Method and device for detecting surface defects of metal plate, electronic equipment and storage medium

Abstract

The application provides a metal plate surface defect detection method and device, electronic equipment and storage medium, wherein the method comprises the steps of: irradiating the metal plate surface with at least two different direction interference systems to generate at least two interference images containing interference fringes extending along a preset direction; wherein the interference fringes of each interference image are independent of each other; dividing the interference images into several detection areas; wherein the extension direction of the detection area is perpendicular to the corresponding interference fringe direction; performing gray value analysis on the detection area to determine the surface defect condition of the metal plate; the detection area includes all detection areas of at least one interference image. The application generates interference fringe images extending in different directions through multi-directional interference imaging technology, and divides detection areas perpendicular to the interference fringe direction to form a detection grid that is orthogonally covered, so as to realize rapid detection of surface defects.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of metal plate detection, and in particular to a metal plate surface defect detection method and device, electronic equipment and storage medium. BACKGROUND

[0002] In modern metal processing and manufacturing industry, the surface quality of metal plates directly affects the performance, appearance and market competitiveness of products. Surface scratches, cracks, pits and other defects are common quality defects in the production process of metal plates, which not only affect the appearance of products, but also may reduce the mechanical properties and service life of materials. Therefore, efficient and accurate surface defect detection technology has become an indispensable part of the production process of metal plates.

[0003] Metal plate surface defect detection is a key link of industrial quality control, but the existing technology has significant limitations. For example, a steel plate surface defect detection method based on visual features is disclosed in Chinese patent CN117036357A. This method introduces the distribution characteristics of reflective and backlit pixels on the surface of the steel plate, proposes a defect classification method based on visual features, effectively solves the contradiction between static scanning and continuous production, and improves the accuracy of defect recognition through gray value sequence analysis. However, this technical solution still has limitations and relies only on single-direction reflective / backlight pixel distribution analysis, which cannot cover defects extending in multiple directions such as longitudinal cracks and transverse scratches, resulting in high missed detection rate and difficulty in dealing with the diversified characteristics of complex defects. SUMMARY

[0004] In order to achieve the above-mentioned purposes and other advantages according to the present application, according to the first aspect of the embodiments of the present application, a metal plate surface defect detection method is provided, comprising the following steps:

[0005] Irradiate the surface of the metal plate by at least two different direction interference systems to generate at least two interference images containing interference fringes arranged along the preset arrangement direction of the corresponding interference system;

[0006] Divide the interference images along the direction perpendicular to the arrangement direction of the interference fringes into a plurality of detection regions, so that each detection region covers at least part of each interference fringe;

[0007] Obtain the gray value sequence of the detection region along the arrangement direction of the interference fringes to calculate the gray deviation value between the gray value in the gray value sequence and the gray value in the preset standard gray value sequence of the detection region;

[0008] Determine whether the detection region has surface defects according to the gray deviation value;

[0009] The detection regions with surface defects between the interference images are superimposed in position to locate the defect regions on the surface of the metal plate.

[0010] Optionally, the interference images comprise a first interference image and a second interference image, directions of interference fringes of the first interference image and the second interference image being independent of each other; a detection region of the first interference image is denoted as a first detection region; and a detection region of the second interference image is denoted as a second detection region.

[0011] Before the step of "superimposing the detection regions with surface defects between the interference images in position to locate the defect regions on the surface of the metal plate", the method further comprises the steps of:

[0012] determining whether the first interference image has the first detection region with the surface defect;

[0013] if the first interference image has the first detection region with the surface defect, it is determined that the metal plate has the surface defect, and a gray deviation value is calculated according to a gray value sequence of the second detection region to determine the second detection region with the surface defect in the second interference image;

[0014] if the first interference image does not have the first detection region with the surface defect, it is determined that the metal plate does not have the surface defect.

[0015] Optionally, the step of "superimposing the detection regions with surface defects between the interference images in position to locate the defect regions on the surface of the metal plate" comprises the steps of:

[0016] constructing a plane reference system according to the first interference image, the second interference image and a position mapping relationship of the surface of the metal plate;

[0017] locating the first detection region with the surface defect and the second detection region with the surface defect in the plane reference system;

[0018] marking a region, in which the first detection region with the surface defect and the second detection region with the surface defect are superimposed, as a surface defect region in the plane reference system;

[0019] obtaining a position coordinate of the surface defect region on the surface of the metal plate according to the position mapping relationship between the plane reference system and the surface of the metal plate.

[0020] Optionally, a direction perpendicular to a direction of interference fringes in the interference image is defined as a sampling direction.

[0021] The step of "obtaining a gray value sequence of the detection region along the direction of the interference fringes to calculate a gray deviation value between a gray value in the gray value sequence and a gray value in a preset standard gray value sequence of the detection region" comprises the steps of:

[0022] obtaining a preset standard gray value sequence, wherein the standard gray value sequence comprises n gray values;

[0023] dividing the detection area into n gray sampling areas along the arrangement direction of the interference fringes;

[0024] obtaining the gray values of the gray sampling areas to generate a gray value sequence corresponding to the detection area, the gray value sequence comprising n gray values;

[0025] calculating the deviation values of the gray values in the gray value sequence from the corresponding gray values in the preset standard gray value sequence.

[0026] Optionally, after the step of "performing superposition on the detection areas with surface defects between the interference images to locate the defect areas on the surface of the metal plate", the method further comprises the steps of:

[0027] obtaining a polarized image of the defect areas on the surface of the metal plate;

[0028] inputting the polarized image into a pre-trained surface defect type analysis model to output the defect types of the defect areas.

[0029] According to a second aspect of the embodiments of the present application, a method for detecting surface defects of a metal plate is also provided, comprising the following steps:

[0030] irradiating the surface of the metal plate by a first interference system to generate a first interference image of interference fringes arranged in a preset arrangement direction of the first interference system;

[0031] dividing the first interference image into a plurality of first detection areas along a direction perpendicular to the arrangement direction of the interference fringes in the first interference image; wherein each first detection area covers at least a part of each interference fringe in the first interference image;

[0032] obtaining a gray value sequence of the first detection areas along the arrangement direction of the interference fringes to calculate the gray deviation values of the gray values in the gray value sequence from the gray values in the preset standard gray value sequence of the detection area;

[0033] determining whether the first detection area has surface defects according to the gray deviation values;

[0034] If the first interference image has the first detection region with the surface defect, the second interference image is acquired, and the second detection region with the surface defect is detected in the second interference image, so as to perform position superposition on the first detection region with the surface defect and the second detection region with the surface defect, and locate the defect region on the metal plate surface; wherein the second interference image is generated by irradiating the metal plate surface by a second interference system, and interference fringes in the second interference image and the first interference image are arranged in mutually independent directions; the second interference image is divided into a plurality of second detection regions along a direction perpendicular to the arrangement direction of the interference fringes; and each second detection region covers at least a partial region of each interference fringe in the second interference image.

