CN118710724B - An AI visual weld position automatic search method - Google Patents

Abstract

The invention discloses an automatic searching method for AI visual weld joint positions, which comprises the following steps: s1, obtaining a steel bar welding track line G1 of a to-be-welded joint position HD 1; s2, the midpoint of the connecting segment extends to the two ends respectively and is obtained from a track type tableObtaining a welding seam position X1 of a to-be-welded point HD1 in mm; s3, adjusting the posture of the welding gun according to the two steel bar distribution postures; s4, if the welding line position search is successful, starting the next welding point position search; otherwise, the weld position shifts according to the weld position shift data. Compared with the prior art, when the reinforcing mesh is standardized, the length of the welding seam and the starting point of the welding seam are obtained in the track type table, the position of the welding seam is the middle part of the connecting section, and the searching efficiency of the position of the welding seam is improved. When the reinforcing mesh has errors, the welding seam position which can be welded by the robot is obtained by shifting according to the found welding seam position shifting data, and the automatic searching efficiency of the welding seam position is improved.

Description

An AI visual weld position automatic search method

Technical Field

The present invention relates to the technical field of steel mesh welding, and in particular to an AI visual weld position automatic search method.

Background Art

In the construction industry, a large number of prefabricated steel mesh structures are widely used. In order to ensure that the main steel mesh is subjected to force in the component, multiple main steel bars are connected by full penetration welding. The weld has the following characteristics: large size, many styles of steel mesh and small quantity, poor consistency, many welding points, symmetrical welding on both sides of the points, short weld length, many weld forms, uneven weld gaps of raw material ribbed steel bars, and unclear weld center lines. Due to the above weld characteristics, the weld position of the automated robot welding needs to be realized by manual teaching of the weld position through online programming. The steel mesh is long and large, and the number of welds is large and short, resulting in low efficiency of manual teaching of welds. In addition, the steel mesh has many specifications and small batches, so it is necessary to manually teach the weld position of each mesh; and the weld position error of the same batch of steel mesh is large due to production and assembly errors, so it is necessary to manually teach the weld position of each mesh. Based on the above reasons, the manual teaching method of steel mesh robot welding is inefficient and repetitive, and manual teaching of weld position seriously limits the application of robot welding in the field of steel mesh welding.

Summary of the invention

The present invention provides an AI visual weld position automatic search method to solve the problem of low efficiency of robot searching welds due to large weld position errors caused by manufacturing and assembly errors of steel meshes.

The present invention provides an AI visual weld position automatic search method, comprising the following steps:

S1: Obtain a steel bar welding trajectory line G1 of a welding point HD1; the steel bar welding trajectory line G1 includes a connecting section and steel bars around the connecting section, and the connecting section is a section of the steel bar welding trajectory line G1 where two steel bars are arranged side by side;

S2: Find the steel bar welding trajectory line LE that matches the steel bar welding trajectory line G1 in the trajectory type table, then obtain the weld length L1mm corresponding to the steel bar welding trajectory line LE in the trajectory type table, and then extend from the midpoint of the connection segment to both ends. mm to obtain the weld position X1 of the weld point HD1; obtain the weld starting point O corresponding to the steel bar welding trajectory line LE from the trajectory type table, then the weld starting point O is the weld starting point of the weld position X1;

S3: When the two steel bars at the weld position X1 are located on the same horizontal plane, go to S4; otherwise, the weld position X1 is calculated by the weld trajectory posture planning unit to obtain the weld position X1 with welding gun posture adjustment data;

S4: Whether the robot can perform welding according to the weld position X1, if yes, the weld position search is successful, and after the weld position HD1 is successfully welded, go to step S1 and start the next welding position search; otherwise, go to step S5;

S5: Case set B includes multiple cases, and the second weld position of each case is obtained by offsetting the first weld position according to the weld position offset data. The robot performs welding according to the second weld position of the case. The first weld position is the midpoint of the steel bar connection section in the case, extending amm to both ends. 2a is the weld length of the case. The gaps at both ends of weld position X1 are compared with the gaps at both ends of the first weld position of each case, and the case AL1 with the closest gap is found. Then, weld position X1 is offset according to the weld position offset data of case AL1 to obtain weld position Y1, and weld position Y1 is changed to weld position X1, and then go to step S3.

