An intelligent coiling method for a finishing heavy coiler

Abstract

The present application relates to the field of intelligent metallurgical technology, in particular to a finishing heavy coiling unit intelligent coiling method, comprising the following steps: coiling the crown block to the first positioning saddle; positioning the outer ring head of the steel coil at the ground roller saddle, identifying the information, and detecting and processing the inner ring appearance, and using the uncoiling robot to perform uncoiling operation on the steel coil; moving the steel coil to the second positioning saddle by the transport trolley, automatically measuring the coil diameter of the steel coil, intelligently re-measuring the centering, and automatically coiling. The finishing heavy coiling unit intelligent coiling method shortens the single-coil operation time of traditional manual coiling, reduces the coil-changing downtime, and effectively improves the overall productivity of the finishing heavy coiling unit through the automatic connection process combined with visual algorithm processing technology; through the synergistic effect of high-definition industrial cameras, industrial cameras, 3D laser line scanning cameras, and laser sensors, the outer ring head positioning of the steel coil, the inner ring appearance detection, the coil diameter measurement, and the centering re-checking are controlled to improve the centering accuracy.

Description

Technical Field

[0001] This invention relates to the field of intelligent metallurgical technology, specifically to an intelligent winding method for a finishing and rewinding unit. Background Technology

[0002] Finishing and recoiling mills are crucial for improving steel coil quality and meeting the personalized needs of downstream users in steel production. The inlet coiling process, as the first core step in the mill's operation, directly impacts the production quality and capacity of the entire subsequent process through its efficiency, precision, and stability. Intelligent improvements to the inlet coiling process essentially aim to address the pain points of traditional manual or semi-automated coiling methods, adapting to the high-efficiency, high-quality, low-consumption, and safe requirements of modern steel production.

[0003] Currently, traditional coiling mostly relies on manual operation. Manually adjusting the alignment accuracy between the steel coil and the uncoiler requires repeated trial and error, and the coiling time for a single coil can be as long as 5 to 10 minutes. Moreover, manual visual adjustment of the coil may cause the steel coil to collide and rub against the equipment, resulting in surface defects, and may even lead to crushing or squeezing accidents.

[0004] By adopting an intelligent system that uses laser ranging and visual positioning technology, the centering accuracy can be controlled within ±1mm, and the winding time can be shortened to 1-2 minutes. This significantly reduces downtime for changing rolls and lowers the potential dangers caused by manual operation. Therefore, the intelligent improvement of the inlet winding is an inevitable requirement for finishing and rewinding units to adapt to efficient production, high-quality control, safe production and intelligent transformation. It is also an important manifestation of the steel industry's upgrade from scale expansion to quality and efficiency. Summary of the Invention

[0005] (a) Technical problems to be solved To address the shortcomings of existing technologies, this invention provides an intelligent winding method for finishing and rewinding units. This method has the advantage of achieving intelligent winding based on laser ranging and visual positioning technology, and solves the problems of high mechanical injury risk, low visual accuracy, and poor stability of traditional manual or semi-automatic winding methods.

[0006] (II) Technical Solution To achieve the aforementioned goal of intelligent winding based on laser ranging and visual positioning technology, this invention provides the following technical solution: a finishing and rewinding unit, comprising a ground track installed on the ground, a 3D laser line scanning camera and an industrial camera arranged on the right side of the ground track, a transport trolley arranged on the top of the ground track, a de-bundling robot movably installed on the front side of the ground track, and a saddle assembly installed on the rear side of the ground track. The saddle assembly includes a first positioning saddle, a ground roller saddle, and a second positioning saddle arranged sequentially from right to left. A steel coil is placed on top of the first positioning saddle, a high-definition industrial camera is installed on the rear side of the ground roller saddle, a mandrel is arranged on the left side of the ground track, an automatic diameter measuring device is installed on the left side of the second positioning saddle, a suspension track is installed directly above the saddle assembly, and an overhead crane is suspended on the suspension track.

[0007] A method for intelligent winding of a finishing and rewinding unit includes the following steps: S1. Lower the overhead crane into the first positioning saddle; S2. The steel coil is positioned at the ground roller saddle, the outer ring is identified, the information is identified and the inner ring is inspected, and the steel coil is unbundled using a debundling robot. S3. The steel coil is transferred to the second positioning saddle by a transport trolley, the diameter of the steel coil is automatically measured, intelligent re-measurement is performed after centering, and the coil is automatically loaded.

