Workpiece processing method, apparatus, system, storage medium, and computer program product

Abstract

The present disclosure relates to a workpiece processing method, device, system, storage medium and computer program product, and relates to the field of intelligent manufacturing. The workpiece processing method comprises: in a body coordinate system of a workpiece processing device, according to model point cloud data of the workpiece, performing multiple point cloud registrations on acquired measurement point cloud data of the workpiece to obtain multiple transformation matrices between the measurement point cloud data and the model point cloud data, the measurement point cloud data comprising a first reference point on the workpiece determined in a body coordinate system of the workpiece; processing the first reference point according to the multiple transformation matrices to obtain a second reference point; updating the body coordinate system of the workpiece with the second reference point as the origin; and determining position information of a processing path in the updated body coordinate system of the workpiece, for controlling the workpiece processing device to process the workpiece. The technical solution of the present disclosure can reduce the positioning error of the workpiece, thereby improving the processing precision.

Description

Technical Field

[0001] This disclosure relates to the field of intelligent manufacturing, and in particular to a workpiece processing method, workpiece processing device, workpiece processing system, computer-readable storage medium, and computer program product. Background Technology

[0002] In the field of intelligent manufacturing, high-precision grinding of various workpieces is a core process for ensuring product quality. These workpieces are generally characterized by their large size, heavy weight, complex surface structure, and irregular weld seam morphology. When ordinary tooling is used to fix the workpieces, positioning errors and posture deviations are inevitable.

[0003] In related technologies, in order to improve the positioning accuracy of workpieces, feature points of the workpieces are extracted by taking a single photograph, and the workpiece pose is then calculated based on these points. Summary of the Invention

[0004] The inventors of this disclosure have discovered the following problems in the above-mentioned related technologies: the actual error in workpiece positioning is large, making it difficult to achieve high-precision machining.

[0005] In view of this, this disclosure proposes a workpiece machining method that can reduce workpiece positioning error and thus improve machining accuracy.

[0006] According to some embodiments of this disclosure, a workpiece machining method is provided, comprising: in the volume coordinate system of a workpiece machining device, performing multiple point cloud registrations on acquired measurement point cloud data of the workpiece based on model point cloud data of the workpiece to obtain multiple transformation matrices between the measurement point cloud data and the model point cloud data, wherein the measurement point cloud data includes a first reference point determined on the workpiece in the volume coordinate system of the workpiece; processing the first reference point according to the multiple transformation matrices to obtain a second reference point; updating the volume coordinate system of the workpiece with the second reference point as the origin; and determining the position information of the machining path in the updated volume coordinate system of the workpiece for controlling the workpiece machining device to machine the workpiece.

[0007] In some embodiments, performing multiple point cloud registrations on the acquired measurement point cloud data of the workpiece based on the model point cloud data of the workpiece includes: acquiring the current measurement point cloud data of the workpiece in the first pose of the workpiece processing device, the current measurement point cloud data including a first reference point; registering the current measurement point cloud data with the model point cloud data to obtain the current transformation matrix; updating the first pose based on the current transformation matrix; repeating the above steps until the maximum number of iterations is reached, and outputting multiple transformation matrices.

[0008] In some embodiments, updating the first pose according to the current transformation matrix includes updating the first pose and the first reference point according to the current transformation matrix.

[0009] In some embodiments, updating the first pose and the first reference point according to the current transformation matrix includes updating the body coordinate system of the workpiece according to the updated first reference point.

[0010] In some embodiments, updating the first pose and the first reference point according to the current transformation matrix includes: updating the first pose according to the product of the current transformation matrix and the current coordinate information of the first pose; and updating the first reference point according to the product of the current transformation matrix and the current coordinate information of the first reference point.

[0011] In some embodiments, registering the current measured point cloud data with the model point cloud data to obtain the current transformation matrix includes: registering the current measured point cloud data and the model point cloud data according to the feature information of the current measured point cloud data to obtain the feature point correspondence between the current measured point cloud data and the model point cloud data; and determining the current transformation matrix according to the feature point correspondence.

[0012] In some embodiments, determining the current transformation matrix based on the correspondence of feature points includes: obtaining the current transformation matrix by using an iterative nearest-point algorithm based on the correspondence of feature points.

[0013] In some embodiments, acquiring current measurement point cloud data of a workpiece in a first pose of the workpiece processing device, wherein the current measurement point cloud data includes a first reference point, includes preprocessing the current measurement point cloud data, wherein the preprocessing includes at least one of filtering, noise reduction, and downsampling.

[0014] In some embodiments, the first reference point includes the starting point of the workpiece machining path.

[0015] In some embodiments, processing a first reference point according to multiple transformation matrices to obtain a second reference point includes: fusing the multiple transformation matrices to obtain a composite transformation matrix; and processing the first reference point according to the composite transformation matrix to obtain the second reference point.

[0016] In some embodiments, the workpiece processing method further includes: in response to a mode selection instruction, determining whether the current operating mode of the workpiece processing device is a teaching mode or an offline programming mode, wherein the teaching mode includes controlling the workpiece processing actuator to run along the processing path via a teaching pendant, and the offline programming mode includes controlling the workpiece processing actuator to run along the processing path according to offline generated code; and processing the workpiece according to the current operating mode.

