Defect identification method and device after strip steel pickling and heavy brushing and electronic equipment
By using image processing and model recognition technology based on steel grade quality, defects that have a small impact on quality after pickling and rebrushing of steel strips are screened out and eliminated, solving the problem of missed defects in the detection system and achieving efficient and accurate defect identification.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-02
- Publication Date
- 2026-03-10
AI Technical Summary
In existing technologies, the surface texture of steel strip after pickling and rebrushing is too heavy, resulting in too many defects being detected by the inspection system. Quality inspectors cannot detect key defects in time, leading to the omission of serious defects.
Based on the preset steel quality, the target steel grade of the strip to be inspected is determined. Through image preprocessing, edge recognition algorithm, defect type database and trained re-brush level defect recognition model, defects with little impact on strip quality are screened and eliminated, thereby improving the accuracy and efficiency of key defect identification.
This improved the accuracy of defect identification after pickling and rebrushing of strip steel, reduced the rate of missed defects, and ensured the efficiency and accuracy of strip steel quality inspection.
Smart Images
Figure CN121640159A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of strip steel production technology, and in particular to a method, apparatus and electronic equipment for identifying defects in strip steel after pickling and rebrushing. Background Technology
[0002] Surface defects in strip steel are a common problem in steel production. They can be caused by a variety of factors, including production processes, raw material quality, and equipment condition. These defects include porosity, iron oxide scale indentation, zinc slag defects, air knife streaks, and incomplete plating. These defects not only affect the appearance quality of the strip steel but may also reduce its mechanical properties and corrosion resistance, thereby affecting the performance and lifespan of the final product. Traditional steel companies rely on personnel to uncoil and inspect for surface defects. This inspection method is prone to batch quality defects due to limited uncoiling length, low defect detection rate, and untimely defect feedback. It can even lead to defective steel coils entering the market and causing quality problems.
[0003] For high-end pickled steel sheet production lines, major steel companies have installed surface inspection systems. These systems can automatically and dynamically detect and classify defects on the upper and lower surfaces of the strip during the rolling process. To improve the quality of automotive steel sheets, a series of modifications have been made to the pickling and rinsing section, circulation system, and misting system, including the addition of re-brushing equipment. This equipment adds multiple sets of brush rollers with bristles, water washing, and mist extraction, and brushes the strip surface at a 40-60° angle and under compressive stress. This can remove 4-8μm surface defects from the strip and improve the surface quality of subsequent processes, such as galvanized automotive outer panels. However, the surface after pickling and re-brushing will have a certain degree of texture.
[0004] When the surface texture of the strip steel is too heavy, it will cause the pickling surface inspection system to detect too many defects, and the quality inspectors will not be able to find key defects in time, resulting in serious defects being missed. Summary of the Invention
[0005] This application provides a method, apparatus, and electronic device for identifying defects in steel strip after pickling and rebrushing. The embodiments provided by this application solve the technical problem in the prior art where the number of defects detected by the pickling surface inspection system is too large, and quality inspectors cannot detect key defects in time, resulting in serious defects being missed. The embodiments provided by this application can screen out and remove defects on the surface of the steel strip to be inspected after pickling that have a small impact on the quality of the steel strip to be inspected, thereby improving the accuracy and efficiency of identifying key defects that affect the quality of the steel strip to be inspected and reducing the probability of defects being missed.
[0006] In a first aspect, this application provides a method for identifying defects in steel strip after pickling and rebrushing, the method comprising: Based on the preset steel grade quality, the target steel grade corresponding to the strip to be tested is determined; After pickling and rebrushing the strip steel of the target steel grade, a target image of the surface of the strip steel to be inspected after pickling and rebrushing and the rebrushing level corresponding to the target image are determined. The target image is used to characterize the image determined by image preprocessing of the initial image of the surface of the strip steel to be inspected. Based on the re-brush level and the preset edge recognition algorithm, the target image is used to identify defects and determine the defect image to be identified corresponding to the strip to be detected. Based on a preset defect type database and a preset defect recognition algorithm, the defect images to be identified are filtered to determine the target defect image corresponding to the strip to be detected. Based on the trained re-brush level defect recognition model and preset defect screening rules, defects in the target defect image are screened to determine the key defects on the surface of the strip steel to be inspected that affect the quality of the strip steel to be inspected.
[0007] In one feasible implementation, the image preprocessing includes grayscale processing. The step of determining the target image of the strip surface after pickling and rebrushing of the target steel grade and the corresponding rebrushing level of the target image includes: An initial image of the surface of the strip steel to be inspected after pickling and rebrushing of the target steel grade. The initial image is subjected to grayscale processing to delete image points with abnormal grayscale values. The target image of the strip surface to be inspected after pickling and rebrushing and the rebrushing level corresponding to the target image are determined. The grayscale processing is performed using the median grayscale value of the pixels.
[0008] In one feasible implementation, the step of performing defect identification on the target image based on the repainting level and a preset edge recognition algorithm to determine the defect image to be identified corresponding to the strip to be detected includes: Based on the reflow level corresponding to the target image and the target steel grade corresponding to the strip to be detected, the background texture parameters and sensitivity parameters corresponding to the preset edge recognition algorithm are determined. Based on the background texture parameters, the sensitivity parameters, and the preset edge recognition algorithm, the target image is used to identify defects and determine the defect image to be identified corresponding to the strip to be detected.