[0035] If the first interference image has no first detection region with the surface defect, it is determined that the metal plate has no surface defect.

[0036] Optionally, the step of "performing position superposition on the first detection region with the surface defect and the second detection region with the surface defect, and locating the defect region on the metal plate surface" specifically includes the steps of:

[0037] constructing a plane reference system according to the first interference image, the second interference image, and the position mapping relationship between the metal plate surface;

[0038] locating the first detection region with the surface defect and the second detection region with the surface defect in the plane reference system;

[0039] marking a region in which the first detection region with the surface defect and the second detection region with the surface defect are superimposed in the plane reference system as a surface defect region;

[0040] obtaining the position coordinates of the surface defect region on the metal plate surface according to the position mapping relationship between the plane reference system and the metal plate surface.

[0041] According to a third aspect of the embodiments of the present application, a metal plate surface defect detection device is also provided, which includes:

[0042] an interference image generation unit configured to irradiate a metal plate surface by at least two interference systems in different directions respectively, so as to generate at least two interference images containing interference fringes arranged along a preset arrangement direction of the corresponding interference system; wherein the arrangement directions of the interference fringes in the interference images are mutually independent;

[0043] a detection region division unit configured to divide the interference image into a plurality of detection regions along a direction perpendicular to the arrangement direction of the interference fringes, so that each detection region covers at least a partial region of each interference fringe;

[0044] The gray value sequence generation unit is configured to obtain a gray value sequence of the detection area along the arrangement direction of the interference fringes, and calculate a gray deviation value between a gray value in the gray value sequence and a gray value in a preset standard gray value sequence of the detection area;

[0045] The defect detection unit is configured to determine whether the detection area has a surface defect according to the gray deviation value;

[0046] The defect positioning unit is configured to perform superposition on the positions of the detection areas having the surface defects in the interference images to locate the defect area on the surface of the metal plate.

[0047] According to a third aspect of the embodiments of the present application, an electronic device is also provided, which includes a memory and a processor; the memory is configured to store one or more computer instructions, and the one or more computer instructions are executed by the processor to implement the method steps described above.

[0048] According to a fourth aspect of the embodiments of the present application, a storage medium is also provided, which stores computer instructions; when the computer instructions are executed by a processor, the method steps described above are implemented.

[0049] In view of the deficiencies in the prior art of metal plate surface defect detection, the present application generates interference fringe images extending in different directions through multi-directional interference imaging technology, forms a multi-directional coverage detection mechanism, effectively improves the recognition ability of multi-directional defects such as cracks and scratches, and avoids the problem of missed detection caused by the parallelism of the defect direction and the interference fringes. By dividing the detection area along the arrangement direction of the interference fringes into a plurality of gray sampling units, a gray value sequence is generated and a deviation calculation is performed with a standard sequence, thereby realizing accurate positioning of the defect. Further, through the position superposition of the detection areas having defects in the multi-directional interference images, the recognition accuracy of the defect area is ensured. At the same time, the polarization image of the defect area is input into a pre-trained surface defect type analysis model to realize automatic classification of the defect type, thereby breaking through the limitations of traditional manual experience judgment. Through the collaborative application of multi-directional interference detection and intelligent analysis, the overall scheme realizes efficient screening, accurate area positioning and accurate type recognition of the metal plate surface defects, and meets the needs of high-precision and high-efficiency detection of industrial production lines. BRIEF DESCRIPTION OF DRAWINGS

[0050] The accompanying drawings, which are included to provide a further understanding of the present application and constitute a part of this application, illustrate certain illustrative embodiments of the present application and together with the description serve to explain the present application. In the drawings:

[0051] Figure 1 A flowchart of a metal plate surface defect detection method according to Embodiment 1 of the present application is shown in the figure;

[0052] Figure 2A schematic diagram of the principle of the interference fringes provided for the present application;

[0053] Figure 3 A schematic diagram of a detection region division method provided for Embodiment 1 of the present application;

[0054] Figure 4 A flowchart of a process of "obtaining a sequence of gray scale values of the detection region along the arrangement direction of the interference fringes to calculate a gray scale deviation value between the gray scale value in the sequence of gray scale values and a preset standard sequence of gray scale values of the detection region" provided for Embodiment 1 of the present application;

[0055] Figure 5 A flowchart of a process of "performing position superposition on the detection region with surface defects between the interference images to locate the defect region on the surface of the metal plate" provided for Embodiment 1 of the present application;

[0056] Figure 6 A flowchart of a process of "performing position superposition on the detection region with surface defects between the interference images to locate the defect region on the surface of the metal plate" provided for Embodiment 1 of the present application;

[0057] Figure 7 A flowchart of a process of "performing position superposition on the detection region with surface defects between the interference images to locate the defect region on the surface of the metal plate" provided for Embodiment 1 of the present application;

[0058] Figure 8 A flowchart of a process of a metal plate surface defect detection method provided for Embodiment 2 of the present application;

[0059] Figure 9 A flowchart of a process of "performing position superposition on the first detection region with surface defects and the second detection region with surface defects to locate the defect region on the surface of the metal plate" provided for Embodiment 2 of the present application;

[0060] Figure 10 A schematic diagram of a planar reference system of a first interference image and a second interference image provided for Embodiment 2 of the present application;

[0061] Figure 11 A schematic diagram of a metal plate surface defect detection device provided for Embodiment 3 of the present application;

[0062] Figure 12 A schematic diagram of an electronic device provided for Embodiment 4 of the present application;

[0063] Figure 13 A schematic diagram of a storage medium provided for Embodiment 5 of the present application. DETAILED DESCRIPTION

[0064] In order to make the person skilled in the art better understand the technical scheme of the present application, the technical scheme in the embodiments of the present application will be described clearly and completely below in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by the person skilled in the art without creative labor should belong to the protection scope of the present application. Embodiment 1

[0065] According to a first aspect of the embodiments of the present application, a method for detecting surface defects of a metal plate is provided, as shown in Figure 1 The method comprises the following steps:

[0066] S11, irradiating the surface of the metal plate by at least two different direction interference systems respectively, to generate at least two interference images containing interference fringes arranged along the preset arrangement direction of the corresponding interference system; wherein the arrangement directions of the interference fringes between the interference images are independent of each other.

[0067] It should be noted that the above-mentioned interference system includes but is not limited to Michelson interferometer, Fizeau interferometer or other optical devices based on the principle of light wave interference, which is composed of laser light source, beam splitter, mirror and lens and other optical elements. The preset arrangement direction of the interference fringes is determined by the relative position adjustment of the optical elements, and remains fixed during operation, which can be set to any angle within the range of 0° to 360°.