[0017] According to some other embodiments of this disclosure, a workpiece processing apparatus is provided, comprising: a point cloud registration module, configured to perform multiple point cloud registrations on the acquired measurement point cloud data of the workpiece based on the model point cloud data of the workpiece in the volume coordinate system of the workpiece processing apparatus, to obtain multiple transformation matrices between the measurement point cloud data and the model point cloud data, wherein the measurement point cloud data includes a first reference point determined on the workpiece in the volume coordinate system of the workpiece; a processing module, configured to process the first reference point according to the multiple transformation matrices to obtain a second reference point; an updating module, configured to update the volume coordinate system of the workpiece with the second reference point as the origin; and a determining module, configured to determine the position information of the processing path in the updated volume coordinate system of the workpiece, for controlling the workpiece processing apparatus to process the workpiece.

[0018] In some embodiments, the point cloud registration module is used to acquire the current measurement point cloud data of the workpiece in the first pose of the workpiece processing device. The current measurement point cloud data includes a first reference point. The current measurement point cloud data is registered with the model point cloud data to obtain the current transformation matrix. The first pose is updated according to the current transformation matrix. The above steps are repeated until the maximum number of iterations is reached, and multiple transformation matrices are output.

[0019] In some embodiments, the point cloud registration module is used to update the first pose and the first reference point according to the current transformation matrix.

[0020] In some embodiments, the point cloud registration module is used to update the volume coordinate system of the workpiece based on the updated first reference point.

[0021] In some embodiments, the point cloud registration module is used to update the first pose based on the product of the current transformation matrix and the current coordinate information of the first pose; and to update the first reference point based on the product of the current transformation matrix and the current coordinate information of the first reference point.

[0022] In some embodiments, the point cloud registration module is used to register the current measured point cloud data and the model point cloud data according to the feature information of the current measured point cloud data and the feature information of the model point cloud data, so as to obtain the feature point correspondence between the current measured point cloud data and the model point cloud data; and determine the current transformation matrix according to the feature point correspondence.

[0023] In some embodiments, the point cloud registration module is used to obtain the current transformation matrix based on the correspondence of feature points using an iterative nearest-point algorithm.

[0024] In some embodiments, the point cloud registration module is used to preprocess the current measurement point cloud data, and the preprocessing includes at least one of filtering, noise reduction, and downsampling.

[0025] In some embodiments, the first reference point includes the starting point of the workpiece machining path.

[0026] In some embodiments, the processing module is used to fuse multiple transformation matrices to obtain a comprehensive transformation matrix; and to process the first reference point according to the comprehensive transformation matrix to obtain a second reference point.

[0027] In some embodiments, the processing module is used to obtain a comprehensive transformation matrix based on the concatenation of multiple transformation matrices.

[0028] In some embodiments, the determining module is configured to determine, in response to a mode selection instruction, whether the current operating mode of the workpiece processing device is a teaching mode or an offline programming mode. The teaching mode includes controlling the workpiece processing actuator to run along the processing path via a teaching pendant, and the offline programming mode includes controlling the workpiece processing actuator to run along the processing path according to offline generated code. The workpiece is processed according to the current operating mode.

[0029] According to further embodiments of the present disclosure, a workpiece processing apparatus is provided, comprising: a memory; and a processor coupled to the memory, the processor being configured to execute the workpiece processing method of any of the above embodiments based on instructions stored in the memory device.

[0030] According to further embodiments of this disclosure, a workpiece processing system is provided, comprising: a workpiece processing apparatus as described in any of the above embodiments; and a fixing device for fixing the workpiece.

[0031] According to further embodiments of the present disclosure, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the workpiece processing method of any of the above embodiments.

[0032] According to further embodiments of the present disclosure, a computer program product is also provided, including instructions that, when executed by a processor, cause the processor to perform a workpiece machining method according to any of the above embodiments.

[0033] In the above embodiments, the workpiece is accurately positioned based on multiple point cloud registrations to update the workpiece's volume coordinate system, and machining is performed according to the machining path in the updated workpiece coordinate system. This reduces workpiece positioning errors, thereby improving machining accuracy. Attached Figure Description

[0034] The accompanying drawings, which form part of this specification, illustrate embodiments of this disclosure and, together with the specification, serve to explain the principles of this disclosure.

[0035] This disclosure will become clearer with reference to the accompanying drawings and the following detailed description, wherein: Figure 1 Flowcharts illustrating some embodiments of the workpiece processing method of this disclosure; Figure 2Flowcharts illustrating other embodiments of the workpiece processing method of this disclosure; Figure 3 Block diagrams illustrating some embodiments of the workpiece processing apparatus of this disclosure; Figure 4 Block diagrams illustrating other embodiments of the workpiece processing apparatus of this disclosure; Figure 5 Block diagrams showing further embodiments of the workpiece processing apparatus of this disclosure Figure 6 Block diagrams illustrating some embodiments of the workpiece machining system of this disclosure; Figure 7 Schematic diagrams showing some embodiments of the workpiece processing system of this disclosure are provided. Detailed Implementation

[0036] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the present disclosure.