[0009] In one feasible implementation, the step of filtering the defect images to be identified based on a preset defect type database and a preset defect identification algorithm to determine the target defect image corresponding to the strip to be detected includes: Based on a preset defect identification algorithm, the grayscale range of the defects corresponding to the strip steel to be detected is determined; Based on a preset defect type database and the grayscale range of the defects, the defect images to be identified are filtered to determine the target defect image corresponding to the strip to be detected.
[0010] In one feasible implementation, the step of filtering the defect images to be identified based on a preset defect type database and the grayscale range of the defects to determine the target defect image corresponding to the strip to be detected includes: Based on a preset defect type database, candidate defect images containing defects are determined from the defect images to be identified; Based on the grayscale range of the defect, the candidate defect images containing the defect are filtered to determine the target defect image corresponding to the strip to be detected.
[0011] In one feasible implementation, the step of filtering defects in the target defect image based on a trained re-brush level defect recognition model and preset defect filtering rules to determine key defects on the surface of the strip steel to be inspected that affect the quality of the strip steel includes: The target defect image is input into the trained re-brush level defect recognition model to identify at least one brush mark defect in the target defect image and the re-brush level corresponding to each brush mark defect, and the brush mark defects with a mild re-brush level are removed. Based on preset defect screening rules, the target defect image after removing minor brush marks is screened to identify key defects on the surface of the strip steel that affect the quality of the strip steel to be inspected.
[0012] In one feasible implementation, the step of filtering defects in the target defect image based on a trained re-brush level defect recognition model and preset defect filtering rules to determine key defects on the surface of the strip steel to be inspected that affect the quality of the strip steel includes: The target defect image is input into the trained re-brush level defect recognition model to identify at least one brush mark defect in the target defect image and the re-brush level corresponding to each brush mark defect, and the brush mark defects with a mild re-brush level are removed. Based on preset defect screening rules, the target defect image after removing minor brush marks is screened to identify key defects on the surface of the strip steel that affect the quality of the strip steel to be inspected.
[0013] In one feasible implementation, the trained repainting level defect recognition model is determined in the following manner: Obtain sample images of steel strip defects with different brush mark levels; The sample strip defect image is input into the initial re-brush level defect recognition model, and the initial re-brush level defect recognition model is trained to determine the trained re-brush level defect recognition model.
[0014] In a second aspect, this application provides a defect identification device for steel strip after pickling and rebrushing, the defect identification device comprising: The first determining module is used to determine the target steel grade corresponding to the strip to be inspected based on the preset steel grade quality. The second determining module is used to determine, after pickling and rebrushing the strip steel of the target steel grade, the target image of the surface of the strip steel to be inspected and the rebrushing level corresponding to the target image, wherein the target image is used to characterize the image determined by image preprocessing of the initial image of the surface of the strip steel to be inspected. The third determining module is used to perform defect identification on the target image based on the re-brush level and the preset edge recognition algorithm, and determine the defect image to be identified corresponding to the strip to be detected; The fourth determining module is used to filter the defect images to be identified based on a preset defect type database and a preset defect identification algorithm, and determine the target defect image corresponding to the strip to be detected. The fifth determination module is used to filter defects in the target defect image based on the trained re-brush level defect recognition model and preset defect screening rules, and to determine the key defects on the surface of the strip steel to be inspected that affect the quality of the strip steel to be inspected.
[0015] In a third aspect of this application, an electronic device is provided, comprising: a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus, and the machine-readable instructions are executed by the processor to perform the steps of the defect identification method after pickling and rebrushing of steel strip as described above.
[0016] In a fourth aspect of this application, an embodiment of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the steps of the defect identification method for pickled and rebrushed steel strip as described above.
[0017] Compared with the prior art, the defect identification method, apparatus, and electronic equipment for pickled and rebrushed strip steel provided in this application have the following advantages: Based on a preset steel grade quality, the embodiments of this application determine the target steel grade corresponding to the strip steel to be inspected. After pickling and rebrushing the strip steel of the target steel grade, the target image of the surface of the strip steel to be inspected after pickling and rebrushing and the rebrushing level corresponding to the target image are determined. Then, based on the rebrushing level and a preset edge recognition algorithm, defect identification is performed on the target image to determine the defect image to be identified for the strip steel to be inspected. Next, based on a preset defect type database and a preset defect identification algorithm, the defect images to be identified are filtered to determine the target defect image corresponding to the strip steel to be inspected. Based on a trained rebrushing level defect identification model and preset defect filtering rules, defects in the target defect images are filtered to determine the key defects on the surface of the strip steel to be inspected that affect its quality. This application can filter and remove defects on the surface of the strip steel to be inspected after pickling that have a minor impact on its quality, improving the accuracy and efficiency of identifying key defects that affect the quality of the strip steel to be inspected and reducing the probability of missed defects. Attached Figure Description
[0018] Figure 1 The flowchart illustrates a defect identification method for steel strip after pickling and rebrushing, as provided in the embodiments of the application. Figure 2 This illustration shows a schematic diagram of the initial image in a defect identification method for steel strip after pickling and rebrushing, provided in an embodiment of this application. Figure 3 This illustration shows a schematic diagram of grayscale processing to determine the target image in a defect identification method for steel strip after pickling and rebrushing, provided in an embodiment of this application. Figure 4 This illustration shows a schematic diagram of the unit detection frame before being changed in a defect identification method for steel strip after pickling and rebrushing, provided in an embodiment of this application. Figure 5 This illustration shows a schematic diagram of the modified unit detection frame in a defect identification method for steel strip after pickling and rebrushing, provided in an embodiment of this application. Figure 6 This illustration shows a structural diagram of the tile after modification in a defect identification method for steel strip after pickling and rebrushing, provided in an embodiment of this application. Figure 7 This paper shows a structural block diagram of a defect identification device for steel strip after pickling and rebrushing, provided in an embodiment of this application. Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown.