[0068] The interference image refers to the image formed by the reflected light of the metal plate surface captured by the camera system, wherein the interference fringes extend along the preset arrangement direction of the interference system, which appears as a parallel line structure, and its form is modulated by the deformation or defects of the metal plate surface.

[0069] In this embodiment, the light source of the interference system can be an LED light source, which has the advantages of low energy consumption, high efficiency, long service life and flexibility. However, there are certain limitations in generating high-precision structured light patterns, especially cosine fringe structured light periods at the micron and sub-micron levels. These limitations mainly come from the inherent physical size of the LED lamp bead and the diffusion characteristics of the secondary diffusion plate. Preferably, the light source of the interference system is set to laser, specifically by using a green laser with a wavelength of 532 nm as the light source.

[0070] In an alternative embodiment, after the laser beam is expanded, filtered and collimated by the expansion system and the filtering system, it is subjected to shearing interference by a double-plate shearing interference system, forming a fringe interference structured light on the interference wave surface, as shown in Figure 2The interference fringes shown have high flatness and contrast. The collimation and noise conditions are also good, the fringes are flat and not deformed, and there is essentially no noise phenomenon. Good interference fringe illumination will improve the overall quality of the picture and increase the visibility of defects.

[0071] In the interference images generated by different interference systems, the arrangement directions of the interference fringes are independent of each other, that is, the interference fringe directions of each image are not parallel and not collinear in space. Specifically, the interference fringes generated by the first interference system are arranged along a first preset direction, and the interference fringes generated by the second interference system are arranged along a second preset direction, wherein the included angle between the first preset direction and the second preset direction is not equal to 0° or 180°, thereby covering all possible non-parallel angle combinations, and ensuring that defects of any orientation on the metal plate surface can be effectively detected.

[0072] S12, divide the interference image into a plurality of detection regions along a direction perpendicular to the arrangement direction of the interference fringes, so that each detection region covers at least part of each interference fringe.

[0073] In this embodiment, the interference image is divided into a plurality of detection regions perpendicular to the arrangement direction of the interference fringes, each detection region covers part of the interference fringes in the vertical direction, and each interference fringe is at least partially located in a detection region. Specifically, the arrangement direction of the interference fringes is preset and fixed by the interference system, for example, when the interference fringes extend along the horizontal direction, the division direction is the vertical direction, and the interference image is divided into a plurality of strip-shaped regions.

[0074] In an optional embodiment, as shown in the figure, the interference fringes in the interference image 10 extend along a preset first direction (x direction in the figure), wherein the x direction is defined as the arrangement direction of the interference fringes (i.e. the interference fringes are perpendicular to the x axis) ; the interference image is divided along a second direction perpendicular to the x direction (i.e. the extension direction of the interference fringes, y direction in the figure), wherein the y direction is perpendicular to the x direction (the included angle is 90°), and the determination of the y direction is not dependent on the absolute coordinate system, but is dynamically determined by the actual arrangement direction of the interference fringes. Figure 3 Figure 3 The width of the detection region a101 (represented by the red area in the interference image 10) is set to be an integer multiple or a fractional multiple of the interference fringe spacing, so that each interference fringe is covered by at least one detection region a101 in the y axis direction (i.e. part of each interference fringe is located in the detection region a101). It should be noted that there can be overlap or gap between adjacent detection regions a101, but all interference fringes are covered within the range of the detection region set, so that the characteristic information of the interference fringes is completely preserved when the gray value sequence is extracted along the x direction. Figure 2 Figure 3

[0075] ​​​The division mode is suitable for interference images with an interference fringe arrangement direction at any angle (including horizontal, inclined or non-parallel configuration), and through dynamic adjustment of the detection region width, adaptive coverage of different interference fringe densities is achieved, effectively supporting the universality of multi-directional interference detection.

[0076] S13, a gray value sequence of the dry detection region along the interference fringe arrangement direction is obtained to calculate a gray deviation value between the gray value in the gray value sequence and a gray value in a preset standard gray value sequence of the detection region.

[0077] In this embodiment, in combination with the foregoing embodiment, for each detection region divided in S12, a continuous gray value sequence is extracted along the interference fringe arrangement direction. Optionally, the preset standard gray value sequence is obtained through experiments based on a normal metal plate surface defect-free interference image used for illumination in S11. The sequence is generated by statistical analysis of a large number of normal sample interference image gray data, including mean, median, Gaussian distribution fitting, linear regression model, neural network model or any other statistical or machine learning method, and the construction process does not depend on specific algorithms, preset thresholds, image resolutions or environmental conditions. The calculation of the gray deviation value uses any mathematical measurement method, including absolute difference, relative difference, root mean square error, Pearson correlation coefficient, mutual information metric or any other algorithm that can quantify the difference in gray distribution, to generate a list of gray deviation values. This calculation method is fully compatible with the independence design of the interference fringe arrangement direction in S11 and the division logic of the detection region in S12, ensuring the comparability of the standard sequence and the image to be measured based on the same sampling conditions, and is suitable for all metal plate materials and surface states. The gray deviation calculation does not depend on the specific angle of the interference fringes, the image resolution, the noise level or the lighting conditions, thereby realizing universal detection of defects on the metal plate surface at any orientation and effectively supporting the multi-directional interference detection system.

[0078] In an optional embodiment, a direction perpendicular to the direction of the interference fringes in the interference image is defined as the sampling direction; as Figure 4 As shown in the above S13 step of "obtaining a gray value sequence of the dry detection region along the interference fringe arrangement direction to calculate a gray deviation value between the gray value in the gray value sequence and the gray value in the preset standard gray value sequence of the detection region", the specific steps include the following steps:

[0079] S131, a preset standard gray value sequence is obtained, wherein the standard gray value sequence includes n gray values.

[0080] S132, the detection region is divided into n gray sampling regions along the interference fringe arrangement direction.

[0081] S133, the gray value of the gray sampling region is obtained to generate a gray value sequence corresponding to the detection region including n gray values.

[0082] S134, calculate the deviation value of each gray value in the gray value sequence and its corresponding preset standard gray value sequence.

[0083] In the above steps S131 and S132, the value of n is dynamically determined based on the interference image resolution and the detection accuracy requirement, for example, n is set to 100 in a high-resolution image, and n is set to 50 in a low-resolution image. The standard sequence is generated by an average value method, and the length thereof strictly matches the number of subsequent gray sampling regions. Alternatively, the sampling region segmentation point position is segmented according to the segmentation manner of the obtained standard gray value sequence, and the segmentation point position is adaptively adjusted according to the interference fringe spacing, so as to ensure that each gray sampling region covers the feature region (such as the center line region) of the interference fringe, and the segmentation number accurately corresponds to the n value, thereby avoiding sampling distortion caused by changes in fringe density.

