Non-contact hot-rolled steel coil diameter and tower defect on-line detection method and system

Abstract

The present application relates to a non-contact hot-rolled steel coil diameter and tower-shaped defect on-line detection method and system, an image acquisition device is installed on one side of the steel coil diameter, and both sides are lighted by using a light source, an industrial camera takes real-time images of the hot-rolled steel coil during the coiling process at a certain frame rate, the image data is transmitted to an image processing unit through a network cable, the real-time coil diameter and tower-shaped detection results of the steel coil are obtained through image processing and analysis, and are displayed on a computer control interface, and are fed back to the coiling tension control system to adjust the stable control of the coiling tension. The present application has the advantages that the non-contact real-time detection of the coil diameter and tower shape can be realized, the coil diameter detection result can also be fed back to the coiling tension control system for adjustment, the step of detecting the tower-shaped defect after the coiling is completed is saved, the calculation is simple, efficient and accurate, the visual image data is clear at a glance, and is convenient for future maintenance.

Description

Technical Field

[0001] This invention belongs to the field of sheet and strip coiling technology, and more specifically, relates to a non-contact online detection method and system for hot-rolled steel coil diameter and tower-shaped defects. Background Technology

[0002] In the production of sheet and strip rolling, coiling is the final and crucial step in hot rolling. Its proper functioning and the quality of the coiling directly affect the production and product quality of the hot continuous rolling mill. For the coiling machine's control system, whether it's speed control or indirect tension control, accurate real-time coil diameter values ​​are essential to ensure constant tension control during the coiling process and guarantee the quality of the steel coil.

[0003] Traditional indirect coil diameter calculation methods, such as speed calculation and coil count calculation, are significantly affected by slippage or uneven strip thickness control, impacting the accuracy of coil diameter calculations. Furthermore, tower-shaped defects are among the most common coil defects. Currently, these defects are primarily detected offline by manual visual inspection or using laser or camera detection instruments during coil transportation after uncoiling or after hot-rolled strip coiling. This process is wasteful of manpower and resources, inefficient, and disconnected from the coiling process, making it difficult to establish a systematic control and interaction mechanism.

[0004] In addition, some existing technologies use methods to detect the tower shape of steel coils by calculating the height difference between the highest height at the end of the steel coil and the normal height to obtain tower shape data. Such a solution first requires the normal height to be defined in advance, and secondly, it can only identify the tower shape, but cannot specifically distinguish whether the tower shape is an inner tower shape or an outer tower shape, nor can it identify the location of the tower shape defect between the layers of the steel coil. Summary of the Invention

[0005] To address the problems of traditional methods for measuring coil diameter and detecting pyramidal defects, the present invention aims to design an online detection method and system for hot-rolled steel coil diameter and pyramidal defects based on machine vision. This method enables online detection of coil diameter and pyramidal defects during the hot-rolled steel coil winding process in a non-contact manner. It is simple to calculate, highly accurate, low in cost, and easy to maintain.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] This invention provides a non-contact online detection method for diameter and conical defects in hot-rolled steel coils. An image acquisition device is installed on one radial side of the steel coil, and a light source is used for double-sided illumination. The method includes:

[0008] The image acquisition device is used to acquire image data of hot-rolled steel coils during the coiling process;

[0009] The hot-rolled steel coil image data is received via a communication network cable through the I / O interface;

[0010] By acquiring the ROI region, segmenting image features and background regions, and highlighting image edge features and filtering edge region noise through image grayscale transformation, contrast enhancement, and Canny edge filtering;

[0011] The Canny filter operator is used to extract the feature edges of the image, and then the shape or length feature filtering operator is used to filter out the upper and lower edges and end face edge features of the steel coil. Finally, the actual distance between the upper and lower edges of the steel coil is output through the system calibration results, which is the actual coil diameter value.

[0012] By analyzing the axial offset of the end face edge coordinates of each layer of strip steel during the coiling process, it is possible to identify whether it is a tower-shaped defect.

[0013] Furthermore, the range of the acquired steel coil image area includes the rectangular area enclosed by the upper and lower edges of the steel coil and the end face edge on one side.

[0014] Furthermore, by analyzing the axial offset of the end face edge coordinates of each layer of strip steel during the coiling process, it is possible to identify whether a tower-shaped defect exists, including:

[0015] By dynamically comparing the axial offset of the outermost strip end face edge contour with the edge contour of the upper or inner strip end face when the steel coil is rolled in real time, tower-shaped defects can be accurately identified.

