Non-standard part multifunctional machining system and method based on three-dimensional visual driving

Abstract

The invention provides a non-standard part multifunctional machining system and method based on three-dimensional visual driving, and the system comprises a sorting and feeding module which is used for transporting non-standard parts to be machined; the machining module is used for machining the to-be-machined part; the control module is used for controlling operation of the sorting and feeding module and the processing module; based on the three-dimensional point cloud recognition technology, the non-standard part multifunctional machining work station driven by the measured data of the three-dimensional sensor and the system of the work station are designed, the work station does not depend on information such as part drawings and models, and the work station is driven by the three-dimensional point cloud data collected in real time. Various machining procedures such as out-of-order sorting, edge milling, edge grinding, edge beveling, surface grinding and polishing and surface spraying of the non-standard parts are achieved; the problem of automatic machining of non-standard parts is solved, meanwhile, normal machining work of standard parts can be compatible, and the problem that a single work station cannot give consideration to various machining work is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of parts processing, and in particular to a non-standard parts multifunctional processing system and method based on three-dimensional vision drive. Background Art

[0002] The key technologies of the multifunctional processing workstation for non-standard parts lie in the positioning of parts, the processing and recognition of three-dimensional point clouds of parts, and the planning and generation of processing trajectories.

[0003] At present, the mainstream non-standard parts automated processing equipment mainly relies on the import of part drawings and models or the input of part templates. The former cannot be applied to the non-standard manufacturing industry, because most non-standard parts, especially elbow plates, supplementary plates, non-standard special-shaped steel plates and other parts, do not have corresponding models, and cannot be input into models or drawings. In addition, the method of relying on model input is cumbersome in the model programming process, which only transfers the work pressure of on-site staff to the programming pressure of programmers, and the automation level is low; the latter relies on on-site part template input, and there are many types of non-standard parts, so a large number of templates need to be manually input before the equipment is operated, which increases labor costs and time. When there are new parts in the incoming material frame and they cannot be matched in the template library, it will cause equipment errors. Therefore, it is necessary to strictly require that the incoming parts have been entered into the parts library, and the requirements for the previous process are high.

[0004] At present, most automatic parts processing equipment only has processing functions, and the loading process requires manual loading. There are many types of non-standard parts, most of which are steel parts, which are heavy and stacked in disorder. The manual sorting and loading work intensity is high. In addition, most parts processing equipment at present cannot achieve fast flipping. If double-sided grinding and milling are required, robots are required to perform complex flipping operations, which greatly reduces the efficiency of automated production.

[0005] Patent document CN114147584A discloses an intelligent grinding device and method for the free edge of a hull part. The technical method adopts an earlier two-dimensional image recognition method to perform contour recognition in a two-dimensional image. The contour recognition of a two-dimensional image has extremely strict requirements on the on-site acquisition environment. Factors such as on-site lighting, smoke, and dust will affect the acquisition quality of the two-dimensional image, and thus affect the accuracy of contour recognition. At the same time, a two-dimensional image can only provide two-dimensional coordinate information. In order to ensure the quality of grinding, it is necessary to provide the robot with accurate three-dimensional coordinate information of the parts. Therefore, in terms of grinding quality, the two-dimensional image recognition method has obvious defects.

[0006] Patent document CN116551472A discloses a method and system for automatic grinding of ship parts. The technical method is to manually enter template library information, mark the part contour and free edge information in the template library, and then use the template matching method to identify and grind the free edges. The entry of the template library increases labor costs, reduces the efficiency of automated operations, and has poor adaptability to new parts.

[0007] Patent document CN113963129A discloses a point cloud-based template matching and online recognition method for small ship components, which is used in small ship component grinding equipment. This method is also based on template matching of three-dimensional point clouds. It uses ICP point cloud matching to align the parts to be polished with the templates in the parts library. Manual entry of the template library is still required, which increases labor costs, reduces the efficiency of automated operations, and is not adaptable to new parts. Summary of the invention

[0008] In view of the defects in the prior art, the purpose of the present invention is to provide a non-standard parts multifunctional processing system and method based on three-dimensional vision drive.

