Welding robot teaching-system-free programming method for arbitrary placement of workpieces

Abstract

The embodiment of the invention provides a teaching-system-free programming method for a welding robot aiming at arbitrary placement of workpieces. The method comprises the following steps: a 3D camera photographs to obtain an overall point cloud in a scene; according to the features and the communication form of the workpiece point cloud, point cloud processing is carried out, welding seam recognition clustering is carried out, and welding seam feature types and positions are obtained; according to production takt efficiency consideration, a welding path is planned, and welding starting and ending point coordinates and a robot locating method are determined; robot action instructions are matched through the robot pose library and the welding seam type features; matching the external shaft position according to the welding seam position and coordinate system transformation; and the upper computer generates a robot program file and issues welding operation. Through a specific programming method and a hardware system, welding seams of different specifications and types can be recognized for disordered workpieces, self-adaptive programming can be carried out, a robot can automatically find positions to carry out high-precision welding, manual intervention is not needed, and high-efficiency automatic manufacturing is achieved.

Description

Technical Field

[0001] The present application relates to the technical field of welding robot systems, and in particular to a teaching-free system programming method for welding robots with arbitrarily placed workpieces. Background Art

[0002] With the rapid advancement of science and technology, industrial robots have become an important part of today's industrial production. Compared with the limitations of manual operation, they can complete various complex and dangerous tasks more accurately and highly consistently, greatly improving production efficiency and improving product quality.

[0003] However, in the prior art, most of the programming methods for industrial robots are still based on teaching programming, but the manual teaching process is cumbersome and inefficient. It is difficult to obtain a consistent path and ensure welding quality for complex welding trajectories. With the expansion of robot application areas, offline programming has gradually become a more popular programming method. However, offline programming requires scene construction, and using the model of the welding workpiece for simulation and trajectory adjustment can save a lot of labor and maintenance costs. However, the offline programming method relies too much on pre-set programs for operation, cannot adapt to scene changes, and requires a high time investment offline, which cannot meet the work rhythm requirements. Summary of the invention

[0004] In view of the above problems, the present invention proposes a teaching-free robot system programming method. When the welding workpieces are placed in disorder in a large scene, adaptive programming can be performed according to the feedback of the visual system. Only welding positioning is required without reprogramming, thus realizing intelligent robot welding.

[0005] The purpose of the present invention is to provide a teaching-free system programming method for a welding robot with arbitrarily placed workpieces, which can identify welds of different specifications and types for disordered workpieces and perform adaptive programming. The robot automatically locates its position for high-precision welding without manual intervention, thereby achieving high-efficiency automated manufacturing.

[0006] An embodiment of the present invention provides a method for programming a teaching-free system of a welding robot for arbitrarily placing workpieces, the method comprising:

[0007] S1: 3D camera takes pictures to obtain the overall point cloud in the scene;

[0008] S2: According to the features and connectivity of the workpiece point cloud, the point cloud is processed and weld identification and clustering is performed to obtain the weld feature type and position;

[0009] S3: Based on the production cycle efficiency, plan the welding path, determine the welding start and end point coordinates and the robot positioning method;

[0010] S4: Match the robot action instructions through the robot posture library and the weld type characteristics, and match the external axis position according to the weld position and coordinate system transformation;

[0011] S5: The host computer generates the robot program file and sends the welding operation.

[0012] In some embodiments, the S1: taking photos with a 3D camera to obtain an overall point cloud in the scene includes:

[0013] The workpieces in the scene can be placed arbitrarily, the entire scene can be scanned by a 3D camera, and the scanning results can be processed into point clouds.

[0014] In some embodiments, the 3D camera uses micro-electro-mechanical system (MEMS) structured light projection to perform point cloud input on the entire scanning plane and identify workpiece information.

[0015] In some embodiments, point cloud information of the entire scene is obtained by superimposing multiple scanning planes and merging point clouds.

[0016] In some embodiments, the step S2: processing the point cloud and performing weld identification clustering according to the features and connectivity of the workpiece point cloud to obtain the weld feature type and location includes:

[0017] The point cloud is downsampled and filtered, and the top point cloud is clustered, compressed, and reconstructed to obtain a planar point cloud;

[0018] Based on the obtained planar point cloud, feature extraction is performed to identify intersections and end forms, classify weld types, perform feature extraction, and form a list of weld position and coordinate data.