[0016] Furthermore, the machine vision software used for system calibration, image processing, and analysis is the Halcon algorithm library.

[0017] Furthermore, it also includes: feeding back the roll diameter detection results to the winding machine tension control system to provide feedback adjustment for the stable control of the winding tension.

[0018] The present invention also provides a non-contact online detection system for diameter and tower-shaped defects of hot-rolled steel coils, the system comprising: an image acquisition unit and an image processing unit;

[0019] The image acquisition unit includes a CMOS industrial camera and a support frame for fixing the CMOS industrial camera; the support frame is fixed to the wall surrounding the winding machine or the metal structure of the equipment.

[0020] The image processing unit and the image acquisition unit communicate via an I / O interface. The image processing unit employs a non-contact online detection method for hot-rolled steel coil diameter and tower-shaped defects, as described above, to achieve online detection of hot-rolled steel coil diameter and tower shape.

[0021] Furthermore, the system also includes a computer control module for receiving the real-time coil diameter value and tower shape discrimination result of the steel coil obtained by the image processing unit;

[0022] The computer control module includes a data storage unit, a feedback unit, and an image acquisition system control unit. The data storage unit stores the measured coil diameter value, the coordinate values ​​of the coil end face edge, and the tower shape discrimination result at each moment. The feedback unit transmits the real-time coil diameter value to the winding tension control system for online adjustment of the winding tension stability. The image acquisition system control unit allows the user to control parameters such as the frame rate and gain of the image acquisition device according to the actual working conditions.

[0023] Furthermore, the system also includes a display module, used to receive real-time images of the steel coil acquired by the image acquisition unit, and simultaneously receive the coil diameter, end face coordinates, and tower shape discrimination results obtained after processing by the image processing unit, and display the original image data, coil diameter results, steel coil end face point data offset changes, and tower shape defect discrimination results.

[0024] Furthermore, the image acquisition unit also includes a camera cooling device, which is a water-cooled cylindrical device, in which the camera is arranged without affecting the lens field of view.

[0025] Furthermore, the image acquisition unit also includes: an LED strip light source; the LED strip light source illuminates the back to highlight the upper and lower edges of the steel coil, and simultaneously illuminates the front to highlight the dynamic changes in the edge of the steel coil end face.

[0026] Advantages and positive effects of the present invention:

[0027] (1) The online detection method and system for hot-rolled steel coil diameter and tower shape provided by this invention detects both the tower shape and coil diameter during the hot-rolled strip coiling process (when the strip is coiled into a coil on the coiler), and also includes the detection of the dynamic coil diameter during coiling. The purpose is to control the coil shape quality by adjusting the strip coiling tension in real time through the detection of the coil diameter, and to reflect the quality of the coil shape by the detection of the tower shape, and to control the tower shape defects caused by tension fluctuations. This solves the technical problems of offline detection schemes, such as wasted manpower and resources, low efficiency, and the disconnect between the detection and coiling processes, making it difficult to form a systematic control and interaction. In addition, this invention realizes non-contact online detection of coil diameter and tower shape through the two-dimensional detection principle of machine vision, avoiding the complex calculation process in traditional methods, reducing the cumulative error that may exist in each measurement step, and the device is simple, easy to maintain, low in cost, and highly accurate in detection.

[0028] (2) The present invention can adjust the strip coiling tension by real-time measurement of the dynamic coil diameter during the coiling process and feeding it back to the coiling tension control system, thereby ensuring that the strip is coiled with constant tension, improving the strip coiling yield, ensuring excellent coil quality, and feeding back the axial offset of each layer contour edge to the coiling control system in real time to adjust the tower shape caused by unstable tension, and the offset reflects the control effect, thus realizing intelligent coiling control.

[0029] (3) The present invention has a unique approach to detecting tower-shaped defects. It uses an online detection method to identify and record the coordinates of the end edge points of each layer of strip steel during the coiling process. By analyzing the axial offset of the end face edge point coordinates of each layer of strip steel during the coiling process, it detects whether it is a tower-shaped defect and related data. This includes real-time dynamic comparison of the axial offset of the end face edge contour of the outermost layer of strip steel with the end face edge contour of the layer above or inner layer of strip steel during the coiling process, which accurately identifies tower-shaped defects and greatly improves the accuracy of tower-shaped defect identification.