[0037] At the same time, it should be understood that, for ease of description, the dimensions of the various parts shown in the accompanying drawings are not drawn according to actual scale.

[0038] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit this disclosure or its application or use.

[0039] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0040] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0041] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.

[0042] As mentioned earlier, the relevant technology has insufficient positioning accuracy for the workpiece, resulting in a deviation between the planned machining path and the actual position of the workpiece. This deviation will be directly transmitted to the machining execution stage, ultimately reducing the machining accuracy of the workpiece.

[0043] To address the aforementioned issues, this disclosure proposes a workpiece machining method that accurately positions the workpiece through multiple point cloud registrations and updates the workpiece's volume coordinate system. By dynamically updating the workpiece coordinate system, the workpiece positioning accuracy is effectively transferred to the machining path, ensuring a precise match between the machining path and the actual position of the workpiece, thereby improving machining accuracy.

[0044] Figure 1 Flowcharts illustrating some embodiments of the workpiece processing method of this disclosure are shown.

[0045] like Figure 1 As shown, in step 110, in the body coordinate system of the workpiece processing device, the obtained measurement point cloud data of the workpiece is registered multiple times based on the model point cloud data of the workpiece to obtain multiple transformation matrices between the measurement point cloud data and the model point cloud data. The measurement point cloud data includes the first reference point on the workpiece determined in the body coordinate system of the workpiece.

[0046] For example, point cloud registration involves unifying the spatial coordinates of workpiece point cloud data collected at different times and from different perspectives through matrix operations.

[0047] For example, the body coordinate system of a workpiece machining device includes the robot base coordinate system, the robot tool coordinate system, and the camera coordinate system. The robot base coordinate system is the robot's fixed reference coordinate system, used to describe the position and orientation of each joint and end effector relative to the robot body. The robot tool coordinate system is fixed to the robot's end effector, such as a grinding tool, and is used to describe the tool's position and orientation relative to the robot's end effector normal. The camera coordinate system is fixed to the camera and is used to describe the position and orientation of the object relative to the camera.

[0048] For example, measurement point cloud data is acquired using a structured light camera. The structured light camera projects a structured light grating onto the surface of a workpiece mounted on a standard fixture, according to a preset shooting range, to acquire the three-dimensional information of the workpiece surface. The model point cloud data includes point cloud data of a standard workpiece pre-stored in the workpiece processing device.

[0049] In step 120, the first reference point is processed according to multiple transformation matrices to obtain the second reference point. For example, depending on the workpiece type (such as excavator boom weld, mine car frame weld, pump truck boom weld, or casting), at a certain height at the theoretical design position of the workpiece (such as the weld start end, casting reference hole position, etc.), the workpiece processing device is guided by a teach pendant to record the coordinates and theoretical normal vector of the first reference point. The first reference point can serve as the reference for subsequent matrix transformations.

[0050] In step 130, the workpiece's volume coordinate system is updated with the second reference point as the origin. For example, the workpiece's volume coordinate system is fixed on the workpiece and is used to describe the position and orientation of each point on the workpiece relative to a fixed point on the workpiece. It can be understood that the second reference point is obtained by transforming the first reference point. The first and second reference points essentially correspond to specific points on the workpiece; only the spatial position information in the coordinate system is changed.

[0051] In step 140, the position information of the machining path is determined in the updated workpiece body coordinate system to control the workpiece machining device to process the workpiece. For example, for workpieces of the same type, the machining path is generally the same.

[0052] The above embodiments propose a workpiece machining method. Through multiple point cloud registrations, multiple transformation matrices are calculated between the measured point cloud data and the model point cloud data. Each transformation matrix includes information representing the current workpiece positioning error. By fusing these multiple matrices, the first reference point and the workpiece pose can be jointly calculated and positioned, thereby reducing the workpiece positioning error. In this way, by dynamically updating the workpiece's body coordinate system, the positioning accuracy is effectively transferred to the work trajectory, ensuring a precise match between the work trajectory and the actual workpiece pose, thus improving machining accuracy.

[0053] The following examples illustrate the process of multiple point cloud registrations in step 110.

[0054] In some embodiments, the current measurement point cloud data of the workpiece is acquired in the first pose of the workpiece processing device. The current measurement point cloud data includes a first reference point. The current measurement point cloud data is registered with the model point cloud data to obtain the current transformation matrix. The first pose is updated according to the current transformation matrix. The above steps are repeated until the maximum number of iterations is reached, and multiple transformation matrices are output.

[0055] For example, in the field of intelligent manufacturing, pose is a collective term for position and attitude. In three-dimensional space, pose precisely describes the state of a rigid body in space, typically including six degrees of freedom. Position is determined by the coordinate values ​​of the workpiece on the X, Y, and Z axes in the coordinate system, while attitude is determined by the angles of rotation of the workpiece around the X, Y, and Z axes.

[0056] For example, the first pose of the workpiece processing device is updated based on the transformation matrix obtained from each point cloud registration, and the point cloud data for the next registration is collected for registration with the model point cloud data.