[0019] In some embodiments, the step S3: planning a welding path, determining the welding start and end point coordinates and the robot positioning method based on production cycle efficiency considerations, includes:

[0020] According to the production cycle requirements and the robot's shortest path planning, the starting point data of the weld is deeply searched through the nearest point search algorithm, and is reasonably arranged and combined with the robot's installation position and path.

[0021] In some embodiments, the S3: planning the welding path, determining the welding start and end point coordinates and the robot positioning method based on the production cycle efficiency considerations, further includes:

[0022] According to the characteristic type of the weld, the reasonable welding process and posture of the robot are planned for each weld. The robot posture and the positioning method for the robot to accurately locate the arc starting point are determined by presetting the welding process and moving trajectory.

[0023] In some embodiments, the step S4: matching robot motion instructions with the weld type features through the robot posture library, and matching the external axis position according to the weld position and coordinate system transformation, includes:

[0024] According to the starting and ending coordinates of the weld, the coordinate system of the robot coordinate system and the workpiece coordinate system are changed through the preset robot posture database to obtain the precise position of the robot's external axis.

[0025] An embodiment of the present invention provides a teaching-free system for a welding robot, which is used to execute the method described in any of the above embodiments.

[0026] In some embodiments, the system comprises:

[0027] 3D camera, used to take pictures and obtain the overall point cloud of the scene;

[0028] A point cloud processing unit is used to process the point cloud and perform weld identification and clustering according to the features and connectivity of the workpiece point cloud, and obtain the weld feature type and position;

[0029] Path planning unit, used to plan welding paths according to production cycle efficiency, determine welding start and end point coordinates and robot positioning method;

[0030] The robot motion instruction matching unit is used to match the robot motion instruction with the weld type feature through the robot posture library; and to match the external axis position according to the weld position and coordinate system transformation;

[0031] The host computer is used to generate robot program files and issue welding operations.

[0032] Compared with the prior art, the programming method of the present invention not only improves the overall system, but also includes many improvements in details. Specifically, it has the following beneficial effects:

[0033] 1. The 3D camera of the present invention uses MEMS structured light to scan the entire point cloud of the scene. One shot can cover a plane. Multiple sensors are arranged horizontally to overlap the shooting range, reducing the area that cannot be reached by structured light.

[0034] 2. In the point cloud processing and clustering of the present invention, outliers are removed through filtering and clipping, and a weld coordinate system is established with independent welds, so that the weld results obtained by scanning are highly accurate;

[0035] 3. The present invention performs feature recognition on the rib plate, can identify the type and end form of the rib plate, and finely divides different positioning methods. Through the sensor on the robot head, the weld is precisely positioned, and a variety of positioning modes can be adopted to adaptively obtain the welding start and end points;

[0036] 3. In the present invention, the welding path planning adopts the nearest point search algorithm according to the requirements of scientific production scheduling and robot movement sequence, and combines the welding process such as welding retention and corner wrapping to design the corresponding path, thus ensuring high-quality and efficient welding. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The drawings illustrate generally, by way of example and not limitation, various embodiments discussed herein.

[0038] Figure 1 This is a flow chart of a method for programming a teaching-free system of a welding robot for arbitrarily placing workpieces according to an embodiment of the present invention;

[0039] Figure 2 is a schematic diagram of the workpiece;

[0040] Figure 3 Schematic diagram of the robot system;

[0041] Figure 4 This is a schematic diagram of the welding path planning results. DETAILED DESCRIPTION

[0042] In order to enable a more detailed understanding of the features and technical contents of the embodiments of the present application, the implementation of the embodiments of the present application is described in detail below in conjunction with the accompanying drawings. The attached drawings are for reference only and are not used to limit the embodiments of the present application.

[0043] In the embodiments of the present application, it should be noted that, unless otherwise specified and limited, the term "connection" should be understood in a broad sense. For example, it can be an electrical connection or a connection between two components. It can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to the specific circumstances.

[0044] It should be noted that the terms "first\second\third" involved in the embodiments of the present application are only used to distinguish similar objects, and do not represent a specific order for the objects. It is understandable that the specific order or sequence of "first\second\third" can be interchanged where permitted. It should be understood that the objects distinguished by "first\second\third" can be interchanged where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein.