[0084] In an optional embodiment, the gray value in the above step S133 is extracted from the center line pixel sequence of each sampling region, and is sequentially combined to form n gray value sequences, which completely retain the local feature distribution of the interference fringe, and are consistent with the length and position of the standard sequence.

[0085] The deviation value in the above step S134 can be calculated by basic subtraction and absolute value operation.

[0086] In a specific example, the preset standard gray value sequence is A1, A2,..., A k ,..., A n (based on normal metal plate sample statistics); the gray value sequence generated by the detection region is B1, B2,..., B k ,..., B n ; and the calculation deviation value sequence is |B1 - A1|, |B2 - A2|,..., |B k - A k |,..., |B n - A n |.

[0087] S14, determining whether the detection region has surface defects according to the gray deviation value;

[0088] The gray deviation value is compared with a preset threshold value. If any gray deviation value exceeds the preset threshold value, it is determined that the detection region has surface defects, otherwise it is determined to be defect-free. The preset threshold value is dynamically set based on the metal plate material, surface state or detection environment, for example, a lower threshold value is used for a stainless steel surface, and a higher threshold value is used for an aluminum alloy surface, so as to ensure that the determination result matches the actual defect feature.

[0089] In one specific embodiment, the list of gray scale deviation values generated in S134 is compared with a preset determination threshold T, and if all the deviation values are smaller than the preset determination threshold T, it is determined that the detection region is defect-free; if there is any |B k - A k | > T(k = 1, 2,..., n), it is determined whether there is a defect in the detection region. Optionally, the determination threshold T is dynamically set based on the metal plate material and surface state, for example, T = 20 for a stainless steel surface and T = 35 for an aluminum alloy surface, to ensure the accuracy of defect determination. In this embodiment, the deviation value is calculated by basic subtraction and absolute value operation, which only requires CPU basic instructions, has high calculation efficiency, and can quickly screen the detection region that may have a defect.

[0090] S16, superimpose the positions of the detection regions in which surface defects exist between the interference images to locate the defect region of the metal plate surface.

[0091] It should be noted that superimposition refers to spatial alignment and intersection calculation of the detection regions in which defects exist in multiple interference images in a unified coordinate system. Specifically, the defect regions of each interference image are mapped to the reference coordinate system of the metal plate surface through coordinate transformation, so that the defect positions detected in different directions are spatially matched. When the defect regions in multiple directions overlap in space, the overlapping region is the actual defect position. This superimposition process reduces the defect position range from a wide detection region in a single direction to an accurate region confirmed by multiple directions through complementary detection. For example, the defect regions detected by two different direction interference systems partially overlap, and the size of the overlapping part is smaller than that of any single region, thereby eliminating the false judgment and range deviation that may be caused by single direction detection. This superimposition mechanism does not depend on the number of specific directions, interference angles or image parameters, and is applicable to any number of interference images (including two or more), any interference fringe arrangement angle, and all metal plate materials and surface states. By accurately reducing the defect position range, this method ensures high precision and reliability of defect positioning.

[0092] In this embodiment, the detection areas identified as defective in S14 from multiple interferometric images generated in S11 are spatially superimposed in a unified coordinate system. Specifically, the defective areas of each interferometric image are mapped to the reference coordinate system of the metal plate surface using any applicable spatial coordinate transformation method, such as translation, rotation, affine transformation, perspective transformation, or any other geometric mapping method. The intersection or union of these areas is calculated as the defect localization result. This method is applicable to any number of interferometric images, including two, three, or more, with arbitrary interference fringe arrangement angles including horizontal, tilted, non-parallel configurations, or arbitrary angle combinations. It is applicable to all metal plate materials, including stainless steel, aluminum alloy, copper alloy, titanium alloy, and composite materials, and surface conditions, including polishing, sandblasting, coating, oxidation, electroplating, and corrosion. This method does not rely on specific coordinate system transformation algorithms, preset conditions, or image parameters, and can effectively eliminate misjudgments that may occur from single-direction detection, ensuring the accuracy and reliability of defect localization. By superimposing defect information from multiple directions, this method can accurately identify defects of arbitrary orientation on the metal plate surface, including cracks, scratches, pits, oxidation points, pores, inclusions, etc., and is applicable to defects of different sizes from micrometers to millimeters, thus covering all possible defect detection scenarios.

[0093] In an optional embodiment, the interference image includes a first interference image and a second interference image whose interference fringe arrangement directions are independent of each other; wherein, the detection area of ​​the first interference image is denoted as the first detection area; and the detection area of ​​the second interference image is denoted as the second detection area.

[0094] like Figure 5 As shown, before step S16, which involves "overlaying the positions of the detection areas with surface defects between the interference images to locate the defect areas on the metal plate surface," the following step is also included:

[0095] S151. Determine whether a first detection area with surface defects is detected in the first interference image; if a first detection area with surface defects is detected in the first interference image, proceed to step S152; if no first detection area with surface defects is detected in the first interference image, proceed to step S153.

[0096] S152. Determine that there is a surface defect in the metal plate, and calculate the gray value deviation value according to the gray value sequence of the second detection area to determine the second detection area in the second interference image where the surface defect is detected.

[0097] S153. Determine that the metal plate has no surface defects.

[0098] In this embodiment, step S151 determines whether there is a defect detection region in the first interference image according to the S14 determination result. If the first interference image contains a detection region determined by S14 to have defects, S152 is executed; otherwise, S153 is directly executed. In step S152, when S151 determines that the first interference image has a defect region, it is immediately determined that the metal plate has a surface defect. At the same time, the second interference image is processed using the gray scale deviation calculation method of S13-S14 to determine the defect position in the second detection region, providing a data basis for subsequent superposition positioning of S16. This processing process is independent of the interference fringe angle and image parameters, ensuring the timeliness of defect determination. In step S153, when S151 determines that the first interference image has no defect region, it is directly determined that the metal plate has no surface defect, and the second interference image does not need to be processed. This process significantly improves detection efficiency by prioritizing the verification of the defect state of the first interference image, and is applicable to any number of interference images (including two or more), effectively avoiding redundant calculations.

[0099] In one specific embodiment, step S16 of "performing superposition on the position of the detection region with surface defects between the interference images to locate the defect region on the surface of the metal plate" specifically includes the steps of: Figure 6

[0100] S161, constructing a plane reference system according to the first interference image, the second interference image, and the position mapping relationship of the surface of the metal plate;

[0101] In this embodiment, a unified plane reference system is constructed based on the geometric relationship between the first interference image and the second interference image and the surface of the metal plate. This reference system is established by any applicable spatial coordinate transformation method, including translation, rotation, affine transformation, or perspective transformation, to ensure that the reference system accurately corresponds to the physical position of the surface of the metal plate. This construction process is applicable to any number of interference images, including two or more image systems, and any interference fringe arrangement angle, and is independent of specific coordinate systems or image parameters, providing a general spatial reference for subsequent defect positioning.