Preferably, in step S5, the specific steps of finding the case AL1 with the closest gap are: first find the case group ALZ with the same weld length as the weld position X1, and then find the difference between the gaps at both ends of the first weld position and the gaps at both ends of the weld position X1 in the case group ALZ; in the case group ALZ, the case AL1 with the smallest sum of the gap differences between the first weld position and the weld position X1 is the closest case.

Preferably, in the specific steps of finding the case AL1 with the closest gap: the difference between the gap at one end of the first weld position and the gap at one end of the weld position is JXX1, the difference between the gap at the other end of the first weld position and the gap at the other end of the weld position is JXX2, the sum of JXX1 and JXX2 is the sum of the gap differences between the first weld position and the weld position X1, and in the case group ALZ, the case AL1 with the smallest sum of JXX1 and JXX2 is the closest case.

Preferably, in step S2, the trajectory type table includes multiple types of steel bar welding trajectory lines, multiple weld lengths and multiple weld starting points, and the steel bar welding trajectory lines, weld lengths and weld starting points correspond to each other one by one; the specific steps of finding the steel bar welding trajectory line LE are: matching the steel bar welding trajectory line G1 with the steel bar welding trajectory lines in the trajectory type table, and determining which type of steel bar welding trajectory line the steel bar welding trajectory line G1 belongs to, and the type of steel bar welding trajectory line to which the steel bar welding trajectory line G1 belongs is the matching steel bar welding trajectory line LE.

Preferably, in step S2, the starting point of the weld in the steel bar welding trajectory line is one of the left end, right end, upper end or lower end of the weld.

Preferably, after welding of welding point HD1 is completed, welding point HD1 is added to case set B as a new case.

Preferably, before step S1, the following steps are also included: the component scanning unit Q scans the workpiece to be welded to obtain a regional image of all the points to be welded; point cloud processing is performed on the regional image to obtain the steel bar welding trajectory lines of all the points to be welded, and then the steel bar welding trajectory lines execute steps S1-S5 in sequence.

Preferably, in step S4, whether the robot can perform welding according to the weld position X1 is determined based on the following criteria: if the largest gap in the weld position X1 exceeds a preset value, the robot cannot perform welding according to the weld position X1.

Preferably, after the maximum gap in the weld position X1 in step S4 does not exceed the preset value, if the robot cannot weld according to the weld position X1 due to an abnormal event, it is determined whether the abnormal event meets one of the abnormality library. If so, the abnormal event is ignored and the robot welds normally; otherwise, an error is reported to notify manual intervention.

Preferably, the abnormal library includes rust on the steel bar, uneven weld flow on the steel bar, a part or all of the ribs on the steel bar being worn away, and small impurities existing on the steel bar.

Compared with the prior art, in the present invention, when the steel mesh is manufactured and assembled in accordance with the specification, the weld length and the weld starting point are obtained in the trajectory type table, and the weld position is the middle of the connecting section. This setting can effectively improve the efficiency of the weld position search. When the steel mesh has a large weld position error due to manufacturing and assembly errors, the weld position in the middle of the connecting section is offset according to the weld position offset data obtained from the closest case found in case set B to obtain the weld position where the robot can weld. This setting ensures that the robot can automatically search for the weld position. The two work together to enable the robot to automatically search for welds and can effectively improve the search efficiency of the weld position. The present invention can ensure the intelligent and precise positioning of the weld position, which is conducive to improving the operating efficiency of the welding robot.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

Fig. 1 is a top view of the present invention;

FIG2 is a schematic structural diagram of a steel mesh according to the present invention;

FIG3 is a schematic structural diagram of a region image in a steel mesh of the present invention;

FIG4 is a schematic structural diagram of the present invention in which the weld position is in the middle of the connecting section;

FIG5 is a schematic diagram of the structure of the present invention after the weld position in the middle of the connecting section is offset;

FIG6 is a schematic structural diagram of a steel bar welding trajectory line 1 of the present invention;

FIG7 is a schematic structural diagram of a second steel bar welding trajectory line according to the present invention;

FIG8 is a schematic structural diagram of a steel bar welding trajectory line 3 of the present invention;

FIG. 9 is a schematic structural diagram of adjusting the welding gun posture according to the present invention.