[0019] Figure 7 and Figure 8 The correspondence between the figure labels and figure titles in the accompanying drawings is as follows: 700 Defect identification device after pickling and rebrushing of steel strip; 710 First determination module; 720 Second determination module; 730 Third determination module; 740 Fourth determination module; 750 Fifth determination module; 800 Electronic device; 810 Processor; 820 Memory; 830 Bus. Detailed Implementation
[0020] To better understand the technical solutions provided in the embodiments of this specification, the technical solutions of the embodiments of this specification will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of this specification and the specific features in the embodiments are detailed descriptions of the technical solutions of the embodiments of this specification, rather than limitations on the technical solutions of this specification. In the absence of conflict, the embodiments of this specification and the technical features in the embodiments can be combined with each other.
[0021] In this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, without necessarily requiring or implying any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. The term "two or more" includes two or more cases.
[0022] First, the applicable application scenarios of this application will be introduced. The embodiments provided in this application are applicable to the field of strip steel production technology.
[0023] Currently, when the surface texture of the strip steel is too heavy, it will cause the pickling surface inspection system to detect too many defects, making it impossible for quality inspectors to detect key defects in time, resulting in serious defects being missed.
[0024] In existing technologies, newer hot and cold rolling production lines are equipped with surface inspection equipment. Based on the camera sensor type, this can be categorized into surface scanning and line scanning. Pickling units more often install surface scanning inspection instruments, which can better detect pickling defects such as color differences. Surface scanning uses LED infrared strobe light sources or ultraviolet light sources; while line scanning uses higher brightness linear array light sources. Surface scanning systems offer greater stability, thus having significant advantages in continuous casting and hot rolling.
[0025] Based on this, the embodiments of this application provide a method, device, and electronic device for identifying defects after pickling and rebrushing of steel strip. The embodiments provided by this application solve the technical problem in the prior art where the number of defects detected by the pickling surface inspection system is too large, and quality inspectors cannot find key defects in time, resulting in serious defects being missed. The embodiments provided by this application can screen out and remove defects on the surface of the steel strip to be inspected after pickling that have a small impact on the quality of the steel strip to be inspected, improve the accuracy and efficiency of identifying key defects that affect the quality of the steel strip to be inspected, and reduce the probability of defects being missed.
[0026] Figure 1 This is a flowchart illustrating a defect identification method for steel strip after pickling and rebrushing, as provided in the embodiments of the application. Figure 1 As shown, the defect identification method for steel strip after pickling and rebrushing includes the following steps: S101. Based on the preset steel grade quality, determine the target steel grade corresponding to the strip to be tested.
[0027] In this step, the embodiment provided in this application uses surface scanning to identify defects in the strip steel to be inspected after pickling and rebrushing. Specifically, it can rely on surface scanning machine vision inspection equipment. This application uses rebrushing mode messages to distinguish the materials corresponding to the strip steel to be inspected. Specifically, based on the preset steel grade quality, it determines the target steel grade that matches the preset steel grade quality of the strip steel to be inspected. This application can identify defects in strip steel to be inspected for different target steel grades, which improves the accuracy of defect identification in strip steel to be inspected and reduces the influence of defect identification results between strip steel to be inspected for different target steel grades.
[0028] Understandably, the embodiments provided in this application can classify the strip installation quality (i.e., steel grade) based on the differences in Si, P, or other special production composition in the steel, which lead to differences in the surface quality of the strip. Specifically, the steel grade can be divided into different material groups such as low, normal, and high, thereby generating different strips of low, normal, and high quality. S102. After pickling and rebrushing the strip steel of the target steel grade, determine the target image of the surface of the strip steel after pickling and rebrushing and the rebrushing level corresponding to the target image, wherein the target image is used to characterize the image determined by image preprocessing of the initial image of the surface of the strip steel.
[0029] In this step, after determining the target steel grade corresponding to the strip steel to be tested, the embodiment provided in this application first determines whether the production line of the strip steel to be tested has added special production line processes such as edge trimming, re-brushing, and coating. If so, after the pickling and re-brushing process of the strip steel to be tested for the target steel grade, the re-brushing level corresponding to the strip steel to be tested under the target steel grade is determined, and the re-brushing level is determined by adding "special material parameters".
[0030] It is understandable that, in the embodiments provided in this application, after determining the target steel grade and the re-brushing level of the strip to be tested, if it is necessary to optimize the special material of the re-brushing level of the strip to be tested in the subsequent process, the optimization process of this parameter will not affect the other normal steel grades.
[0031] In this application, the brushing levels of the embodiments can be set to different pressures and speeds: A (light): 800 rpm + 2 bar; B (medium): 1000 rpm + 5 bar; C (heavy): 1200 rpm + 10 bar. The brush roller pressure of level C is high and the speed is fast, which makes the physical gap between the bristles and the texture formed on the roller surface more consistent, while level A has fine brush marks.
