Metal plate electric arc welding intelligent detection method

By combining machine vision and tensile testing, the problems of low efficiency and high cost in metal sheet welding inspection have been solved, achieving efficient and accurate welding quality assessment, reducing production costs and improving resource utilization.

CN120948219BActive Publication Date: 2025-12-23NANTONG BAOJIA PRECISION MASCH CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202511481351.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-10-16
Publication Date
2025-12-23
Estimated Expiration
2045-10-16

AI Technical Summary

Technical Problem

Existing technologies for inspecting welded metal sheets are characterized by low efficiency, large errors, and high costs, making it difficult to meet the real-time inspection needs of large-scale production lines.

Method used

By employing machine vision combined with image acquisition and stretching detection methods, the angle and distribution position of the bonding wire are obtained, and image recognition and repair processing are performed to ensure the consistency of image acquisition and the accuracy of bonding wire detection. The strength and stability of the bonding wire are evaluated using the stretching operation.

Benefits of technology

It improves the accuracy and efficiency of welding inspection, reduces scrap rate, lowers production costs, and enhances resource utilization and inspection flexibility.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120948219B_ABST
    Figure CN120948219B_ABST
Patent Text Reader

Abstract

The present application relates to the technical fields of welding detection, and discloses a kind of metal plate arc welding intelligent detection method, comprising: the metal plate of welding completion is placed in detection table, and the image of upper surface is collected, and it is judged as one or two class products;Repair and detect again to two class products, until it is changed to one class product;Collect one class product image, calculate weld line angle and divide weld line group, determine the number of weld line and tensile test sequence;Obtain weld line position, determine tensile direction and tensile site, and execute tensile operation in turn;Collect image after each tensile, compare initial image and image to be measured, and according to the result, keep or reject product;Determine tensile direction according to the angle of weld line in weld line group;The force condition of each weld line angle determines, adjust tensile direction, make it perpendicular to weld line arrangement angle, ensure that weld line is stressed in the best direction, avoid offsetting pulling force, can more accurately evaluate weld line strength and stability.
Need to check novelty before this filing date? Find Prior Art

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of welding detection, in particular to a metal plate arc welding intelligent detection method. BACKGROUND

[0002] The welding of metal plates is an important manufacturing process in modern industrial production and is widely used in the automotive, shipbuilding, aerospace, construction and other industries. During the welding process of metal plates, the quality of the welding line has a direct impact on the performance, service life and safety of the final product. Therefore, it is crucial to detect the welded metal plates.

[0003] In the prior art, the welding detection of metal plates is usually carried out by manual detection or simple instrument detection. However, with the expansion of production scale and the improvement of product quality requirements, the traditional detection methods have gradually exposed many shortcomings.

[0004] Firstly, traditional manual detection relies on human vision and experience, which has the disadvantages of low efficiency, high error rate and inability to quantify. In addition, manual detection usually requires a large amount of time and human resources, especially in large-scale production lines, which cannot meet the requirements of rapid and stable quality detection.

[0005] Secondly, although the existing instrument detection method improves the detection accuracy and efficiency to a certain extent, the prior art is difficult to judge the quality of the welding by a simple method. The current detection means mainly relies on X-ray imaging or ultrasonic detection. Although these methods can detect internal defects of welding, the equipment cost is high and the operation is complex, which is not suitable for real-time detection in large-scale production lines.

[0006] Therefore, the present application aims to provide a method that combines machine vision, according to the distribution of the welding line, by stretching the welding line at a certain angle and direction to detect welding defects. SUMMARY

[0007] The present application provides a metal plate arc welding intelligent detection method, which solves the problems mentioned in the background art.

[0008] The present application provides the following technical solution: a metal plate arc welding intelligent detection method, comprising:

[0009] Place the metal plate on the detection table;

[0010] Obtain the image acquisition position and the image acquisition direction;

[0011] At the image acquisition position, collect the image of the upper surface of the metal plate in the image acquisition direction, denoted as the first detection image;

[0012] According to the first detection image, the metal plate is marked as a first type product or a second type product;

[0013] The second type product is repaired, and after the repair, detection is performed. According to the detection result, it is judged whether to perform re-repair processing until the second type product is converted into the first type product;

[0014] At the image acquisition position, the image of the upper surface of the first type product is collected in the image acquisition direction, and is recorded as an initial image;

[0015] According to the initial image, the angle of each weld line on the upper surface of the first type product is obtained;

[0016] According to the angle of each weld line on the upper surface of the first type product, the weld lines on the upper surface of the first type product are divided into different weld line groups;

[0017] The number of weld lines existing in each weld line group is obtained respectively, and the execution order of the tensile detection is obtained according to the number of weld lines;

[0018] The distribution position of the weld line in the weld line group on the upper surface of the first type product is obtained, and the tensile direction and the tensile position of the tensile detection are obtained according to the distribution position;

[0019] For the first type product, the tensile operation is sequentially performed on each weld line group at the tensile position of each weld line group in the tensile direction according to the execution order of the tensile detection;

[0020] In the tensile detection, after each tensile operation is completed, the image of the upper surface of the first type product is collected at the image acquisition position in the image acquisition direction, and is recorded as a to-be-detected image;

[0021] The to-be-detected image and the initial image are compared, and the first type product is retained or rejected according to the comparison result.