[0030] (4) This invention focuses on innovation in equipment and tower-shaped detection principles. Specifically, it uses a novel arrangement of industrial cameras and light sources to analyze the axial offset of the edge coordinates of each strip's end face during the coiling process, identifying whether it is a tower-shaped defect and related data. This includes: accurately identifying tower-shaped defects by real-time dynamic comparison of the axial offset between the outermost strip's end face edge contour and the edge contour of the previous or innermost strip's end face during coiling. Unlike existing tower-shaped detection methods that focus on algorithm research (i.e., using Sobel operators, convolutional neural networks, etc., for traditional surface defect detection, specifically acquiring an overall image of the coil's end face, performing image preprocessing through filtering and image segmentation, and then using a neural network to classify and identify defect features), this invention, through innovation in the tower-shaped detection principle, only requires two-dimensional measurement data from visual technology. Therefore, from the perspective of tower-shaped defect detection principles, this invention is simpler and more convenient. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 This is a schematic diagram illustrating the principle of online detection of hot-rolled steel coil diameter and tower shape based on machine vision in an embodiment of the present invention.

[0033] Figure 2This is a schematic diagram of the machine vision image acquisition system in an embodiment of the present invention;

[0034] Figure 3 This is a schematic diagram of the image acquisition position in an embodiment of the present invention;

[0035] Figure 4 This is a schematic diagram of the camera cooling device in an embodiment of the present invention;

[0036] Figure 5 This is a schematic diagram illustrating the principle of roll diameter measurement in an embodiment of the present invention;

[0037] Figure 6 This is a schematic diagram of the tower-shaped detection principle in an embodiment of the present invention;

[0038] Among them, 1-support frame; 2-strip light source; 3-steel coil; 4-strip light source; 5-support frame; 6-industrial lens; 7-CMOS industrial camera; 8-water outlet; 9-cooling device; 10-support frame; 11-water inlet; 12-support frame; 13-strip light source. Detailed Implementation

[0039] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0040] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0041] This invention provides a non-contact online detection method for coil diameter and pyramid shape defects during the hot-rolled steel coil winding process. It is used for online measurement of coil diameter and real-time detection of pyramid shape defects during the hot-rolled strip coiling process. The method includes an industrial camera, industrial lens, light source, support frame, cooling device, host computer, and various cables. The main scheme involves placing an image acquisition device, consisting of an industrial camera and lens, radially on the underground coiler drum. The device is fixed by the support frame to prevent significant vibration and displacement of the camera due to environmental influences, reducing external interference. The image distance is sufficient to ensure that the maximum diameter of the steel coil is within the lens's field of view. A strip-shaped single white light source is selected to illuminate the steel coil from both the back and front sides, highlighting the edge features of the coil. The light source is also fixed by the support frame. The industrial camera acquires images of the coiled shape in real time, which are transmitted to a computer via a communication cable. The image processing unit processes and analyzes the images to obtain the real-time detection results of the coil diameter and pyramid shape.

[0042] like Figure 1 As shown in this embodiment of the invention, a non-contact online detection method for coil diameter and pyramid shape defects during the coiling process of hot-rolled steel coils firstly involves an image acquisition unit capturing images of the hot-rolled steel coils during the coiling process at a certain frame rate. Then, the image data is transmitted to an image processing unit on a computer via a communication network cable. After processing and analysis, the detection results of coil diameter and pyramid shape are promptly sent to the computer control software. The detection results are then displayed in real time on the software interface, and the coil diameter measurement results are simultaneously transmitted to the coiling machine control system. The accurate coil diameter obtained through real-time measurement is used to provide feedback and adjust the stability of the tension control, forming a closed-loop control to ensure the coiling quality of the steel coils.

[0043] To make it easier to understand, each step is explained in detail below.

[0044] S1. Installation and fixing of the image acquisition unit.

[0045] The image acquisition unit specifically includes a CMOS industrial camera, an LED strip light source, and a fixing device. The CMOS industrial camera is fixed to the support frame via a threaded adapter plate, and the LED light source is also fixed to the support frame via a threaded connection. The support frame is fixed to the wall or metal structure of the equipment around the coiler. The specific fixing position must ensure that the maximum diameter of the steel coil after winding is within the field of view of the industrial lens. The fixing position of the strip light source for backlighting must ensure that the upper and lower horizontal edge features of the steel coil are highlighted. The position of the front-lit light source only needs to ensure that the edge features of the vertical side of the end face of the steel coil are highlighted.

