A round steel end face quality evaluation and optimal identification object screening method and system

Abstract

The application discloses a round steel end face quality evaluation and optimal marking object screening method and system, and belongs to the technical field of metallurgical rolling mill machine vision detection. The application collects the end face image and three-dimensional point cloud data of a bundle of round steels through 3D vision, divides a single end face after pretreatment, and performs itemized quantitative scoring from five dimensions of flatness, smoothness, integrity, end face inclination and effective marking area. The comprehensive quality score is calculated through a weighting formula and is sorted in descending order. The optimal marking object is screened in combination with a qualified threshold, low-score whole bundle alarm is given, and a robot positioning and laser power adaptive adjustment are simultaneously linked. The application realizes automatic quantitative evaluation of the round steel end face quality, unifies the judgment standard, has high screening efficiency, is suitable for production line rhythm, can effectively avoid end face cracking and marking defects, and improves the finished product qualified rate and traceability reliability.

Description

Technical Field

[0001] This invention relates to the field of machine vision inspection and automated laser marking technology in metallurgical steel rolling, and in particular to a method and system for quality assessment of round steel end faces and selection of optimal marking objects. Background Technology

[0002] In metallurgical steel rolling production, finished round steel products need to be marked on their end faces to achieve product information traceability. The round steel specifications on-site cover Φ20-Φ50, and the steel types include carbon structural steel, high-carbon steel, and low-alloy steel, with a carbon content ranging from 0.05% to 1.2%. Traditionally, the marking method involves spraying paint and welding metal labels onto the ends of bundled round steel. Currently, this is gradually being upgraded to robots equipped with adjustable-power lasers for end face marking. However, due to the rolling and lifting processes, the end faces of multiple round steel bars within a bundle generally suffer from defects such as unevenness, oxidation peeling, end face defects, end face skew, and surface contamination. Usually, only one end face can be selected for marking.

[0003] The existing vision system can only locate the spatial position of round steel bars and lacks the function of quantitative inspection of end face quality, which has many defects: First, the system defaults to selecting round steel bars in the middle of the bundle, and cannot distinguish the quality of the end face. It is easy to select defective end faces, resulting in incomplete and insufficient depth of laser-engraved characters, and the product information cannot be effectively traced. Second, some production lines rely on manual visual selection of steel materials, which has low screening efficiency, inconsistent evaluation standards, and large differences in screening results between different shifts. It also cannot match the production cycle of ≤35s / piece. Third, the lack of quantitative rating data for end face quality means that the laser cannot adaptively match the output power. When marking inferior end faces, the characters are easy to become blurry. High-power marking can also easily cause cracks in the end face of the round steel bars, which damages the quality of the finished product. Summary of the Invention

[0004] To address the technical problems of inability to automatically quantify and rate the quality of bundled round steel end faces, inconsistent screening standards, and unstable labeling quality in existing technologies, this invention provides a method and system for evaluating the quality of round steel end faces and selecting the optimal labeling object. It abandons random steel selection and manual steel selection modes, realizes automatic sorting of the quality of round steel within a bundle and selection of the optimal end face, and simultaneously relies on the end face evaluation results to adaptively adjust the laser power, ensuring that the entire solution is compatible with the production line's production rhythm and improving the labeling quality and finished product qualification rate.

[0005] The technical solution adopted by the present invention, a method and system for quality assessment and optimal identification object selection of round steel end face, is as follows: A method for evaluating the quality of round steel end faces and selecting optimal identification objects includes the following steps: S1. Collect two-dimensional images and three-dimensional morphological data of the end face of bundled round steel bars, and simultaneously obtain round steel production parameters; S2. Preprocess the two-dimensional image to segment the independent end face regions of a single round steel bar, and extract the spatial position and morphological parameters of each end face. S3. Based on multiple end-face quality evaluation dimensions, quantitative scores are given to the end face of a single round steel bar to obtain individual quality scores. S4. Calculate the comprehensive quality score of the end face of a single round steel bar using a weighted fusion algorithm, and sort all round steel bars in the bundle in descending order of comprehensive quality score; S5. Select the first-ranked round steel as the candidate identification object, set the qualified score threshold, and if the comprehensive quality score of the candidate object reaches the threshold, it is determined to be a qualified identification object; otherwise, an abnormal alarm signal is output. S6. Send the location data and end-face quality parameters of the qualified marked object to the control system, and link the marking execution mechanism to complete the adaptive marking operation.

[0006] A further improvement of the technical solution of the present invention is that: in step S1, a 2D image and 3D topography data of the end face are acquired by a 3D vision camera, wherein the 3D topography data is 3D point cloud data of the end face; the round steel production parameters include the diameter of the round steel, the steel grade and the carbon content.