[0022] Optionally, the image acquisition position and the image acquisition direction are obtained, comprising:

[0023] After the metal plate is placed on the detection table, the center position of the upper surface of the metal plate is obtained, and is recorded as a center point;

[0024] A perpendicular line perpendicular to the upper surface of the metal plate is made through the center point, and is recorded as a reference vertical line;

[0025] The intersection point of the reference vertical line and the detection table is obtained, and is recorded as a reference intersection point;

[0026] The direction along the reference vertical line from the center point to the reference intersection point is taken as the image acquisition direction;

[0027] On the reference vertical line, a point position is obtained, so that when the image of the upper surface of the metal plate is collected at the point position in the image acquisition direction, the collected image completely contains the upper surface of the metal plate.

[0028] Optionally, the marking the metal plate as a first-class product or a second-class product according to the first detection image comprises:

[0029] identifying and detecting the soldering line position in the first detection image through image recognition technology;

[0030] if the soldering line in the first detection image is identified and detected to have a welding defect, the metal plate is marked as a second-class product;

[0031] if all the soldering lines in the first detection image do not have welding defects, the metal plate is marked as a first-class product.

[0032] Optionally, the second-class product is subjected to repair processing, and after the repair, detection is performed to determine whether to perform re-repair processing until the second-class product is converted into a first-class product, comprising:

[0033] the soldering line with a welding defect in the second-class product is subjected to repair processing, and the second-class product after the repair processing is recorded as a to-be-tested repair plate;

[0034] the to-be-tested repair plate is placed on a detection table, and an upper surface image of the to-be-tested repair plate is collected at an image collection position and in an image collection direction, recorded as a second detection image;

[0035] if all the soldering lines in the second detection image do not have welding defects, the second-class product is re-marked as a first-class product;

[0036] if a soldering line in the second detection image has a welding defect, the soldering line with the welding defect is marked;

[0037] the marked soldering line in the second-class product is subjected to re-repair processing, and after the re-repair processing, detection is performed again to determine whether the soldering line has a welding defect, and whether the re-repair processing is performed again is determined;

[0038] when it is detected that all the soldering lines do not have welding defects, the second-class product is re-marked as a first-class product.

[0039] Optionally, the angle of each soldering line on the upper surface of the first-class product is obtained according to the initial image, comprising:

[0040] an edge line of the upper surface of the first-class product is obtained;

[0041] any straight edge line is taken as a reference straight line;

[0042] for each soldering line on the upper surface of the first-class product, two end points of each soldering line are obtained;

[0043] the perpendicular distance between the two end points of each soldering line and the reference straight line is obtained and compared.

[0044] In the two end points, the one with shorter perpendicular distance to the reference straight line is selected as the reference end point of the soldering line;

[0045] In the reference straight line, one of the two points is recorded as the first reference point, and the other is recorded as the second reference point;

[0046] The direction from the first reference point to the second reference point is recorded as the ray direction;

[0047] For each soldering line, from the reference end point of the soldering line, a ray parallel to the reference straight line is drawn in the ray direction on the upper surface of the product, which is recorded as the reference ray;

[0048] The angle value between each soldering line and the reference ray of the soldering line is obtained as the angle of each soldering line.

[0049] Optionally, the soldering lines on the upper surface of the product are divided into different soldering line groups according to the angles of the soldering lines on the upper surface of the product, comprising:

[0050] Obtain the angle of each soldering line on the upper surface of the product;

[0051] Divide each soldering line according to the angle of the soldering line, so that the soldering lines with the same angle are divided into the same soldering line group;

[0052] Iterate through all the soldering lines existing on the upper surface of the product to obtain several soldering line groups.

[0053] Optionally, the number of soldering lines existing in each soldering line group is obtained respectively, and the execution order of the tensile detection is obtained according to the number of soldering lines, comprising:

[0054] The number of soldering lines existing in each soldering line group is obtained respectively, and the number of soldering lines in each soldering line group is recorded as the number of soldering lines in each soldering line group;

[0055] Arrange the soldering line groups in order from more to less according to the number of soldering lines in each soldering line group;

[0056] Number the soldering line groups in order according to the arrangement order of the soldering line groups;

[0057] The numbering order of the soldering line groups is used as the execution order of the tensile detection.