[0008] Preferably, in S2, a ground roller is arranged on the top of the ground roller saddle, and a motor is installed and connected to one end of the ground roller. The ground roller is rotated by the motor. When the steel coil reaches the top of the ground roller saddle, the L1 signal triggers the start of the head positioning detection and label identification processing of the outer ring of the steel coil. The high-definition industrial camera takes pictures continuously during the rotation of the steel coil and uses LED light source to supplement the lighting according to the on-site environment. The high-definition industrial camera transmits the pictures to the upper winding control system server via gigabit Ethernet.

[0009] Preferably, the coiling control system uses a visual detection algorithm to perform lead positioning detection on the outer ring of the steel coil. When the lead reaches the appropriate position, it controls the ground roller to stop rotating to ensure that the lead of the outer ring of the steel coil stays in the designated position. During the lead detection process, the coiling control system uses a target positioning algorithm to perform label target detection on the outer surface of the steel coil. After extracting the label area, it uses a visual algorithm to perform label content recognition. The label content is the coil number sequence of the steel coil.

[0010] Preferably, in S2, the steel coil specification information and trigger signal are sent to the machine vision system via L2 to start the detection program of the inner ring morphology of the steel coil. The system controls the industrial camera on the end face of the steel coil to capture high-definition images of the inner ring area of ​​the steel coil, and the industrial camera transmits the images to the system server via industrial Ethernet.

[0011] Preferably, the system server detects the inner ring shape of the steel coil and its deviation from the center of the coil using a visual algorithm. If the deviation exceeds the threshold or other abnormalities occur that prevent the coil from being wound, an abnormal signal is transmitted to the LI system to achieve interlock control and issue a voice alarm. If the offset does not exceed the threshold, proceed to the next step and use the unbundling robot to remove the outer strapping of the steel coil.

[0012] Preferably, in S3, three laser sensors are used to measure the diameter of the steel coil. After the measurement is completed, the height of the steel coil is adjusted according to the coordinate position set by the uncoiler drum for centering. After centering is completed, the L1 dry contact signal triggers and starts the upper coil height verification and detection process. The upper coil control system controls the 3D laser line scanning camera to scan the surface of the steel coil.

[0013] Preferably, the system uses 3D reconstruction technology to redraw the 3D shape of the cross-section of the steel coil surface, and compares the calculated center height of the steel coil with the preset center height of the uncoiler drum in the system to identify the height deviation. If the deviation exceeds the set threshold, an abnormal signal is sent to the LI system to achieve a chain stop of the winding action; if the deviation does not exceed the set threshold, automatic winding is performed.

[0014] (III) Beneficial Effects Compared with the prior art, the present invention provides an intelligent winding method for finishing and rewinding units, which has the following beneficial effects: 1. The intelligent winding method of this finishing and rewinding unit, through the automated connection process between the overhead crane, transport trolley and each saddle, combined with L1 signal triggering, gigabit Ethernet data transmission and visual algorithm processing technology, shortens the single-roll operation time of traditional manual winding, significantly reduces the downtime for changing rolls, and the entire winding process does not require manual intervention and repeated trial and error, realizing continuous operation, effectively improving the overall capacity of the finishing and rewinding unit, and adapting to the high-efficiency operation requirements of modern steel production.

[0015] 2. The intelligent winding method of this finishing and rewinding unit, through the synergistic effect of high-definition industrial cameras, industrial cameras, 3D laser line scanning cameras and three laser sensors, uses visual inspection algorithms, target positioning algorithms and three-dimensional reconstruction technology to accurately control the positioning of the outer ring of the steel coil, the detection of the inner ring shape, the measurement of the coil diameter and the centering verification, thereby controlling the centering accuracy and avoiding problems such as collision and friction between the steel coil and the equipment caused by traditional manual visual adjustment, thus reducing the generation of surface defects of the steel coil.