[0032] Here, we take one embodiment as an example: the target steel grade corresponding to the strip steel to be tested in the embodiment provided in this application is determined to be 64AO2. After determining that the strip steel to be tested of the target steel grade is pickled and rebrushed, the strip steel to be tested after pickling and rebrushing is marked as 64AO2A, 64AO2B, or 64AO2C, where A / B / C specifically represent the rebrushing level of the strip steel to be tested after pickling.
[0033] For example, the image preprocessing in the embodiments provided in this application includes grayscale processing. After pickling and rebrushing the strip steel of the target steel grade, an initial image of the surface of the strip steel to be inspected after pickling and rebrushing is obtained. Grayscale processing is performed on the initial image to delete image points with abnormal grayscale values in the initial image, and a target image of the surface of the strip steel to be inspected after pickling and rebrushing and the rebrushing level corresponding to the target image are determined. The grayscale processing is performed using the median grayscale value of the pixels.
[0034] It should be noted that the embodiment provided in this application involves acquiring an initial image of the surface of the strip steel to be inspected after pickling and rebrushing of the target steel grade using an image acquisition device. Then, a median grayscale value algorithm is used to process the initial image in grayscale, and image points with abnormal grayscale values in the initial image are deleted. The initial image after deletion is then determined as the target image of the surface of the strip steel to be inspected after pickling and rebrushing.
[0035] It is understood that the grayscale processing of the initial image in the embodiments provided in this application is for the purpose of noise reduction of the background texture of the initial image. Specifically, it can be: to perform grayscale averaging on the area of the initial image with a horizontal and vertical size of i*j pixels to remove isolated image points with grayscale abnormalities. Here, the size of i*j needs to be matched with the defect (i is the number of horizontal pixels and j is the number of vertical pixels).
[0036] In order to avoid removing normal defects (defects that are very dark or very bright), the embodiments provided in this application use the median gray value for grayscale calculation, thereby reducing the probability of removing very dark or very bright defects and affecting the overall average grayscale.
[0037] Figure 2 This illustration shows a schematic diagram of the initial image in a defect identification method for steel strip after pickling and rebrushing, provided in an embodiment of this application.
[0038] Figure 3 This illustration shows a schematic diagram of grayscale processing for determining the target image in a defect identification method for steel strip after pickling and rebrushing, provided in an embodiment of this application. Figure 3 As shown, in the embodiments provided in this application, the spacing of the brush bristles is set to 0.6mm, 2.8mm and 3.7mm, and the initial image height is set to 70mm. This application averages the brightness and darkness grayscale values, and the texture of the bright and dark areas of the image is reduced after grayscale averaging.
[0039] In the above, this application sets the value of i to 0.6mm*2, 2.8mm*2, and 3.7mm*2, and j to 70mm. Combining the fact that the size of one pixel is 0.5mm and 1mm in the horizontal and vertical directions, the value of i is converted to the number of pixels as 2.5 / 11.2 / 14.8 (rounded to the integers 3 / 11 / 15), and the value of j is 70.
[0040] S103. Based on the re-brush level and the preset edge recognition algorithm, perform defect recognition on the target image to determine the defect image to be identified corresponding to the strip to be detected.
[0041] In this step, the embodiments provided in this application first perform a preliminary quick defect screening and identification on the target image after determining the target image, to determine whether there are defects in the target image, and identify the target image with defects as the defect image to be identified corresponding to the strip steel to be detected, and directly remove the target image without defects, so as to avoid calculating a large number of redundant identifications when further identifying key defects, thereby affecting the identification efficiency.
[0042] It should be noted that the methods for defect identification of the target image in the embodiments provided in this application include, but are not limited to, border detection (BD), detection module (Observation Device, OD), and edge region detection (ERD), etc.
[0043] For example, this application specifically determines the image of the defect to be identified corresponding to the strip to be inspected in the following way: Based on the reflow level corresponding to the target image and the target steel grade corresponding to the strip to be detected, the background texture parameters and sensitivity parameters corresponding to the preset edge recognition algorithm are determined; based on the background texture parameters, the sensitivity parameters and the preset edge recognition algorithm, the target image is used to identify defects, and the defect image to be identified corresponding to the strip to be detected is determined.
[0044] As described above, different target re-brush levels correspond to different background texture parameters and sensitivity parameters. This application determines that the sensitivity parameter range corresponding to the Low-Normal-High level is (0-255); the background texture parameter range is (0-100%). For target images with many textures or pseudo-defects at the re-brush level, the embodiments provided in this application need to reduce the sensitivity parameter obtained by defect identification and increase the background texture parameter.
[0045] Here, the background texture parameters and sensitivity parameters of the strip steel to be detected under different target repaint levels and different target steel grades are shown in Table 1: Table 1
[0046] S104. Based on a preset defect type database and a preset defect recognition algorithm, the defect images to be identified are filtered to determine the target defect image corresponding to the strip to be detected.
[0047] In this step, after initially determining the defect image to be identified from the target image, the embodiment provided in this application begins to perform a secondary screening of the aforementioned defect image to be identified. This screening of the image requires comparing all suspected defects in the defect image to be identified with preset defects in the preset defect type database, identifying the suspected defects in the defect image to be identified that match the preset defect type database, and removing the defect images to be identified that do not match the preset defect type database. The defect images to be identified remaining after removal are determined to be the target defect images corresponding to the strip steel to be inspected, so as to reduce the number of false defects.