[0058] Optionally, the distribution position of the soldering line in the soldering line group on the upper surface of the product is obtained, and the tensile direction and the tensile position of the tensile detection are obtained according to the distribution position, comprising:

[0059] For each soldering line group on the upper surface of the product, on the upper surface of the product:

[0060] making a vertical line perpendicular to all the welding lines existing in the welding group, and marking the vertical line as a marker vertical line;

[0061] making two straight lines parallel to the marker vertical line, and making the two straight lines intersect with two welding lines respectively, and making the area between the two straight lines completely cover all the welding lines existing in the welding group, and marking the two straight lines as marker straight lines;

[0062] taking one of the two marker straight lines as a first marker straight line, and taking the other as a second marker straight line;

[0063] making a vertical line connecting the first marker straight line and the second marker straight line, and perpendicular to the first marker straight line and the second marker straight line, and marking the vertical line as a connecting vertical line;

[0064] obtaining the length of the connecting vertical line, and marking the length as a vertical line length;

[0065] setting a stretching interval;

[0066] starting from the position of the first marker straight line, making a straight line perpendicular to the connecting vertical line every other stretching interval, and marking the straight line as a stretching line;

[0067] starting from the first marker straight line, numbering each stretching line in turn;

[0068] obtaining the intersection point of each stretching line and the upper surface edge line of the product, as the stretching position of each stretching line;

[0069] if the stretching line and the upper surface edge line of the product only have one intersection point, the stretching line is deleted;

[0070] if the stretching line and the upper surface edge line of the product have two intersection points, taking one of the intersection points as a first stretching position, and the other as a second stretching position;

[0071] obtaining the part of each stretching line on the upper surface of the product, and obtaining the midpoint of the part, and marking the midpoint as the stretching midpoint of each stretching line;

[0072] for each stretching line, the direction from the stretching midpoint to the first stretching position along the stretching line is taken as the first stretching direction of the stretching line, and the direction from the stretching midpoint to the second stretching position is taken as the second stretching direction of the stretching line.

[0073] Optionally, the stretching detection is performed on each welding group in turn according to the execution order of the stretching detection, at the stretching position of each welding group, in the stretching direction, including:

[0074] setting a stretching time and a stretching force;

[0075] For the solder wire group existing on the surface of a type of product, according to the execution sequence of the tensile detection, the tensile operation is executed on each tensile group in turn;

[0076] For each solder wire group, the tensile operation is executed on each tensile wire in turn according to the numbering sequence of the tensile wires of the solder wire group;

[0077] The tensile operation on each tensile wire is specifically that a tensile force with the same size is applied to each tensile wire at a first tensile position in a first tensile direction and at a second tensile position in a second tensile direction at the same time, and the tensile force is kept for a period of time;

[0078] When the period of time for keeping the tensile force is equal to the tensile time, the tensile force is stopped, and one time of the tensile operation is completed.

[0079] Optionally, the comparison between the to-be-tested image and the initial image includes reserving or rejecting the type of product according to a comparison result.

[0080] When each to-be-tested image is collected, the collected to-be-tested image is compared with the initial image to obtain a similarity between the to-be-tested image and the initial image;

[0081] A similarity threshold is set;

[0082] The similarity between the to-be-tested image and the initial image is compared with the similarity threshold;

[0083] If the similarity between the to-be-tested image and the initial image is less than the similarity threshold, the tensile detection is stopped, and the type of product subjected to the tensile detection is rejected;

[0084] If the similarity between the to-be-tested image and the initial image is greater than or equal to the similarity threshold, the tensile detection is continuously executed on the type of product, and in the process of executing the tensile detection, each collected to-be-tested image is compared with the initial image to obtain a similarity between the to-be-tested image and the initial image, until the tensile detection on the type of product is completed.

[0085] When in the process of the tensile detection, the similarity between all the to-be-tested images and the initial image is greater than or equal to the similarity threshold, the type of product subjected to the tensile detection is reserved.

[0086] The present application has the following beneficial effects:

[0087] 1. The method of determining the tensile direction according to the angle of the solder wire in the solder wire group can improve the pertinence of the tensile detection, the angle of each solder wire determines the stress condition of the solder wire on the product surface, by adjusting the tensile direction, it is ensured that the tensile operation is perpendicular to the arrangement angle of the solder wire, so that each solder wire receives the tensile force in the most suitable direction for the angle, unnecessary offset tensile force is avoided when the solder wire is tensiled, so that the strength and stability of the solder wire are more accurately evaluated.

[0088] 2. By obtaining the center point of the metal plate and the reference intersection point with the testing station, the direction and position of image acquisition can be determined. This ensures that the angle and range of each image acquisition are consistent, improving the stability and consistency of image acquisition. This effectively reduces errors in subsequent image recognition processes and avoids inaccurate detection results due to angle deviations. It also helps to ensure accurate detection results of the weld wire when comparing multiple image acquisitions.

[0089] 3. By repeatedly repairing defective Category II products until they are transformed into qualified Category I products, this step helps to minimize the scrap rate and improve resource utilization efficiency. Compared with directly scrapping defective products, the repair process not only reduces material waste but also lowers the company's production costs and increases the recycling rate of resources.