[0102] S162, locating the first detection region with surface defects and the second detection region with surface defects in the plane reference system;

[0103] In this embodiment, the detection regions in the first interference image and the second interference image that are determined by S14 to have defects are mapped to the plane reference system through coordinate transformation, accurately determining the spatial position of the defect region in the reference system. This mapping process is independent of the specific angle, image resolution, or surface state of the interference fringes, and is applicable to all metal plate materials, including stainless steel, aluminum alloy, copper alloy, titanium alloy, etc., and all surface states, including polishing, sandblasting, coating, oxidation, electroplating, corrosion, etc., ensuring the accuracy and consistency of defect region positioning. ​

[0104] S163, mark the area where the first detection area of the surface defect in the planar reference system and the second detection area where the surface defect exists are superimposed as a surface defect area;

[0105] In this embodiment, the intersection of the two defect areas in the planar reference system is calculated as the final defect area. This intersection area accurately represents the actual defect position on the metal plate surface, effectively eliminating false positives and range deviations that may be caused by single-direction detection. The superposition process complements the defect information from multiple directions, reducing the defect position range from a broad area in a single direction to an accurate area confirmed by multiple directions. This method is applicable to any number of interference images and any interference fringe arrangement angle, ensuring high precision and reliability of defect positioning.

[0106] S164, according to the position mapping relationship between the planar reference system and the metal plate surface, obtain the position coordinates of the surface defect area on the metal plate surface.

[0107] In an alternative embodiment, as shown in FIG. 16, after step S16 of "performing position superposition on the detection areas where the surface defects exist between the interference images to locate the defect area on the metal plate surface", the method further comprises the steps of: Figure 7

[0108] S171, obtain a polarized image of the defect area on the metal plate surface;

[0109] In this embodiment, based on the defect area position determined in S16, a high-precision photograph of the defect area is taken using a polarized imaging system. The polarized imaging system includes a polarized light source and a polarized camera, which can obtain image data of the defect area under different polarization angles. This method is applicable to all metal plate materials including stainless steel, aluminum alloy, copper alloy, titanium alloy, composite materials, etc., and all surface states including polishing, sandblasting, coating, oxidation, electroplating, corrosion, etc., and can effectively capture the subtle features of defects. The acquisition of the polarized image does not depend on specific light source or camera parameters, ensuring the stability and consistency of image quality, and providing high-quality input data for defect type analysis.

[0110] S172, input the polarized image into a pre-trained surface defect type analysis model to output the defect type of the defect area.

[0111] ​In this embodiment, the surface defect type analysis model is pre-trained by deep learning technology, and contains multiple defect type recognition capabilities including cracks, scratches, pits, oxidation points, pores, inclusions, etc. The model is trained based on a large number of labeled polarization image data and can accurately identify the defect type. The polarization image input is processed by the model to output the defect type information. This process is applicable to any defect size from microns to millimeters, is independent of specific defect types or metal plate materials, and ensures the accuracy and comprehensiveness of defect type analysis. The model training process is independent of specific image parameters or environmental conditions, so that the method can maintain high precision in various detection scenarios, cover all possible defect types and metal plate application fields, maximize the protection range and support the complete application system of metal plate surface defect detection. Embodiment 2

[0112] According to a second aspect of the embodiments of the present application, a metal plate surface defect detection method is provided, as shown in Figure 8 The method comprises the following steps:

[0113] S21, irradiating the metal plate surface by a first interference system to generate a first interference image of interference fringes arranged in a preset arrangement direction of the first interference system.

[0114] The first interference system is composed of laser light source, beam splitter, mirror and lens, etc. optical elements, including Michelson interferometer, Fizeau interferometer or other optical devices based on the principle of light wave interference. The arrangement direction of the interference fringes can be set arbitrarily within the range of 0° to 360° by adjusting the relative positions of the optical elements, and remains fixed during operation. The first interference image captured by the camera system is an image formed by the reflected light of the metal plate surface, in which the interference fringes extend along the preset arrangement direction of the first interference system and appear as parallel line structures, and their morphology is modulated by the deformation or defects of the metal plate surface. This method is applicable to all metal plate materials including stainless steel, aluminum alloy, copper alloy, titanium alloy, composite materials, etc. and all surface states including polishing, sandblasting, coating, oxidation, electroplating, corrosion, etc. The setting of the interference system type and arrangement direction is independent of specific metal plate characteristics or detection environment, ensuring the wide applicability of the method. The acquisition process of the first interference image is independent of image resolution and lighting conditions, and can maintain stable and reliable image quality in various detection scenarios, providing high-quality input data for subsequent defect detection. This implementation supports the detection of any metal plate material and surface state through a universal interference system architecture, covering all possible defect detection scenarios.

[0115] S22, dividing the first interference image into a plurality of first detection regions along a direction perpendicular to the arrangement direction of the interference fringes thereof; wherein each first detection region covers at least part of the area of each interference fringe in the first interference image.

[0116] The first interference image is divided into a plurality of first detection regions perpendicular to the arrangement direction of the interference fringes, each first detection region covering a partial area of the interference fringes in the vertical direction, ensuring that each interference fringe is at least partially located in a first detection region. This division method is applicable to any angle of the arrangement direction of the interference fringes, including horizontal tilt non-parallel configuration or any angle combination, and does not depend on the specific direction of the interference fringes or the image coordinate system. By dynamically adjusting the width of the first detection region, this method can adaptively cover images with different interference fringe densities, and is applicable to all metal plate materials, including stainless steel, aluminum alloy, copper alloy, titanium alloy, composite materials, all surface states, including polishing, sandblasting, coating, oxidation, electroplating, corrosion, and all image resolutions. The width of the first detection region is set to an integer multiple or a fractional multiple of the interference fringe pitch, so that each interference fringe is covered by at least one first detection region in the vertical direction, thereby completely retaining the local feature distribution of the interference fringes. There is allowed to be overlap or gap between adjacent first detection regions, but all interference fringes are covered within the range of the first detection region set, ensuring the integrity of the gray value sequence extraction.

[0117] S23, obtaining a gray value sequence of the first detection region along the arrangement direction of the interference fringes to calculate a gray deviation value of the gray value in the gray value sequence from the gray value in the preset standard gray value sequence of the detection region.