[0007] A further improvement of the technical solution of the present invention is that: in step S2, the preprocessing includes noise reduction, image enhancement and binarization; the single end face region is segmented by contour extraction combined with minimum circumcircle fitting algorithm, and the extracted morphological parameters include theoretical end face area, end face spatial coordinates and end face tilt angle.

[0008] A further improvement of the technical solution of the present invention is that: in step S3, the multiple end face quality evaluation dimensions include flatness, smoothness, integrity, end face tilt and effective imprinting area, and the score of each item ranges from 0 to 100 points.

[0009] A further improvement of the technical solution of this invention is as follows: the flatness is calculated based on the standard deviation of the end face elevation of the three-dimensional point cloud, and the smaller the standard deviation, the higher the flatness score; the smoothness is extracted by the gray-level co-occurrence matrix to extract image texture features, and the score is calculated based on the proportion of oxide scale, stains, and pitting; the integrity is converted into a score by the ratio of the actual effective end face area to the theoretical standard end face area; the end face tilt is calculated based on the angle between the end face plane normal and the center of the round steel, and the smaller the angle, the higher the score; the effective engraving area is scored based on the normalized area of ​​the safety engraving area at the center of the end face.

[0010] A further improvement to the technical solution of the present invention is that, in step S4, the calculation formula of the weighted fusion algorithm is: Q=0.3×A+0.25×B+0.2×B+0.15×D+0.1×E In the formula, Q is the overall quality score, A is the flatness score, B is the smoothness score, C is the integrity score, D is the end face tilt score, and E is the effective imprinted area score.

[0011] A further improvement of the technical solution of the present invention is that: in step S5, the pass score threshold is 85 points.

[0012] A further improvement of the technical solution of the present invention is that: in step S6, the marking execution mechanism includes a six-axis robot and an adjustable power laser; the control system sends position data to the six-axis robot to control its motion trajectory, and at the same time adaptively matches the output power of the laser according to the end face quality parameters and the round steel production parameters.

[0013] A system for evaluating the quality of round steel end faces and selecting optimal marking objects, used to implement the above method, includes a 3D vision acquisition module, an image processing and calculation module, a PLC control module, a six-axis robot, and an adjustable power laser marking mechanism; wherein, the signal output terminal of the 3D vision acquisition module is connected to the signal input terminal of the image processing and calculation module, the signal output terminal of the image processing and calculation module is connected to the signal input terminal of the PLC control module, and the PLC control module is connected to the six-axis robot and the adjustable power laser marking mechanism respectively.

[0014] The technological advancements achieved by this invention due to the adoption of the above technical solutions are as follows: This invention replaces the manual screening method, and the end face inspection and sorting of the whole bundle of round steel takes ≤5s, which can meet the production cycle requirement of ≤35s for the total marking time of a single piece, thus improving the efficiency of automated production.

[0015] This invention establishes a unified end-face quality evaluation standard through five quantitative scoring indicators, avoiding human judgment errors. The end-face marking characters after optimization are complete and clear, ensuring traceability of the entire product process.

[0016] This invention relies on the adaptive matching of laser output power based on end-face quality scoring to eliminate cracking defects on the end face of round steel caused by power mismatch, while avoiding the problem of shallow engraving on inferior end faces, thus improving the finished product qualification rate.

[0017] This invention provides an early warning for the output of unqualified end faces in bundles, which can help avoid batch marking defects and reduce production waste.

[0018] Meanwhile, this invention can be adapted to various types of round steel with specifications of Φ20-Φ50 and carbon content of 0.05%-1.2%, making it highly versatile and directly adaptable to the technical transformation and implementation of existing production lines. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments. In the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessarily obscuring the concepts of this invention.

[0020] This embodiment provides a method and system for evaluating the quality of round steel end faces and selecting the optimal marking objects. This embodiment is applied to the automatic marking production line of round steel in a steel rolling workshop, processing bundled round steel with specifications of Φ20-Φ50 and carbon content of 0.05%-1.2%. The entire system consists of a 3D vision system, a PLC control system, a six-axis robot, and an adjustable power laser marking mechanism. The specific process is as follows: First, bundles of round steel bars are conveyed to the marking station via a chain, where a locking mechanism secures the bundles to prevent movement during marking. A 3D vision camera mounted directly above the station is activated, simultaneously acquiring two-dimensional end-face images and three-dimensional point cloud data of the entire bundle. At the same time, the vision system reads parameters such as diameter, steel type, and carbon content of the current bundle of round steel bars from the MES (Manufacturing Execution System) via industrial Ethernet.