[0045] The following is combined with Figures 1 to 3 The embodiments of the present invention are described in detail.

[0046] The embodiment of the present invention provides a method for programming a welding robot teaching-free system for arbitrarily placing workpieces, the steps are as follows: Figure 1 As shown, the process includes the following steps S1 to S5.

[0047] S1: The 3D camera takes a photo to obtain the overall point cloud in the scene.

[0048] S2: According to the features and connectivity of the workpiece point cloud, the point cloud is processed and weld identification and clustering is performed to obtain the weld feature type and location.

[0049] S3: Based on the production cycle efficiency, plan the welding path, determine the welding start and end point coordinates and the robot positioning method.

[0050] S4: Match the robot action instructions with the weld type features through the robot posture library, and match the external axis position according to the weld position and coordinate system transformation.

[0051] S5: The host computer generates the robot program file and sends the welding operation.

[0052] In some embodiments, S1: taking a photo with a 3D camera to obtain an overall point cloud in a scene includes:

[0053] Arrange the objects in the scene arbitrarily. For example, the objects are Figure 2 As shown, this is only an example and does not specifically limit the form of the workpiece. The entire scene is scanned by a 3D camera, and the scan results are processed into point clouds.

[0054] In some embodiments, the 3D camera uses micro-electro-mechanical system (MEMS) structured light projection to input point cloud of the entire scanning plane and identify workpiece information.

[0055] In some embodiments, point cloud information of the entire scene is obtained by superimposing multiple scanning planes and merging point clouds.

[0056] In some embodiments, S2: based on the features and connectivity of the workpiece point cloud, the point cloud is processed and weld identification and clustering is performed to obtain the weld feature type and position, including:

[0057] The point cloud is downsampled and filtered, and the top point cloud is clustered, compressed, and reconstructed to obtain a planar point cloud.

[0058] Based on the obtained planar point cloud, feature extraction is performed to identify intersections and end forms, classify weld types, perform feature extraction, and form a list of weld position and coordinate data.

[0059] In some embodiments, S3: based on the production cycle efficiency, planning the welding path, determining the welding start and end point coordinates and the robot positioning method, including:

[0060] According to the production cycle requirements and the robot's shortest path planning, the starting point data of the weld is deeply searched through the nearest point search algorithm, and is reasonably arranged and combined with the robot's installation position and path.

[0061] In some embodiments, S3: planning a welding path based on production cycle efficiency, determining welding start and end point coordinates and a robot positioning method, further includes:

[0062] According to the characteristic type of the weld, the reasonable welding process and posture of the robot are planned for each weld. The robot posture and the positioning method for the robot to accurately locate the arc starting point are determined by presetting the welding process and moving trajectory.

[0063] In some embodiments, S4: matching robot motion instructions with the weld type features through the robot posture library, and matching the external axis position according to the weld position and coordinate system transformation, including:

[0064] According to the starting and ending coordinates of the weld, the coordinate system of the robot coordinate system and the workpiece coordinate system are changed through the preset robot posture database to obtain the precise position of the robot's external axis.

[0065] An embodiment of the present invention provides a teaching-free system for a welding robot, which is used to execute a method as described in any of the above embodiments.

[0066] In some embodiments, Figure 3 As shown, the system includes:

[0067] A 3D camera, or three-dimensional camera, is used to take pictures and obtain the overall point cloud in the scene.

[0068] The point cloud processing unit is used to process the point cloud and perform weld identification and clustering according to the features and connectivity of the workpiece point cloud to obtain the weld feature type and position.

[0069] The path planning unit is used to plan the welding path according to the production cycle efficiency, determine the welding start and end point coordinates and the robot positioning method.

[0070] The robot motion instruction matching unit is used to match the robot motion instruction with the weld type feature through the robot posture library, and to match the external axis position according to the weld position and coordinate system transformation.

[0071] The host computer is used to generate robot program files and issue welding operations.

[0072] In the embodiment of the present application, the workpiece is placed at any position, and there is no need to import or pre-set a three-dimensional model. Only a 3D line laser camera is used to collect point clouds in the entire welding scene.

[0073] Since the target object is basically composed of a base plate and reinforcing ribs, but the length, angle and end form of each rib plate are not fixed, and the placement in the scene is also not fixed. The 3D camera uses MEMS structured light projection to input point clouds of the entire scanning plane and identify the workpiece information.

