Automatic detection system and detection method

Abstract

The invention provides an automatic detection system and a detection method. The system comprises a transportation component used for bearing a target component; the detection device is arranged outside the transportation component in a covering manner so as to detect the target component transported to the detection area by the transportation component; wherein the detection device comprises a light source module and a shooting element, the light source module comprises a light emitting part used for emitting 405nm ultraviolet light and 650nm red light, the shooting element is a 2048-pixel linear array camera, and when a light source emitted by the light source module irradiates a target part of a target part, shooting is carried out through the shooting element to obtain a preprocessed picture; the processing unit is connected with the shooting element and is used for carrying out polarization difference processing on the preprocessed picture so as to identify the defect on the target part; the problem that hidden defects such as microscopic cracks and stress whitening are difficult to capture by using a common camera when the defects of the fan blade are detected in the prior art is solved.

Description

Technical Field

[0001] This application relates to the field of air conditioning technology, and more specifically, to an automatic detection system and detection method. Background Technology

[0002] In modern manufacturing, especially in the production of air conditioning cross-flow fan blades, traditional manual inspection or existing automated inspection technologies have always been important means of product quality control. Manual inspection relies on the operator's experience and intuition to make a preliminary assessment of the product's appearance and functionality, but this method has significant limitations, such as low inspection efficiency, high labor intensity, and uncertainty caused by subjective judgment. With technological advancements, the introduction of two-dimensional vision systems has enabled automated inspection. Utilizing cameras and image processing algorithms, defects such as cracks, scratches, or color differences on the product surface can be inspected without contact, greatly improving the speed and consistency of inspection. Automated optical inspection systems typically combine high-performance vision sensors, complex optical components, and precise mechanical design to adapt to products moving rapidly on the production line. These systems can capture product images under static or dynamic conditions using high-speed cameras and use image processing algorithms for defect identification, thereby achieving real-time quality monitoring of products. However, despite the significant improvements brought by two-dimensional vision systems, their performance is limited in certain application scenarios, especially when dealing with the inspection of complex curved surfaces and highly reflective surfaces.

[0003] However, existing technologies for automated optical inspection of air conditioning cross-flow fan blades face multiple technical challenges, particularly in handling specular and diffuse reflection noise, which severely impacts the accuracy and consistency of inspection results. Traditional two-dimensional vision inspection systems are limited by the characteristics of the fan blade materials—the strong specular reflection of the metal core and the complex diffuse reflection of the plastic blades—making it difficult to capture microscopic cracks ≤0.3mm and latent defects such as stress whitening while maintaining the image signal-to-noise ratio. In such cases, images captured by ordinary cameras in dynamic environments often contain a large amount of noise, reducing the accuracy of crack and defect identification and resulting in a false positive rate of over 30%. Summary of the Invention

[0004] The main objective of this invention is to provide an automatic detection system and method to solve the problem in the prior art that when detecting defects in wind turbine blades, it is difficult to capture hidden defects such as micro-cracks and stress whitening using ordinary cameras.

[0005] To achieve the above objectives, according to one aspect of the present invention, an automatic detection system is provided for detecting defects on a target component. The automatic detection system includes a transport component that is movably disposed and is used to carry the target component.

[0006] A detection device is installed outside the transport component to form a detection zone with at least a portion of the transport component, for detecting target components transported from the transport component to the detection zone;

[0007] The detection device includes a light source module and an imaging element. The light source module includes a light-emitting component for emitting 405nm ultraviolet light and 650nm red light. The imaging element is a 2048-pixel line array camera, which takes a picture when the light source emitted by the light source module illuminates the target part of the target component to obtain a pre-processed image.

[0008] The processing unit, connected to the imaging element, performs polarization difference processing on the pre-processed image to identify defects on the target component.

[0009] Furthermore, the testing device also includes a support body, which is disposed outside the transport component;

[0010] A clamping assembly is disposed on a support body. The clamping assembly is movably disposed in the vertical direction, and at least a portion of the clamping assembly is movably disposed in the horizontal direction. When the clamping assembly is in a first position, the at least portion of the clamping assembly is controlled to move in the horizontal direction to clamp the target component and drive the target component to rotate.

[0011] The imaging element is mounted on the clamping assembly, and the positive light emission direction of the imaging element is oriented towards the transport component.