[0118] The step S23 extracts a continuous gray value sequence along the arrangement direction of the interference fringes for the first detection region divided in S22, which completely retains the local feature distribution of the interference fringes. The preset standard gray value sequence is generated based on a large number of normal metal plate samples through statistical analysis, including mean, median, Gaussian distribution fitting, linear regression model, neural network model, or any other statistical or machine learning method, and the construction process does not depend on specific algorithms, preset thresholds, image resolutions, or environmental conditions. The calculation of the gray deviation value uses basic subtraction and absolute value operation, which is realized by a single basic arithmetic operation, and only requires CPU basic instructions, with high calculation efficiency and can be completed in milliseconds. This calculation method is applicable to all metal plate materials, including stainless steel, aluminum alloy, copper alloy, titanium alloy, composite materials, all surface states, including polishing, sandblasting, coating, oxidation, electroplating, corrosion, and all arrangement angles of the interference fringes. The calculation of the gray deviation value does not depend on the specific angle of the interference fringes, image resolution, noise level, or lighting conditions, thereby realizing universal detection of defects on the metal plate surface in any orientation.

[0119] The step S23 can calculate the gray deviation value between the gray value in the gray value sequence and the gray value in the standard gray value sequence of the detection region according to the implementation of the steps S131-S134, which will not be described herein. The calculation of the gray deviation value adopts the basic subtraction and absolute value operation, and is realized by a single basic arithmetic operation. The process only needs the basic instruction of the CPU, and has high calculation efficiency and can be completed in milliseconds.

[0120] S24, determining whether the first detection region has the surface defect according to the gray deviation value; if the first interference image has the first detection region with the surface defect, performing the step S25; if the first interference image does not have the first detection region with the surface defect, performing the step S26.

[0121] The determination is based on the comparison between the gray deviation value list generated in the step S23 and the preset threshold value. If any deviation value exceeds the preset threshold value, it is determined that there is a defect, otherwise it is determined that there is no defect. The preset threshold value is dynamically set based on the surface state of the metal plate material or the detection environment, and is suitable for all metal plate materials including stainless steel, aluminum alloy, copper alloy, titanium alloy and composite materials, and all surface states including polishing, sand blasting, coating, oxidation, electroplating and corrosion. The determination method is independent of the interference fringe angle image resolution or the detection environment, and ensures the accuracy and consistency of the defect determination, covering all possible defect detection scenarios.

[0122] S25, obtaining a second interference image, and detecting a second detection region with a surface defect in the second interference image to perform superposition on the first detection region with the surface defect and the second detection region with the surface defect to locate the defect region on the surface of the metal plate; wherein the second interference image is generated by irradiating the surface of the metal plate by a second interference system, and the arrangement direction of the interference fringes in the second interference image and the first interference image is independent of each other; the second interference image is divided into a plurality of second detection regions along a direction perpendicular to the arrangement direction of the interference fringes; and the second detection region covers at least part of the area of each interference fringe in the second interference image.

[0123] The superposition process maps the defect regions of the two interference images to a unified spatial reference through coordinate transformation, and accurately narrows down the defect position range to the intersection region by using the complementary information in two directions, thereby realizing the millisecond-level defect positioning.

[0124] S26, determining that the metal plate does not have the surface defect.

[0125] When the step S24 determines that the first interference image has no defect region, the step S26 is directly performed, and the second interference image does not need to be processed. The flow significantly improves the detection efficiency by preferentially verifying the defect state of the first interference image, and avoids redundant calculation.

[0126] In an optional embodiment, as Figure 9As shown, the step of "positioning the defect area of the metal plate surface by performing superposition of the first detection area with surface defects and the second detection area with surface defects in position, in the step S25 above, specifically includes the steps of:

[0127] S251. Constructing a plane reference system according to the first interference image, the second interference image, and the position mapping relationship of the metal plate surface;

[0128] S252. Positioning the first detection area with surface defects and the second detection area with surface defects in the plane reference system;

[0129] S253. Marking the superimposed area of the first detection area with surface defects and the second detection area with surface defects in the plane reference system as the surface defect area;

[0130] S254. Obtaining the position coordinates of the surface defect area of the metal plate surface according to the position mapping relationship of the plane reference system and the metal plate surface.

[0131] In one specific embodiment, as shown, Figure 10 the interference fringe extension direction of the first interference image 11 is 0°, and its coordinate system is (x 11 ,y 11 ); the interference fringe extension direction of the second interference image 12 is 90°, and its coordinate system is (x 12 ,y 12 ); and the unified coordinate system of the metal plate surface 3 is (x3, y3). When performing the step S251 above, the x3 axis corresponds to the direction of the x 11 axis of the first interference image 11, and the y3 axis corresponds to the direction of the y 11 axis of the first interference image 11; the x3 axis corresponds to the direction of the y 12 axis of the second interference image 12, and the y3 axis corresponds to the direction of the x 12 axis of the second interference image 12.

[0132] When performing the step S252, in the plane reference system (x3, y3), the first detection area a111 Figure 10 with surface defects of the first interference image 11 (the red area in the first interference image 11) is represented as a set of longitudinal coordinate ranges P 11 ={[Y 11 , Y 12 ], [Y 21 , Y 22 ],..., [Y i1 , Y i2 ],..., [Y n1 , Y n2 ]}. i1Y i2 is the end coordinate; the second detection area a121 Figure 10 of the second interferometric image 12 in which a surface defect exists (the red area in the second interferometric image 12) is represented as a horizontal coordinate range set Q 12 = {[X 11 , X 12 ], [X 21 , X 22 ],..., [X j1 , X j2 ],..., [X m1 , X m2} where X j1 is the start coordinate of the jthdefect area in the x3direction, and X j2 is the end coordinate.

[0133] When step S253 is performed, the final defect area is the intersection of the first defect area set P 11 and the second defect area set Q 12 , i.e., all the combined areas of x3in [X j1 , X j2 and y3in [Y i1 , Y i2 . For example, if P 11 contains defect areas [111,115] and [366,380], and Q 12 contains defect areas [101,135] and [256,260], then the defect area in the (x3, y3) coordinate system includes areas of x3∈ [101,135] and y3∈ [111,115], x3∈ [101,135] and y3∈ [366,380], x3∈ [256,260] and y3∈ [111,115], and x3∈ [256,260] and y3∈ [366,380].

[0134] When step S254 is performed, the defect area coordinates are converted into physical coordinates in the (x3, y3) coordinate system, and the defect position range (x3∈ [X j1 , X j2 ], y3∈ [Y i1 , Y i2 ]) is outputted, achieving millisecond-level accurate positioning.