[0057] In the above embodiments, an iterative method is used, referencing the measurement results of the actual pose of the workpiece in the previous point cloud registration, to adjust the first pose of the workpiece processing device. This continuously reduces the minute errors remaining after each positioning, effectively improving the workpiece positioning accuracy.

[0058] The following examples illustrate the implementation process of the iteration.

[0059] In some embodiments, the first pose and the first reference point are updated based on the current transformation matrix.

[0060] For example, based on the registration results each time, not only the first pose but also the first reference point is updated, ensuring that the pose of the workpiece processing device remains unchanged relative to the reference point. In other words, the acquisition device always takes pictures of the workpiece from a preset shooting angle. In the above embodiment, by iterating over the reference point and pose, the original first reference point is always correctly tracked regardless of how the device moves, effectively suppressing the interference of movement on the positioning of the first reference point. This ensures the stability of the first reference point throughout the entire processing flow, thereby improving workpiece positioning accuracy.

[0061] The iterative process of the workpiece coordinate system is illustrated below through some examples.

[0062] In some embodiments, the body coordinate system of the workpiece is updated based on the updated first reference point.

[0063] For example, by taking the updated first reference point as the origin and assigning the three directional coordinate values ​​of the updated first reference point to the corresponding axes of the workpiece coordinate system, the updated workpiece body coordinate system can be obtained. It can be understood that regardless of how the first reference point is updated, it always corresponds to a specific point on the workpiece; only the spatial position information of the first reference point in the coordinate system changes.

[0064] In the above embodiments, by dynamically updating the workpiece's volume coordinate system, the spatial positioning of the first reference point can be made more closely match the current actual pose, thereby reducing the system deviation introduced by workpiece displacement or attitude changes and improving the accuracy of point cloud registration. The following examples illustrate the implementation process of iterating the reference point and pose.

[0065] In some embodiments, the first pose is updated based on the product of the current transformation matrix and the current coordinate information of the first pose; the first reference point is updated based on the product of the current transformation matrix and the current coordinate information of the first reference point.

[0066] In some embodiments, the current transformation matrix is:

[0067] R1 is a 3×3 initial rotation matrix, representing the rotational attitude of the workpiece in space; T1 is a 3×1 initial translation vector, representing the positional offset of the workpiece in space.

[0068] In some embodiments, the first reference point P is subjected to homogeneous coordinate transformation using the current transformation matrix M1 to obtain the first-stage guidance start point P1, calculated using the following formula: Similarly, the first pose is updated based on the current transformation matrix M1.

[0069] The above embodiments illustrate the update process of the first pose and the first reference point. Using the first reference point as a fixed benchmark, the workpiece pose deviation after each registration is fed back to the workpiece processing device, achieving successive correction of the first pose. Furthermore, the position of the first reference point is recalibrated based on newly acquired point cloud data, ensuring that the first reference point for subsequent registrations always remains consistent with the actual workpiece pose. This update mechanism shortens the computation cycle of a single registration, substantially improving the overall solution efficiency.

[0070] The following examples illustrate the relevant processing of point cloud registration.

[0071] In some embodiments, registration is performed based on the feature information of the current measured point cloud data and the feature information of the model point cloud data to obtain the correspondence between the feature points of the current measured point cloud data and the model point cloud data; the current transformation matrix is ​​determined based on the correspondence between the feature points. For example, algorithms such as curvature analysis, scale-invariant feature transformation, and accelerated robust feature extraction can be used for feature extraction, and the feature information is used to characterize the geometric and texture features of the workpiece.

[0072] In the above embodiments, feature extraction constructs a highly discriminative feature point correspondence. High-quality feature matching can effectively reduce mismatched point pairs, accelerate the convergence process of point cloud registration, and improve solution accuracy. Therefore, it can reduce point cloud registration errors and improve workpiece positioning accuracy.

[0073] In some embodiments, the current transformation matrix is ​​obtained using an iterative nearest-point algorithm based on the correspondence of feature points. For example, the iterative nearest-point algorithm is a registration method for accurately aligning two sets of 3D point cloud data. It calculates the rotation matrix and translation vector between the two sets of 3D point cloud data. By gradually approximating the optimal solution until the error is less than a threshold or the maximum number of iterations is reached, the two sets of point cloud data are aligned in space.

[0074] For example, a combination of the iterative nearest point algorithm and the absolute orientation algorithm can be used, or the Kalman filter algorithm, particle filter algorithm, etc. can be used to replace the iterative nearest point algorithm to achieve point cloud registration.

[0075] In the above embodiments, the iterative nearest-point algorithm demonstrates strong processing capabilities for irregular welds and complex curved surface workpieces. Through this algorithm, not only is accurate conversion of 3D point cloud data achieved, but also accurate calculation of workpiece pose deviations. This reduces point cloud registration errors, thereby improving workpiece positioning accuracy.

[0076] The following examples illustrate the relevant processing of point cloud data.

[0077] In some embodiments, the current measured point cloud data is preprocessed, including at least one of filtering, noise reduction, and downsampling. For example, filtering may employ methods such as Gaussian filtering, statistical filtering, pass-through filtering, and bilateral filtering; downsampling may employ methods such as voxel grid downsampling and random sample consensus algorithm; and noise reduction may employ methods such as smoothing or outlier removal.