[0012] Furthermore, the clamping assembly includes a clamping beam arranged in a horizontal direction, wherein the imaging element is disposed on the clamping beam;

[0013] Two support components are movably mounted at both ends of the clamping beam;

[0014] Two first telescopic elements are both set on the clamping crossbeam. The two first telescopic elements are set one-to-one with the two support components. Each first telescopic element is connected to its corresponding support component so that the two first telescopic elements can drive the two support components to move closer to or further away from each other in order to clamp or release the target component.

[0015] Furthermore, the clamping assembly also includes a clamping beam, which is arranged horizontally, wherein the imaging element is disposed on the clamping beam;

[0016] Two support components are respectively located at both ends of the clamping beam;

[0017] The first telescopic element is disposed on the clamping beam and is connected to one of the supporting components to drive the supporting component to move horizontally relative to the clamping beam in order to clamp or release the target component.

[0018] Furthermore, the clamping assembly also includes a rotating assembly, which includes two rotating parts rotatably disposed on two support parts, so that when the clamping assembly is in the first position, the two support parts are controlled to move toward each other, or one of the support parts is controlled to move toward the other support part, so that at least a portion of the rotating parts contacts the target part to clamp the target part and drive the target part to rotate.

[0019] Furthermore, the automatic detection system also includes a limiting component having a limiting surface, on which the target part is placed to limit its movement; and / or,

[0020] The support body includes a main body and a second telescopic element disposed on the main body. The telescopic end of the second telescopic element is connected to the clamping assembly, so as to drive the clamping assembly to move in the vertical direction via the second telescopic element; and / or,

[0021] The automatic detection system also includes two detection components, which are respectively mounted on the two support components of the clamping assembly, with the detection end of the detection component facing the support component opposite to it.

[0022] According to another aspect of the present invention, a detection method is provided for detecting defects on a target component, the detection direction being applicable to the aforementioned automatic detection system, and the detection method comprising:

[0023] When the target component is determined to be in the detection area, the light source module and the imaging element are turned on. When the light source of the light source module illuminates the target component, the imaging element is controlled to take a picture of the target component to obtain a pre-processed image.

[0024] The pre-processed image is input into the processing unit for polarization difference processing to identify defects on the target component;

[0025] The light source module includes light-emitting components for emitting 405nm ultraviolet light and 650nm red light.

[0026] Furthermore, after the step of inputting the preprocessed image into the processing unit for polarization difference processing, the detection method also includes:

[0027] Construct the U-Net model;

[0028] The differential image obtained after polarization difference processing is input into the U-Net model for segmentation to output multiple segmentation masks. The segmentation masks include crack regions and whitening defect regions. Features are extracted and classified from the crack regions and whitening defect regions to determine the type of crack and the type of whitening defect. Among them, the types of cracks include primary cracks and secondary cracks, and the types of whitening defects include primary defects and secondary defects. When the type of crack is a secondary crack or the type of whitening defect is a secondary defect, the target part is marked as defective.

[0029] Further, the steps of feature extraction and classification of the cracked region to determine the type of crack include:

[0030] The extracted crack features are input into a CBAM attention module for identification and classification. When the crack length is greater than 0.3 mm and the crack width is less than or equal to 0.5 mm, the crack type is determined to be a secondary type; otherwise, the crack type is determined to be a primary type.

[0031] Furthermore, features are extracted and classified from the whitening defect areas to determine the types of whitening defects, including:

[0032] The extracted whitening defect features are input into a CBAM attention module for identification and classification. When the area of ​​the whitening defect is greater than 2 mm... 2 If the whitening defect is found to be of the second-level defect type, then the whitening defect is determined to be of the first-level defect type.

[0033] By integrating multispectral (405nm ultraviolet and 650nm red light) illumination with polarization differential image processing, the technical solution of this invention effectively overcomes imaging interference caused by specular reflection from metal core surfaces and diffuse reflection from plastic curved surfaces, significantly improving detection accuracy. The detection rate of cracks and whitening defects reaches over 99.7%. The system employs a 2048-pixel linear array camera, coupled with microsecond-level synchronous triggering technology, ensuring clear and unblurred images even at high-speed fan rotation of 120rpm. This fully meets the production line cycle time requirement of ≤5 seconds per unit, achieving high-speed dynamic detection. Attached Figure Description

[0034] The accompanying drawings, which form part of this application, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0035] Figure 1 An isometric view of the detection system according to an embodiment of this application is shown;

[0036] Figure 2A front view of the detection system according to an embodiment of this application is shown.