[0046] like Figure 2The diagram shows the structure of a machine vision image acquisition system. The industrial lens 6 and industrial camera 7 constitute the core components for image acquisition. Their installation position ensures that when the steel coil 3 is rolled to its maximum diameter, the upper and lower edges and one end face of the steel coil remain within the camera's field of view. The bar light sources 4 and 13 are fixed to one side of the camera, positioned above and below it respectively, illuminating the steel coil 3 from the front at a certain angle to highlight the edge features of the steel coil's end face. The bar light source 2 is positioned opposite the camera, illuminating the steel coil from the back. The purpose of the bar light source 2's back illumination is to highlight the upper and lower edges of the steel coil; therefore, its size should always be larger than the maximum diameter of the steel coil. Supports 1, 5, 10, and 12 in the diagram respectively fix the bar light source 2, bar light source 4, industrial camera 7, and bar light source 13.

[0047] like Figure 3 The diagram shows the image acquisition positions. The industrial lens 6 and industrial camera 7 primarily capture the upper and lower edges and end face edges of the steel coil. Therefore, they are positioned coplanar with the end face of the steel coil 3, ensuring the end face is centered in the field of view and the upper and lower edges fall within it. The bar light sources 4 and 13 highlight the end face features and axial offset of the steel coil's edges. Therefore, they are positioned on the same side as the camera for frontal lighting, with the end face of the steel coil 3 centered on the light source. The bar light sources 4 and 13 are located on the upper and lower sides of the camera, illuminating the steel coil 3 at a certain angle to ensure uniform lighting and reduce grayscale differences. The bar light source 2 is intended to highlight the upper and lower edges of the steel coil. Therefore, it is positioned opposite the camera for back lighting, ensuring the light source size is larger than the upper and lower edges of the steel coil 3.

[0048] like Figure 4 The diagram shows a camera cooling device. The industrial lens 6 and the industrial camera 7 are installed in the cylindrical cooling device 9, ensuring that the field of view of the industrial lens is not affected. This cooling device uses water cooling. Cold water is introduced into the cylindrical cooling device through the inlet 11 and flows out through the outlet 8, repeatedly circulating to achieve the purpose of cooling and reducing temperature, protecting the camera and lens hardware from the impact of high temperature of the steel coil on their working performance and lifespan.

[0049] S2, Data transmission.

[0050] After acquiring an image, the CMOS industrial camera transmits the image data to the image processing unit via a gigabit communication network cable through the I / O interface; the LED light source is controlled by the PC unit to switch on and off and adjust its illumination brightness via its own power line.

[0051] S3, System Calibration.

[0052] After the camera is fixed in place, use the Halcon calibration assistant to perform calibration. First, load the description file of the calibration board, input the width and height of a single pixel of the industrial camera and the focal length of the lens. Then, place the calibration board near the center area of ​​the roll and adjust the lens to make the calibration board focus accurately. Take 15-20 images of the calibration board in various positions and angles within the lens's field of view. The calibration board should occupy 1 / 3 to 1 / 2 of the image field of view. After completing the calibration, export the code.

[0053] S4. Image Processing and Analysis.

[0054] This invention utilizes Halcon machine vision software to process and analyze images. First, a mapping map is generated, and the `map_image` operator is used to correct radial distortion. Then, the image is converted to grayscale using `rgb1_to_gray` to improve feature contrast. Next, `gen_rectangle1` is used to obtain the region of interest (ROI), specifically the rectangular area containing the edge of one side face and the top and bottom edges of the steel coil, reducing environmental interference in the overall image. The `reduce_domain` operator is used to segment the image background and the ROI, retaining only the ROI. Edge enhancement and filtering are performed using the `emphasize` and `mean_image` operators to highlight edge features and filter noise. Next, the `edges_sub_pix` operator is used to extract sub-pixel edges (XLD) from the grayscale image. Finally, various operators such as `select_shape_xld` are used to filter the desired outer contour edges of the steel coil, namely the side face edge and the top and bottom contour edges.