[0048] For example, based on a preset defect recognition algorithm, the grayscale range of the defect corresponding to the strip to be detected is determined; based on a preset defect type database and the grayscale range of the defect, the image of the defect to be identified is filtered to determine the target defect image corresponding to the strip to be detected.
[0049] Understandably, the embodiments provided in this application use a preset defect recognition algorithm to annotate defects in the defect image to be identified. Specifically, it can detect and identify defects based on unit detection boxes (tiles), identify the grayscale range of the defects corresponding to the detected strip steel, compare and match the information of each preset defect in the preset defect type database with the grayscale range of the identified defects, and remove the defect images to be identified where the defects do not match, thereby determining and generating the target defect image corresponding to the strip steel to be detected.
[0050] It should be noted that the unit detection frame (tile) in the embodiments provided in this application can be customized and set according to the morphology and characteristics of the defects to be detected. Here, the tile in this application can be set as strip, sheet, or oblique shape, etc.
[0051] The typical tile unit detection box is defined as follows: { x = 4; y = 8; overlap = { x = 4; y = 4;};}; Where x is usually 4-16; y is usually 4-32; overlap is usually set to {x=4; y=4}.
[0052] Considering the characteristics of texture (vertical), and since tiles also have a similar vertical orientation, defect detection is affected by vertical texture. Therefore, the x and y values are set to x > y, shifting to horizontal detection. Specific changes to the horizontal and vertical orientation of tiles are as follows: Figure 4 and Figure 5 As shown.
[0053] Figure 4 This illustration shows a schematic diagram of the unit detection frame before being changed in a defect identification method for steel strip after pickling and rebrushing, provided in an embodiment of this application. Figure 5 The diagram shows a modified unit detection frame in a defect identification method for steel strip after pickling and rebrushing, as provided in an embodiment of this application.
[0054] In the above, the overlap adjustment is specifically: overlap={ x=4; y=4 :}. The overlap adjustment is used to indicate that 4 pixels are superimposed in both the x and y directions. Here, reducing the number of superimposed pixels in the y direction is also to reduce the vertical height of the tile, so the final setting is: tile= { x = 8; y = 4; overlap={ x=4; y=2:};}.
[0055] Here, the structural diagram of the tile after the change is as follows: Figure 6 As shown. Figure 6 This illustration shows a structural diagram of the tile after modification in a defect identification method for steel strip after pickling and rebrushing, provided in an embodiment of this application.
[0056] For example, based on a preset defect type database, candidate defect images containing defects are determined from the defect images to be identified; based on the grayscale range of the defects, the candidate defect images containing defects are filtered to determine the target defect image corresponding to the strip to be detected.
[0057] Understandably, this application determines whether there are candidate defects in the image to be identified that match the preset defect type database based on the preset defect type database. If there are, the image to be identified with the candidate defects is determined as a candidate defect image. Then, the grayscale range of each candidate defect in the candidate defect image is determined, such as the bright pixel range and the dark pixel range. Then, the bright pixel range and the dark pixel range are compared with the preset dark pixel grayscale deviation and the preset bright pixel grayscale deviation, and the target defect image corresponding to the strip to be detected that meets the requirements is selected.
[0058] It should be noted that when the tile is determined to have 32 pixels, the preset pixel grayscale deviation threshold is [20, 25, 30, 35]. The corresponding number of bright pixels exceeding the standard should not exceed [12, 9, 6, 4], the corresponding number of dark pixels exceeding the standard should not exceed [12, 9, 6, 3], and the total number of bright and dark pixels exceeding the standard should not exceed [16, 13, 9, 6]. Otherwise, the tile is considered to contain candidate defects. This setting can reduce the texture candidate defects caused by re-brushing to a certain extent, and determine the image of the defect to be identified where the above candidate defects are located as the target defect image.
[0059] Here, in a candidate defect image of this application, if the number of bright pixels or dark pixels exceeds 12 points, or if both bright and dark pixels exceed the limit, but the total number of pixels exceeding the limit is 16 points, it is assumed that the tlie module has a candidate defect.
[0060] In the above-described embodiments, the embodiments provided in this application, through comparison, determine that the grayscale of the texture defects is relatively light and the contrast is not large. The average grayscale of the dark color difference is 95-105, the average grayscale of the bright color difference is 140-150, the grayscale of the steel background is 120, the grayscale deviation of the dark pixels is about 15-25, and the grayscale deviation of the bright pixels is about 20-30. Therefore, the preset dark pixel grayscale deviation and the preset bright pixel grayscale deviation of this application need to be reduced by about 15-25 for dark pixels and by 20-30 for bright pixels. Therefore, the preset pixel grayscale deviation threshold is set to above 20, specifically [20, 25, 30, 35].
[0061] The preset defect type database can be customized and used according to different application scenarios. This preset defect type database is used to store and record various types of preset defects.
[0062] S105. Based on the trained re-brush level defect recognition model and preset defect screening rules, the defects in the target defect image are screened to determine the key defects on the surface of the strip steel to be inspected that affect the quality of the strip steel to be inspected.
[0063] In this step, the embodiment provided in this application, after determining the target defect image, inputs the target defect image into the trained re-brush level defect recognition model to determine the brush mark defect corresponding to the target defect image and the brush mark level. Then, the brush mark defects in the identified target defect image are filtered according to the preset defect screening rules to determine the key defects on the surface of the strip steel to be inspected that affect the quality of the strip steel to be inspected.