[0090] 4. By using a reference ray, the angle of each weld wire is obtained with a unified reference standard and grouped according to the angle. This method allows weld wires with the same angle to be grouped together, ensuring that the weld wires in each group are inspected under the same conditions, reducing inspection errors caused by uneven distribution or inconsistent direction of the weld wires. At the same time, grouping weld wires with the same angle into a weld wire group allows for subsequent stretching operations to be performed on the weld wire group. This enables all weld wires in a weld wire group to be inspected at once, reducing the number of subsequent stretching operations, simplifying the stretching inspection operation, and facilitating the completion of the inspection of all weld wires in a short time.

[0091] 5. Determine the execution order of tensile testing based on the number of weld wires. Prioritize tensile testing on the group of weld wires with the largest number of weld wires. This ensures that the largest number of weld wires are detected as quickly as possible during the tensile testing process, making it easier to identify weld wires with welding defects. This helps improve the efficiency of tensile testing, shorten the testing time, and increase production efficiency.

[0092] 6. By performing a tensile test on the weld wire, it is possible to determine whether the weld is strong enough under a certain tensile force, thereby judging the quality of the weld wire. At the same time, each tensile operation in the tensile test process applies the same tensile force to each weld wire, ensuring the consistency and rationality of the test.

[0093] 7. Set a similarity threshold. When the similarity between the image to be tested and the initial image is lower than the threshold, stop the stretching detection and remove the product. By comparing the similarity, the false detection rate can be effectively reduced, and misjudgment caused by slight image differences can be avoided. The setting of the similarity threshold not only improves the flexibility of detection, but can also be adjusted according to the actual situation of different products, thereby optimizing the detection process. Attached Figure Description

[0094] Figure 1The schematic diagram of the position of the reference ray of the present application.

[0095] Figure 2 The schematic diagram of the position of the stretch line of the present application. DETAILED DESCRIPTION

[0096] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all the other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the present application.

[0097] Embodiment one, a metal plate arc welding intelligent detection method, comprising:

[0098] Placing the metal plate on the detection table;

[0099] Obtaining the image acquisition position and the image acquisition direction;

[0100] At the image acquisition position, the metal plate upper surface image is collected in the image acquisition direction, which is recorded as the first detection image; in the embodiment, the metal plate upper surface refers to the surface of the metal plate where the welding line exists, which is kept on the same horizontal plane, and the welding line on it is located on the same plane;

[0101] According to the first detection image, the metal plate is marked as a first type of product or a second type of product;

[0102] The second type of product is repaired and detected after the repair, and whether to perform the repair again is judged according to the detection result, until the second type of product is converted into the first type of product;

[0103] At the image acquisition position, the first type of product upper surface image is collected in the image acquisition direction, which is recorded as the initial image;

[0104] According to the initial image, the angle of each welding line on the first type of product upper surface is obtained;

[0105] According to the angle of each welding line on the first type of product upper surface, the welding lines on the first type of product upper surface are divided into different welding line groups;

[0106] The number of welding lines existing in each welding line group is obtained respectively, and the execution order of the stretch detection is obtained according to the number of welding lines;

[0107] The distribution positions of the welding wires in the welding wire group on the surface of the product are acquired, and the stretching direction and the stretching position of the stretching detection are acquired according to the distribution positions; the stretching direction is determined according to the angle of the welding wire in the welding wire group, which can improve the accuracy of the stretching detection, the angle of each welding wire determines the stress condition of the welding wire on the surface of the product, by adjusting the stretching direction, it is ensured that the stretching operation is perpendicular to the arrangement angle of the welding wire, so that each welding wire receives the stretching force in the most suitable direction according to the angle, unnecessary offset force is avoided on the welding wire during stretching, so as to more accurately evaluate the strength and stability of the welding wire.

[0108] For the product, the stretching operation is sequentially performed on each welding group at the stretching position of each welding group in the stretching direction according to the execution sequence of the stretching detection.

[0109] In the stretching detection, after each stretching operation is completed, the image of the upper surface of the product is collected at the image collection position in the image collection direction, which is recorded as a to-be-tested image.

[0110] The to-be-tested image and the initial image are compared, and the product is retained or rejected according to the comparison result.

[0111] The image collection position and the image collection direction are acquired, including:

[0112] After the metal plate is placed on the detection table, the center position of the upper surface of the metal plate is acquired, which is recorded as a center point.

[0113] A perpendicular line to the upper surface of the metal plate is made through the center point, which is recorded as a reference vertical line.

[0114] The intersection point of the reference vertical line and the detection table is acquired, which is recorded as a reference intersection point.

[0115] The direction along the reference vertical line from the center point to the reference intersection point is taken as the image collection direction.