[0009] A non-standard parts multifunctional processing system based on three-dimensional vision drive provided by the present invention comprises:

[0010] Sorting and loading module, used to transport non-standard parts to be processed;

[0011] Processing module, used for processing the parts to be processed;

[0012] A control module, used to control the operation of the sorting and loading module and the processing module;

[0013] The sorting and loading module comprises:

[0014] Loading and sorting tray, used to place the cut parts to be processed;

[0015] The handling robot is used to perform three-dimensional scanning on the loading and sorting trays and transport the parts to be processed to the processing area;

[0016] The 3D vision sensor in the loading area is used to obtain the point cloud data of the parts to be processed;

[0017] Magnetic handling device, used to absorb and release the parts to be processed;

[0018] The processing module comprises:

[0019] The magnetic reversible processing platform is used to turn over the parts to be processed that require double-sided processing;

[0020] Processing robots are used to perform corresponding parts processing work according to the processing path provided by the industrial computer;

[0021] The three-dimensional visual sensor in the processing area is used to perform three-dimensional scanning on the parts to be processed on the magnetic flip processing platform to obtain the three-dimensional point cloud data of the parts to be processed;

[0022] Replaceable processing tools, by replacing different processing tools to process the parts to be processed

[0023] Finished product trays are used to store finished parts after processing.

[0024] Preferably, the control module includes an industrial computer on which the visual software and the host computer system are installed;

[0025] The visual software is used to process 3D point clouds, identify the coordinates of the center of mass of parts, and provide them to the handling robot for the handling point and posture; it automatically identifies the part contour point cloud and surface point cloud, the edge of the part to be processed, and the surface point cloud information of the part that needs to be polished or sprayed, and sends the generated processing trajectory to the polishing robot;

[0026] The host computer system is used to control the program operation of the robot.

[0027] Preferably, the control module also includes a PLC control cabinet for starting, stopping and monitoring the system.

[0028] Preferably, the three-dimensional visual sensor in the loading area and the magnetic suction handling device are installed at the end of the three-dimensional visual sensor in the loading area.

[0029] Preferably, the processing area three-dimensional vision sensor and the replaceable processing tool are installed on the flange end of the processing robot.

[0030] Preferably, the magnetically reversible processing platform includes a first magnetic processing station and a second magnetic processing station; the magnetization of the stations is controlled by PLC to absorb the parts to be processed; the first magnetic processing station can be connected to the rotating axis and flipped by the PLC host computer system.

[0031] Preferably, the replaceable machining tool can replace tools including a floating force-controlled grinding head, a floating force-controlled milling tool, a floating force-controlled grinding wheel, a bevel cutting tool and a spraying tool.

[0032] Preferably, it is possible to complete random sorting, edge milling, edge grinding, edge beveling, surface grinding and polishing, and surface spraying of the parts to be processed.

[0033] A multifunctional processing method for non-standard parts based on three-dimensional vision drive provided by the present invention comprises:

[0034] Step S1: Move the handling robot to the top of the loading and sorting tray, and use the loading area 3D vision sensor at the end to scan the loading and sorting tray to obtain original point cloud data;

[0035] Step S2: Send the original point cloud data to the control module, filter out the background point cloud and interference point cloud of the loading and sorting tray through the point cloud preprocessing algorithm, and retain the part point cloud above the loading and sorting tray;

[0036] Step S3: The 3D point cloud processing algorithm based on the visual software automatically segments the point cloud of the randomly stacked parts and identifies the topmost graspable parts; calculates the centroid coordinates of the part point cloud, converts the coordinates into the robot coordinate system, and sends them to the handling robot;

[0037] Step S4: Move the handling robot to the top of the part according to the part coordinates, control the magnetic suction handling device to magnetize, and adsorb the part onto the magnetic suction handling device;