[0135] In combination with the foregoing embodiments, in a preferred implementation, in step S23, the first detection area in the first interferometric image 11 has a coordinate range in the y 11 direction of [Y low , Y high .] along the direction of interference fringe arrangement (x 11 Image width W pixels, sequence length n (n is the preset sequence point number, for example, n = W / 10, that is, each sequence point covers W / n pixels). Sequence index k (k = 1, 2,..., n) and x 11 Coordinate interval mapping relationship is x 11 ∈ [(k-1) × (W / n) +1, k × (W / n)]). The gray value sequence B k is calculated as y 11 The coordinate range in the y low direction is [Y high , Y 11 ], and the gray value in the x k direction coordinate range is [(k-1) × (W / n) + 1, k × (W / n)] area. The preset standard gray value sequence A k (namely, the gray value sequence generated based on the normal metal plate sample statistics) is consistent with the sequence length n, and the deviation value sequence is |B k - A k |.

[0136] Step S24 compares the gray value deviation list generated in S23 with the threshold T. If all deviation values are less than T, it is determined that there is no defect. If there is a gray value deviation |B k - A m | > T (k = 1, 2,..., n), the sequence number k of the gray value deviation is recorded. The preset threshold T is dynamically set based on the metal plate material and surface state, for example, T = 20 for a stainless steel surface and T = 35 for an aluminum alloy surface, to ensure that the defect determination matches the actual characteristics. The effective defect area index set K = {k1, k2,..., km} is obtained, where m is the index number (m ≤ n) that satisfies |B k - A k | > T.

[0137] In the execution of step S25, the second interference image 12 (the interference fringe extension direction is 90°) is obtained, based on the mapping relationship established in S23, the index k corresponds to the second detection area with the horizontal coordinate range x3 ∈ [(k-1) × (W / n) + 1, k × (W / n)] in the plane reference system (x3, y3). The gray value deviation is calculated. This index scheme accurately locates the horizontal coordinate range of the second detection area through k, reduces the processing of invalid areas, and the number of overlapping areas may be reduced from n 2 to n × m, significantly reducing the number of area overlaps in step S253 in the foregoing embodiment, to improve the detection efficiency. Embodiment 3

[0138] According to a third aspect of the embodiments of the present application, as shown in Figure 11 A metal sheet surface defect detection device 90 is also provided, which comprises:

[0139] An interference image generation unit 91 is configured to irradiate the metal sheet surface by at least two different direction interference systems respectively to generate at least two interference images containing interference fringes arranged along the preset arrangement direction of the corresponding interference system; wherein the interference fringe arrangement directions of the interference images are independent of each other.

[0140] A detection region division unit 92 is configured to divide the interference images along the direction perpendicular to the interference fringe arrangement direction into a plurality of detection regions, so that each detection region covers at least part of each interference fringe.

[0141] A gray value sequence generation unit 93 is configured to obtain the gray value sequence of the detection region along the interference fringe arrangement direction to calculate the gray deviation value of the gray value in the gray value sequence and the gray value in the preset standard gray value sequence of the detection region.

[0142] A defect detection unit 94 is configured to determine whether the detection region has a surface defect according to the gray deviation value.

[0143] A defect positioning unit 95 is configured to perform position superposition on the detection region having a surface defect between the interference images to position the defect region of the metal sheet surface.

[0144] The detection device is used to perform all or part of the steps of the metal sheet surface defect detection method provided by the foregoing method embodiments, and will not be described here. Embodiment 4

[0145] According to a fourth aspect of the embodiments of the present application, an electronic device 50 is also provided, as shown in Figure 12 The electronic device 50 comprises a processor 51 and a memory 52 for storing processor executable instructions; wherein the processor 51 can comprise one or more processing cores, such as a 4-core processor, an 8-core processor, etc.

[0146] The memory 52 can comprise one or more computer readable storage media, which can be tangible and non-transitory. The memory 52 can further comprise a high-speed random access memory and a non-volatile memory, such as one or more disk storage devices, flash storage devices. In some embodiments, the non-transitory computer readable storage medium in the memory 52 is used to store at least one instruction for being executed by the processor 51 to implement all or part of the steps of the metal sheet surface defect detection method provided in the embodiments of the present application.

[0147] In some embodiments, the electronic device 50 can also optionally include a peripheral device interface 53 and at least one peripheral device. Specifically, the peripheral device includes at least one of a power supply 55. Embodiment 5

[0148] According to a fifth aspect of the embodiments of the present application, a storage medium 60 is also provided, as shown in the figure, which includes a stored computer program, wherein the computer program, when running, controls the device where the storage medium 60 is located to perform all or part of the steps of the metal sheet surface defect detection method provided by the above method embodiments. Figure 13

[0149] Wherein, the program instruction is stored in a computer readable storage medium (which can be CD ROM, U disk, mobile hard disk, etc.) or network, including a plurality of computer program instructions to make a computer equipment (which can be a personal computer, server, or network equipment, etc.) execute the above method according to the embodiments of the present application.

[0150] Through the above description of the embodiments, those skilled in the art can easily understand that the example embodiments described herein can be implemented by software, or by software combined with necessary hardware. Although the embodiments of the present application have been disclosed as above, it is not limited to the application and implementation listed in the specification and embodiments, and it can be fully applied to various fields suitable for the present application, and additional modifications can be easily realized by those skilled in the art, therefore, the present application is not limited to specific details and figures shown and described herein, without departing from the general concept defined by the claims and equivalent scope.

[0151] The device, electronic equipment, non-volatile computer storage medium and method provided by the embodiments of the present application are corresponding, therefore, the device, electronic equipment, non-volatile computer storage medium also has similar beneficial technical effects as the corresponding method, since the beneficial technical effects of the method have been described in detail above, therefore, the beneficial technical effects of the corresponding device, electronic equipment, non-volatile computer storage medium will not be described here.

[0152] Those skilled in the art also know that in addition to implementing the controller in the form of pure computer readable program code, the same function can also be realized by logically programming the method steps to make the controller in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers and embedded microcontrollers, etc. Therefore, such a controller can be considered as a hardware component, and the means included therein for realizing various functions can also be considered as structures within the hardware component. Or even, the means for realizing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0153] ​The systems, apparatuses, modules, or units disclosed in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer can be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0154] For the sake of clarity, the above apparatuses are described with reference to functional blocks and various units that perform the functions. Of course, the functions of the various units can be combined in one or more software and / or hardware modules.

[0155] Those skilled in the art should understand that the embodiments of the present specification can be provided as a method, a system, or a computer program product. Therefore, the embodiments of the present specification can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects.

[0156] The present specification is described with reference to flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present specification. It should be understood that each flow and / or block in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one or more flows and / or blocks. Figure 1 The apparatus that implements the functions specified in one or more flows and / or blocks.