Reference numerals:

1. Work platform, 2. Robot, 3. Operating table, 4. Fixture, 5. Steel mesh, 6. Area picture, 7. Weld position X1, 8. Steel bar welding trajectory line G1, 51. Steel bar, 100. Connection section.

DETAILED DESCRIPTION

In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

Referring to Figures 1-2, this embodiment provides an AI visual weld position automatic search method, including an operating table 3, a steel mesh 5 is installed on the operating table 3 through a fixture 4, a component scanning unit P scans the operating table 3, obtains the overall model C of the steel mesh 5 and the rough coordinates of all the points to be welded, and the component scanning unit Q moves to each of the points to be welded according to the rough coordinates to scan and obtain the regional image 6 of all the points to be welded, performs point cloud processing on the regional image 6 to obtain the steel bar welding trajectory lines of all the points to be welded, and then sequentially performs steps S1-S5 on the steel bar welding trajectory lines of all the points to be welded to obtain the weld position of each point to be welded. The component scanning unit P and the component scanning unit Q both include a global shutter CMOS camera, a nano red line laser, and a hardware triggered scanning synchronization module. The component scanning unit P is arranged on the working platform 1, and is used to obtain the point cloud data of the surrounding environment of the steel mesh 5. In order to shoot the operating table 3 and the steel mesh 5 as a whole, the component scanning unit P will be set relatively high, and the data accuracy obtained will be relatively low accordingly. The component scanning unit Q is arranged at the end of the robot 2 arm, and is used to obtain the point cloud data of the points to be welded. This arrangement can obtain high-precision point cloud data of the points to be welded. This embodiment also includes the following steps:

Step S1: Obtain the steel bar welding trajectory line G18 of the welding point HD1; referring to FIG3, the steel bar welding trajectory line G18 includes a connecting section 100 and steel bars 51 around the connecting section 100, and the connecting section 100 is a section of the steel bar welding trajectory line G18 where two steel bars 51 are arranged side by side; this step is set to facilitate matching the steel bar welding trajectory line G18 with the trajectory type table in subsequent steps.

The track type table includes multiple types of steel bar welding track lines, multiple weld lengths and multiple weld starting points, and the steel bar welding track lines, weld lengths and weld starting points correspond to each other. In the early stage, the steel bar welding track lines of the steel mesh 5 are classified, specifically according to the distribution pattern of the two steel bars 51 where the weld is located; then the weld length and weld starting point of each type of steel bar welding track line are designed, and the weld starting point in the steel bar welding track line is one of the left end, right end, upper end or lower end of the weld. After the track type table is established in the system, various types of steel bar welding track lines and their corresponding data can be continuously added in the later stage. Attached Figures 6-8 are three different types of steel bar welding track lines. The weld in Figure 6 is welded from left to right, and its weld starting point is the left end; the weld in Figure 7 is welded from left to right, and its weld starting point is the left end; the weld in Figure 8 is welded from right to left, and its weld starting point is the right end.

Step S2: Find the steel bar welding trajectory line LE that the steel bar welding trajectory line G18 matches in the trajectory type table. Specifically: match the steel bar welding trajectory line G18 with the steel bar welding trajectory line in the trajectory type table, and determine which type of steel bar welding trajectory line G18 belongs to. The type of steel bar welding trajectory line to which the steel bar welding trajectory line G18 belongs is the matching steel bar welding trajectory line LE. Then obtain the weld length L1mm corresponding to the steel bar welding trajectory line LE in the trajectory type table, and then extend from the midpoint of the connecting segment 100 to each end. mm to obtain the weld position X17 of the weld point HD1; obtain the weld starting point O corresponding to the steel bar welding trajectory line LE from the trajectory type table, and the weld starting point O is the weld starting point of the weld position X17; assuming that the weld starting point O is the right end, the weld starting point of the weld position X17 is at the right end, and welding should be performed from right to left. Referring to FIG. 4, the weld position X17 is the middle part of the connecting section 100, and the weld starting point of the weld position X17 is at the left end.

