Method, apparatus, and medium for assembling industrial parts
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- MIRACLE AUTOMATION ENG CO LTD
- Filing Date
- 2025-04-30
- Publication Date
- 2026-06-26
AI Technical Summary
In existing technologies, the accuracy of identifying complex industrial parts is poor, resulting in low precision in robot grasping and assembly, which affects the operating efficiency and reliability of automated production lines.
By detecting RGB and depth images, visual and depth features are extracted using a pre-trained object detection model. Combined with a pose estimation model, the part categories and positions are accurately identified. Based on the 6D pose, the robot's grasping posture and path are determined, enabling precise grasping and assembly.
It enables precise identification and assembly of industrial parts, improving the operational efficiency and reliability of automated production lines.
Smart Images
Figure CN120244532B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of robot control technology in industrial automation, and in particular to a method, equipment and medium for assembling industrial parts. Background Technology
[0002] In manufacturing, the assembly of complex industrial parts is a critical step in the production process. However, current technologies still rely on traditional visual recognition methods for identifying industrial components composed of multiple parts, which suffer from poor accuracy. This results in lower precision in robot grasping and assembly, impacting the operational efficiency and reliability of automated production lines. Summary of the Invention
[0003] In response to the aforementioned problems and technical requirements, the applicant proposes an assembly method, equipment, and medium for industrial parts to solve the problem of poor assembly accuracy caused by poor part identification accuracy in the prior art, thereby achieving accurate identification and assembly of industrial parts.
[0004] An embodiment of the application provides a method for assembling industrial parts, the method comprising:
[0005] The detection process includes RGB and depth images of industrial parts, obtaining the part category of each industrial part and the corresponding detection box in the RGB image. The industrial parts include at least three industrial parts.
[0006] Based on the detection box and RGB image, the segmentation mask of each industrial part is obtained, and the 3D model corresponding to each part category is obtained;
[0007] The segmentation mask, RGB image, depth image and 3D model are input into the pre-trained pose estimation model to obtain the 6D pose of each industrial part output by the pose estimation model, and the robot's grasping posture is determined based on the 6D pose. The pose estimation model is trained based on segmentation mask samples, RGB image samples, depth image samples, 3D model samples and 6D pose samples.
[0008] Based on the assembly sequence and 6D pose of industrial parts, the robot's grasping posture and grasping path are determined, and the robot is controlled to complete the grasping and assembly of industrial parts based on the grasping path and grasping posture.
[0009] According to the assembly method for industrial parts provided in the embodiments of this application, the detection includes RGB images and depth images of the industrial parts, obtaining the part category of each industrial part and the corresponding detection box in the RGB image, including:
[0010] The RGB image and the depth image are input into a pre-trained object detection model. The object detection model extracts visual features from the RGB image and depth features from the depth image. The visual features and the depth features are fused to obtain fused features. Based on the fused features, the part category and detection box output by the object detection model are obtained and output.
[0011] The target detection model is trained based on RGB image samples, depth image samples, part category samples, and detection box samples.
[0012] The visual features include color features, edge features, and texture features.
[0013] According to the assembly method for industrial parts provided in the embodiments of this application, after extracting visual features from the RGB image and depth features from the depth image using a target detection model, the method further includes:
[0014] Determine whether the visual feature belongs to a known feature;
[0015] If it is determined that the visual feature belongs to a known feature, the step of fusing the visual feature and the depth feature to obtain the fused feature is performed.
[0016] If the visual feature is determined not to be a known feature, the target detection model is optimized based on the incremental learning algorithm, and the optimized target detection model is used to detect and output the part category and detection box.
[0017] According to the assembly method for industrial parts provided in the embodiments of this application, before determining the robot's grasping path based on the assembly sequence and 6D pose of the industrial parts, the method further includes:
[0018] Obtain the assembly requirements corresponding to industrial components;
[0019] Based on assembly requirements and part categories, determine the assembly sequence and assembly location of each industrial part, wherein the assembly location is the position of a certain industrial part.
[0020] Based on the assembly sequence and 6D pose of industrial parts, the robot's grasping path is determined, including:
[0021] Based on the assembly sequence and 6D pose, a gripping path is determined for other industrial parts to be gripped and placed at the assembly position, wherein the other industrial parts are the remaining industrial parts after removing the industrial parts at the assembly position from the industrial components.