[0016] 3. The intelligent winding method of this finishing and rewinding unit is based on unmanned operation of the winding process. It eliminates the need for repeated manual adjustments and visual inspections, reducing the risk of injury such as crushing and squeezing caused by traditional manual close-range participation in steel coil adjustment and positioning. Through the interlocking control and voice alarm mechanism of the L1 and L2 systems, the operation can be automatically stopped when there is a deviation from the threshold or abnormal situation, further ensuring production safety. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall structure of the finishing and rewinding unit of the present invention; Figure 2 This is a flowchart of the intelligent winding method for the rewinding unit of the present invention.

[0018] In the diagram: 1. Ground track; 2. 3D laser line scan camera; 3. Industrial camera; 4. Transport trolley; 5. Unbundling robot; 6. Saddle assembly; 601. First positioning saddle; 602. Ground roller saddle; 603. Second positioning saddle; 7. Steel coil; 8. High-definition industrial camera; 9. Mandrel; 10. Automatic diameter measuring device; 11. Suspension track; 12. Overhead crane. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments and accompanying drawings. 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 are within the scope of protection of the present invention. Example 1

[0020] In this embodiment, a steel coil 7 with an inner diameter of 508mm and a coil diameter of 1750mm is selected. The intelligent winding method of this finishing and rewinding unit achieves the winding steps of the steel coil 7 as follows: S1. After the steel coil 7 is placed on the first positioning saddle 601 by the overhead crane 12, it is transported to the ground roller saddle 602 by the transport trolley 4.

[0021] S2. When the steel coil 7 reaches the ground roller saddle 602, the L1 signal triggers the start of the outer ring lead positioning detection and tag recognition of the steel coil 7. The high-definition industrial camera 8 takes pictures continuously during the rotation of the steel coil. At the same time, LED light source is used to supplement the light according to the on-site conditions. Then the pictures are transmitted to the server via gigabit Ethernet. The system completes the outer ring lead positioning detection of the steel coil through a visual detection algorithm. When the lead reaches the appropriate position, the ground roller is stopped by the command to ensure that the outer ring lead stays in the designated position. During the lead inspection process, the system uses a target localization algorithm to detect the label on the outer surface of the steel coil. After extracting the label area, a vision algorithm is used to identify the label content, mainly the coil number. The steel coil specification information and trigger signal are sent to the machine vision system via L2 to initiate the inner ring morphology inspection process of the steel coil. The industrial camera 3 takes pictures and LED supplementary light is used to transmit the images to the system and initiate image processing. The inner ring parameters are detected, and the inner diameter is found to be 628mm, the center offset is 4mm, the lateral offset is 4mm, and the longitudinal offset is 0mm. The detection results are compared with the preset feature values ​​of the mandrel to determine that the conditions for coiling are met. Then, the unbundling robot is used to remove the steel coil strapping.

[0022] S3. After removing the strapping, the trolley transports the coil to the second positioning saddle 603, triggering the automatic diameter measurement signal. The diameter of the steel coil is measured to be 1751.5mm using three laser photoelectric switches. After the measurement is completed, the height of the steel coil is adjusted according to the coordinate position set by the uncoiler drum for centering. After centering is completed, the L1 dry contact signal triggers the start of the upper coil height verification and inspection process. The system controls the 3D laser line scanning camera 2 to scan the surface of the steel coil 7. The 3D reconstruction technology is used to redraw the 3D shape of the cross section of the steel coil surface. The center height of the steel coil is obtained through mathematical calculation. The height is compared with the preset center height of the uncoiler drum in the system. The calculated height deviation is 0.12mm. The deviation does not exceed the set threshold, so the automatic winding program is started. Example 2

[0023] In this embodiment, the high-definition industrial camera 8 continuously takes pictures during the rotation of the steel coil, and simultaneously uses LED light source to supplement lighting according to the site conditions. Then, the pictures are transmitted to the server via gigabit Ethernet. The system completes the positioning detection of the outer ring of the steel coil through a visual inspection algorithm. The specific steps are as follows: 1. Image Acquisition Trigger: When the steel coil reaches the ground roller saddle, the L1 signal triggers the high-definition industrial camera to start acquisition. The camera resolution is set to 1920×1080, the frame rate is 25fps, and the LED light source is used to supplement the light with a brightness of 500 lux to ensure uniform image brightness. The acquired image is transmitted to the server in real time via gigabit Ethernet and stored as a grayscale image I(x,y), where x is the horizontal pixel coordinate, x∈[0,1919]; y is the vertical pixel coordinate, y∈[0,1079].