[0021] After data acquisition, the vision control computer preprocesses the original 2D image: first, Gaussian filtering removes image noise; then, histogram equalization enhances the image; and finally, adaptive threshold binarization is performed to highlight the boundary between the end face region and the background. Subsequently, a contour extraction algorithm is used to identify all circular contours in the image. For each contour, minimum circumcircle fitting is performed to separate the independent end face region of each round steel bar within the bundle. For each round steel bar, its theoretical end face area, the 3D spatial coordinates of the end face center, and the tilt angle of the end face plane relative to the axis are calculated.

[0022] Next, the end face of a single round steel bar will be quantitatively scored based on five indicators, with each indicator ranging from 0 to 100 points: Flatness A: Extract the 3D point cloud height data of the corresponding end face area, calculate the elevation standard deviation of all points. A full score of 100 points is obtained when the standard deviation is 0. For every 0.1mm increase in the standard deviation, the corresponding score is deducted. The larger the standard deviation, the lower the score.

[0023] For surface finish B, a grayscale image of a single end face is extracted, and texture features are extracted through the grayscale co-occurrence matrix. The pixel percentage of oxide scale, stains, and pits in the end face is counted. A score of 100 is obtained when the percentage is 0. For every 1% increase in the defect percentage, the corresponding score is deducted.

[0024] Completeness C: The actual area corresponding to the effective pixels in the statistical end face area is divided by the theoretical standard end face area of ​​the round steel of this specification. The ratio is multiplied by 100 to obtain the completeness score. Missing edges or corners will directly reduce the effective area and lower the score.

[0025] The end face inclination D is calculated by fitting the end face plane with a 3D point cloud and calculating the angle between the normal of the end face plane and the center of the round steel. A score of 100 points is obtained when the angle is 0°. The corresponding score is deducted for every 1° increase in the angle.

[0026] Effective engraving area E: Based on the center of the end face, delineate a safe engraving area without defects, calculate the area of ​​the safe area and normalize it according to the maximum usable area of ​​this specification to obtain the effective engraving area score.

[0027] After completing the individual scoring, the overall quality score Q is calculated using a weighted formula: Q = 0.3 × A + 0.25 × B + 0.2 × C + 0.15 × D + 0.1 × E. The Q values ​​of all round steel bars in the bundle are sorted in descending order, and the round steel bar with the highest ranking is selected as the candidate for marking.

[0028] The system is set to a pass threshold of 85 points. If the Q value of the first round steel is ≥85 points, it is judged as the best marking steel material. If the highest Q value of the whole bundle of round steel is lower than 85 points, the vision control computer sends a fault alarm signal to the PLC, the production line suspends the marking operation, and prompts manual intervention.

[0029] After screening, the vision-based industrial control computer transmits the 3D spatial coordinates, end-face comprehensive score, and steel grade data of the selected round steel bars to the PLC control system. The PLC generates motion trajectory instructions based on the coordinate data and sends them to the controller of the six-axis robot. This controller moves the robot's end-effector laser marking head to the front of the target round steel bar's end face. Simultaneously, the PLC, combining the end-face quality score, round steel diameter, and carbon content parameters, automatically calculates the appropriate laser output power and sends a power adjustment command to the laser, achieving adaptive dimming marking. After marking, the six-axis robot returns to its origin position, and the conveyor chain moves the next bundle of round steel bars into place, entering the next round of inspection and screening cycle.

[0030] In this embodiment, the detection and screening calculation of the entire bundle's end face takes no more than 5 seconds. Combined with robot positioning and laser marking processes, the total time for a single item can be controlled within 35 seconds, perfectly matching the production line's pace. Simultaneously, the laser power is automatically adapted to end faces of different qualities, effectively avoiding end face cracking caused by high power and blurred lettering caused by low power.

[0031] In the above embodiments, a method and system for evaluating the quality of round steel end faces and selecting the optimal marking objects are provided. This invention replaces the manual screening mode, and the time for sorting and inspecting the end faces of a whole bundle of round steel is ≤5s, which can meet the production cycle requirement of a total marking time of ≤35s for a single piece, thus improving the efficiency of automated production. This invention establishes a unified end face quality evaluation standard through quantitative scoring of five indicators, avoiding human judgment errors. The marking characters on the end faces after optimization are complete and clear, ensuring traceability of the entire product process. This invention relies on the end face quality scoring to adaptively match the laser output power, eliminating the cracking defects of the round steel end faces caused by power mismatch, while avoiding the problem of shallow marking on inferior end faces, thus improving the finished product qualification rate. This invention provides early warning for the output of the whole bundle of unqualified end faces, which can avoid the problem of batch marking defects in advance and reduce production waste. At the same time, this invention can be adapted to a variety of round steel with specifications of Φ20-Φ50 and carbon content of 0.05%-1.2%, which is highly versatile and can be directly adapted to the technical transformation and implementation of existing production lines.