[0022] According to the assembly method of industrial parts provided in the embodiments of this application, different industrial parts correspond to different gripping postures and different gripping paths.
[0023] The robot, controlled by grasping path and grasping posture, completes the grasping of industrial parts, including:
[0024] Perform the following gripping operation on each other industrial part:
[0025] The robot is controlled to grasp other industrial parts using the current grasping posture corresponding to the current other industrial parts; the robot is controlled to place other industrial parts at the assembly position using the current grasping path corresponding to the current other industrial parts.
[0026] Determine whether all other industrial parts in the industrial component have been picked up; if so, determine that the industrial component is assembled; otherwise, determine the next other industrial part based on the assembly sequence and perform a picking operation on the next other industrial part.
[0027] The other industrial parts are the remaining industrial parts after removing the industrial parts at the assembly position from the industrial components.
[0028] The assembly method for industrial parts provided in the embodiments of this application further includes:
[0029] During the process of gripping industrial parts by controlling the gripping path and gripping posture, the gripping force and displacement of the robot are monitored in real time.
[0030] Based on the grasping force, the displacement, and the 6D pose of the industrial part at the assembly position, the robot's grasping pose is adjusted in real time.
[0031] According to the assembly method for industrial parts provided in the embodiments of this application, controlling a robot to complete the assembly of industrial parts includes:
[0032] Determine the assembly connection positions between other current industrial parts and the target industrial part;
[0033] The assembly position is determined based on the assembly connection position, the current pose of other industrial parts, and the 6D pose of the target industrial part.
[0034] The robot is controlled to place other industrial parts in the assembly position and complete the assembly of other industrial parts.
[0035] The other industrial parts are the remaining industrial parts after removing the industrial parts at the assembly position from the industrial components, and the industrial parts at the assembly position are the target industrial parts.
[0036] The assembly method for industrial parts according to the embodiments of this application, after obtaining the detection frame based on the fusion features, further includes:
[0037] Redundant detection boxes are removed based on the nonmaximum suppression algorithm.
[0038] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the assembly method of the industrial parts as described above.
[0039] This application also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the assembly method for industrial parts as described above.
[0040] The assembly method, equipment, and medium for industrial parts provided in this application obtain the part category, corresponding detection box in the RGB image, and segmentation mask for each industrial part by detecting RGB and depth images containing industrial parts. The industrial parts include at least three industrial parts. A 3D model corresponding to each part category is obtained. The segmentation mask, RGB image, depth image, and 3D model are input into a pre-trained pose estimation model to obtain the 6D pose of each industrial part output by the pose estimation model. The robot's grasping posture is determined based on the 6D pose. Therefore, this application can accurately obtain the position and posture (6D pose) of each industrial part. Furthermore, based on the assembly sequence and 6D pose of the industrial parts, the robot's grasping posture and grasping path are determined. The robot is then controlled to complete the grasping and assembly of the industrial parts based on the grasping path and grasping posture. Thus, this application accurately determines the corresponding grasping posture and grasping path for each industrial part, ensuring accurate identification, grasping, and assembly of each industrial part, further improving the operating efficiency and reliability of automated production lines. Attached Figure Description
[0041] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0042] Figure 1 This is one of the flowcharts illustrating the assembly method of industrial parts provided in the embodiments of this application;
[0043] Figure 2 This is a second schematic flowchart of the assembly method for industrial parts provided in the embodiments of this application;
[0044] Figure 3 This is the third flowchart illustrating the assembly method for industrial parts provided in the embodiments of this application;
[0045] Figure 4 This is the fourth flowchart illustrating the assembly method for industrial parts provided in the embodiments of this application;
[0046] Figure 5 This is the fifth flowchart illustrating the assembly method for industrial parts provided in the embodiments of this application;
[0047] Figure 6 This is the sixth flowchart illustrating the assembly method for industrial parts provided in the embodiments of this application;
[0048] Figure 7 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation
[0049] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.