[0024] 2. Image Preprocessing: First, Gaussian filtering is used to remove salt-and-pepper noise from the industrial environment. Then, histogram equalization is used to enhance image contrast and eliminate the effects of uneven lighting. Gaussian filtering is used to smooth images and suppress noise; its formula is: , in For Gaussian kernel function, = 1.5 represents the filtering standard deviation, which adapts to the noise intensity of industrial scenarios; is the filtered image; represents the convolution operation; Histogram equalization enhances edge features by adjusting the pixel gray-level distribution, and its formula is: where is the number of pixels with gray level i, N = 1920×1080 = 2073600 represents the total number of pixels in the image; CDF(k) is the cumulative distribution function of gray level k; r is the original pixel gray value and r ∈ [0, 255]; s is the gray value of the pixel after equalization.

[0025] 3. Edge detection: The Canny algorithm is used to detect the edges on the surface of the steel coil, focusing on capturing the discontinuous edge features at the leading end. The leading end is the starting end of the outer circle of the steel coil, and the edge shows a sudden change characteristic. The steps of Canny edge detection are as follows: 1) Gradient calculation: The Sobel operator is used to calculate the gradients in the horizontal and vertical directions: , , , where is the horizontal Sobel operator ([ [1, 0, -1], [2, 0, -2], [1, 0, -1] ]), is the vertical Sobel operator ([ [1, 2, 1], [0, 0, 0], [-1, -2, -1] ]); is the gradient amplitude; is the gradient direction.

[0026] 2) Non-maximum suppression: Retain the local maximum points in the gradient direction and eliminate non-edge points.

[0027] 3) Double-threshold screening: Set the low threshold T1 = 50 and the high threshold T2 = 150. Retain the strong edges where G(x, y) ≥ T2 and the weak edges where T1 ≤ G(x, y) < T2 and are connected to the strong edges. Finally, obtain the edge image E(x, y), where E(x, y) = 1 is an edge point and E(x, y) = 0 is a non-edge point.

[0028] 4. Leading-end feature extraction: Detect the continuous edges of the steel coil circumference through the Hough transform, and screen out the areas where the edge gap exceeds the threshold, which are the candidate positions of the leading end. At this time, the outer circle of the steel coil is approximately an arc in the image. Detect the continuous edges through the Hough circle transform, and the formula is: where (a, b) is the center coordinate and r is the radius of the circle; By statistically analyzing the voting peak in the parameter space of (a, b, r), the fitted circle of the steel coil peripheral edge is obtained.

[0029] 5. Positioning Judgment and Feedback: Calculate the deviation between the candidate position and the preset baseline. The preset baseline is the vertical centerline of the steel coil center, corresponding to image coordinates x=960. If the deviation is within the allowable range, send a command to control the ground roller to stop rotating, completing the head positioning. The process for determining the positioning of the leader is as follows: Calculate the continuity of the fitted circle edge. If the gap d of a certain edge segment (d is the pixel distance between adjacent edge points) satisfies d>T, where T=50 pixels, corresponding to an actual physical distance of 5mm, then this gap is the leader position P(x0,y0). Calculate the deviation Δx=|x0-960| between P(x0,y0) and the baseline x=960. If Δx<5 pixels and corresponds to an actual distance of 0.5mm, then the positioning is successful. Example 3

[0030] In this embodiment, during the lead detection process, the system completes the target detection of the label on the outer surface of the steel coil through a target localization algorithm. After extracting the label area, it completes the label content recognition through a visual algorithm. The specific steps are as follows: 1. Image Preprocessing: The steel coil surface image I(x,y) captured by the high-definition industrial camera is first denoised using Gaussian filtering, then converted into a binary image using OTSU adaptive thresholding to highlight the label area. Industrial labels are mostly white with black text or black with white text. OTSU adaptive thresholding automatically determines the optimal threshold to separate the label from the background. The formula is:

[0031]

[0032] Where k is the candidate threshold. grayscale value and The percentage of pixels ≤k grayscale value >k pixel percentage; 、 These are the average grayscale values ​​of the two types of pixels, respectively. The optimal threshold is defined as B(x,y), where B(x,y) is a binary image, 1 represents the foreground label, and 0 represents the background.