[0064] For example, the target defect image is input into the trained re-brush level defect recognition model to identify at least one brush mark defect in the target defect image and the re-brush level corresponding to each brush mark defect, and brush mark defects with a re-brush level of mild are removed; based on preset defect screening rules, the target defect image after removing mild brush mark defects is screened to determine the key defects on the surface of the strip steel to be inspected that affect the quality of the strip steel to be inspected.
[0065] In this step, the embodiments provided in this application can input the target defect image into a trained re-brush level defect recognition model. Through self-learning (decision tree, neural network, etc.), the model can automatically distinguish the brush mark defects corresponding to the target defect image and the brush mark level. Brush mark defects with a re-brush level of mild are removed. Then, the target defect image after removing mild brush mark defects is filtered according to a preset defect screening rule to remove false defects (i.e., non-key defects) and to determine the key defects on the surface of the strip steel that affect the quality of the strip steel to be inspected.
[0066] Understandably, the logic behind setting preset defect screening rules is as follows: Defects that are large but have low density should be removed: such as defects with area A > 400 mm2 and density A_r < 30%; defects that are too small should be removed: such as defects with area < 4 mm2; defects whose average gray value is not within the threshold range should be removed: such as removing defects with an average gray value of 95-105 for dark colors and 140-150 for light colors.
[0067] For example, the embodiments provided in this application determine the trained repainting level defect recognition model in the following manner: Acquire sample strip defect images of brush mark defects at different re-brushing levels; input the sample strip defect images into an initial re-brushing level defect recognition model, train the initial re-brushing level defect recognition model, and determine the trained re-brushing level defect recognition model.
[0068] Here, the sample strip defect images of different brush mark defects of different re-brushing levels in this application can be: sample strip defect images of three different degrees of severity: brush mark A, brush mark B, and brush mark C; the recognition accuracy of the trained re-brushing level defect recognition model is about 85-90%.
[0069] In the above, the type of the initial re-refresh level defect identification model in the embodiments provided in this application can be customized and used according to different application scenarios. The initial re-refresh level defect identification model in this application can be set as: CBE self-learning model.
[0070] The defect identification method for strip steel after pickling and rebrushing provided in this application differs from existing defect identification methods in that it determines the target steel grade of the strip steel to be inspected based on a preset steel grade quality. After pickling and rebrushing the strip steel of the target steel grade, it determines the target image of the strip steel surface after pickling and rebrushing and the rebrushing level corresponding to the target image. Then, based on the rebrushing level and a preset edge recognition algorithm, it performs defect identification on the target image to determine the defect image to be identified for the strip steel to be inspected. Next, based on a preset defect type database and a preset defect identification algorithm, it filters the defect images to be identified to determine the target defect image corresponding to the strip steel to be inspected. Finally, it identifies the defect images based on the trained rebrushing level defects. The identification model and preset defect screening rules are used to screen defects in the target defect image to identify key defects on the surface of the strip steel to be inspected that affect its quality. This application can screen and remove defects on the surface of the strip steel to be inspected after pickling that have a minor impact on its quality, thereby improving the accuracy and efficiency of identifying key defects that affect the quality of the strip steel to be inspected and reducing the probability of missed defects. This application performs initial and secondary screening of the target image where key defects are located through image preprocessing and optimization of low detection sensitivity, removes minor false defects, and further accurately locks down key defects through self-learning, thereby improving the accuracy of key defect detection. It has universality and reduces the occurrence of blindly reducing the number of defects.
[0071] Figure 7 This diagram illustrates a structural block diagram of a defect identification device for steel strip after pickling and rebrushing, provided in an embodiment of this application. Figure 7 As shown, the defect identification device 700 for steel strip after pickling and rebrushing includes: The first determining module 710 is used to determine the target steel grade corresponding to the strip to be inspected based on the preset steel grade quality.
[0072] The second determining module 720 is used to determine, after pickling and rebrushing the strip steel of the target steel grade, a target image of the surface of the strip steel to be inspected and the rebrushing level corresponding to the target image, wherein the target image is used to characterize the image determined by image preprocessing of the initial image of the surface of the strip steel to be inspected.
[0073] The third determining module 730 is used to perform defect identification on the target image based on the re-brush level and the preset edge recognition algorithm, and to determine the defect image to be identified corresponding to the strip to be detected.
[0074] The fourth determining module 740 is used to filter the defect images to be identified based on a preset defect type database and a preset defect identification algorithm, and determine the target defect image corresponding to the strip to be detected.
[0075] The fifth determination module 750 is used to screen defects in the target defect image based on the trained re-brush level defect recognition model and preset defect screening rules, and to determine the key defects on the surface of the strip steel to be inspected that affect the quality of the strip steel to be inspected.
[0076] For example, the image preprocessing includes grayscale processing, and the second determining module 720 is specifically used for: An initial image of the surface of the strip steel to be inspected after pickling and rebrushing of the target steel grade.
[0077] The initial image is subjected to grayscale processing to delete image points with abnormal grayscale values. The target image of the strip surface to be inspected after pickling and rebrushing and the rebrushing level corresponding to the target image are determined. The grayscale processing is performed using the median grayscale value of the pixels.
[0078] For example, the third determining module 730 is specifically used for: Based on the re-brush level corresponding to the target image and the target steel grade corresponding to the strip to be detected, the background texture parameters and sensitivity parameters corresponding to the preset edge recognition algorithm are determined.