[0116] A point position is acquired on the reference vertical line, so that when the image of the upper surface of the metal plate is collected at the point position in the image collection direction, the collected image completely contains the upper surface of the metal plate; by acquiring the center point of the metal plate and the reference intersection point with the detection table, the direction and position of the image collection are determined, which can ensure that the angle and range of each image collection are consistent, improve the stability and consistency of the image collection, effectively reduce the error in the subsequent image recognition process, avoid inaccurate detection results due to angle deviation, and facilitate accurate detection results of the welding wire when multiple image collections are compared.

[0117] The metal plate is marked as a product or a product according to the first detection image, including:

[0118] The image recognition technology is prior art;

[0119] If the welding defects exist in the welding lines in the first detection image, the metal plate is marked as a second-class product; the welding defects in the embodiment refer to the welding defects that can be detected by image recognition, such as uneven welding lines and incomplete welding;

[0120] If all the welding lines in the first detection image do not have welding defects, the metal plate is marked as a first-class product.

[0121] The second-class product is repaired and detected after repair, and whether the re-repair processing is performed is determined according to the detection result, until the second-class product is converted into a first-class product, comprising:

[0122] The welding lines with welding defects in the second-class product are repaired, and the second-class product after repair is recorded as a to-be-tested repair plate;

[0123] The to-be-tested repair plate is placed on the detection table, and the image of the upper surface of the to-be-tested repair plate is collected at the image collection position and in the image collection direction, and recorded as a second detection image;

[0124] If all the welding lines in the second detection image do not have welding defects, the second-class product is marked as a first-class product again;

[0125] If a welding line in the second detection image has a welding defect, the welding line with the welding defect is marked;

[0126] The marked welding line in the second-class product is repaired again, and whether the welding defects exist in all the welding lines is detected again after the repair processing, and whether the repair processing is performed again is determined;

[0127] When it is detected that all the welding lines do not have welding defects, the second-class product is marked as a first-class product again; by repairing the second-class product with defects for multiple times until it is converted into a qualified first-class product, this step helps to minimize the waste rate and improve the utilization efficiency of resources. Compared with directly scrapping the products with defects, the repair processing not only reduces the waste of materials, but also reduces the production cost of enterprises and increases the recycling rate of resources.

[0128] The angle of each welding line on the upper surface of the first-class product is obtained according to the initial image, comprising:

[0129] An edge line on the upper surface of the first-class product is obtained;

[0130] Referring to Figure 1 , any straight edge line is taken as a reference straight line;

[0131] For each welding line existing on the upper surface of the product, two end points of each welding line are obtained;

[0132] The perpendicular distance between the two end points of each welding line and the reference straight line is obtained and compared;

[0133] Among the two end points, the one with the shorter perpendicular distance to the reference straight line is selected as the reference end point of the welding line;

[0134] On the reference straight line, any two points are selected, one of which is recorded as a first reference point and the other is recorded as a second reference point;

[0135] The direction from the first reference point to the second reference point is recorded as the ray direction;

[0136] For each welding line, a ray parallel to the reference straight line is drawn in the ray direction from the reference end point of the welding line on the upper surface of the product, which is recorded as the reference ray;

[0137] The angle value between each welding line and the reference ray of the welding line is obtained as the angle of each welding line.

[0138] The welding lines on the upper surface of the product are divided into different welding line groups according to the angle of each welding line on the upper surface of the product, comprising:

[0139] The angle of each welding line on the upper surface of the product is obtained;

[0140] Each welding line is divided according to the angle of the welding line, so that welding lines with the same angle are divided into the same welding line group;

[0141] All welding lines existing on the upper surface of the product are traversed for division to obtain a plurality of welding line groups; by means of the reference ray, a unified reference standard is used to obtain the angle of each welding line, and the welding lines are grouped according to the angle, which can divide the welding lines with the same angle into the same group, ensure that the welding lines in each group are detected under the same condition, and reduce the detection error caused by uneven distribution or inconsistent direction of the welding lines; at the same time, the welding lines with the same angle are divided into a welding line group, and subsequent stretching operation is performed on the welding line group, which can detect all welding lines in one welding line group at a time, can reduce the number of subsequent stretching operations, simplify the stretching detection operation, and is conducive to completing the detection of all welding lines in a short time.

[0142] The number of welding lines existing in each welding line group is obtained respectively, and the execution order of the stretching detection is obtained according to the number of welding lines, comprising:

[0143] The number of welding lines existing in each welding line group is obtained respectively and recorded as the number of welding lines of each welding line group;

[0144] arranging the wire groups in order according to the number of wires in each wire group from more to less;

[0145] numbering the wire groups in order according to the arrangement order of the wire groups;

[0146] using the numbering order of the wire groups as the execution order of the tensile test; determining the execution order of the tensile test according to the number of wires, and preferentially performing the tensile test on the wire group with the most wires, so that the wire group with the most wires can be detected at the fastest speed in the process of the tensile test, which facilitates the discovery of the wire group with welding defects, and is beneficial to improving the efficiency of the tensile test, shortening the detection time, and improving the production efficiency.