[0033] 2. Label Target Localization: Using Haar features and the Adaboost target localization algorithm, based on a preset label template (rectangle size 100×50 pixels), candidate label regions in the image are detected. The steps of Haar feature and Adaboost target localization are as follows: 1) Haar Rectangular Feature Extraction: This extract describes the rectangular outline features of the label. The formula is:

[0034] Where p is the pixel value, and the black and white blocks are complementary regions of the rectangular features. The label outline is represented by calculating the difference between the sum of their pixels.

[0035] 2) Adaboost classifier training and testing:

[0036]

[0037] in Let be the classification error rate in the t-th iteration. Let be the weight of the t-th weak classifier, and T=1000 be the number of iterations; H(x) is the t-th weak classifier; H(x) is the strong classifier. When H(x)=1, it is determined to be a labeled region, and when H(x)=-1, it is a non-labeled region.

[0038] 3. Label Region Extraction: Morphological dilation (removing holes) and erosion (removing burrs) are performed on the candidate regions to obtain the complete label region R(x1,y1,x2,y2) ((x1,y1) is the coordinate of the top left corner, and (x2,y2) is the coordinate of the bottom right corner). The morphological operation is as follows: Expansion: B dilate =B⊕S (S is a 3×3 rectangular structural element, used to fill the holes in the label area). Corrosion: B erode =B dilate ⊖S (Remove edge burrs to obtain a neat label area).

[0039] 4. Character Segmentation: Project the label area R vertically to find the blank spaces between characters, and segment the volume number characters into 6 independent characters C1, C2, ..., C6 (C... j For the j-th character (j∈[1,6]), the binary image B of the label is processed by vertical projection and character segmentation. erode Perform vertical projection and calculate the sum of pixels in each column as follows:

[0040] in For the sum of pixels in the y-th column, when =0 represents the gap between characters, and the independent character C is obtained by segmenting based on this. j .

[0041] 5. Character Recognition: Based on a template matching algorithm, the segmented characters are matched against a standard digit template library to output the complete volume number. Template matching character recognition enables volume number extraction. A preset standard digit template library T0~T9 (corresponding to digits 0~9, size 20×30 pixels) is used, and the character matching degree is determined by a correlation coefficient. The formula is:

[0042] Volume Number =

[0043] in For characters With template The correlation coefficient, R∈[0,1], the closer R is to 1, the higher the matching degree; The result is the recognition result of the j-th character; the match is considered valid when R≥0.85. Example 4

[0044] In this embodiment, the system controls the 3D laser line scanning camera 2 to scan the surface of the steel coil 7, and uses 3D reconstruction technology to redraw the 3D topography of the cross-section of the steel coil surface. The center height of the steel coil is obtained through mathematical calculation. The specific steps are as follows: 1. Point cloud data acquisition: The L1 dry contact signal triggers the 3D laser line scanning camera to start scanning. The scanning frequency is 50Hz, with 1024 points per scan line and a scanning time of 0.5 seconds. The point cloud data P={(X i ,Y i Z i )|i=1,2,...,N} (N=50×1024=51200, (X i ,Y i Z i () represents the three-dimensional coordinates in the camera coordinate system.

[0045] 2. Point Cloud Denoising: Outliers (laser reflection interference points in industrial environments) are removed using statistical filtering to obtain a clean point cloud P'. Statistical filtering denoising removes outliers and retains valid points on the steel coil surface. The formula is: , , ,

[0046] Where k=10 is the number of nearest neighbors. Let be the average distance from the i-th point to its 10 nearest neighbors; The average distance is the mean. The standard deviation is denoted as ; for > Points that are outliers are removed.