[0037] The above figures include the following reference numerals:

[0038] 1. Target component; 2. Transport component; 3. Light source module; 4. Imaging element; 5. Support body; 51. Main body; 52. Second telescopic element; 53. Support foot; 6. Clamping assembly; 61. Clamping beam; 62. Support component; 63. First telescopic element; 64. Rotating assembly; 641. Rotating component; 7. Limiting assembly. Detailed Implementation

[0039] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0040] In modern manufacturing, especially in the production of air conditioning cross-flow fan blades, traditional manual inspection or existing automated inspection technologies have always been important means of product quality control. Manual inspection relies on the operator's experience and intuition to make a preliminary assessment of the product's appearance and functionality, but this method has significant limitations, such as low inspection efficiency, high labor intensity, and uncertainty caused by subjective judgment. With technological advancements, the introduction of two-dimensional vision systems has enabled automated inspection. Utilizing cameras and image processing algorithms, defects such as cracks, scratches, or color differences on the product surface can be inspected without contact, greatly improving the speed and consistency of inspection. Automated optical inspection systems typically combine high-performance vision sensors, complex optical components, and precise mechanical design to adapt to products moving rapidly on the production line. These systems can capture product images under static or dynamic conditions using high-speed cameras and use image processing algorithms for defect identification, thereby achieving real-time quality monitoring of products. However, despite the significant improvements brought by two-dimensional vision systems, their performance is limited in certain application scenarios, especially when dealing with the inspection of complex curved surfaces and highly reflective surfaces.

[0041] However, existing technologies for automated optical inspection of air conditioning cross-flow fan blades face multiple technical challenges, particularly in handling specular and diffuse reflection noise, which severely impacts the accuracy and consistency of inspection results. Traditional two-dimensional vision inspection systems are limited by the characteristics of the fan blade materials—the strong specular reflection of the metal core and the complex diffuse reflection of the plastic blades—making it difficult to capture microscopic cracks ≤0.3mm and latent defects such as stress whitening while maintaining the image signal-to-noise ratio. In such cases, images captured by ordinary cameras in dynamic environments often contain a large amount of noise, reducing the accuracy of crack and defect identification and resulting in a false positive rate of over 30%.

[0042] The main objective of this invention is to provide an automatic detection system and method to solve the problem in the prior art that when detecting defects in wind turbine blades, it is difficult to capture hidden defects such as micro-cracks and stress whitening using ordinary cameras.

[0043] Example 1

[0044] like Figure 1 and Figure 2 As shown, this application embodiment first provides an automatic detection system for detecting defects on a target component 1. In this embodiment, the target component 1 is an axial flow fan. The automatic detection system includes a support frame, a transport component 2, a detection device, a processing unit, a limiting component 7, and a detection element. The support frame is located on a support base, which in this embodiment can be the ground. The transport component 2, which in this embodiment is a belt, is mounted on the support frame. The limiting component 7 is mounted on the transport component 2, and includes two limiting platforms. Both limiting platforms have semi-circular limiting surfaces, the shape of which is adapted to the shape of the axial flow fan's housing. The openings of the two semi-circular limiting surfaces of the two limiting platforms face each other. The two limiting platforms are respectively located on both sides of the transport component 2 along its width direction to form the limiting component 7 for limiting the target component 1. By setting the limiting component 7, defects can be avoided during the detection of the target component 1. The target component 1 moves relative to the transport component 2. At least part of the detection device is mounted on the transport component 2. The detection device includes a support body 5, a light source module 3, and an imaging element 4. The support body 5 includes a main body 51 and a second telescopic element 52. The main body 51 is placed on the bottom surface by support feet 53. There are two support feet 53, which are respectively located on both sides of the main body 51. The support body 5 and the transport component 2 together form a detection area. The detection device also includes a clamping assembly 6 mounted on the support body 5. When the transport component 2 moves the target component 1 to the detection area, the clamping assembly 6 clamps the target component 1 and drives the target component 1 to rotate. Then, the light source module is controlled to emit 405nm ultraviolet light and 650nm red light. The emitted light source illuminates the blades of the axial flow fan. At the same time, the imaging element 4 takes a picture. The processing unit performs polarization difference processing on the image obtained by the imaging element 4, thereby enabling the identification of defects on the target component 1.