[0147] the distribution positions of the wires in the wire group on the upper surface of the product are obtained, and the tensile direction and the tensile position of the tensile test are obtained according to the distribution positions, and the method comprises the following steps:

[0148] for each wire group on the upper surface of the product, on the upper surface of the product:

[0149] a perpendicular line perpendicular to all the welding lines in the welding group is drawn, and the perpendicular line is recorded as a marking perpendicular line;

[0150] two straight lines parallel to the marking perpendicular line are drawn, the two straight lines intersect with two wires respectively, and the area between the two straight lines completely covers all the welding lines in the welding group, and the two straight lines drawn are recorded as marking straight lines;

[0151] one of the two marking straight lines is recorded as a first marking straight line, and the other is recorded as a second marking straight line;

[0152] referring to Figure 2 a perpendicular line connecting the first marking straight line and the second marking straight line and perpendicular to the first marking straight line and the second marking straight line is drawn, and the perpendicular line is recorded as a connecting perpendicular line;

[0153] the length of the connecting perpendicular line is obtained, and is recorded as a perpendicular line length;

[0154] a tensile interval is set; the tensile interval can be determined according to the distance between the two marking straight lines and the number of wires in the welding group, for example, the ratio of the distance between the two marking straight lines to the number of wires in the welding group can be taken as the tensile interval;

[0155] from the position of the first marking straight line, a straight line perpendicular to the connecting perpendicular line is drawn every tensile interval, and the straight line is recorded as a tensile line;

[0156] from the first marking straight line, each tensile line is numbered in order;

[0157] the intersection point of each tensile line and the edge line of the upper surface of the product is obtained as the tensile position of each tensile line;

[0158] If the stretching line and the edge line of the upper surface of the product of the first type have only one intersection point, the stretching operation cannot be performed on the one intersection point, and the stretching line is deleted;

[0159] If the stretching line and the edge line of the upper surface of the product of the first type have two intersection points, one of the intersection points is recorded as the first stretching position, and the other is recorded as the second stretching position;

[0160] The part of the line segment of each stretching line on the upper surface of the product of the first type is obtained, and the midpoint of the part of the line segment is obtained and recorded as the stretching midpoint of each stretching line;

[0161] For each stretching line, the direction from the stretching midpoint to the first stretching position along the stretching line is taken as the first stretching direction of the stretching line, and the direction from the stretching midpoint to the second stretching position is taken as the second stretching direction of the stretching line.

[0162] The stretching operation is sequentially performed on each welding group at the stretching positions of each welding group in the order of the stretching detection, including:

[0163] The stretching time and the stretching force are set. The stretching time can be arbitrarily set, and the stretching force is set according to the maximum tensile force that the welding line can theoretically withstand after welding, for example, 80% of the maximum tensile force can be taken as the set stretching force;

[0164] For the welding group existing on the upper surface of the product of the first type, the stretching operation is sequentially performed on each stretching group in the order of the stretching detection;

[0165] For each welding group, the stretching is sequentially performed on each stretching line in the order of the numbering of the stretching lines of the welding group;

[0166] The stretching of each stretching line specifically includes simultaneously applying the same stretching force to the first stretching position in the first stretching direction and to the second stretching position in the second stretching direction, and maintaining the stretching for a period of time;

[0167] When the time of maintaining the stretching is equal to the stretching time, the stretching is stopped, and one stretching operation is completed. Through the stretching, the detection of the welding line is completed, which can detect whether the welding is firm enough under the action of a certain stretching force, and further judge the quality of the welding line. At the same time, each stretching operation in the stretching detection process is performed by applying the same stretching force to each welding line, which ensures the consistency and rationality of the detection.

[0168] The comparison between the to-be-detected image and the initial image includes retaining or rejecting the product of the first type according to the comparison result, including:

[0169] When each of the collected images to be detected is collected, the collected image to be detected is compared with the initial image to obtain the similarity between the image to be detected and the initial image;

[0170] A similarity threshold is set, which is used to determine whether the image to be detected is similar to the initial image, and can be set according to production needs. The greater the similarity threshold is set, the more similar the image to be detected is to the initial image, that is, the smaller the deformation allowed by the wire after stretching is required.

[0171] The similarity between the image to be detected and the initial image is compared with the similarity threshold. The image similarity acquisition and similarity comparison technology is prior art.

[0172] If the similarity between the image to be detected and the initial image is less than the similarity threshold, the stretching detection is stopped, and the products of the type subjected to the stretching detection are rejected.

[0173] If the similarity between the image to be detected and the initial image is greater than or equal to the similarity threshold, the stretching detection of the products of the type is continued, and during the stretching detection, each of the collected images to be detected is compared with the initial image to obtain the similarity between the image to be detected and the initial image, until the stretching detection of the products of the type is completed.