[0074] By superimposing multiple scanning planes and merging the point clouds, the point cloud information of the entire scene can be obtained. However, the overall point cloud often has a large amount of data and will produce shadows due to occlusion. The conventional approach is to downsample and filter the point cloud, and cluster, compress, and reconstruct the top point cloud to obtain a planar point cloud.

[0075] Based on the obtained plane point cloud, feature extraction is completed through point cloud processing methods such as standard Hough transform, and the intersection and end form are autonomously identified without human intervention, the weld types are specifically divided, and some types of features are extracted to meet the subsequent robot positioning accuracy.

[0076] After the weld type recognition and feature extraction are performed in the large scene plane point cloud, a list of weld position and coordinate data is formed;

[0077] According to the production cycle requirements and the robot's shortest path planning, the starting point data of the weld is deeply searched through the nearest point search algorithm, and is reasonably arranged and combined with the robot's installation position and path.

[0078] Furthermore, according to the characteristic type of the weld, a reasonable welding process and posture of the robot are planned for each weld, and the robot posture and the positioning method for the robot to accurately locate the arc starting point are determined by presetting the welding process and movement trajectory.

[0079] Furthermore, according to the starting and ending coordinates of the weld, the coordinate system of the robot coordinate system and the workpiece coordinate system are changed through the preset robot posture database to obtain the precise position of the robot's external axis.

[0080] The present invention provides a method for programming a welding robot teaching-free system for arbitrarily placing workpieces, the steps are as follows: Figure 1 As shown, it is applicable to Figure 3 The system architecture shown in the figure can realize the adaptive intelligent programming of the welding robot system for any placement of the workpiece by visually identifying the weld features of the workpiece and establishing a weld coordinate system data set. Specifically:

[0081] First, in step 1, different workpieces are randomly placed in a large scene and the whole scene is scanned by a 3D camera. The scan results are processed into point clouds. A huge amount of point cloud data can be obtained based on structured light photography. By combining the point cloud data features of structured light photography and the type of workpiece ribs, an efficient recognition method is designed.

[0082] Step 2: Downsample, compress and filter the point cloud, identify the rib type for the top point cloud, and compress the height to filter the scene ground and the bottom plate of the workpiece. In the identification of the rib, the complete point cloud of the rib is reconstructed according to the characteristics and connectivity of the rib, and the rib is automatically clustered and identified. The identified rib ends are matched with the point cloud according to the feature set of the workpiece end to obtain the feature type of the rib.

[0083] Step 3: According to the direction of the rib plate and the start and end coordinates of the weld, the production cycle requirements and the robot's shortest path planning, the nearest point search algorithm is used to conduct a deep search of the weld's starting point data, and a reasonable arrangement and combination with the robot's installation position and path is performed. According to the welding requirements of the workpiece, the accessibility and safety points of the welding operation sequence are considered, and the paths are calculated separately when the dual robots are operating separately.

[0084] Step 4: Deploy the robot posture library in the host computer, pre-set and save the robot's six-axis posture and welding gun angle when it reaches a specific point, and obtain the robot's corresponding welding action instructions by matching the robot posture library with the weld type characteristics; at the same time, according to the weld position and coordinate point, according to the coordinate system transformation algorithm, the robot base coordinate system and the workpiece scene world coordinate system are transformed, the robot base movement distance is calculated, and the external axis position of the corresponding weld is matched. The matching effect is as follows: Figure 4 shown.

[0085] Step 5: After processing the above-mentioned job files in the weld, the host computer generates the welding job and the corresponding robot program, and sends the program to the robot through the TCP protocol for teaching-free robot operation.

[0086] The 3D camera of the present invention scans the entire point cloud of the scene using MEMS structured light for surface scanning. One shot can cover a plane. Multiple sensors are arranged horizontally to overlap the shooting range, reducing the area that cannot be reached by structured light.

[0087] In the point cloud processing and clustering of the present invention, outliers are removed through filtering and clipping, and a weld coordinate system is established with independent welds, so that the weld results obtained by scanning are highly accurate;

[0088] In the present invention, the rib plate is feature-recognized, the type and end form of the rib plate can be identified, different positioning methods are carefully divided, and the weld is precisely positioned through the sensor on the robot head. A variety of positioning modes can be adopted to adaptively obtain the welding start and end points.