[0038] Step S5: Move the handling robot carrying the parts to the top of the magnetic reversible processing platform, control the magnetic handling device to demagnetize, place the parts on the first magnetic processing station, control the magnetic reversible processing platform to magnetize, and absorb the parts to be processed;

[0039] Step S6: Move the processing robot to the top of the magnetic flip processing platform, and use the processing area 3D vision sensor at the end to perform 3D scanning on the front of the part to be processed to obtain original point cloud data;

[0040] Step S7: Send the original point cloud data to the control module, filter out the background point cloud and interference point cloud through the point cloud preprocessing algorithm, and retain the point cloud of the part to be processed;

[0041] Step S8: extracting the outer contour of the part to be processed based on the three-dimensional point cloud processing algorithm and performing point cloud analysis, and automatically distinguishing the area to be processed based on the part processing rules;

[0042] Step S9: extract the free edges to be processed, generate a free edge processing path, including path point coordinates and processing posture, and send it to the processing robot;

[0043] Step S10: The processing robot performs processing according to the processing path provided by the system;

[0044] Step S11: Process the other side to complete the processing.

[0045] Preferably, the step S11 includes the following sub-steps:

[0046] Step S11.1: The flipping mechanism of the magnetic flippable processing platform flips and demagnetizes one side of the platform, flips the part to be processed onto the processing platform on the other side, controls the platform on one side to reset, and magnetizes the platform on the other side, and adsorbs the part onto the second magnetic processing station;

[0047] Step S11.2: Move the processing robot to the top of the magnetic reversible processing platform, use the processing area 3D vision sensor at the end to perform 3D scanning on the back side of the part to be processed, and repeat steps S7 to S10;

[0048] Step S11.3: After the front and back sides of the part are processed, the handling robot is moved to the top of the magnetic reversible processing platform, and the three-dimensional visual sensor of the end processing area is used to perform three-dimensional scanning and photography of the finished part, and steps S2-S3 are repeated to obtain the centroid coordinates of the finished part;

[0049] Step S11.4: The handling robot uses a magnetic handling device to absorb the finished parts according to the coordinates calculated by the system and places the parts into the finished product tray;

[0050] Step S11.5: Move the transport robot to the top of the loading and sorting tray, and repeat steps S1 to S11 until it is detected that there are no parts in the loading and sorting tray.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] 1. Based on the 3D point cloud recognition technology, the present invention designs a multifunctional processing workstation for non-standard parts driven by measured data of 3D sensors and its system. It can be driven by 3D point cloud data collected in real time without relying on part drawings, models and other information, to realize various processing procedures such as random sorting, edge milling, edge grinding, edge beveling, surface grinding and polishing, surface spraying, etc. of non-standard parts; it not only solves the problem of automated processing of non-standard parts, but also is compatible with the normal processing of standard parts, and solves the problem that a single workstation cannot take into account multiple processing operations.

[0053] 2. The present invention realizes double-sided processing of parts at the same processing station by designing a reversible magnetic processing platform. The traditional flipping process generally requires an additional flipping table. The robot grabs the part and places it on the flipping table, grabs the part from the other side, and then places it back on the processing platform to flip the part. The entire flipping process takes about 2 minutes. The reversible magnetic processing platform of the present invention only takes 10 seconds to flip, which increases the efficiency by 12 times, effectively solving the problem of complicated and inefficient flipping process of non-standard parts.

[0054] 3. The equipment in the present invention adopts a modular design, so the workstation solution is flexible and changeable. The loading module and the processing module can be expanded according to the production capacity requirements to achieve 1-to-1 or many-to-many module pairing to improve production capacity and efficiency, or based on a collaborative robot and a mobile base, a processing module is carried to form a small mobile polishing workstation, making the processing mode more flexible; at the same time, since the two modules can operate independently, the automatic loading module can be replaced by manual loading according to the budget to reduce costs. Removing the automatic loading module can directly save the hardware cost of a robot and an industrial camera, reducing the overall cost by about 40%.