[0050] This application provides an assembly method for industrial parts. This method can be applied to smart terminals and servers. This application uses the application of this method in a server as an example for illustration; this is for illustrative purposes only and is not intended to limit the scope of protection of this application. Furthermore, some other descriptions in the embodiments are also illustrative and will not be described in detail thereafter. The specific implementation of this method is as follows: Figure 1 As shown:
[0051] Step 101: Detect RGB and depth images containing industrial parts to obtain the part category of each industrial part and the corresponding detection box in the RGB image.
[0052] The industrial components include at least three industrial parts.
[0053] Step 102: Obtain the segmentation mask for each industrial part based on the detection box and RGB image, and obtain the 3D model corresponding to each part category.
[0054] Step 103: Input the segmentation mask, RGB image, depth image and 3D model into the pre-trained pose estimation model to obtain the 6D pose of each industrial part output by the pose estimation model, and determine the robot's grasping posture based on the 6D pose.
[0055] The pose estimation model is trained based on segmentation mask samples, RGB image samples, depth image samples, 3D model samples, and 6D pose samples.
[0056] Among them, the pose estimation model includes the FoundationPose model, which is used for 6D pose estimation and tracking.
[0057] Step 104: Based on the assembly sequence and 6D pose of the industrial parts, determine the robot's grasping posture and grasping path, and control the robot to complete the grasping and assembly of the industrial parts based on the grasping path and grasping posture.
[0058] The assembly method for industrial parts provided in this application involves detecting RGB and depth images containing industrial components to obtain the component category, the corresponding detection box in the RGB image, and the segmentation mask for each industrial component. The industrial components include at least three components. A 3D model corresponding to each component category is obtained. The segmentation mask, RGB image, depth image, and 3D model are input into a pre-trained pose estimation model to obtain the 6D pose of each industrial component output by the pose estimation model. Based on the 6D pose, the robot's grasping posture is determined. Therefore, this application can accurately obtain the position and posture (6D pose) of each industrial component. Furthermore, based on the assembly sequence and 6D pose of the industrial components, the robot's grasping posture and grasping path are determined. Based on the grasping path and grasping posture, the robot is controlled to complete the grasping and assembly of the industrial components. Thus, this application accurately determines the corresponding grasping posture and grasping path for each industrial component, ensuring accurate identification, grasping, and assembly of each industrial component, further improving the operational efficiency and reliability of automated production lines.
[0059] In one specific embodiment, the detection includes RGB images and depth images of industrial parts, and the specific implementation of obtaining the part category of each industrial part and the corresponding detection box in the RGB image is as follows: Figure 2 As shown:
[0060] Step 201: Input the RGB image and depth image into the pre-trained object detection model.
[0061] Step 202: Extract visual features from the RGB image and depth features from the depth image using the object detection model; fuse the visual features and depth features to obtain fused features; and obtain and output the part category and detection box output by the object detection model based on the fused features.
[0062] The target detection model was trained based on RGB image samples, depth image samples, part category samples, and detection box samples.
[0063] Visual features include color features, edge features, and texture features.
[0064] Specifically, the object detection model is trained using a mixed-precision training method and a distributed training method. The mixed-precision training method balances computational speed and accuracy, while the distributed training method improves training efficiency and scalability. The combination of these two training methods ensures the accuracy of the object detection model.
[0065] Specifically, the object detection model includes the YOLOv8 model.
[0066] Specifically, this application makes some improvements to the model structure of the existing YOLOv8 model. A convolutional block attention module is introduced in the neck structure to enhance feature representation through spatial and channel attention and improve feature fusion effect. The backbone network adopts the CAFormer module to enhance feature extraction capability, adapt to complex industrial scenarios, and improve the accuracy of model prediction.
[0067] In one specific embodiment, after obtaining the detection boxes based on the fused features, redundant detection boxes are removed based on the non-maximum suppression algorithm.
[0068] Specifically, during the prediction process, the target detection model may generate multiple detection boxes for the same industrial part. Based on feature fusion, multiple corresponding detection boxes can be obtained. Furthermore, a non-maximum suppression algorithm is used to remove redundant detection boxes, specifically by suppressing detection boxes with high overlap, thereby improving the accuracy and efficiency of the detection results.