[0089] In the present invention, welding path planning adopts the nearest point search algorithm according to the requirements of scientific production scheduling and robot movement sequence, and combines the welding process such as welding retention and corner wrapping to design the corresponding path, thus ensuring high-quality and efficient welding.

[0090] The technical solutions described in the embodiments of the present application can be combined arbitrarily without conflict.

[0091] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for programming a welding robot teaching-free system for arbitrarily placed workpieces, characterized in that: The method comprises: S1: 3D camera takes pictures to obtain the overall point cloud in the scene; S2: According to the features and connectivity of the workpiece point cloud, the point cloud is processed and weld identification and clustering is performed to obtain the weld feature type and position; S3: Based on the production cycle efficiency, plan the welding path, determine the welding start and end point coordinates and the robot positioning method; S4: Match the robot action instructions through the robot posture library and the weld type characteristics, and match the external axis position according to the weld position and coordinate system transformation; S5: The host computer generates the robot program file and sends the welding operation.

2. The method for programming a welding robot teaching-free system for arbitrarily placing workpieces according to claim 1, characterized in that: S1: 3D camera takes pictures to obtain the overall point cloud in the scene, including: The workpieces in the scene can be placed arbitrarily, the entire scene can be scanned by a 3D camera, and the scanning results can be processed into point clouds.

3. The method for programming a welding robot teaching-free system for arbitrarily placing workpieces according to claim 2, characterized in that: The 3D camera projects micro-electromechanical system (MEMS) structured light, performs point cloud input on the entire scanning plane, and identifies workpiece information.

4. The method for programming a welding robot teaching-free system for arbitrarily placing workpieces according to claim 2, characterized in that: By superimposing multiple scanning planes and merging point clouds, point cloud information of the entire scene is obtained.

5. The method for programming a welding robot teaching-free system for arbitrarily placing workpieces according to claim 1, characterized in that: S2: According to the features and connectivity of the workpiece point cloud, the point cloud is processed and weld identification and clustering is performed to obtain the weld feature type and position, including: The point cloud is downsampled and filtered, and the top point cloud is clustered, compressed, and reconstructed to obtain a planar point cloud; Based on the obtained planar point cloud, feature extraction is performed to identify intersections and end forms, classify weld types, perform feature extraction, and form a list of weld position and coordinate data.

6. The method for programming a welding robot teaching-free system for arbitrarily placing workpieces according to claim 5, characterized in that: S3: Based on the production cycle efficiency, the welding path is planned, the welding start and end point coordinates and the robot positioning method are determined, including: According to the production cycle requirements and the robot's shortest path planning, the starting point data of the weld is deeply searched through the nearest point search algorithm, and is reasonably arranged and combined with the robot's installation position and path.

7. The method for programming a welding robot teaching-free system for arbitrarily placing workpieces according to claim 6, characterized in that: S3: planning the welding path, determining the welding start and end point coordinates and the robot positioning method according to the production cycle efficiency considerations, further includes: According to the characteristic type of the weld, the reasonable welding process and posture of the robot are planned for each weld. The robot posture and the positioning method for the robot to accurately locate the arc starting point are determined by presetting the welding process and moving trajectory.

8. The method for programming a welding robot teaching-free system for arbitrarily placing workpieces according to claim 1, characterized in that: S4: matching robot motion instructions through the robot posture library and the weld type features, matching the external axis position according to the weld position and coordinate system transformation, including: According to the starting and ending coordinates of the weld, the coordinate system of the robot coordinate system and the workpiece coordinate system are changed through the preset robot posture database to obtain the precise position of the robot's external axis.

9. A welding robot teaching-free system, characterized in that: The system is used to execute the method according to any one of claims 1 to 8.

10. The teaching-free welding robot system according to claim 9, characterized in that: The system comprises: 3D camera, used to take pictures and obtain the overall point cloud of the scene; A point cloud processing unit is used to process the point cloud and perform weld identification and clustering according to the features and connectivity of the workpiece point cloud, and obtain the weld feature type and position; Path planning unit, used to plan welding paths according to production cycle efficiency, determine welding start and end point coordinates and robot positioning method; The robot motion instruction matching unit is used to match the robot motion instruction with the weld type feature through the robot posture library; and to match the external axis position according to the weld position and coordinate system transformation; The host computer is used to generate robot program files and issue welding operations.

Images