[0078] In the above embodiments, by preprocessing the measured point cloud data before point cloud registration, including filtering, noise reduction, and downsampling of the original point cloud data, the quality of the point cloud data is effectively enhanced. High-quality data input improves the reliability and stability of the point cloud registration algorithm, thereby improving the accuracy of point cloud registration and ultimately enhancing the workpiece positioning accuracy.

[0079] The setting of reference points is illustrated below through some examples.

[0080] In some embodiments, the first reference point includes the starting point of the workpiece machining path. For example, to simplify on-site operations, the first reference point is set as the starting point of the machining path. It is understood that the first reference point can be set as other fixed feature points on the workpiece, such as the weld end of an excavator boom, a reference hole in a mining truck frame, or a positioning pin on a pump truck boom. As long as the first reference point can be accurately extracted from the point cloud data and is fixedly associated with the workpiece, it can be used as the reference point for matrix transformation.

[0081] In the above embodiments, by using the first reference point as the starting point of the machining path, the operation is simplified while ensuring the consistency of the machining process, which is beneficial to improving machining efficiency. Furthermore, since the first reference point coincides with the starting point of the machining path, errors caused by coordinate transformation are avoided, thereby improving the machining accuracy of the workpiece.

[0082] In some embodiments, multiple transformation matrices are fused to obtain a composite transformation matrix; based on the composite transformation matrix, a first reference point is processed to obtain a second reference point. For example, the composite transformation matrix is ​​obtained by weighted averaging of multiple transformation matrices.

[0083] In the above embodiments, a cascaded correction matrix link is constructed by fusing multiple transformation matrices. Each matrix operation corrects the residual error of the previous registration, forming a progressive error elimination strategy with step-by-step convergence, thereby reducing the positioning error caused by ordinary tooling.

[0084] The following examples illustrate the working mode of the workpiece processing apparatus of this disclosure.

[0085] In some embodiments, in response to a mode selection instruction, the current operating mode of the workpiece processing apparatus is determined to be either a teaching mode or an offline programming mode. The teaching mode includes controlling the workpiece processing actuator to run along the processing path via a teaching pendant, while the offline programming mode includes controlling the workpiece processing actuator to run along the processing path based on offline generated code. The workpiece is then processed according to the current operating mode. For example, this disclosure provides two programming methods that can be flexibly selected based on the on-site operation scenario; both methods are executed based on the updated workpiece coordinate system.

[0086] In some embodiments, in teach-in mode, the workpiece processing actuator is controlled by a teach pendant to move along a preset path and the coordinate information of key points is recorded. Motion parameters, such as speed, acceleration, and grinding pressure, are adjusted according to actual conditions, and the program is edited and verified. In response to the expected results being met, the workpiece processing device begins the grinding operation.

[0087] In some embodiments, the offline programming mode utilizes offline programming software to generate precise operation instructions for the workpiece machining actuator, ensuring that the trajectory conforms to actual requirements. A simulation environment is established by importing the workpiece model data and coordinate system information. Then, a preset machining path is drawn within the simulation environment, ultimately generating the corresponding control program. The simulation process includes detailed checks and improvements to the control program. Finally, the control program is imported into the workpiece machining device for verification and the machining task is initiated.

[0088] In the above embodiments, the operating mode of the workpiece processing device can be flexibly selected according to the on-site operation scenario. The offline programming mode is suitable for efficient execution in large-scale batch production, while the teaching mode is suitable for workpiece processing that requires precise technology. By flexibly selecting different programming modes, the production efficiency and accuracy requirements under different application scenarios can be significantly improved.

[0089] In some of the embodiments below, the flow of a workpiece processing method is illustrated by way of example.

[0090] Figure 2 Flowcharts illustrating some other embodiments of the workpiece processing method of this disclosure are shown.

[0091] like Figure 2 As shown, in step 210, preliminary preparation and initial trajectory planning are carried out.

[0092] In some embodiments, the workpiece processing apparatus is equipped with an industrial robot, a structured light camera, a standard adjustable fixture, and a grinding tool. Before processing, the robot tool coordinate system is calibrated to ensure calibration accuracy, and the camera's extrinsic parameters are also calibrated. Subsequently, at the theoretical starting point of the excavating arm's weld seam, the robot is guided by a teach pendant to record the first reference point and its theoretical normal vector. Based on this, the weld seam grinding trajectory is planned, such as a continuous curved surface trajectory, and the movement speed and grinding pressure are set. According to the workpiece processing requirements (such as a smooth weld seam transition or polishing of a casting surface), the initial settings for the processing path parameters, movement speed, grinding pressure, etc., are pre-planned, and the distribution range of the processing path in the workpiece's theoretical coordinate system is defined.

[0093] For example, the coordinates of the first reference point are determined as follows:

[0094] Theoretical normal vector of the first reference point:

[0095] The theoretical normal vector of the first reference point indicates the orientation information of the workpiece in three-dimensional space at the first reference point.

[0096] In step 220, a first-stage point cloud data acquisition and transformation matrix solution is performed.