S3: When the two steel bars 51 at the weld position X17 are located on the same horizontal plane, go to S4; otherwise, it means that the two steel bars 51 are up and down. If the height difference between the two exceeds a certain range, the welding gun will hit the higher steel bar 51 during welding, causing welding interruption. Therefore, the posture of the welding gun needs to be adjusted to prevent the gun from hitting the gun. The weld position X17 is calculated by the weld trajectory posture planning unit to obtain the weld position X17 with welding gun posture adjustment data;

There are two ways to adjust the welding gun posture. One is: the weld trajectory posture planning unit calculates the rotation angle of the welding gun based on the spatial data of the two steel bars 51 at the weld position X17, so that the welding gun is always perpendicular to the plane where the two steel bars 51 are located.

Another is: the weld trajectory posture planning unit calculates the maximum rotation angle θ°C of the welding gun based on the spatial data of the two steel bars 51 at the weld position X17, and the welding gun rotates by θ°C to weld the spot HD1 to be welded. In this process, the posture of the welding gun is no longer adjusted. Specifically, the welding gun rotates by θ°C to the side of the lower steel bar 51. Referring to FIG. 9, when the steel bar 51 on the left is higher than the steel bar 51 on the right, the welding gun rotates by θ°C to the right before welding, and then welding is performed.

S4: Whether robot 2 can weld according to weld position X17, if yes, weld position search is successful, after welding of weld point HD1 is successful, go to step S1 and start the next weld position search; otherwise, execute step S5; in this step, if the welding gun can weld according to weld position X17, step S5 can be omitted, which can effectively improve the efficiency of weld position search. Setting step S5 is that the steel mesh 5 is fixed on the operating table 3 by a clamp 4, and this method is multi-point fixation. Referring to FIG1, there may be only one clamp 4 to fix the steel bar 51 in the area of the weld point, and the steel bar 51 is unevenly stressed, and the steel bar 51 will become bent, so that the gap between the two steel bars 51 is different at each place, so there will be a situation where the welding gun cannot weld at the weld point or the welding effect is not good. When welding cannot be performed or the welding effect is not good, it is necessary to execute step S5, and refer to the successful case to offset the weld position so that the welding gun can weld or improve the welding effect.

Case set B includes multiple cases. The second weld position of each case is obtained by offsetting the first weld position according to the weld position offset data. The second weld position is the welding position of robot 2. The first weld position is the midpoint of the connecting section 100 of the steel bar 51 in the case, extending amm to both ends. 2a is the weld length of the case.

S5: Compare the gaps at both ends of weld position X17 with the gaps at both ends of the first weld position of each case, find the case AL1 with the closest gap, then offset weld position X17 according to the weld position offset data of case AL1 to obtain weld position Y1, convert weld position Y1 to weld position X17, and go to step S3. The specific steps of finding the case AL1 with the closest gap are: first find the case group ALZ with the same weld length as weld position X17, then find the gap difference between the gaps at both ends of the first weld position and the gap difference between the gaps at both ends of weld position X17 in the case group ALZ, and the case AL1 with the smallest sum of the gap differences between the first weld position and the gaps at both ends of weld position X17 in the case group ALZ is the closest case. Assume that the difference between the gap at one end of the first weld position and the gap at one end of the weld position is JXX1, the difference between the gap at the other end of the first weld position and the gap at the other end of the weld position is JXX2, and the sum of JXX1 and JXX2 is the sum of the gap differences between the first weld position and the weld position X17. In case group ALZ, the case AL1 with the smallest sum of JXX1 and JXX2 is the closest case. Referring to FIG5, for example: in case group ALZ, the gap of steel bar 51 at one end of the first weld position closest to case AL1 is 8, and the gap of steel bar 51 at the other end is -3, the gap of steel bar 51 at one end of the welding position X1 of the weld point to be welded is 8.5, and the gap of steel bar 51 at the other end is -3.2, and the gap difference is |8.5-8|+|-3.2+3|=0.7, that is, the gap difference of other cases in case group ALZ is greater than 0.7. Since the second weld position of AL1 is obtained by shifting the first weld position 30 mm to the left, the weld position X17 is shifted 30 mm to the left to obtain the weld position Y1.