[0055] S5. Real-time detection of steel coil diameter and tower shape.

[0056] The desired steel coil diameter can be converted into the distance between the top and bottom edges of the rectangle representing the steel coil in the image. If the industrial lens is not a telecentric lens, this distance can be converted to the steel coil diameter using simple mathematical relationships such as similar triangles. More specifically, according to the camera imaging principle, the lens images the light reflected from the object onto the image plane. Since light travels in straight lines, the size of the steel coil is much larger than the lens size, and the steel coil is a cylindrical object. Therefore, the distance d between the top and bottom edges of the steel coil identified in the image is not the actual steel coil diameter D, but rather the distance between the outline edges of the steel coil where it is tangent to the lens. Therefore, to avoid measurement errors, the distance d between the top and bottom edges of the steel coil in the image needs to be converted to the actual steel coil diameter D. The specific steps are: before operation, use a laser rangefinder or similar device to measure the distance L from the camera to the center of the roll. The conversion relationship between the steel coil diameter D and the "false diameter d" can be obtained using the principle of similar triangles, with the mathematical relationship L / (D / 2)=D / √(D^2-d^2), thus obtaining the actual steel coil diameter D. In Halcon, the required distance between the top and bottom edges of the steel coil can be obtained by measuring the two-dimensional dimensions of the image, in pixels. Then, by converting the image coordinate system to the world coordinate system through the calibration results, the true diameter of the steel coil can be obtained, in mm.

[0057] This invention employs a clever method for detecting tower-shaped defects in steel coils. After identifying the edge contour of the coil's end face, relevant operators are used to output the coordinate files of each point on the edge contour. The tower shape can be categorized into three types: convex, concave, and interlayered. The extreme cases of convex and concave tower shapes are illustrated here to explain the detection principle. If an convex tower defect occurs, the coordinates of the points on the coil's end face shift along the axial direction as the coil diameter increases (unlike normal coiling where the contour coordinates do not shift along the axial direction). In this case, the contour edges at the end face will be displayed layer by layer on the image. If a concave tower defect occurs, the coordinates of the points on the coil's end face shift along the axial direction as the coil diameter increases, and each layer of coiled steel in the image will cover the end edge of the previous layer, ensuring that only one end edge contour always exists in the image. By investigating a large number of steel coils with a pyramidal shape, an axial offset standard for judging pyramidal defects was determined. It is stipulated that if the offset of the contour coordinates of the end point of the steel coil exceeds this standard value, it is judged to be pyramidal.

[0058] like Figure 5 The diagram shows the principle of coil diameter measurement. The industrial lens 6 and industrial camera 7 acquire images of the steel coil 3 in real time. The acquired image area is shown in the right figure, ensuring that one end face and the top and bottom edges of the steel coil 3 are within the field of view. At this point, the measurement of the steel coil diameter is transformed into the measurement of the distance between the top and bottom edges of the steel coil. Through two-dimensional measurement based on machine vision, the true distance between the top and bottom edges in the steel coil image is obtained, thus efficiently and conveniently acquiring the true coil diameter value.

[0059] like Figure 6 The diagram illustrates the principle of tower-shaped defect detection. Industrial lens 6 and industrial camera 7 acquire images of steel coil 3 in real time. The acquired image area is shown in the right figure, primarily ensuring that one end face and the top and bottom edges of steel coil 3 are within the field of view. After image processing to filter out accurate feature edges of the steel coil end face, relevant operators are used to output a data file of the coordinates of each point on the edge contour of the steel coil end face. The specific tower-shaped defect detection principle is as follows: Tower shapes are classified into three types: convex, concave, and interlayered. The extreme cases of convex and concave tower shapes are listed here to illustrate the tower-shaped defect detection principle. If an convex tower-shaped defect occurs, such as... Figure 6 As shown in Figure ①, as the diameter of the steel coil increases, the coordinates of the points on the end face of the steel coil shift along the axial direction (during normal coiling, the contour coordinates at the end face do not shift along the axial direction). At this time, the contour edges at the end face will be displayed layer by layer on the image; if an inwardly concave, tower-shaped defect occurs, such as... Figure 6 As shown in Figure ②, due to the increasing diameter of the steel coil, the coordinates of the points on the coil's end face shift along the axial direction. Furthermore, each layer of coiled steel in the image covers the end edge of the layer above, resulting in only one end edge contour always appearing in the image. By investigating a large number of coils exhibiting a pyramidal shape, a standard for identifying pyramidal defects was determined. It was stipulated that if the offset of the coil's end point contour coordinates exceeds this standard value, it is considered pyramidal. By comparing the image-processed output data of the coil's end face edge contour, the axial offset of the coil's end edge can be obtained, enabling efficient identification of pyramidal defects.