[0097] In some embodiments, to improve efficiency in actual operation, point cloud registration is performed twice, and the processing method includes polishing.

[0098] In some embodiments, the structured light camera takes an initial photograph of the workpiece, acquiring raw point cloud data. The raw point cloud data undergoes pass-through filtering for denoising and voxel grid downsampling to obtain denoised point cloud data. Weld edge feature points are extracted from the denoised point cloud data and matched with the workpiece model point cloud data. The initial transformation matrix M1 is obtained by combining the iterative nearest-point algorithm and the absolute orientation algorithm.

[0099] Using the initial transformation matrix M1, perform a homogeneous coordinate transformation on the first reference point to calculate the updated first reference point:

[0100] Based on the updated coordinates of the first pose, the robot adjusts its pose and moves to that position, precisely guiding the structured light camera to the optimal shooting angle of the first reference point, in preparation for the second photo.

[0101] In step 230, two-stage point cloud data acquisition and transformation matrix solving are performed.

[0102] In some embodiments, the structured light camera is controlled to take a second picture at a preset shooting position, and the structured light grating is projected again to accurately acquire updated 3D point cloud data of the first reference point and the surrounding working area, providing data support for high-precision feature extraction. Subsequently, the acquired measurement point cloud data undergoes refined preprocessing, employing a bilateral filtering algorithm to further remove noise points while preserving feature details, such as weld fusion lines and surface textures of castings. High-precision feature points of the workpiece are extracted based on curvature analysis, and the processed measurement point cloud data is registered with the model point cloud data to eliminate minor residual errors from the initial positioning.

[0103] In some embodiments, the iterative nearest-point algorithm is used to obtain the homogeneous transformation matrix M2 describing the deviation between the actual and theoretical poses of the workpiece:

[0104] In step 240, the final reference point and workpiece coordinate system update are determined.

[0105] In some embodiments, the comprehensive transformation matrix M_final is obtained based on the above transformation matrices M1 and M2: M_final=M2×M1=

[0106] The final second reference point is determined based on the comprehensive transformation matrix. :

[0107] Using the final second reference point P_start_final as the origin, and combining the attitude information obtained by solving the comprehensive transformation matrix M_final, the body coordinate system of the workpiece is updated.

[0108] In step 250, the grinding trajectory is programmed and the high-precision operation is executed.

[0109] In some embodiments, a teach pendant can guide the robot to move along the planned curved weld path in the updated workpiece coordinate system, recording a turning point at regular intervals to capture the coordinate information of key points. Parameters such as the robot's motion speed, acceleration, and grinding pressure are set, and the trajectory smoothness is optimized to ensure a shock-free and perfectly aligned grinding process. Subsequently, the workpiece processing device starts the grinding operation until grinding is complete.

[0110] Figure 3 Block diagrams illustrating some embodiments of the workpiece processing apparatus of this disclosure are shown.

[0111] This disclosure provides a workpiece processing apparatus 3, comprising: a point cloud registration module 31, used to perform multiple point cloud registrations on the acquired measurement point cloud data of the workpiece based on the model point cloud data of the workpiece in the volume coordinate system of the workpiece processing apparatus, so as to obtain multiple transformation matrices between the measurement point cloud data and the model point cloud data, wherein the measurement point cloud data includes a first reference point determined on the workpiece in the volume coordinate system of the workpiece; a processing module 32, used to process the first reference point according to the multiple transformation matrices to obtain a second reference point; an updating module 33, used to update the volume coordinate system of the workpiece with the second reference point as the origin; and a determining module 34, used to determine the position information of the processing path in the updated volume coordinate system of the workpiece, for controlling the workpiece processing apparatus to process the workpiece.

[0112] In some embodiments, the point cloud registration module 31 is used to acquire the current measurement point cloud data of the workpiece in the first pose of the workpiece processing device. The current measurement point cloud data includes a first reference point. The current measurement point cloud data is registered with the model point cloud data to obtain the current transformation matrix. The first pose is updated according to the current transformation matrix. The above steps are repeated until the maximum number of iterations is reached, and multiple transformation matrices are output.

[0113] In some embodiments, the point cloud registration module 31 is used to update the first pose and the first reference point according to the current transformation matrix.

[0114] In some embodiments, the point cloud registration module 31 is used to update the body coordinate system of the workpiece based on the updated first reference point.

[0115] In some embodiments, the point cloud registration module 31 is used to update the first pose based on the product of the current transformation matrix and the current coordinate information of the first pose; and to update the first reference point based on the product of the current transformation matrix and the current coordinate information of the first reference point.

[0116] In some embodiments, the point cloud registration module 31 is used to register the current measured point cloud data and the model point cloud data according to the feature information of the current measured point cloud data and the feature information of the model point cloud data, so as to obtain the feature point correspondence between the current measured point cloud data and the model point cloud data; and determine the current transformation matrix according to the feature point correspondence.

[0117] In some embodiments, the point cloud registration module 31 is used to obtain the current transformation matrix based on the correspondence of feature points using an iterative nearest point algorithm.