[0079] Based on the background texture parameters, the sensitivity parameters, and the preset edge recognition algorithm, the target image is used to identify defects and determine the defect image to be identified corresponding to the strip to be detected.
[0080] For example, the fourth determining module 740 is specifically used for: Based on a preset defect identification algorithm, the grayscale range of the defects corresponding to the strip steel to be detected is determined.
[0081] Based on a preset defect type database and the grayscale range of the defects, the defect images to be identified are filtered to determine the target defect image corresponding to the strip to be detected.
[0082] For example, the step of filtering the defect images to be identified based on a preset defect type database and the grayscale range of the defects to determine the target defect image corresponding to the strip to be detected includes: Based on a preset defect type database, candidate defect images containing defects are identified from the defect images to be identified.
[0083] Based on the grayscale range of the defect, the candidate defect images containing the defect are filtered to determine the target defect image corresponding to the strip to be detected.
[0084] For example, the fifth determining module 750 is specifically used for: The target defect image is input into the trained re-brush level defect recognition model to identify at least one brush mark defect in the target defect image and the re-brush level corresponding to each brush mark defect, and the brush mark defects with a mild re-brush level are removed.
[0085] Based on preset defect screening rules, the target defect image after removing minor brush marks is screened to identify key defects on the surface of the strip steel that affect the quality of the strip steel to be inspected.
[0086] For example, the trained repainting level defect recognition model is determined in the following way: Obtain sample strip defect images of brush mark defects at different brushing levels.
[0087] The sample strip defect image is input into the initial re-brush level defect recognition model, and the initial re-brush level defect recognition model is trained to determine the trained re-brush level defect recognition model.
[0088] The defect identification device 700 for strip steel after pickling and rebrushing provided in this application embodiment, compared with the defect identification device in the prior art, determines the target steel grade corresponding to the strip steel to be inspected based on a preset steel grade quality, and after pickling and rebrushing the strip steel of the target steel grade, determines the target image of the surface of the strip steel to be inspected after pickling and rebrushing and the rebrushing level corresponding to the target image. Then, based on the rebrushing level and a preset edge recognition algorithm, it performs defect identification on the target image to determine the defect image to be identified corresponding to the strip steel to be inspected. Next, based on a preset defect type database and a preset defect identification algorithm, it filters the defect images to be identified to determine the target defect image corresponding to the strip steel to be inspected, and based on the trained rebrushing level... The defect identification model and preset defect screening rules are used to screen defects in the target defect image to identify key defects on the surface of the strip steel to be inspected that affect its quality. This application can screen and remove defects on the surface of the strip steel to be inspected after pickling that have a minor impact on its quality, thereby improving the accuracy and efficiency of identifying key defects that affect the quality of the strip steel to be inspected and reducing the probability of missed defects. This application performs initial and secondary screening of the target image where key defects are located through image preprocessing and optimization of low detection sensitivity, removes minor false defects, and further accurately locks down key defects through self-learning, thereby improving the accuracy of key defect detection. It has universality and reduces the occurrence of blindly reducing the number of defects.
[0089] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 8 As shown, the electronic device 800 includes a processor 810, a memory 820, and a bus 830.
[0090] The memory 820 stores machine-readable instructions executable by the processor 810. When the electronic device 800 is running, the processor 810 and the memory 820 communicate via the bus 830. When the machine-readable instructions are executed by the processor 810, they can perform the operations described above. Figures 1 to 6 The steps of the defect identification method after pickling and rebrushing of the strip steel in the method embodiment shown are described in detail in the method embodiment, and will not be repeated here.
[0091] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can perform the above-described actions. Figures 1 to 6 The steps of the defect identification method after pickling and rebrushing of the strip steel in the method embodiment shown are described in detail in the method embodiment, and will not be repeated here.
[0092] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.
[0093] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.
[0094] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-readable program code.
[0095] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0096] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0097] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0098] This application also provides a computer program product, which includes computer software instructions that, when executed on a processing device, cause the processing device to execute a process for identifying defects after pickling and rebrushing of steel strip.
[0099] A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a server or data center that integrates one or more available media. The available medium may be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).
[0100] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.
[0101] In the several embodiments provided in this application, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between devices or units, and may be electrical, mechanical, or other forms.
[0102] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0103] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.
[0104] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0105] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.
[0106] Although preferred embodiments have been described in this specification, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this specification.
[0107] Obviously, those skilled in the art can make various modifications and variations to this specification without departing from its spirit and scope. Therefore, if such modifications and variations fall within the scope of the claims and their equivalents, this specification is also intended to include such modifications and variations.
Claims
1. A method for identifying defects after pickling and brushing of a steel strip, characterized in that, The defect identification method after pickling and heavy brushing of the strip steel comprises the following steps: Based on the preset steel quality, the target steel grade corresponding to the to-be-detected strip steel is determined; After pickling and heavy brushing of the to-be-detected strip steel of the target steel grade, the target image of the surface of the to-be-detected strip steel after pickling and heavy brushing and the heavy brushing level corresponding to the target image are determined, wherein the target image is used to represent the image determined after the initial image of the surface of the to-be-detected strip steel is preprocessed via image preprocessing; Based on the heavy brushing level and a preset edge identification algorithm, the target image is subjected to defect identification to determine the to-be-identified defect image corresponding to the to-be-detected strip steel; Based on a preset defect type database and a preset defect identification algorithm, the to-be-identified defect image is screened to determine the target defect image corresponding to the to-be-detected strip steel; Based on the trained heavy brushing level defect identification model and a preset defect screening rule, the defects in the target defect image are screened to determine the key defects on the surface of the to-be-detected strip steel affecting the quality of the to-be-detected strip steel.