[0060] The online detection method for the diameter and pyramid shape of hot-rolled steel coils provided by this invention achieves non-contact online detection of the diameter and pyramid shape through the two-dimensional detection principle of machine vision. This avoids the complex calculation process in traditional methods, reduces the cumulative error that may exist in each measurement step, and the device is simple, easy to maintain, low in cost, and highly efficient and accurate in detection.

[0061] Corresponding to the above-mentioned non-contact online detection method for hot-rolled steel coil diameter and tower-shaped defects, the present invention also provides a non-contact online detection system for hot-rolled steel coil diameter and tower-shaped defects, the system comprising: an image acquisition unit and an image processing unit;

[0062] The image acquisition unit includes: a CMOS industrial camera and a support frame for fixing the CMOS industrial camera; the support frame is fixed to the wall or metal structure of the equipment surrounding the winding machine. It also includes: a camera cooling device, which is a water-cooled cylindrical device, allowing the camera to be placed inside the cylinder without affecting the lens's field of view. Furthermore, it includes: an LED strip light source; the LED strip light source provides backlighting to highlight the upper and lower edges of the steel coil, and simultaneously provides front lighting to highlight the dynamic changes in the edge of the steel coil's end face.

[0063] The image processing unit and the image acquisition unit communicate data through an I / O interface. The image processing unit uses the above-mentioned non-contact online detection method for hot-rolled steel coil diameter and tower-shaped defects to realize online detection of hot-rolled steel coil diameter and tower shape.

[0064] In another embodiment, the system further includes a computer control module for receiving the real-time coil diameter value and tower shape discrimination result obtained by the image processing unit.

[0065] The computer control module includes a data storage unit, a feedback unit, and an image acquisition system control unit. The data storage unit stores the measured coil diameter value, the coordinate values ​​of the coil end face edge, and the tower shape discrimination results at each moment. The feedback unit transmits the real-time coil diameter value to the winding tension control system for online adjustment of the winding tension stability. The image acquisition system control unit allows the user to control parameters such as the frame rate and gain of the image acquisition device according to the actual working conditions.

[0066] In another embodiment, the system further includes a display module, which is used to receive real-time images of the steel coil acquired by the image acquisition unit, and simultaneously receive the coil diameter, end face coordinates and tower shape discrimination results acquired by the image processing unit, and display the original image data, coil diameter results, steel coil end face point data offset changes and tower shape defect discrimination results.

[0067] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention 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 or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A non-contact online detection method for diameter and pyramidal defects in hot-rolled steel coils, characterized in that, An image acquisition device is installed on one radial side of the steel coil, and a light source is used for double-sided illumination; the method includes: The image acquisition device is used to acquire image data of hot-rolled steel coils during the coiling process; The hot-rolled steel coil image data is received via a communication network cable through the I / O interface; By acquiring the ROI region, segmenting image features and background regions, and highlighting image edge features and filtering edge region noise through image grayscale transformation, contrast enhancement, and Canny edge filtering; The Canny filter operator is used to extract the feature edges of the image, and then the shape or length feature filtering operator is used to filter out the upper and lower edges and end face edge features of the steel coil. Finally, the actual distance between the upper and lower edges of the steel coil is output through the system calibration results, which is the actual coil diameter value. By analyzing the axial offset of the edge coordinates of each layer of strip steel during the coiling process, it is possible to identify whether a tower-shaped defect exists. The tower-shaped defect includes three types: outward convexity, inward concavity, and interlayer overlap. If an outward convex tower-shaped defect occurs, the coordinates of the points on the end face of the coil shift along the axial direction as the coil diameter increases, resulting in the outline edges of the end face appearing layer by layer in the image. If an inward concavity tower-shaped defect occurs, the coordinates of the points on the end face of the coil shift along the axial direction as the coil diameter increases, and each layer of strip steel coiling in the image covers the end edge of the previous layer, ensuring that only one end edge outline always exists in the image. Through investigation of a large number of coils exhibiting tower-shaped defects, an axial offset standard for determining tower-shaped defects is established. If the offset of the end point outline coordinates of the coil exceeds this standard value, it is determined to be a tower-shaped defect. The dynamic roll diameter is measured in real time during the winding process and fed back to the winding tension control system. The system provides feedback adjustment for the stable control of the winding tension, thereby regulating the strip winding tension and ensuring that the strip is wound with constant tension. At the same time, the axial offset of the contour edge of each layer is fed back to the winding control system in real time to regulate the tower shape caused by unstable tension. The offset reflects the control effect, thus realizing intelligent winding control.