[0118] In some embodiments, the point cloud registration module 31 is used to preprocess the current measurement point cloud data, and the preprocessing includes at least one of filtering, noise reduction, and downsampling.

[0119] In some embodiments, the first reference point includes the starting point of the workpiece machining path.

[0120] In some embodiments, the processing module 32 is used to fuse multiple transformation matrices to obtain a comprehensive transformation matrix; and to process the first reference point according to the comprehensive transformation matrix to obtain a second reference point.

[0121] In some embodiments, the processing module 32 is used to obtain a comprehensive transformation matrix based on the cascading of multiple transformation matrices.

[0122] In some embodiments, the determining module 34 is configured to determine, in response to a mode selection instruction, whether the current working mode of the workpiece processing device is a teaching mode or an offline programming mode. The teaching mode includes controlling the workpiece processing actuator to run along the processing path via a teaching pendant, and the offline programming mode includes controlling the workpiece processing actuator to run along the processing path according to offline generated code. The workpiece is processed according to the current working mode.

[0123] Various embodiments of the workpiece processing apparatus 3 can be similarly described in the various embodiments of the aforementioned workpiece processing method, and will not be repeated here.

[0124] Figure 4 Block diagrams showing other embodiments of the workpiece processing apparatus of this disclosure are shown.

[0125] like Figure 4 As shown, the workpiece processing apparatus 4 of this embodiment includes a memory 41 and a processor 42 coupled to the memory 41. The processor 42 is configured to execute the workpiece processing method of any embodiment of this disclosure based on instructions stored in the memory 41.

[0126] The memory 41 may include, for example, system memory, fixed non-volatile storage media, etc. The system memory stores, for example, the operating system, application programs, boot loader, database, and other programs.

[0127] Figure 5 Block diagrams showing further embodiments of the workpiece processing apparatus of this disclosure are shown.

[0128] like Figure 5 As shown, the workpiece processing apparatus 5 of this embodiment includes a memory 510 and a processor 520 coupled to the memory 510. The processor 520 is configured to execute the workpiece processing method of any of the foregoing embodiments based on instructions stored in the memory 510.

[0129] The memory 510 may include, for example, system memory, fixed non-volatile storage media, etc. The system memory may store, for example, the operating system, application programs, boot loader, and other programs.

[0130] The workpiece processing device 5 may also include an input / output interface 530, a network interface 540, and a storage interface 550. These interfaces 530, 540, and 550, as well as the memory 510 and processor 520, can be connected via, for example, a bus 560. The input / output interface 530 provides a connection interface for input / output devices such as monitors, mice, keyboards, touchscreens, microphones, and speakers. The network interface 540 provides a connection interface for various networked devices. The storage interface 550 provides a connection interface for external storage devices such as SD cards and USB flash drives.

[0131] Figure 6 Block diagrams illustrating some embodiments of the workpiece processing system of this disclosure are shown.

[0132] like Figure 6 As shown, the workpiece processing system 6 includes the workpiece processing device 61 and the fixing device 62 in any of the above embodiments, for fixing the workpiece.

[0133] In some embodiments, the fixing device 62 includes general-purpose tooling. This tooling is suitable for clamping and positioning various types of engineering machinery workpieces without requiring high-precision positioning capabilities. The general-purpose tooling is designed with an adjustable structure, offering strong versatility and low cost, and can adapt to different specifications and types of engineering machinery workpieces.

[0134] Figure 7 Schematic diagrams showing some embodiments of the workpiece processing system of this disclosure are provided.

[0135] like Figure 7 As shown, the workpiece processing system 7 includes an industrial robot 71, a data acquisition device mounting bracket 72 and data acquisition device 73, and a grinding actuator 74.

[0136] In some embodiments, the acquisition device 73 includes a structured light camera. The structured light camera has high resolution and can operate stably under complex lighting conditions in a workshop and with strong reflections from the workpiece.

[0137] This disclosure also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the workpiece processing method of any of the above embodiments.

[0138] This disclosure also provides a computer program product including instructions that, when executed by a processor, cause the processor to perform a workpiece processing method according to any of the above embodiments.

[0139] Those skilled in the art will understand that embodiments of this disclosure can be provided as methods, systems, or computer program products. Therefore, this disclosure can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this disclosure can take the form of a computer program product embodied on one or more computer-usable non-transitory storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0140] So far, some embodiments of the workpiece processing method according to this disclosure have been described in detail. To avoid obscuring the concept of this disclosure, some details known in the art have not been described. Those skilled in the art can fully understand how to implement the technical solutions disclosed herein based on the above description.

[0141] The methods and systems of this disclosure may be implemented in many ways. For example, they may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above-described order of steps for the methods is for illustrative purposes only, and the steps of the methods of this disclosure are not limited to the specific order described above unless otherwise specifically stated. Furthermore, in some embodiments, this disclosure may also be implemented as a program recorded on a recording medium, the program including machine-readable instructions for implementing the methods according to this disclosure. Thus, this disclosure also covers recording media storing programs for performing the methods according to this disclosure.

[0142] While specific embodiments of this disclosure have been described in detail by way of examples, those skilled in the art should understand that the examples are for illustrative purposes only and are not intended to limit the scope of this disclosure. Those skilled in the art should understand that modifications can be made to the above embodiments without departing from the scope and spirit of this disclosure.