[0055] Other beneficial effects of the present invention will be explained in the specific implementation manner through the introduction of specific technical features and technical solutions. Through the introduction of these technical features and technical solutions, those skilled in the art should be able to understand the beneficial technical effects brought about by the technical features and technical solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0057] Figure 1 It is a schematic diagram of the system structure of the present invention.

[0058] Figure 2 The figure is a flow chart of the method of the present invention.

[0059] Description of Reference Numerals

[0060] Loading and sorting tray 1 Processing area 3D vision sensor 7

[0061] Handling robot 2 Replaceable processing tools 8

[0062] 3D vision sensor in loading area 3 Finished product tray 9

[0063] Magnetic handling device 4 Industrial computer 10

[0064] Magnetic reversible processing platform 5 PLC control cabinet 11

[0065] Processing robot 6 DETAILED DESCRIPTION

[0066] The present invention is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several changes and improvements can also be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0067] Example 1

[0068] Reference Figure 1 As shown, a multifunctional processing system for non-standard parts based on 3D vision drive includes a processing workstation.

[0069] The workstation consists of a sorting and loading module and a processing module, wherein the sorting and loading module consists of a loading and sorting tray 1, a three-dimensional visual sensor in the loading area 3, a handling robot 2, and a magnetic handling device 4; the processing module consists of a magnetic flippable processing platform 5, a three-dimensional visual sensor in the processing area 7, a processing robot 6, replaceable processing tools 8, and a finished product tray 9. The two modules are jointly controlled by a PLC control cabinet 11, an industrial computer 10, and visual recognition software and host computer software.

[0070] In the sorting and loading module, the loading area three-dimensional vision sensor 3 and the magnetic suction handling device 4 are installed at the flange end of the handling robot 2, and the replaceable processing tool 8 and the processing area three-dimensional vision sensor 7 in the processing module are installed at the flange end of the processing robot 6.

[0071] The loading and sorting tray 1 is used to place the non-standard parts that need to be processed after cutting, and they are manually stacked and placed in the loading and sorting tray 1 according to their types.

[0072] The end of the handling robot 2 is equipped with the three-dimensional visual sensor 3 of the loading area and the magnetic handling device 4, which are used to perform three-dimensional scanning on the loading sorting tray 1 and transport the parts to the processing area. The magnetic flip processing platform 5 includes two magnetic processing stations AB. The station magnetization and demagnetization are controlled by PLC to adsorb the parts to be processed on the station. Station A is connected to the rotatable shaft and flipped through PLC and the host computer system to achieve rapid flipping of parts that need double-sided processing.

[0073] A three-dimensional visual sensor 7 and a replaceable processing tool 8 of the processing area are installed at the end of the processing robot 6, which perform three-dimensional scanning of the workpiece to be processed on the grinding platform, and perform corresponding parts processing work through corresponding processing tools, such as a floating force-controlled grinding head, a floating force-controlled milling tool, a floating force-controlled grinding wheel, a bevel cutting tool, a spraying tool, etc.

[0074] The finished product tray 9 is used to store the finished product parts that have been processed.

[0075] The industrial computer 10 is installed with visual software and a host computer system. The visual software is used to process three-dimensional point clouds, identify the coordinates of the centroid of parts, provide the handling point and posture to the handling robot 2, and automatically identify the contour point cloud and surface point cloud of parts. It can automatically identify the edge of the processing and the surface point cloud information of the parts that need to be polished and sprayed, generate the processing trajectory and send it to the polishing robot. The host computer system controls the operation of the robot program.

[0076] The PLC system is responsible for the start and stop of the entire equipment, and the overall process monitoring and control, including the magnetization, demagnetization, and flipping operations of the magnetically-absorbable flippable processing platform 5 .