2. The strip steel defect recognition method after pickling and brushing according to claim 1, characterized by, The image preprocessing comprises grayscale processing, and the determination of the target image of the surface of the to-be-detected strip steel after pickling and heavy brushing and the heavy brushing level corresponding to the target image after pickling and heavy brushing of the to-be-detected strip steel of the target steel grade comprises the following steps: After pickling and heavy brushing of the to-be-detected strip steel of the target steel grade, the initial image of the surface of the to-be-detected strip steel after pickling and heavy brushing is determined; The initial image is subjected to grayscale processing, and the image points with abnormal grayscale values in the initial image are deleted to determine the target image of the surface of the to-be-detected strip steel after pickling and heavy brushing and the heavy brushing level corresponding to the target image, wherein the grayscale processing is calculated by using the median grayscale value of pixels.
3. The strip steel post pickling and brushing defect recognition method according to claim 1, characterized by, The defect identification of the target image based on the heavy brushing level and a preset edge identification algorithm to determine the to-be-identified defect image corresponding to the to-be-detected strip steel comprises the following steps: Based on the heavy brushing level corresponding to the target image and the target steel grade corresponding to the to-be-detected strip steel, the background texture parameter and the sensitivity parameter corresponding to the preset edge identification algorithm are determined; Based on the background texture parameter, the sensitivity parameter and the preset edge identification algorithm, the target image is subjected to defect identification to determine the to-be-identified defect image corresponding to the to-be-detected strip steel.
4. The strip steel post pickling and brushing defect recognition method according to claim 1, characterized by, The screening of the to-be-identified defect image based on the preset defect type database and the preset defect identification algorithm to determine the target defect image corresponding to the to-be-detected strip steel comprises the following steps: Based on the preset defect identification algorithm, the grayscale range of the defects corresponding to the to-be-detected strip steel is determined; Based on the preset defect type database and the grayscale range of the defects, the to-be-identified defect image is screened to determine the target defect image corresponding to the to-be-detected strip steel.
5. The strip steel pickling post-brushing defect recognition method according to claim 4, characterized by, The screening of the to-be-identified defect image based on the preset defect type database and the grayscale range of the defects to determine the target defect image corresponding to the to-be-detected strip steel comprises the following steps: Based on the preset defect type database, the candidate defect image with defects in the to-be-identified defect image is determined; Based on the gray scale range of the defect, the candidate defect image with the defect is screened to determine a target defect image corresponding to the to-be-detected strip steel.
6. The strip steel post pickling and brushing defect recognition method according to claim 1, characterized by, The trained heavy brushing level defect recognition model and a preset defect screening rule are used to screen the defects in the target defect image to determine key defects on the surface of the to-be-detected strip steel that affect the quality of the to-be-detected strip steel, including: The target defect image is input into the trained heavy brushing level defect recognition model to identify at least one brush mark defect in the target defect image and a heavy brushing level corresponding to each brush mark defect, and the brush mark defect with a light heavy brushing level is removed; The target defect image after the light brush mark defect is removed is screened based on the preset defect screening rule to determine the key defects on the surface of the to-be-detected strip steel that affect the quality of the to-be-detected strip steel.
7. The strip steel pickling post-brushing defect recognition method according to claim 6, characterized by, The trained heavy brushing level defect recognition model is determined in the following manner: Sample strip steel defect images of brush mark defects with different heavy brushing levels are obtained; The sample strip steel defect images are input into an initial heavy brushing level defect recognition model to train the initial heavy brushing level defect recognition model and determine the trained heavy brushing level defect recognition model.
8. A strip steel pickling post-brushing defect recognition device characterized by, The strip steel pickling and heavy brushing defect recognition device includes: A first determination module is configured to determine a target steel grade corresponding to a to-be-detected strip steel based on a preset steel grade quality. A second determination module is configured to determine a target image on the surface of the to-be-detected strip steel after pickling and heavy brushing of the to-be-detected strip steel of the target steel grade and a heavy brushing level corresponding to the target image, wherein the target image is an image determined after an initial image on the surface of the to-be-detected strip steel is preprocessed. A third determination module is configured to perform defect recognition on the target image based on the heavy brushing level and a preset edge recognition algorithm to determine a to-be-recognized defect image corresponding to the to-be-detected strip steel. A fourth determination module is configured to screen the to-be-recognized defect image based on a preset defect type database and a preset defect recognition algorithm to determine a target defect image corresponding to the to-be-detected strip steel. A fifth determination module is configured to screen defects in the target defect image based on the trained heavy brushing level defect recognition model and a preset defect screening rule to determine key defects on the surface of the to-be-detected strip steel that affect the quality of the to-be-detected strip steel.
9. An electronic device, comprising: The processor, the memory, and the bus, the memory stores machine readable instructions executable by the processor, when the electronic device is running, the processor and the memory communicate through the bus, the machine readable instructions are executed by the processor to perform the steps of the strip steel pickling and heavy brushing defect recognition method in any one of claims 1-7. The computer readable storage medium stores a computer program, the computer program is executed by the processor to perform the steps of the strip steel pickling and heavy brushing defect recognition method in any one of claims 1-7.
10. A computer-readable storage medium, characterized in that,