[0069] Specifically, the probability of an industrial part being included in the detection frame is determined, and the probabilities are sorted from high to low; the highest-ranked frame is used as the reference frame, and the intersection-union ratio (IUR) between the reference frame and each detection frame is calculated; detection frames with IUR greater than a preset IUR are deleted.
[0070] In one specific embodiment, after extracting visual features from the RGB image and depth features from the depth image using an object detection model, part categories and detection boxes are predicted, specifically as follows: Figure 3 As shown:
[0071] Step 301: Determine whether the visual feature belongs to the known features. If yes, proceed to step 302; otherwise, proceed to step 303.
[0072] Step 302: Fuse the visual features and the depth features to obtain the fused features.
[0073] Furthermore, based on the fusion features, the part categories and detection boxes output by the target detection model are obtained and output.
[0074] Step 303: Optimize the target detection model based on the incremental learning algorithm, and use the optimized target detection model to detect and output the part category and detection box.
[0075] Specifically, when it is determined that the visual features do not belong to the known features, the target detection model is optimized based on the incremental learning algorithm, visual features and deep features, and the visual features and the deep features are fused to obtain fused features; and the optimized target detection model and fused features are used to detect and output the part category and detection box.
[0076] Among them, known features are those that the object detection model has seen during the training phase.
[0077] Specifically, incremental learning algorithms include: neural networks that employ self-organizing incremental learning.
[0078] This application, when visual features are determined to be unrecognized, collects new data samples for online learning and updates the model parameters of the object detection model in real time. Through an incremental learning algorithm, the weights of the object detection model are dynamically adjusted, enabling the model to adapt more quickly to new environments and changes in workpieces. By continuously adapting to new environments and workpiece changes in this way, the accuracy and robustness of detection are effectively improved.
[0079] In one specific embodiment, before determining the robot's grasping path based on the assembly sequence and 6D pose of the industrial parts, the assembly requirements corresponding to the industrial parts are obtained; based on the assembly requirements and part categories, the assembly sequence and assembly position of each industrial part are determined. Furthermore, based on the assembly sequence and 6D pose, the grasping path for other industrial parts to be grasped and placed at the assembly positions is determined.
[0080] The assembly location refers to the position of a specific industrial part.
[0081] Other industrial parts are the industrial parts remaining after removing the industrial parts at the assembly position from the industrial components.
[0082] In one specific embodiment, different industrial parts correspond to different gripping postures and different gripping paths.
[0083] For a detailed implementation of the robot's grasping of industrial parts based on grasping path and grasping posture control, please refer to [link to relevant documentation]. Figure 4 :
[0084] Step 401: Perform the following gripping operation on each other industrial part: control the robot to grip the current other industrial part using the current gripping posture corresponding to the current other industrial part; control the robot to place the current other industrial part at the assembly position using the current gripping path corresponding to the current other industrial part.
[0085] Step 402: Determine whether all other industrial parts in the industrial component have been successfully grabbed. If yes, proceed to step 403; otherwise, proceed to step 404.
[0086] Step 403: Confirm that the industrial component assembly is complete.
[0087] Here, the completion of the assembly of the industrial component refers to the completion of the current assembly of the industrial component.
[0088] Step 404: Determine the next other industrial part based on the assembly sequence, and perform a grab operation on the next other industrial part.
[0089] Other industrial parts are the industrial parts remaining after removing the industrial parts at the assembly position from the industrial components.
[0090] This application can simultaneously identify the position and orientation of multiple industrial parts, and plan the gripping path of each industrial part and control the robot based on the identified position and orientation during the gripping and assembly process, so as to achieve fast and accurate identification, gripping and assembly of multiple industrial parts.
[0091] In one specific embodiment, the method also requires real-time adjustment of the robot's grasping pose, as detailed below. Figure 5 As shown:
[0092] Step 501: During the process of the robot grasping the industrial parts based on the grasping path and grasping posture, the grasping force and displacement of the robot during the grasping operation are monitored in real time.
[0093] Step 502: Based on the gripping force, movement displacement, and 6D pose of the industrial part at the assembly position, adjust the robot's gripping pose in real time.
[0094] In this context, the grasping process refers to the robot grasping industrial parts and placing them into the assembly position.