[0174] When the similarity between all the images to be detected and the initial image is greater than or equal to the similarity threshold during the stretching detection, the products of the type subjected to the stretching detection are retained. The similarity threshold is set, and when the similarity between the image to be detected and the initial image is lower than the threshold, the stretching detection is stopped and the products are rejected. Through the comparison of the similarity, the false detection rate can be effectively reduced, and the misjudgment caused by slight image difference can be avoided. The setting of the similarity threshold not only improves the flexibility of the detection, but also can be adjusted according to the actual situation of different products, so as to optimize the detection process.

[0175] It should be noted that, in the present text, relational terms such as first and second and the like can only be used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the terms "comprises", "comprising" or any other variant thereof are intended to cover non-exclusive inclusions, so that a process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such a process, method, article or equipment.

[0176] The above description is only the preferred embodiments of the present application, and it should be pointed out that, for those skilled in the art, without departing from the technical principles of the present application, a number of improvements and refinements can be made, which should also be considered as the protection scope of the present application.

Claims

1. A metal plate arc welding intelligent detection method, characterized in that, The method comprises the following steps: acquiring an image acquisition position and an image acquisition direction; acquiring an image of the upper surface of the metal plate at the image acquisition position and in the image acquisition direction, and recording the image as a first detection image; labeling the metal plate as a first type of product or a second type of product according to the first detection image; performing a repair process on the second type of product, and performing a detection after the repair, and determining whether to perform a re-repair process according to the detection result, until the second type of product is converted into the first type of product; acquiring an image of the upper surface of the first type of product at the image acquisition position and in the image acquisition direction, and recording the image as an initial image; acquiring the angle of each welding line on the upper surface of the first type of product according to the initial image; dividing the welding lines on the upper surface of the first type of product into different welding groups according to the angle of each welding line on the upper surface of the first type of product; acquiring the distribution position of the welding lines in the welding group on the upper surface of the first type of product, and acquiring the stretching direction and the stretching position of the stretching detection according to the distribution position, and the specific steps comprise the following steps: for each welding group on the upper surface of the first type of product, on the upper surface of the first type of product: draw a perpendicular line perpendicular to all welding lines existing in the welding group, and record the perpendicular line as a marking perpendicular line; draw two straight lines parallel to the marking perpendicular line, so that the two straight lines intersect with two welding lines respectively, and the area between the two straight lines completely covers all welding lines existing in the welding group, and record the two straight lines as marking straight lines; take one of the two marking straight lines as a first marking straight line, and take the other as a second marking straight line; draw a perpendicular line connecting the first marking straight line and the second marking straight line, and record the perpendicular line as a connecting perpendicular line; acquire the length of the connecting perpendicular line, and record the length as a perpendicular line length; set a stretching interval; starting from the position of the first marking straight line, draw a straight line perpendicular to the connecting perpendicular line every stretching interval, and record the straight line as a stretching line; number each stretching line starting from the first marking straight line; acquire the intersection point of each stretching line and the edge line of the upper surface of the first type of product as the stretching position of each stretching line; if the stretching line and the edge line of the upper surface of the first type of product only have one intersection point, the stretching line is deleted; if the stretching line and the edge line of the upper surface of the first type of product have two intersection points, take one of the intersection points as a first stretching position, and take the other as a second stretching position; acquire the part of the line segment of each stretching line on the upper surface of the first type of product, and acquire the midpoint of the part of the line segment as the stretching midpoint of each stretching line; for each stretching line, the direction from the stretching midpoint to the first stretching position along the stretching line is taken as the first stretching direction of the stretching line, and the direction from the stretching midpoint to the second stretching position is taken as the second stretching direction of the stretching line; acquire the number of welding lines existing in each welding group respectively, and acquire the execution order of the stretching detection according to the number of welding lines to perform the stretching detection.

2. The intelligent detection method for metal plate electric arc welding according to claim 1, characterized in that, The acquiring of the image acquisition position and the image acquisition direction comprises the following steps: after the metal plate is placed on the detection table, acquire the center position of the upper surface of the metal plate, and record the center position as a center point; draw a perpendicular line perpendicular to the upper surface of the metal plate through the center point, and record the perpendicular line as a reference perpendicular line; acquire the intersection point of the reference perpendicular line and the detection table, and record the intersection point as a reference intersection point; take the direction from the center point to the reference intersection point along the reference perpendicular line as the image acquisition direction; On the reference vertical line, a point is obtained, so that when an image of the upper surface of the metal plate is collected at the point in the image collection direction, the collected image completely contains the upper surface of the metal plate.

3. The intelligent detection method for metal plate electric arc welding according to claim 1, characterized in that, The metal plate is marked as a first-class product or a second-class product according to the first detection image, including: The position of the welding line in the first detection image is detected by image recognition technology; If the welding line is detected to have a welding defect in the first detection image, the metal plate is marked as a second-class product; If all the welding lines in the first detection image do not have welding defects, the metal plate is marked as a first-class product.