[0045] Furthermore, the clamping component 6 is disposed on the support body 5. The clamping component 6 is movably disposed in the vertical direction, and at least a portion of the clamping component 6 is movably disposed in the horizontal direction, so that when the clamping component 6 is in the first position, at least a portion of the clamping component 6 is controlled to move in the horizontal direction to clamp the target component 1 and drive the target component 1 to rotate.

[0046] The imaging element 4 is mounted on the clamping assembly 6, and the positive light emission direction of the imaging element 4 is directed toward the transport component 2.

[0047] Optionally, the capturing element 4 is a camera.

[0048] Furthermore, the clamping assembly 6 includes a clamping beam 61, which is arranged in a horizontal direction, wherein the imaging element 4 is disposed on the clamping beam 61.

[0049] Two support components 62 are movably disposed at both ends of the clamping beam 61;

[0050] Two first telescopic elements 63 are both set on the clamping beam 61. The two first telescopic elements 63 are set one-to-one with the two support components 62. Each first telescopic element 63 is connected to its corresponding support component 62 so that the two first telescopic elements 63 can drive the two support components 62 to move closer to or further away from each other in order to clamp or release the target component 1.

[0051] Optionally, the first telescopic element 63 is a telescopic rod.

[0052] Furthermore, the clamping assembly 6 also includes a clamping beam 61, which is arranged in a horizontal direction, wherein the imaging element 4 is disposed on the clamping beam 61.

[0053] Two support components 62 are respectively disposed at both ends of the clamping beam 61;

[0054] The first telescopic element 63 is disposed on the clamping beam 61. The first telescopic element 63 is connected to one of the support components 62 to drive the support component 62 to move horizontally relative to the clamping beam 61 in order to clamp or release the target component 1.

[0055] Furthermore, the clamping assembly 6 also includes a rotating assembly 64, which includes two rotating parts 641. The two rotating parts 641 are rotatably disposed on the two supporting parts 62, so that when the clamping assembly 6 is in the first position, the two supporting parts 62 are controlled to move toward each other, or one of the supporting parts 62 is controlled to move toward the other supporting part 62, so that at least a portion of the rotating part 641 contacts the target part 1 to clamp the target part 1 and drive the target part 1 to rotate.

[0056] Optionally, a rotating shaft and a drive motor are provided on the support component 62. The output end of the drive motor is connected to the rotating shaft, and a rotating disk is provided on the rotating shaft. A mounting hole is provided in the middle of the target component 1, and the rotating disk can be inserted into the mounting hole to clamp the target component 1. The drive motor is started, and the rotating shaft is driven to rotate through the drive motor, which in turn drives the rotating disk to rotate, thereby completing the rotation of the target component 1.

[0057] Optionally, each support component 62 is slidably connected to the clamping beam 61. Specifically, the first telescopic element 63 is disposed on the clamping beam 61, and the support component 62 is connected to the clamping beam 61 by a slider. The telescopic end of the first telescopic element 63 is connected to the support component 62, thereby driving the support component 62 to move relative to the clamping beam 61 through the first telescopic element 63.

[0058] Alternatively, both support members 62 may move simultaneously relative to the clamping beam 61.

[0059] Alternatively, one of the support components 62 may move relative to the clamping beam 61.

[0060] Furthermore, the support body 5 includes a main body 51 and a second telescopic element 52 disposed on the main body 51. The telescopic end of the second telescopic element 52 is connected to the clamping assembly 6 so as to drive the clamping assembly 6 to move in the vertical direction through the second telescopic element 52.

[0061] Optionally, the second telescopic element 52 is a telescopic rod.

[0062] Furthermore, the automatic detection system also includes two detection components, which are respectively disposed on the two support components 62 of the clamping assembly 6, with the detection end of the detection component facing the support component 62 disposed opposite to it.