[0095] Specifically, during the gripping process, the gripping position (gripping posture) may change due to the robot's movement and the friction between the robot and the industrial parts. Therefore, the gripping force and displacement are monitored in real time during this process, and the robot's gripping posture is adjusted in real time based on the gripping force, displacement, and 6D pose of the industrial parts at the assembly position to ensure the stability and safety of the gripping process and to ensure the precise assembly between industrial parts.
[0096] In one specific embodiment, the specific implementation of controlling the robot to complete the assembly of industrial parts is as follows: Figure 6 As shown:
[0097] Step 601: Determine the assembly connection positions between the current other industrial parts and the target industrial part.
[0098] Step 602: Determine the assembly position based on the assembly connection position, the current pose of other industrial parts, and the 6D pose of the target industrial part.
[0099] Step 603: Control the robot to place other industrial parts in the assembly position and complete the assembly of other industrial parts.
[0100] Among them, other industrial parts are the industrial parts remaining after removing the industrial parts at the assembly position from the industrial components, and the industrial parts at the assembly position are the target industrial parts.
[0101] Specifically, the assembly connection positions of other industrial parts and the target industrial part are determined based on a preset visual recognition model. The assembly connection positions of different other industrial parts and the target industrial part are different.
[0102] Furthermore, other industrial parts are varied in this application, and the target industrial parts are also varied.
[0103] For example, this industrial component consists of three industrial parts, A, B, and C. The assembly sequence is as follows: assemble A onto C, and then assemble B onto A (or a combination of A and C). In the process of assembling A onto C, other industrial parts are A, and the target industrial part is C; in the process of assembling B onto A (or a combination of A and C), other industrial parts are B, and the target industrial part is A (or a combination of A and C).
[0104] For example, this industrial component consists of four industrial parts: A, B, C, and D. The assembly sequence is as follows: assemble A onto C, then assemble B onto A (or a combination of A and C), and finally assemble D onto B (or a combination of B, A, and C). In the process of assembling A onto C, other industrial parts are A, and the target industrial part is C; in the process of assembling B onto A (or a combination of A and C), other industrial parts are B, and the target industrial part is A (or a combination of A and C); in the process of assembling D onto B (or a combination of B, A, and C), other industrial parts are D, and the target industrial part is B (or a combination of B, A, and C).
[0105] Specifically, as the assembly work progresses, the assembly position will change. Therefore, the assembly process of each other industrial part needs to redetermine the assembly position and perform the assembly operation of the industrial part based on the assembly position and the corresponding gripping path.
[0106] This application solves the problem of accurate identification, grasping and assembly of complex industrial parts by using multimodal feature fusion, deep learning algorithms and pose estimation models, thereby improving the assembly efficiency and reliability of robots in complex industrial scenarios.
[0107] Specifically, by fusing multimodal features and using deep learning algorithms, the robot achieves precise localization and recognition of complex parts. A pre-trained SAM model enables rapid segmentation of complex industrial parts without additional training, obtaining pixel-level positional information. The FoundationPose model acquires 6D pose data of the workpiece. Furthermore, the robot's grasping posture is adjusted specifically to the workpiece's characteristics, thereby controlling the robot to achieve precise recognition and assembly of complex industrial parts.
[0108] Figure 7 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 7 As shown, the electronic device may include a processor 701, a communications interface 702, a memory 703, and a communication bus 704. The processor 701, communications interface 702, and memory 703 communicate with each other via the communication bus 704. The processor 701 can call logical instructions from the memory 703 to execute assembly methods for industrial parts.
[0109] Furthermore, the logical instructions in the aforementioned memory 703 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0110] On the other hand, the present invention also provides a computer program product, the computer program product including a computer program stored on a non-transitory computer-readable storage medium, the computer program including program instructions, and when the program instructions are executed by a computer, the computer is able to execute the assembly method of industrial parts provided by the above methods.
[0111] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the assembly method of the industrial parts provided in the above embodiments.
[0112] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.
[0113] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.
[0114] Finally, it should be noted that the above descriptions are merely preferred embodiments of this application, and this application is not limited to the above embodiments. It is understood that other improvements and variations directly derived or conceived by those skilled in the art without departing from the spirit and concept of this application should be considered to be included within the protection scope of this application.