2. The non-contact online detection method for diameter and tower-shaped defects of hot-rolled steel coils according to claim 1, characterized in that, The area of ​​the acquired steel coil image includes the rectangular region enclosed by the upper and lower edges of the steel coil and the end face edge on one side.

3. The non-contact online detection method for diameter and tower-shaped defects of hot-rolled steel coils according to claim 1, characterized in that, By analyzing the axial offset of the end face edge coordinates of each layer of strip steel during the coiling process, it is possible to identify whether a tower-shaped defect exists, including: By dynamically comparing the axial offset of the outermost strip end face edge contour with the edge contour of the upper or inner strip end face when the steel coil is rolled in real time, tower-shaped defects can be accurately identified.

4. The non-contact online detection method for diameter and tower-shaped defects of hot-rolled steel coils according to claim 1, characterized in that, The machine vision software used for system calibration, image processing, and analysis is the Halcon algorithm library.

5. A non-contact online detection system for diameter and pyramidal defects in hot-rolled steel coils, characterized in that, The system includes: an image acquisition unit and an image processing unit; The image acquisition unit includes: an LED bar light source, a CMOS industrial camera, and a support frame for fixing the CMOS industrial camera; the support frame is fixed to the wall or metal structure of the equipment around the coiler; the CMOS industrial camera is located on one radial side of the steel coil; there are three LED bar light sources, one of which is located on the opposite side of the CMOS industrial camera and is used for backlighting to highlight the upper and lower edge features of the steel coil; the other two LED bar light sources are located on the same side as the CMOS industrial camera and are respectively located on the upper and lower sides of the CMOS industrial camera, and are used for front lighting to highlight the edge features of the steel coil end face; The image processing unit and the image acquisition unit communicate data through an I / O interface. The image processing unit adopts a non-contact online detection method for hot-rolled steel coil diameter and tower-shaped defects as described in any one of claims 1 to 4 to realize online detection of hot-rolled steel coil diameter and tower shape. The system also includes a computer control module for receiving the real-time coil diameter value and tower shape discrimination result of the steel coil obtained by the image processing unit; The computer control module includes: a data storage unit, a feedback unit, and an image acquisition system control unit; The data storage unit is used to store the measured coil diameter value, the coordinate value of the end face edge of the steel coil, and the tower shape discrimination result at each moment; The feedback unit is used to transmit the real-time coil diameter value to the coiling tension control system for online regulation of the coiling tension stability; to measure the dynamic coil diameter in real time during the coiling process and feed it back to the coiling tension control system for feedback adjustment of the coiling tension stability control, thereby regulating the strip coiling tension to ensure that the strip is coiled with constant tension. At the same time, the axial offset of the contour edge of each layer is fed back to the coiling control system in real time to regulate the tower shape caused by unstable tension, and the offset reflects the regulation effect, thereby realizing intelligent coiling control. The image acquisition system control unit is used by the user to control the frame rate and gain of the image acquisition device according to the actual working conditions.

6. The non-contact online detection system for diameter and tower-shaped defects of hot-rolled steel coils according to claim 5, characterized in that, The system further includes a display module, which receives real-time images of the steel coil acquired by the image acquisition unit, and simultaneously receives the coil diameter, end face coordinates, and tower shape discrimination results obtained after processing by the image processing unit, and displays the original image data, coil diameter results, steel coil end face point data offset changes, and tower shape defect discrimination results.

7. The non-contact online detection system for diameter and tower-shaped defects of hot-rolled steel coils according to claim 5, characterized in that, The image acquisition unit also includes a camera cooling device, which is a water-cooled cylindrical device, in which the camera is arranged without affecting the lens field of view.

Images