[0063] The axial fan is placed on the belt (transport component 2) and guided and stabilized by the semi-circular limiting surfaces of the two limiting platforms to ensure that the axial fan will not be laterally displaced during the inspection. As the belt moves, the axial fan is transported to the location of the inspection device. When the axial fan reaches the inspection area, the clamping component 6 is activated. Through the action of the first telescopic element 63, the two supporting components 62 move toward the fan housing to achieve precise clamping. Subsequently, the second telescopic element 52 is activated, driving the entire clamping component 6 and the axial fan set on it to move downward to the optimal inspection height. The rotating component 64 drives the fan to rotate to ensure all-round inspection. The light source module 3 emits 405nm ultraviolet light and 650nm red light during the fan rotation. The light evenly covers the surface of the fan blades. The imaging element 4 takes pictures of the fan blades from multiple angles to capture clear images of the fan blades.

[0064] The processing unit receives the raw image from the imaging element 4, performs polarization difference processing on it to eliminate the interference caused by the specular reflection of the metal core and the diffuse reflection of the plastic curved surface, and then uses a deep learning algorithm to analyze the processed image to identify and locate cracks and whitening defects. Once the processing unit identifies a defect, it will immediately generate feedback, indicating whether the detection result is qualified (OK) or unqualified (NG), and at the same time record detailed information such as the type, location and size of the defect.

[0065] By precisely controlling the speed and position of the axial fan, combined with high-speed imaging and advanced image processing algorithms, the efficiency and accuracy of the detection are ensured. The detection cycle is ≤4.8 seconds, and the detection rate of cracks and whitening defects is ≥99.7%. In this technical solution, the use of multispectral polarization illumination technology of ultraviolet and red light effectively suppresses the specular reflection of the metal shaft core and the diffuse reflection noise of the plastic blades, thereby improving the signal-to-noise ratio of the image and reducing the false judgment rate.

[0066] Example 2

[0067] This application embodiment also provides a detection method applicable to the above-mentioned automatic detection system. The detection method includes controlling the light source module 3 and the imaging element 4 to turn on when the target component 1 is determined to be in the detection area, so that when the light source of the light source module 3 illuminates the target component 1, the imaging element 4 is controlled to take a picture of the target component 1 to obtain a pre-processed image.

[0068] The preprocessed image is input into the processing unit for polarization difference processing to identify defects on the target component 1;

[0069] The light source module 3 includes light-emitting components for emitting 405nm ultraviolet light and 650nm red light.

[0070] Furthermore, after the step of inputting the preprocessed image into the processing unit for polarization difference processing, the detection method also includes:

[0071] Construct the U-Net model;

[0072] The differential image obtained after polarization difference processing is input into the U-Net model for segmentation to output multiple segmentation masks. The segmentation masks include crack regions and whitening defect regions. Features are extracted and classified from the crack regions and whitening defect regions to determine the type of crack and the type of whitening defect. Among them, the types of cracks include primary cracks and secondary cracks, and the types of whitening defects include primary defects and secondary defects. When the type of crack is a secondary crack or the type of whitening defect is a secondary defect, the target component 1 is marked as defective.

[0073] Further, the steps of extracting and classifying features of the crack region to determine the type of crack include: inputting the extracted crack features into a CBAM attention module for identification and classification; when the crack length is greater than 0.3 mm and the crack width is less than or equal to 0.5 mm, the crack type is determined to be a secondary type; otherwise, the crack type is determined to be a primary type.

[0074] Furthermore, feature extraction and classification are performed on the whitening defect region to determine the type of whitening defect. This includes inputting the extracted defect features of the whitening defect into a CBAM attention module for identification and classification. When the area of ​​the whitening defect region is greater than 2 mm... 2 If the whitening defect is found to be of the second-level defect type, then the whitening defect is determined to be of the first-level defect type.

[0075] First, when the target component 1 (axial fan) is transported to the detection area, the automatic detection system triggers the activation of the light source module 3 and the imaging element 4. The light-emitting component of the light source module 3 then emits 405nm ultraviolet light and 650nm red light, which uniformly cover and illuminate the surface of the target component 1. At the same time, the imaging element 4 takes pictures of the fan from multiple angles to capture clear images under the light source to obtain pre-processed images.

[0076] Next, the pre-processed images are input into the processing unit for polarization difference processing. This process can effectively distinguish and eliminate the interference caused by the specular reflection of the metal core and the diffuse reflection of the plastic blades, resulting in purer images of cracks and whitening defects.