Claims
1. A method for assembling industrial parts, characterized in that, The method includes: The detection process includes RGB and depth images of industrial parts, obtaining the part category of each industrial part and the corresponding detection box in the RGB image. The industrial parts include at least three industrial parts. Based on the detection box and RGB image, the segmentation mask of each industrial part is obtained, and the 3D model corresponding to each part category is obtained; The segmentation mask, RGB image, depth image and 3D model are input into the pre-trained pose estimation model to obtain the 6D pose of each industrial part output by the pose estimation model, and the robot's grasping posture is determined based on the 6D pose. The pose estimation model is trained based on segmentation mask samples, RGB image samples, depth image samples, 3D model samples and 6D pose samples. Obtain the assembly requirements corresponding to industrial components; Based on assembly requirements and part categories, determine the assembly sequence and assembly location of each industrial part, wherein the assembly location is the position of a certain industrial part. Based on the assembly sequence and 6D pose of industrial parts, the robot's grasping posture and grasping path are determined, and the robot is controlled to complete the grasping and assembly of industrial parts based on the grasping path and grasping posture. Among these, based on the assembly sequence and 6D pose of industrial parts, the robot's grasping posture and grasping path are determined, including: Based on the assembly sequence and 6D pose, the gripping path for other industrial parts to be gripped and placed at the assembly position is determined, wherein the other industrial parts are the remaining industrial parts after removing the industrial parts at the assembly position from the industrial components. The robot, controlled by grasping path and grasping posture, completes the grasping of industrial parts, including: Perform the following gripping operation on each other industrial part: The robot is controlled to grasp other industrial parts using the current grasping posture corresponding to the current other industrial parts; the robot is controlled to place other industrial parts at the assembly position using the current grasping path corresponding to the current other industrial parts. Determine whether all other industrial parts in the industrial component have been picked up; if so, determine that the industrial component is assembled; otherwise, determine the next other industrial part based on the assembly sequence and perform a picking operation on the next other industrial part. Different industrial parts correspond to different gripping postures, and different industrial parts correspond to different gripping paths; Controlling robots to assemble industrial parts includes: Determine the assembly connection positions between other current industrial parts and the target industrial part; The assembly position is determined based on the assembly connection position, the current pose of other industrial parts, and the 6D pose of the target industrial part. The robot is controlled to place other industrial parts at the assembly position and complete the assembly of other industrial parts. The industrial part at the assembly position is the target industrial part.
2. The assembly method for industrial parts according to claim 1, characterized in that, The detection includes RGB and depth images of industrial parts, obtaining the part category of each industrial part and its corresponding detection box in the RGB image, including: The RGB image and the depth image are input into a pre-trained object detection model. The object detection model extracts visual features from the RGB image and depth features from the depth image. The visual features and the depth features are fused to obtain fused features. Based on the fused features, the part category and detection box output by the object detection model are obtained and output. The target detection model is trained based on RGB image samples, depth image samples, part category samples, and detection box samples. The visual features include color features, edge features, and texture features.
3. The assembly method for industrial parts according to claim 2, characterized in that, After extracting visual features from the RGB image and depth features from the depth image using an object detection model, the method further includes: Determine whether the visual feature belongs to a known feature; If it is determined that the visual feature belongs to a known feature, the step of fusing the visual feature and the depth feature to obtain the fused feature is performed. If the visual feature is determined not to be a known feature, the target detection model is optimized based on the incremental learning algorithm, and the optimized target detection model is used to detect and output the part category and detection box.
4. The assembly method for industrial parts according to any one of claims 1-3, characterized in that, The method further includes: During the process of gripping industrial parts by controlling the gripping path and gripping posture, the gripping force and displacement of the robot are monitored in real time. Based on the grasping force, the displacement, and the 6D pose of the industrial part at the assembly position, the robot's grasping pose is adjusted in real time.
5. The assembly method for industrial parts according to claim 2, characterized in that, After obtaining the detection box based on the fusion features, the method further includes: Redundant detection boxes are removed based on the nonmaximum suppression algorithm.
6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the assembly method for industrial parts as described in any one of claims 1 to 5.
7. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the assembly method for industrial parts as described in any one of claims 1 to 5.