The method of finding the closest case and then offsetting according to the weld position offset data in the case is because the found case is not exactly the same as the weld point HD1, and the weld position in the case cannot be directly moved to the weld point HD1. The weld position X17 can only be offset according to the weld position offset data in the closest case to obtain the weld position belonging to the weld point HD1.

In the present invention, when welding of welding point HD1 is completed, welding point HD1 is added as a new case to case set B, thereby completing the update of case set B. Cases with poor welding effects can be deleted manually or automatically by the system later.

In the present invention, the judgment basis for whether the robot 2 can perform welding according to the weld position X17 in step S4 is: if the maximum gap in the weld position X17 exceeds a preset value, the robot 2 cannot perform welding according to the weld position X17.

After the maximum gap in the weld position X17 in step S4 of the present invention does not exceed the preset value, if the robot 2 cannot weld according to the weld position X17 due to an abnormal event, it is determined whether the abnormal event meets one of the abnormality libraries. If so, the abnormality is ignored and the robot 2 welds normally; otherwise, an error is reported to notify manual intervention. The abnormal library includes rust on the steel bar 51, the weld flow on the steel bar 51 is uneven, the ribs on the steel bar 51 are partially or completely worn away, and small impurities exist on the steel bar 51. Rust on the steel bar 51 causes the color of the steel bar 51 to change, which will cause recognition abnormality. When the robot 2 welds, the weld flow flows to the bottom, which will cause unevenness. Small impurities on the steel bar 51 include grass, paper, etc. These abnormal events will affect the recognition of the robot 2, but will not affect welding. Therefore, these abnormal events can be ignored to ensure normal welding.

In this embodiment, case set B includes multiple successful welding cases, each case includes a steel bar welding trajectory line and weld position offset data and weld effect corresponding to the steel bar welding trajectory line, and the weld position offset data includes weld position offset and offset direction. In the early stage, the welding point picture successfully welded by manual welding or other methods is stored in the system after point cloud processing, and the weld position offset data and weld effect of the welding point are stored in the system to obtain case set B. The initial position (first weld position) of the weld of the welding point in each successful welding case is in the middle of the connecting section 100, and then the final weld position (second weld position) is obtained after the offset to one end. For example, the weld position of the welding point in case 1 is initially in the middle of the connecting section 100, and then it is offset to the right by 10 mm to obtain the final weld position, that is, the weld position of the welding point in case 1 is offset by 10 mm to the right from the middle of the connecting section 100.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims (8)