[0077] Then, the U-Net model was constructed and applied to segment the differential image after polarization difference processing, generating multiple segmentation masks, each of which accurately marked the crack region and the whitening defect region. Subsequently, feature extraction was performed on these regions, including the length and width of the crack, and the area and brightness distribution of the whitening defect.

[0078] For cracked areas, the extracted features are input into a deep learning model with a CBAM attention module for identification and classification. Cracks longer than 0.3 mm and narrower than 0.5 mm are classified as secondary cracks. Similarly, for whitening defect areas, the CBAM attention module is used for feature analysis. Whitening defects with an area greater than 2 mm² are classified as secondary defects. Any axial fan marked as having a secondary crack or a secondary whitening defect will be automatically identified as a defective product by the system and require further processing or rejection.

[0079] Throughout the entire inspection process, from the moment the target component 1 enters the inspection area to the image acquisition element 4 collecting the original image, and then to the processing unit performing image processing and defect classification, the continuity and integrity of the inspection process are ensured. In particular, the combined application of multispectral polarization difference technology and deep learning algorithms significantly improves the accuracy and efficiency of the automatic inspection system in identifying cracks and whitening defects, ultimately achieving comprehensive control over the quality issues of axial flow fans.

[0080] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0081] Unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the invention. It should also be understood that, for ease of description, the dimensions of the various parts shown in the drawings are not drawn to actual scale. Techniques, methods, and devices known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and devices should be considered part of the specification. In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values. It should be noted that similar reference numerals and letters in the following figures denote similar items; therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0082] In the description of this invention, it should be understood that the orientation or positional relationship indicated by directional terms such as "front, back, up, down, left, right", "horizontal, vertical, horizontal" and "top, bottom" is generally based on the orientation or positional relationship shown in the accompanying drawings, and is only for the convenience of describing this invention and simplifying the description. Unless otherwise stated, these directional terms do not indicate or imply that the device or element referred to must have a specific orientation or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on the scope of protection of this invention; the directional terms "inner" and "outer" refer to the inner and outer contours relative to the outline of each component itself.

[0083] For ease of description, spatial relative terms such as "above," "on top of," "on the upper surface of," "above," etc., are used herein to describe the spatial positional relationship of a device or feature as shown in the figures to other devices or features. It should be understood that spatial relative terms are intended to encompass different orientations in use or operation beyond the orientation of the device as described in the figures. For example, if the device in the figures were inverted, a device described as "above" or "on top of" other devices or structures would subsequently be positioned as "below" or "under" other devices or structures. Thus, the exemplary term "above" can include both "above" and "below." The device may also be positioned in other different ways (rotated 90 degrees or in other orientations), and the spatial relative descriptions used herein will be interpreted accordingly.

[0084] Furthermore, it should be noted that the use of terms such as "first" and "second" to define components is merely for the purpose of distinguishing the corresponding components. Unless otherwise stated, the above terms have no special meaning and therefore should not be construed as limiting the scope of protection of this invention.

[0085] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An automatic detection system, characterized in that, The automatic detection system is used to detect defects on the target component (1), and the automatic detection system includes: Transport component (2), which is movably disposed and is used to carry the target component (1); A detection device is provided on the outside of the transport component (2) to form a detection zone with at least a portion of the transport component (2) for detecting the target component (1) transported by the transport component (2) to the detection zone; The detection device includes a light source module (3) and an imaging element (4). The light source module (3) includes a light-emitting component for emitting 405nm ultraviolet light and 650nm red light. The imaging element (4) is a 2048-pixel line array camera. When the light source emitted by the light source module (3) illuminates the target part of the target component (1), the imaging element (4) takes a picture to obtain a pre-processed image. The processing unit is connected to the imaging element (4) to perform polarization difference processing on the preprocessed image to identify the defects on the target component (1).

2. The automatic detection system according to claim 1, characterized in that, The detection device further includes: The supporting body (5) is disposed outside the transport component (2); A clamping assembly (6) is disposed on the support body (5). The clamping assembly (6) is movably disposed in the vertical direction, and at least a portion of the clamping assembly (6) is movably disposed in the horizontal direction. When the clamping assembly (6) is in the first position, at least a portion of the clamping assembly (6) is controlled to move in the horizontal direction to clamp the target component (1) and drive the target component (1) to rotate. The imaging element (4) is disposed on the clamping assembly (6), and the positive light emission direction of the imaging element (4) is directed toward the transport component (2).