[0047] 3. Coordinate Transformation: Transform the point cloud P' in the camera coordinate system to the world coordinate system, with the center of the second positioning saddle as the origin. The X-axis is parallel to the steel coil axis, the Y-axis is parallel to the horizontal direction of the ground, and the Z-axis is perpendicular to the ground and upwards. The coordinate transformation is achieved by obtaining the rotation matrix R and translation vector t through camera calibration. The transformation formula is: , , in( , , (X_w, Y_w, Z_w) are the coordinates in the camera coordinate system, and (X_w, Y_w, Z_w) are the coordinates in the world coordinate system; θ is the rotation angle of the steel coil (feedback in real time by the motor encoder, θ∈[0,2π), with an accuracy of 0.01rad); t is the translation vector of the camera mounting position (obtained through calibration).

[0048] 4. Extraction of point cloud of cross section: Filter the point cloud in the world coordinate system Y=Y0±0.01m (Y0=0m is the Y coordinate of the saddle center) to obtain the point cloud P of the axial cross section of the steel coil. slice ={(X i Z i )|i=1,2,...,M} (M is the number of point clouds in the cross section, M≈1000).

[0049] 5. 3D Topography Redrawing: Perform Delaunay triangulation on P_slice to generate a triangular mesh model, and redraw the 3D topography of the cross-section. (Delaunay triangulation (3D topography redrawing)) Based on the point cloud P slice ={(X i Z i According to the Delaunay empty circle property, triangulation is performed: the circumcircle of any triangle contains no other points. The triangulation formula is as follows: For three points A(X1,Z1), B(X2,Z2), and C(X3,Z3), the equation of its circumcircle is:

[0050] Where (a,b) is the center of the circumcircle and R is the radius; the circle parameters are obtained by solving the system of three coordinate equations. If other points are not inside the circle, then △ABC is a valid triangular mesh.

[0051] Topography Redrawing: Piece together all valid triangular meshes to generate a 3D topography model of the steel coil cross section.

[0052] 6. Center Height Calculation: By fitting the point cloud of the cross-section into a circular curve using the least squares method, the Z-coordinate of the circle center is determined, which is the center height H of the steel coil. The steps for least squares circle fitting and center height calculation are as follows: The cross-section of the steel coil is circular; fit the equation of the circle. The error function minimized by the least squares method is: Take the partial derivatives with respect to a, b, and r and set them to 0. The system of equations obtained is as follows: ,

[0053]

[0054] Center height of steel coil: The center height H=b is the Z coordinate of the center of the circle. At this time, the Z axis of the world coordinate system is perpendicular to the ground.

[0055] The beneficial effects of this invention are: the intelligent winding method of the finishing and rewinding unit, through the automated connection process between the overhead crane, the transport trolley and each saddle, combined with L1 signal triggering, gigabit Ethernet data transmission and visual algorithm processing technology, shortens the single-roll operation time of traditional manual winding, significantly reduces the downtime for changing rolls, and the entire winding process does not require manual intervention and repeated trial and error, realizing continuous operation, effectively improving the overall capacity of the finishing and rewinding unit, and adapting to the high-efficiency operation requirements of modern steel production; By leveraging the synergistic effects of high-definition industrial cameras, industrial cameras, 3D laser line scanning cameras, and three laser sensors, and employing visual inspection algorithms, target positioning algorithms, and 3D reconstruction technology, the system precisely controls the positioning of the outer ring of the steel coil, the detection of the inner ring morphology, the measurement of the coil diameter, and the centering verification process. This controls the centering accuracy and avoids problems such as collisions and friction between the steel coil and the equipment caused by traditional manual visual adjustments, thereby reducing the generation of surface defects on the steel coil. Based on the unmanned operation of the coiling process, there is no need for repeated manual adjustments and visual inspections, which reduces the risk of injury such as crushing and squeezing caused by traditional manual close-range participation in steel coil adjustment and positioning. Through the interlocking control and voice alarm mechanism of the L1 and L2 systems, the operation can be automatically stopped when there is a deviation from the threshold or abnormal situation, further ensuring production safety.

