclosure, and is not intended to limit the present disclosure. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for detecting a spinning beam, comprising: When it is determined that the spinning manifold meets the preset defect detection conditions, an image of the spinning manifold to be detected is obtained, wherein the image to be detected includes at least image content of the spinneret, yarn, wire guide hook, and nozzle in the spinning manifold; Inputting the image to be detected into a target detection model to obtain a target detection result of the spinning manifold; the target detection model is used to detect whether there are defects in the spinning manifold to obtain a target detection result; the target detection result includes at least one of the following: the total number of defects, the defect location, and the defect type; The target detection model includes at least a first network layer, a second network layer, and a third network layer; The first network layer is used to filter the background noise of the input image to be detected to extract a low-level feature map; wherein the first network layer includes at least an initial convolution module, a weight distribution module, a wavelet convolution module and a feature fusion module; the initial convolution module is used to perform a wavelet transform on at least the image to be detected to obtain a first low-frequency feature map; the weight distribution module is used to obtain a target weight factor of the feature map of the input image; the wavelet convolution module is used to perform a wavelet transform on the image to be detected to obtain a global feature map representing low-frequency components and high-frequency components; the feature fusion module is used to fuse the first low-frequency feature map with the global feature map based on the target weight factor to filter the background noise and obtain a low-level feature map; The second network layer is used to perform feature enhancement processing on the low-level feature map to obtain a high-level feature map; The third network layer is used to perform defect recognition based on the high-level feature map to obtain target detection results.

2. The method according to claim 1, wherein When it is determined that the spinning manifold meets the preset defect detection conditions, obtaining the image of the spinning manifold to be detected includes: When it is determined that one of the following conditions is met, the image to be detected of the spinning beam is obtained: The detection time of the spinning box reaches the preset detection time; determining whether there are degraded spindles among the plurality of spindles obtained based on the spinning beam; It is determined that the number of the degraded spindles among the plurality of spindles obtained based on the spinning beam is greater than a first threshold.

3. The method according to claim 1, further comprising: When it is determined that the target detection result of the spinning beam does not meet the preset defect requirements, generating abnormal prompt information corresponding to the target detection result; The abnormal prompt information is used to prompt the defect degree of the spinning manifold.

4. The method according to claim 3, wherein: When it is determined that the target detection result of the spinning beam does not meet the preset defect requirement, generating abnormal prompt information corresponding to the target detection result includes: Generate abnormal prompt information corresponding to the target detection result when at least one of the following is determined to be met: The total number of defects included in the target detection result is greater than a second threshold; The defect position included in the target detection result is within the preset defect position range; The defect types included in the target detection result are within a preset defect type range.

5. The method according to any one of claims 1 to 4, wherein: The initial convolution module is used to perform convolution processing on the image to be detected to obtain a first convolution feature map, and perform wavelet transform on the obtained first convolution feature map to obtain a first low-frequency feature map.

6. The method according to any one of claims 1 to 4, wherein: The weight distribution module is used to perform convolution processing on the image to be detected to obtain a second convolution feature map, perform wavelet transformation on the obtained second convolution feature map to obtain a second low-frequency feature map, and perform nonlinear transformation on the obtained second low-frequency feature map to obtain a target weight factor.

7. The method according to any one of claims 1 to 4, wherein: The feature fusion module is used to perform element-level processing on the first low-frequency feature map and the target weight factor, and to perform element-level processing on the global feature map and the target weight factor, and to perform element-wise addition on the results obtained after the element-level processing to filter out background noise and obtain a low-level feature map.

8. A detection device for a spinning beam, comprising: an acquisition unit, configured to acquire an image of the spinning manifold to be inspected when determining that the spinning manifold meets a preset defect detection condition, wherein the image to be inspected includes at least image contents of a spinneret, a yarn, a yarn guide hook, and an oil nozzle in the spinning manifold; a detection unit, configured to input the image to be detected into a target detection model to obtain a target detection result of the spinning manifold; the target detection model is configured to detect whether the spinning manifold has defects to obtain a target detection result; the target detection result includes at least one of the following: a total number of defects, a defect location, and a defect type; The target detection model includes at least a first network layer, a second network layer, and a third network layer; The first network layer is used to filter the background noise of the input image to be detected to extract a low-level feature map; wherein the first network layer includes at least an initial convolution module, a weight distribution module, a wavelet convolution module and a feature fusion module; the initial convolution module is used to perform a wavelet transform on at least the image to be detected to obtain a first low-frequency feature map; the weight distribution module is used to obtain a target weight factor of the feature map of the input image; the wavelet convolution module is used to perform a wavelet transform on the image to be detected to obtain a global feature map representing low-frequency components and high-frequency components; the feature fusion module is used to fuse the first low-frequency feature map with the global feature map based on the target weight factor to filter the background noise and obtain a low-level feature map; The second network layer is used to perform feature enhancement processing on the low-level feature map to obtain a high-level feature map; The third network layer is used to perform defect recognition based on the high-level feature map to obtain target detection results.

9. The device according to claim 8, wherein The acquisition unit is specifically configured to: When it is determined that one of the following conditions is met, the image to be detected of the spinning beam is obtained: The detection time of the spinning box reaches the preset detection time; determining whether there are degraded spindles among the plurality of spindles obtained based on the spinning beam; It is determined that the number of the degraded spindles among the plurality of spindles obtained based on the spinning beam is greater than a first threshold.

10. The apparatus according to claim 8, further comprising: Abnormal prompt unit; wherein, The abnormality prompt unit is used to generate abnormality prompt information corresponding to the target detection result when determining that the target detection result of the spinning manifold does not meet the preset defect requirements; the abnormality prompt information is used to indicate the defect degree of the spinning manifold.

11. The device according to claim 10, wherein The abnormality prompt unit is specifically used to: Generate abnormal prompt information corresponding to the target detection result when at least one of the following is determined to be met: The total number of defects included in the target detection result is greater than a second threshold; The defect position included in the target detection result is within the preset defect position range; The defect types included in the target detection result are within a preset defect type range.

12. The device according to any one of claims 8 to 11, wherein: The initial convolution module is used to perform convolution processing on the image to be detected to obtain a first convolution feature map, and perform wavelet transform on the obtained first convolution feature map to obtain a first low-frequency feature map.

13. The device according to any one of claims 8 to 11, wherein: The weight distribution module is used to perform convolution processing on the image to be detected to obtain a second convolution feature map, perform wavelet transformation on the obtained second convolution feature map to obtain a second low-frequency feature map, and perform nonlinear transformation on the obtained second low-frequency feature map to obtain a target weight factor.

14. The device according to any one of claims 8 to 11, wherein: The feature fusion module is used to perform element-level processing on the first low-frequency feature map and the target weight factor, and to perform element-level processing on the global feature map and the target weight factor, and to perform element-wise addition on the results obtained after the element-level processing to filter out background noise and obtain a low-level feature map.

15. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 7.

16. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1 to 7.

Images