Claims

1. A workpiece processing method, comprising: In the body coordinate system of the workpiece processing device, based on the model point cloud data of the workpiece, the acquired measurement point cloud data of the workpiece is registered multiple times to obtain multiple transformation matrices between the measurement point cloud data and the model point cloud data. The measurement point cloud data includes a first reference point on the workpiece determined in the body coordinate system of the workpiece. The first reference point is processed according to the plurality of transformation matrices to obtain the second reference point; The body coordinate system of the workpiece is updated with the second reference point as the origin; In the updated body coordinate system of the workpiece, the position information of the machining path is determined, which is used to control the workpiece machining device to process the workpiece.

2. The workpiece processing method according to claim 1, wherein performing multiple point cloud registrations on the acquired measurement point cloud data of the workpiece based on the model point cloud data of the workpiece includes: In the first pose of the workpiece processing device, the current measurement point cloud data of the workpiece is acquired, and the current measurement point cloud data includes the first reference point; The current measured point cloud data is registered with the model point cloud data to obtain the current transformation matrix; Update the first pose based on the current transformation matrix; Repeat the above steps until the maximum number of iterations is reached, and output the multiple transformation matrices.

3. The workpiece processing method according to claim 2, wherein updating the first pose according to the current transformation matrix includes: Update the first pose and the first reference point based on the current transformation matrix.

4. The workpiece machining method according to claim 3, wherein updating the first pose and the first reference point according to the current transformation matrix includes: The body coordinate system of the workpiece is updated based on the updated first reference point.

5. The workpiece machining method according to claim 3, wherein updating the first pose and the first reference point according to the current transformation matrix comprises: The first pose is updated based on the product of the current transformation matrix and the current coordinate information of the first pose; The first reference point is updated based on the product of the current transformation matrix and the current coordinate information of the first reference point.

6. The workpiece processing method according to claim 2, wherein, The step of registering the current measured point cloud data with the model point cloud data to obtain the current transformation matrix includes: The registration is performed based on the feature information of the current measured point cloud data and the feature information of the model point cloud data to obtain the correspondence between the feature points of the current measured point cloud data and the model point cloud data; The current transformation matrix is ​​determined based on the correspondence of the feature points.

7. The workpiece processing method according to claim 6, wherein determining the current transformation matrix based on the feature point correspondence includes: Based on the correspondence of the feature points, the current transformation matrix is ​​obtained by using the iterative nearest point algorithm.

8. The workpiece processing method according to claim 2, wherein, The current measurement point cloud data of the workpiece is acquired in the first pose of the workpiece processing device, and the current measurement point cloud data includes the first reference point, including: The current measured point cloud data is preprocessed, and the preprocessing includes at least one of filtering, noise reduction, and downsampling.

9. The workpiece processing method according to any one of claims 1-8, wherein, The first reference point includes the starting point of the workpiece machining path.

10. The workpiece processing method according to any one of claims 1-8, wherein, The step of processing the first reference point according to the plurality of transformation matrices to obtain the second reference point includes: The multiple transformation matrices are fused to obtain a comprehensive transformation matrix; The first reference point is processed according to the comprehensive transformation matrix to obtain the second reference point.

11. The workpiece processing method according to claim 10, wherein, The process of fusing the multiple transformation matrices to obtain the comprehensive transformation matrix includes: The comprehensive transformation matrix is ​​obtained by cascading the multiple transformation matrices.

12. The workpiece processing method according to any one of claims 1-8, further comprising: In response to a mode selection command, the current working mode of the workpiece processing device is determined to be either a teaching mode or an offline programming mode. The teaching mode includes controlling the workpiece processing actuator to run along the processing path via a teaching pendant, and the offline programming mode includes controlling the workpiece processing actuator to run along the processing path based on offline generated code. The workpiece is processed according to the current working mode.

13. A workpiece processing apparatus, comprising: The point cloud registration module is used to perform multiple point cloud registrations on the acquired measurement point cloud data of the workpiece in the body coordinate system of the workpiece processing device, based on the model point cloud data of the workpiece, so as to obtain multiple transformation matrices between the measurement point cloud data and the model point cloud data, wherein the measurement point cloud data includes a first reference point determined on the workpiece. The processing module is used to process the first reference point according to the plurality of transformation matrices to obtain the second reference point; The update module is used to update the body coordinate system of the workpiece with the second reference point as the origin; The determination module is used to determine the position information of the processing path in the updated body coordinate system of the workpiece, and to control the workpiece processing device to process the workpiece.

14. A workpiece processing apparatus, comprising: Memory; and A processor coupled to the memory, the processor being configured to perform the method of any one of claims 1 to 12 based on instructions stored in the memory.

15. A workpiece machining system, comprising: The workpiece processing apparatus according to claim 13 or 14; A fixing device for fixing the workpiece.

16. A computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the method of any one of claims 1 to 12.

17. A computer program product comprising instructions that, when executed by a processor, cause the processor to perform the method according to any one of claims 1 to 12.

Images