1. An AI visual weld position automatic search method, characterized in that it includes the following steps: S1: Obtain a steel bar welding trajectory line G1 of a welding point HD1; the steel bar welding trajectory line G1 includes a connecting section and steel bars around the connecting section, and the connecting section is a section of the steel bar welding trajectory line G1 where two steel bars are arranged side by side; S2: Find the steel bar welding trajectory line LE that matches the steel bar welding trajectory line G1 in the trajectory type table, then obtain the weld length L1mm corresponding to the steel bar welding trajectory line LE in the trajectory type table, and then extend from the midpoint of the connection segment to both ends. mm to obtain the weld position X1 of the weld point HD1; obtain the weld starting point O corresponding to the steel bar welding trajectory line LE from the trajectory type table, then the weld starting point O is the weld starting point of the weld position X1; S3: When the two steel bars at the weld position X1 are located on the same horizontal plane, go to S4; otherwise, the weld position X1 is calculated by the weld trajectory posture planning unit to obtain the weld position X1 with welding gun posture adjustment data; S4: Whether the robot can perform welding according to the weld position X1, if yes, the weld position search is successful, and after the weld position HD1 is successfully welded, go to step S1 and start the next welding position search; otherwise, go to step S5; S5: Case set B includes multiple cases, and the second weld position of each case is obtained by offsetting the first weld position according to the weld position offset data. The robot performs welding according to the second weld position of the case, and the first weld position is the midpoint of the steel bar connection section in the case extending amm to both ends, 2a is the weld length of the case, and the gaps at both ends of the weld position X1 are compared with the gaps at both ends of the first weld position of each case to find the case AL1 with the closest gap, and then the weld position X1 is offset according to the weld position offset data of case AL1 to obtain the weld position Y1, and the weld position Y1 is changed to the weld position X1, and then go to step S3; in step S5, the specific steps of finding the case AL1 with the closest gap are: first find the case AL1 with the closest gap to the weld position Set case group ALZ with the same weld length as X1, and then find the gap difference between the two ends of the first weld position and the gap difference between the two ends of the weld position X1 in case group ALZ. In case group ALZ, the case AL1 with the smallest sum of the gap difference between the first weld position and the two ends of weld position X1 is the closest case; in the specific steps of finding the case AL1 with the closest gap: the difference between the gap at one end of the first weld position and the gap at one end of the weld position is JXX1, the difference between the gap at the other end of the first weld position and the gap at the other end of the weld position is JXX2, the sum of JXX1 and JXX2 is the sum of the gap difference between the first weld position and the two ends of weld position X1, and in case group ALZ, the case AL1 with the smallest sum of JXX1 and JXX2 is the closest case. 2. The AI visual weld position automatic search method according to claim 1 is characterized in that, in step S2, the trajectory type table includes multiple types of steel bar welding trajectory lines, multiple weld lengths and multiple weld starting points, and the steel bar welding trajectory lines, weld lengths and weld starting points correspond to each other one by one; the specific steps of finding the steel bar welding trajectory line LE are: matching the steel bar welding trajectory line G1 with the steel bar welding trajectory line in the trajectory type table, judging which type of steel bar welding trajectory line G1 belongs to, and the type of steel bar welding trajectory line to which the steel bar welding trajectory line G1 belongs is the matching steel bar welding trajectory line LE. 3. The AI visual weld position automatic search method according to claim 2 is characterized in that, in step S2, the weld starting point in the steel bar welding trajectory line is one of the left end, right end, upper end or lower end of the weld. 4. The AI visual weld position automatic search method according to claim 1 is characterized in that after the welding of welding point HD1 is completed, welding point HD1 is added to case set B as a new case. 5. The AI visual weld position automatic search method according to claim 1 is characterized in that, before step S1, it also includes the following steps: the component scanning unit Q scans the workpiece to be welded to obtain a regional image of all the points to be welded; the regional image is processed by point cloud to obtain the steel bar welding trajectory lines of all the points to be welded, and then the steel bar welding trajectory lines are executed in sequence by steps S1-S5. 6. The AI visual weld position automatic search method according to claim 1 is characterized in that the judgment basis for whether the robot can weld according to the weld position X1 in step S4 is: if the maximum gap in the weld position X1 exceeds a preset value, the robot cannot weld according to the weld position X1. 7. The AI visual weld position automatic search method according to claim 6 is characterized in that, after the maximum gap in the weld position X1 in step S4 does not exceed the preset value, if the robot cannot weld according to the weld position X1 due to an abnormal event, it is determined whether the abnormal event meets one of the abnormality libraries. If so, the abnormal event is ignored and the robot welds normally; otherwise, an error is reported to notify manual intervention. 8. The AI visual weld position automatic search method according to claim 7 is characterized in that the abnormal library includes rust on the steel bars, uneven weld flow on the steel bars, a part or all of the ribs on the steel bars being worn away, and small impurities on the steel bars.