[0077] The workstation can support a variety of non-standard parts processing such as random sorting of parts, edge milling, edge grinding, edge beveling, surface grinding and polishing, surface spraying, etc. Taking part edge milling as an example, its working principle is as follows:

[0078] Workers place the non-standard parts that need to be milled after cutting according to their types on the loading and sorting tray 1. The handling robot 2 moves above the loading and sorting tray 1. The three-dimensional visual sensor 3 in the loading area takes a picture of the loading area to obtain the point cloud of the parts in the loading area. The visual system in the industrial computer 10 automatically identifies the parts that can be transported, calculates the center of mass of the parts, and then the handling robot 2 transports the parts to the processing area A of the magnetic flip processing platform 5 through the magnetic handling device 4.

[0079] The processing robot 6 moves to the top of the processing area A, and the processing area three-dimensional visual sensor 7 performs a three-dimensional scan on the part to be polished to obtain the three-dimensional point cloud of the part to be polished. The visual system in the industrial computer 10 identifies the part contour, and automatically identifies the free edge to be polished according to the milling rule, and sends the processing path to the processing robot 6. In the free edge milling operation, the replaceable processing tool 8 is a floating force-controlled milling tool, and the milling robot uses the floating force-controlled milling tool to mill and polish the free edge. After milling, the magnetic suction flip processing platform 5 flips the part, and the processing area three-dimensional visual sensor 7 scans the back of the part. The processing robot 6 performs free edge milling on the free edge of the back of the part. After completion, the handling robot 2 moves to the top of the processing platform, and uses the loading area three-dimensional visual sensor 3 to scan the finished workpiece. The system automatically identifies the finished workpiece on the processing platform, and the handling robot 2 uses the magnetic suction handling device 4 to transport the finished part to the finished product tray 9. Repeat the above operation until all parts in the loading and sorting tray 1 are processed.

[0080] The present invention is based on 3D point cloud recognition technology and designs a multifunctional processing workstation and system for non-standard parts driven by measured data of 3D sensors. The workstation does not rely on part drawings, models and other information, but is driven by 3D point cloud data collected in real time to realize various processing procedures such as random sorting, edge milling, edge grinding, edge beveling, surface grinding and polishing, and surface spraying of non-standard parts. The workstation not only solves the problem of automated processing of non-standard parts, but is also compatible with the normal processing of standard parts, and solves the problem that a single workstation cannot take into account multiple processing operations.

[0081] Example 2

[0082] Reference Figure 2 As shown, taking the edge milling of non-standard parts as an example, the specific workflow is as follows:

[0083] Step S1: After the system is started, the transport robot 2 moves to the top of the loading and sorting tray 1, and uses the loading area three-dimensional visual sensor 3 at the end to scan the loading and sorting tray 1.

[0084] Step S2: After the three-dimensional visual sensor 3 in the loading area obtains the original point cloud of the loading and sorting tray 1 area, the point cloud data is sent to the system. The system filters out the background point cloud and interference point cloud of the loading and sorting tray 1 through the point cloud preprocessing algorithm, and retains the part point cloud above the loading and sorting tray 1.

[0085] Step S3: Through the three-dimensional point cloud processing algorithm of the visual software, the randomly stacked part point cloud is automatically segmented, the topmost graspable parts are identified, the centroid coordinates of the part point cloud are calculated, and the coordinates are converted into the robot coordinate system and sent to the handling robot 2.

[0086] Step S3.1: The formula for calculating the centroid of the part point cloud is as follows:

[0087]

[0088] Among them, P c is the coordinate of the center of mass of the part, r i =(x i ,y i , z i ), i=1,2,..,n are the coordinates of each point, m i is the corresponding mass of the point. Generally, the mass of the parts to be processed is evenly distributed. i Usually it is taken as 1, and the centroid calculation formula can be converted into the following formula:

[0089]

[0090] Step S3.2: P c The formula for converting to the robot base coordinate system is as follows:

[0091] P bsee =RP c +T

[0092] Among them, P base is the coordinate of the part's center of mass in the robot's base coordinate system. R and T are the rotation matrix parameter and translation matrix parameter after the robot's hand-eye calibration. The two parameters constitute the external parameters of the robot's hand-eye calibration and can be queried in the robot system.