[0157] These computer program instructions can also be stored in a computer-readable memory that can direct the computer or other programmable data processing apparatus to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including instruction apparatus that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one or more flows and / or blocks. Figure 1 The apparatus that implements the functions specified in one or more flows and / or blocks.

[0158] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are performed on the computer or other programmable data processing apparatus to produce a computer-implemented process, so that the instructions executed on the computer or other programmable data processing apparatus provide a process for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one or more flows and / or blocks. Figure 1steps of a function specified in one or more blocks.

[0159] In one typical arrangement, the computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0160] The memory can include non-persistent memory and / or volatile memory, such as random access memory (RAM) and / or cache memory, non-volatile memory, such as read-only memory (ROM), EPROM, and / or flash memory, etc. The memory is an example of computer readable media.

[0161] It is also important to note that the words "comprise," "comprising," "contain," "containing," "include," "including," and "includes," and other variations thereof, are intended to cover a non-exclusive inclusion such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises... a" does not, without more constraints, exclude the existence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0162] The specification can be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. Computer-executable instructions, associated data structures, and program modules represent examples of the program code means for executing steps of the methods disclosed herein. The specification can also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules can be located in both local and remote computer storage media including memory storage devices.

[0163] Embodiments in the specification are described progressively, and the same or similar parts among the embodiments can be mutually referred to. Each embodiment focuses on the difference from other embodiments. In particular, the system embodiments are described relatively simply because they are substantially similar to the method embodiments, and the relevant parts can be referred to the description of the method embodiments.

[0164] The specification only describes the embodiments of the specification and does not limit one or more embodiments of the specification. One or more embodiments of the specification can have various modifications and changes for those skilled in the art. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of one or more embodiments of the specification should be included in the scope of the claims of one or more embodiments of the specification.

Claims

1. A method for detecting surface defects in a metal plate, characterized in that, Includes the following steps: The surface of a metal plate is irradiated by at least two interference systems in different directions to generate at least two interference images containing interference fringes arranged along a preset direction of their respective interference systems; wherein the arrangement direction of the interference fringes between the interference images is independent of each other, and the direction of the interference fringes of each interference image is not parallel and not collinear in space; the interference fringes generated by the first interference system are arranged along a first preset direction, and the interference fringes generated by the second interference system are arranged along a second preset direction, wherein the angle between the first preset direction and the second preset direction is not equal to 0° or 180°; The interference image is divided into several detection regions along a direction perpendicular to the direction of the interference fringe arrangement, so that each detection region at least covers a portion of each interference fringe in the vertical direction; the direction of the interference fringe arrangement is preset and fixed by the interference system, and when the interference fringes extend in the horizontal direction, the division direction is vertical, dividing the interference image into multiple strip-shaped regions; Obtain the gray value sequence of the dry detection area along the direction of the interference fringe arrangement, and calculate the gray value deviation between the gray value in the gray value sequence and the gray value in the preset standard gray value sequence of the detection area; The presence of surface defects in the detection area is determined based on the grayscale deviation value. The detection areas of surface defects in the interference images are superimposed to locate the defect areas on the surface of the metal plate. The interference image includes a first interference image and a second interference image whose interference fringe arrangement directions are independent of each other; wherein, the detection area of ​​the first interference image is denoted as the first detection area; and the detection area of ​​the second interference image is denoted as the second detection area. Before the step of "overlaying the positions of the detection areas of surface defects in the interference images to locate the defect areas on the surface of the metal plate", the method further includes the following step: Determine whether the first interference image detects a first detection area with surface defects; If there is a first detection area with surface defects in the first interference image, it is determined that the metal plate has surface defects, and the gray value deviation value is calculated based on the gray value sequence of the second detection area to determine the second detection area with surface defects detected in the second interference image. If the first detection area of ​​the first interference image does not contain a surface defect, then it is determined that the metal plate does not have a surface defect. The step of "overlaying the detected areas of surface defects between the interference images to locate the defect areas on the surface of the metal plate" specifically includes the following steps: A planar reference system is constructed based on the positional mapping relationship between the first interference image, the second interference image, and the surface of the metal plate; Locate a first detection area with surface defects and a second detection area with surface defects in the planar reference system; The area where the first detection area of ​​the surface defect and the second detection area of ​​the surface defect are superimposed in the planar reference system is marked as the surface defect area; Based on the positional mapping relationship between the planar reference system and the surface of the metal plate, the positional coordinates of the areas on the surface of the metal plate where surface defects exist can be obtained.

2. The detection method according to claim 1, characterized in that, The direction perpendicular to the direction of the interference fringes in the interference image is defined as the sampling direction; The step of "obtaining the gray value sequence of the dry detection area along the direction of the interference fringes, and calculating the gray value deviation between the gray value in the gray value sequence and the gray value in the preset standard gray value sequence of the detection area" specifically includes the following steps: Obtain a preset standard grayscale value sequence, wherein the standard grayscale value sequence includes n grayscale values; The detection area is divided into n grayscale sampling areas along the direction of the interference fringes; Obtain the grayscale values ​​of the grayscale sampling area to generate a grayscale value sequence containing n grayscale values ​​corresponding to the detection area; Calculate the deviation of each gray value in the gray value sequence from its corresponding preset standard gray value sequence.

3. The detection method according to claim 1, characterized in that, Following the step of "overlaying the positions of the detection areas with surface defects between the interference images to locate the defect areas on the surface of the metal plate", the method further includes the following step: Obtain polarization images of the defect areas on the surface of the metal plate; The polarization image is input into a pre-trained surface defect type analysis model to output the defect type of the defect region.

4. A device for detecting surface defects in a metal plate, used to implement the steps of the method as described in any one of claims 1-3, characterized in that, include: An interference image generation unit is used to irradiate the surface of a metal plate with at least two interference systems in different directions to generate at least two interference images containing interference fringes arranged along a preset direction of their corresponding interference systems; wherein the arrangement direction of the interference fringes between the interference images is independent of each other; The detection area division unit divides the interference image into several detection areas along a direction perpendicular to the direction of the interference fringe arrangement, so that each detection area at least covers a portion of each interference fringe. The gray value sequence generation unit is used to acquire the gray value sequence of the dry detection area along the direction of the interference fringe arrangement, and to calculate the gray value deviation between the gray value in the gray value sequence and the gray value in the preset standard gray value sequence of the detection area. The defect detection unit determines whether surface defects exist in the detection area based on the grayscale deviation value. The defect location unit performs position superposition on the detection areas of surface defects in the interference images to locate the defect areas on the surface of the metal plate.

5. An electronic device, characterized in that, The method includes a memory and a processor; wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the steps of the method according to any one of claims 1-3.

6. A storage medium, characterized in that, It stores computer instructions; wherein, when executed by a processor, the computer instructions implement the steps of the method described in any one of claims 1-3.

Images