4. The intelligent detection method for metal plate electric arc welding according to claim 1, characterized in that, The second-class product is repaired and detected after repair, and whether to perform re-repair processing is determined according to the detection result, until the second-class product is converted into a first-class product, including: The welding line with a welding defect in the second-class product is repaired, and the second-class product after repair is recorded as a to-be-tested repaired plate; The to-be-tested repaired plate is placed on the detection table, and an image of the upper surface of the to-be-tested repaired plate is collected at the image collection position in the image collection direction, recorded as a second detection image; If all the welding lines in the second detection image do not have welding defects, the second-class product is re-marked as a first-class product; If a welding line in the second detection image has a welding defect, the welding line with the welding defect is marked; The marked welding line in the second-class product is repaired again, and whether to perform the repair processing operation again is determined according to whether all the welding lines have welding defects after the repair processing operation. When it is detected that all the welding lines do not have welding defects, the second-class product is re-marked as a first-class product.

5. The intelligent detection method for metal plate electric arc welding according to claim 1, characterized in that, The angle of each welding line on the upper surface of the first-class product is obtained according to the initial image, including: An edge line on the upper surface of the first-class product is obtained; Any straight edge line is taken as a reference straight line; For each welding line on the upper surface of the first-class product, two end points of each welding line are obtained; The perpendicular distances between the two end points of each welding line and the reference straight line are obtained and compared; Among the two end points, the one with a shorter perpendicular distance from the reference straight line is selected as the reference end point of the welding line; Two points on the reference straight line are taken, one of which is recorded as a first reference point and the other is recorded as a second reference point; The direction from the first reference point to the second reference point is recorded as the direction of the ray; For each welding line, a ray parallel to the reference straight line is drawn from the reference end point of the welding line on the upper surface of the first-class product in the direction of the ray, recorded as a reference ray; The angle between each welding line and the reference ray of the welding line is obtained as the angle of each welding line.

6. The intelligent detection method for metal plate electric arc welding according to claim 1, characterized in that, The welding lines on the upper surface of the first-class product are divided into different welding line groups according to the angle of each welding line on the upper surface of the first-class product, including: The angle of each welding line on the upper surface of the first-class product is obtained; Each welding line is divided according to the angle of the welding line, so that welding lines with the same angle are divided into the same welding line group; All the welding lines on the upper surface of the first-class product are traversed to obtain a plurality of welding line groups.

7. The intelligent detection method for metal plate electric arc welding according to claim 1, characterized in that, The number of welding lines in each welding line group is obtained, and the execution order of the tensile detection is obtained according to the number of welding lines to perform tensile detection, including: Obtaining the number of solder wires existing in each solder wire group, and recording as the number of solder wires of each solder wire group respectively; Arranging the solder wire groups according to the number of solder wires of each solder wire group from more to less; Numbering the solder wire groups according to the arrangement order of the solder wire groups; Taking the numbering order of the solder wire groups as the execution order of the tensile test.

8. The intelligent detection method for metal plate electric arc welding according to claim 1, characterized in that, Including, for a type of products, performing the tensile operation on each solder wire group in the tensile position according to the execution order of the tensile test in the tensile direction: Setting the tensile time and the tensile force; For the solder wire groups existing on the upper surface of the type of products, performing the tensile operation on each solder wire group according to the execution order of the tensile test; For each solder wire group, performing the tensile operation on each tensile wire according to the numbering order of the tensile wires of the solder wire group; The tensile operation on each tensile wire is specifically that the same tensile force is applied to each tensile wire in the first tensile position in the first tensile direction and in the second tensile position in the second tensile direction, and the tensile operation is maintained for a period of time; When the time of maintaining the tensile operation is equal to the tensile time, the tensile operation is stopped, and one tensile operation is completed.

9. The intelligent detection method for metal plate electric arc welding according to claim 1, characterized in that, Including, in the tensile test, after completing the tensile operation each time, collecting the image of the upper surface of the type of products in the image collection position in the image collection direction, and recording as the to-be-tested image; Comparing the to-be-tested image and the initial image, and reserving or rejecting the type of products according to the comparison result; When the to-be-tested image is collected each time, comparing the collected to-be-tested image with the initial image to obtain the similarity between the to-be-tested image and the initial image; Setting the similarity threshold value; Comparing the similarity between the to-be-tested image and the initial image with the similarity threshold value; If the similarity between the to-be-tested image and the initial image is less than the similarity threshold value, stopping the tensile test and rejecting the type of products subjected to the tensile test; If the similarity between the to-be-tested image and the initial image is greater than or equal to the similarity threshold value, continuing to perform the tensile test on the type of products, and comparing each to-be-tested image collected in the process of performing the tensile test with the initial image to obtain the similarity between the to-be-tested image and the initial image until the tensile test on the type of products is completed; When the similarity between all the to-be-tested images and the initial image in the process of the tensile test is greater than or equal to the similarity threshold value, reserving the type of products subjected to the tensile test.

Citation Information

Patent Citations

  • Diagnostic test method and system for circuit experiment board

    CN116819288A

  • Welding detection and repair method and system

    CN118450621A