[0093] Step S4: the handling robot 2 moves to the top of the part according to the part coordinates, and the system controls the magnetic suction handling device 4 to magnetize and adsorb the part onto the handling device.

[0094] Step S5: The handling robot 2 carries the parts and moves to the top of the processing platform. The system controls the magnetic handling device 4 to demagnetize and places the parts on the processing area A on one side of the magnetic reversible processing platform 5. The system controls the magnetic reversible processing platform 5 to magnetize and adsorb the parts to be milled.

[0095] Step S6: The milling robot moves to the top of the processing platform and uses the processing area three-dimensional vision sensor 7 at the end to perform a three-dimensional scan on the front of the part to be milled.

[0096] Step S7: After the three-dimensional visual sensor 7 in the processing area obtains the original point cloud of the part to be milled, it is sent to the system. The system uses the point cloud preprocessing algorithm to filter out the background point cloud such as the processing platform and the interference point cloud, and retains the point cloud of the part to be milled.

[0097] Step S8: Based on the three-dimensional point cloud processing algorithm, the outer contour of the part to be milled is extracted, and the point cloud analysis is performed on the outer contour. Based on the part milling rules, the free edges that need to be milled and the welding edges that do not need to be milled are automatically distinguished.

[0098] Step S9: extract the free edge to be milled, generate a free edge milling path, including path point coordinates and milling posture, and send it to the milling robot.

[0099] Step S10: The milling robot performs free edge milling according to the milling path provided by the system, and the floating force-controlled milling tool monitors the milling force during the milling process to keep the milling force constant and ensure the milling quality.

[0100] Step S11: After the single-sided free edge milling is completed, the system controls the flipping mechanism of the flipping platform to flip and demagnetize one side of the platform, and flip the part to the other side of the processing platform. The system controls one side of the platform to reset and magnetize the other platform to adsorb the part on the processing area B.

[0101] Step S12: The milling robot moves to the top of the milling platform, and uses the three-dimensional vision sensor at the end to perform a three-dimensional scan on the back side of the part to be milled, and repeats steps S7 to S10.

[0102] Step S13: After the front and back sides of the part are milled, the handling robot 2 moves to the top of the processing platform, uses the terminal 3D visual sensor to perform 3D scanning and photographing of the finished part, and repeats steps S2-S3 to obtain the centroid coordinates of the finished part.

[0103] Step S14: the handling robot 2 uses the magnetic handling device 4 to absorb the finished parts according to the coordinates calculated by the system, and places the parts into the finished product tray 9.

[0104] Step S15: The transport robot 2 moves to the top of the loading and sorting tray 1, and repeats steps S1 to S14 until it is detected that there are no parts in the loading and sorting tray 1, and the workflow ends.

[0105] Those skilled in the art know that, in addition to realizing the system and its various devices, modules, and units provided by the present invention in a purely computer-readable program code, the system and its various devices, modules, and units provided by the present invention can be completely realized in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps to realize the same functions. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered as a hardware component, and the devices, modules, and units included therein for realizing various functions can also be regarded as structures within the hardware component; the devices, modules, and units for realizing various functions can also be regarded as both software modules for realizing the method and structures within the hardware component.

[0106] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. In the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. A multifunctional processing system for non-standard parts based on 3D vision drive, characterized in that: include: Sorting and loading module, used to transport non-standard parts to be processed; Processing module, used for processing the parts to be processed; A control module, used to control the operation of the sorting and loading module and the processing module; The sorting and loading module comprises: A loading and sorting tray (1) is used to place the cut parts to be processed; A handling robot (2) is used to perform three-dimensional scanning on the loading and sorting tray (1) and to carry the parts to be processed to the processing area; A three-dimensional visual sensor (3) in the loading area is used to obtain point cloud data of the parts to be processed; A magnetic suction handling device (4) for absorbing and releasing the parts to be processed; The processing module comprises: A magnetically-attractive reversible processing platform (5) is used to turn over the parts to be processed that require double-sided processing; A processing robot (6) is used to perform corresponding parts processing work according to a processing path provided by an industrial control computer (10); A three-dimensional visual sensor (7) in the processing area is used to perform three-dimensional scanning on the parts to be processed on the magnetic reversible processing platform (5) to obtain three-dimensional point cloud data of the parts to be processed; Replaceable processing tools (8), by replacing different processing tools to process the parts to be processed The finished product tray (9) is used to store the finished product parts that have been processed.