[0032] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the concept and scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the inventive concept should fall within the protection scope of the present invention. All technical contents for which protection is sought in this invention are fully described in the claims.

Claims

1. A method for evaluating the quality of round steel end faces and selecting optimal identification objects, characterized in that, Includes the following steps: S1. Collect two-dimensional images and three-dimensional morphological data of the end face of bundled round steel bars, and simultaneously obtain round steel production parameters; S2. Preprocess the two-dimensional image to segment the independent end face regions of a single round steel bar, and extract the spatial position and morphological parameters of each end face. S3. Based on multiple end-face quality evaluation dimensions, quantitative scores are given to the end face of a single round steel bar to obtain individual quality scores. S4. Calculate the comprehensive quality score of the end face of a single round steel bar using a weighted fusion algorithm, and sort all round steel bars in the bundle in descending order of comprehensive quality score; S5. Select the first-ranked round steel as the candidate identification object, set the qualified score threshold, and if the comprehensive quality score of the candidate object reaches the threshold, it is determined to be a qualified identification object; otherwise, an abnormal alarm signal is output. S6. Send the location data and end-face quality parameters of the qualified marked object to the control system, and link the marking execution mechanism to complete the adaptive marking operation.

2. The method for evaluating the quality of the end face of round steel and selecting the optimal identification object according to claim 1, characterized in that: In step S1, a 2D image and 3D topography data of the end face are acquired by a 3D vision camera. The 3D topography data is 3D point cloud data of the end face. The round steel production parameters include the round steel diameter, steel grade and carbon content.

3. The method for evaluating the quality of the end face of round steel and selecting the optimal identification object according to claim 1, characterized in that: In step S2, the preprocessing includes noise reduction, image enhancement and binarization; the single end face region is segmented by contour extraction combined with minimum circumcircle fitting algorithm, and the extracted morphological parameters include theoretical end face area, end face spatial coordinates and end face tilt angle.

4. The method for evaluating the quality of the end face of round steel and selecting the optimal identification object according to claim 1, characterized in that: In step S3, the multiple end-face quality evaluation dimensions include flatness, smoothness, integrity, end-face tilt and effective imprinting area, with each individual item scoring from 0 to 100 points.

5. The method for evaluating the quality of the end face of round steel and selecting the optimal identification object according to claim 4, characterized in that: The flatness is calculated based on the standard deviation of the end face elevation of the 3D point cloud; the smaller the standard deviation, the higher the flatness score. The smoothness is extracted by the gray-level co-occurrence matrix to extract image texture features, and the score is calculated based on the proportion of oxide scale, stains, and pitting. The integrity is converted into a score by the ratio of the actual effective end face area to the theoretical standard end face area. The end face tilt is calculated based on the angle between the end face plane normal and the axis of the round steel; the smaller the angle, the higher the score. The effective engraving area is scored based on the normalized area of ​​the safety engraving area at the center of the end face.

6. The method for evaluating the quality of the end face of round steel and selecting the optimal identification object according to claim 4, characterized in that: In step S4, the calculation formula for the weighted fusion algorithm is as follows: Q=0.3×A+0.25×B+0.2×C+0.15×D+0.1×E In the formula, Q is the overall quality score, A is the flatness score, B is the smoothness score, C is the integrity score, D is the end face tilt score, and E is the effective imprinted area score.

7. The method for evaluating the quality of the end face of round steel and selecting the optimal identification object according to claim 1, characterized in that: In step S5, the passing score threshold is 85 points.

8. The method for evaluating the quality of the end face of round steel and selecting the optimal identification object according to claim 1, characterized in that: In step S6, the marking actuator includes a six-axis robot and an adjustable power laser; the control system sends position data to the six-axis robot to control its motion trajectory, and at the same time adaptively matches the output power of the laser according to the end face quality parameters and round steel production parameters.

9. A system for evaluating the quality of round steel end faces and selecting optimal identification objects, characterized in that: The method for implementing any one of claims 1-8 includes a 3D vision acquisition module, an image processing and calculation module, a PLC control module, a six-axis robot, and an adjustable power laser marking mechanism; wherein the signal output terminal of the 3D vision acquisition module is connected to the signal input terminal of the image processing and calculation module, the signal output terminal of the image processing and calculation module is connected to the signal input terminal of the PLC control module, and the PLC control module is connected to the six-axis robot and the adjustable power laser marking mechanism respectively.