3. The automatic detection system according to claim 2, characterized in that, The clamping assembly (6) includes: A clamping beam (61) is provided along the horizontal direction, wherein the shooting element (4) is provided on the clamping beam (61); Two support components (62) are movably disposed at both ends of the clamping beam (61); Two first telescopic elements (63) are both disposed on the clamping beam (61). The two first telescopic elements (63) are disposed in a one-to-one correspondence with the two support members (62). Each first telescopic element (63) is connected to the support member (62) corresponding to it, so that the two support members (62) can be driven by the two first telescopic elements (63) to move closer to or further away from each other, so as to clamp or release the target member (1).

4. The automatic detection system according to claim 2, characterized in that, The clamping assembly (6) further includes: A clamping beam (61) is provided along the horizontal direction, wherein the shooting element (4) is provided on the clamping beam (61); Two support components (62) are respectively disposed at both ends of the clamping beam (61); A first telescopic element (63) is disposed on the clamping beam (61). The first telescopic element (63) is connected to one of the supporting components (62) to drive the supporting component (62) to move relative to the clamping beam (61) in the horizontal direction to clamp or release the target component (1).

5. The automatic detection system according to claim 3, characterized in that, The clamping assembly (6) further includes: The rotating assembly (64) includes two rotating parts (641), which are rotatably disposed on the two supporting parts (62) respectively, so as to control the two supporting parts (62) to move towards each other when the clamping assembly (6) is in the first position, or to control one of the supporting parts (62) to move towards the other supporting part (62) so that at least a portion of the rotating part (641) contacts the target part (1) to clamp the target part (1) and drive the target part (1) to rotate.

6. The automatic detection system according to claim 2, characterized in that, The automatic detection system further includes a limiting component (7) having a limiting surface, on which the target component (1) is placed to limit the target component (1); and / or, The support body (5) includes a main body (51) and a second telescopic element (52) disposed on the main body (51). The telescopic end of the second telescopic element (52) is connected to the clamping assembly (6) so as to drive the clamping assembly (6) to move vertically through the second telescopic element (52); and / or, The automatic detection system also includes two detection components, which are respectively disposed on the two support components (62) of the clamping assembly (6), and the detection end of the detection component is disposed toward the support component (62) disposed opposite to it.

7. A detection method for detecting defects on a target component (1), said detection method being applicable to the automatic detection system according to any one of claims 1 to 6, characterized in that, The detection method includes: When it is determined that the target component (1) is in the detection area, the light source module (3) and the imaging element (4) are turned on so that when the light source of the light source module (3) illuminates the target component (1), the imaging element (4) is controlled to take a picture of the target component (1) to obtain a pre-processed image; The preprocessed image is input into the processing unit for polarization difference processing to identify the defects on the target component (1); The light source module (3) includes a light-emitting component for emitting 405nm ultraviolet light and 650nm red light.

8. The detection method according to claim 7, characterized in that, After the step of inputting the preprocessed image into the processing unit for polarization difference processing, the detection method further includes: Construct the U-Net model; The differential image obtained after polarization difference processing is input into the U-Net model for segmentation to output multiple segmentation masks. The segmentation masks include crack regions and whitening defect regions. Features are extracted and classified for the crack regions and whitening defect regions to determine the type of crack and the type of whitening defect. The types of cracks include primary cracks and secondary cracks, and the types of whitening defects include primary defects and secondary defects. When the type of crack is a secondary crack or the type of whitening defect is a secondary defect, the target component (1) is marked as defective.

9. The detection method according to claim 8, characterized in that, The step of extracting and classifying features of the cracked region to determine the type of crack includes: The extracted crack features are input into a CBAM attention module for identification and classification. When the crack length is greater than 0.3 mm and the crack width is less than or equal to 0.5 mm, the crack type is determined to be a secondary type; otherwise, the crack type is determined to be a primary type.

10. The detection method according to claim 8, characterized in that, The step of extracting and classifying features from the whitening defect region to determine the type of whitening defect includes: The extracted whitening defect features are input into a CBAM attention module for identification and classification. When the area of ​​the whitening defect is greater than 2 mm... 2 If the whitening defect is found to be of the second-level defect type, then the whitening defect is determined to be of the first-level defect type.

Images