[0056] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A finishing and rewinding unit, comprising a ground track (1) installed on the ground, characterized in that, A 3D laser line scanning camera (2) and an industrial camera (3) are installed on the right side of the ground track (1). A transport trolley (4) is installed on the top of the ground track (1). A bundling robot (5) is movably installed on the front side of the ground track (1). A saddle group (6) is installed on the rear side of the ground track (1). The saddle group (6) includes a first positioning saddle (601), a ground roller saddle (602), and a second positioning saddle (603) arranged sequentially from right to left. A steel coil (7) is placed on the top of the first positioning saddle (601). A high-definition industrial camera (8) is installed on the rear side of the ground roller saddle (602). A mandrel (9) is installed on the left side of the ground track (1). An automatic diameter measuring device (10) is installed on the left side of the second positioning saddle (603). A suspension track (11) is installed directly above the saddle group (6). A crane (12) is hoisted on the suspension track (11).

2. An intelligent winding method for a finishing and rewinding unit, characterized in that, Includes the following steps: S1. Unwind the overhead crane (12) to the first positioning saddle (601). S2. The steel coil (7) is positioned at the ground roller saddle (602) for outer ring head positioning, information recognition and inner ring morphology detection, and the steel coil (7) is unbundled by the unbundling robot (5); S3. The steel coil (7) is transferred to the second positioning saddle (603) by the transport trolley (4), the diameter of the steel coil (7) is automatically measured, and intelligent re-measurement is performed after centering and the coil is automatically loaded.

3. The intelligent winding method for a finishing and rewinding unit according to claim 2, characterized in that, A ground roller is arranged on the top of the ground roller saddle (602) in S2. A motor is installed and connected to one end of the ground roller. The ground roller is rotated by the motor. When the steel coil (7) reaches the top of the ground roller saddle (602), the L1 signal triggers the start of the head positioning detection and label identification processing of the outer ring of the steel coil (7). The high-definition industrial camera (8) takes pictures continuously during the rotation of the steel coil (7) and uses LED light source to supplement the light according to the on-site environment. The high-definition industrial camera (8) transmits the pictures to the upper winding control system server through gigabit Ethernet.

4. The intelligent winding method for a finishing and rewinding unit according to claim 3, characterized in that, The winding control system completes the lead positioning detection of the outer ring of the steel coil (7) through a visual detection algorithm, and when the lead reaches the appropriate position, it controls the ground roller to stop rotating through an instruction to ensure that the lead of the outer ring of the steel coil (7) stays in the designated position. During the lead detection process, the winding control system completes the label target detection on the outer surface of the steel coil (7) through a target positioning algorithm. After extracting the label area, it completes the label content recognition through a visual algorithm. The label content is the roll number sequence of the steel coil (7).

5. The intelligent winding method for a finishing and rewinding unit according to claim 2, characterized in that, In S2, the steel coil specification information and trigger signal are sent to the machine vision system through L2 to start the detection program of the inner circle morphology of the steel coil (7). The system controls the industrial camera (3) on the end face of the steel coil (7) to capture high-definition images of the inner circle area of ​​the steel coil. The industrial camera (3) transmits the images to the system server through the industrial Ethernet.

6. The intelligent winding method for a finishing and rewinding unit according to claim 5, characterized in that, The system server detects the inner ring shape of the steel coil (7) through a visual algorithm and the deviation from the center of the coil. If the deviation exceeds the threshold or other abnormalities occur that prevent the coil from being wound, the server transmits an abnormal signal to the LI system to achieve interlock control and issues a voice alarm. If the offset does not exceed the threshold, proceed to the next step and use the unbundling robot (5) to remove the outer strapping of the steel coil (7).

7. The intelligent winding method for a finishing and rewinding unit according to claim 2, characterized in that, In S3, three laser sensors are used to measure the diameter of the steel coil (7). After the measurement is completed, the height of the steel coil (7) is adjusted according to the coordinate position set by the uncoiler drum for centering. After centering is completed, the L1 dry contact signal triggers and starts the upper winding height verification and detection process. The upper winding control system controls the 3D laser line scanning camera (2) to scan the surface of the steel coil (7).

8. The intelligent winding method for a finishing and rewinding unit according to claim 7, characterized in that, The system uses three-dimensional reconstruction technology to redraw the three-dimensional shape of the cross section of the steel coil (7), and compares the center height of the steel coil (7) with the preset center height of the uncoiling drum in the system to identify the height deviation. If the deviation exceeds the set threshold, an abnormal signal is sent to the LI system to realize the interlocking stop of the winding action; if the deviation does not exceed the set threshold, automatic winding is performed.

Images