2. The non-standard parts multifunctional processing system based on three-dimensional vision drive according to claim 1 is characterized in that: The control module comprises an industrial computer (10) on which visual software and a host computer system are installed; The visual software is used to process the three-dimensional point cloud, identify the coordinates of the center of mass of the part, and provide the handling point and posture to the handling robot (2); automatically identify the part contour point cloud and surface point cloud, the edge of the part to be processed, and the surface point cloud information of the part to be polished or sprayed, and send the generated processing trajectory to the polishing robot; The host computer system is used to control the program operation of the robot.

3. The non-standard parts multifunctional processing system based on three-dimensional vision drive according to claim 2 is characterized in that: The control module also includes a PLC control cabinet (11) for starting, stopping and monitoring the system.

4. The non-standard parts multifunctional processing system based on 3D vision drive according to claim 1 is characterized in that: The three-dimensional visual sensor (3) in the material loading area and the magnetic suction type handling device (4) are installed at the end of the three-dimensional visual sensor (3) in the material loading area.

5. The non-standard parts multifunctional processing system based on 3D vision drive according to claim 1 is characterized in that: The processing area three-dimensional visual sensor (7) and the replaceable processing tool (8) are installed on the flange end of the processing robot (6).

6. The non-standard parts multifunctional processing system based on 3D vision drive according to claim 1 is characterized in that: The magnetically reversible processing platform (5) comprises a first magnetically reversible processing station and a second magnetically reversible processing station; the magnetization of the stations is controlled by a PLC to absorb the parts to be processed; the first magnetically reversible processing station can be connected to a rotating shaft and can be reversed by a PLC host computer system.

7. The non-standard parts multifunctional processing system based on 3D vision drive according to claim 1 is characterized in that: The replaceable processing tool (8) can replace tools including a floating force-controlled grinding head, a floating force-controlled milling tool, a floating force-controlled grinding wheel, a bevel cutting tool and a spraying tool.

8. The non-standard parts multifunctional processing system based on 3D vision drive according to claim 1 is characterized in that: It can complete the random sorting, edge milling, edge grinding, edge beveling, surface grinding and polishing, and surface spraying of the parts to be processed.

9. A method for multifunctional processing of non-standard parts based on 3D vision drive, based on the multifunctional processing system for non-standard parts based on 3D vision drive according to any one of claims 1 to 8, characterized in that: include: Step S1: Move the transport robot (2) above the loading and sorting tray (1), and use the loading area three-dimensional vision sensor (3) at the end to scan the loading and sorting tray (1) to obtain original point cloud data; Step S2: sending the original point cloud data to the control module, filtering out the background point cloud and interference point cloud of the loading and sorting tray (1) through a point cloud preprocessing algorithm, and retaining the part point cloud above the loading and sorting tray (1); Step S3: The 3D point cloud processing algorithm based on the visual software automatically segments the point cloud of the randomly stacked parts and identifies the topmost graspable parts; Calculate the centroid coordinates of the part point cloud, convert the coordinates to the robot coordinate system, and send them to the handling robot (2); Step S4: the handling robot (2) is moved to the top of the part according to the part coordinates, and the magnetic suction handling device (4) is controlled to be magnetized to adsorb the part onto the magnetic suction handling device (4); Step S5: the handling robot (2) carries the parts and moves to the top of the magnetic reversible processing platform (5), the magnetic handling device (4) is controlled to be demagnetized, the parts are placed on the first magnetic processing station, and the magnetic reversible processing platform (5) is controlled to be magnetized to absorb the parts to be processed; Step S6: Move the processing robot (6) to the top of the magnetic reversible processing platform (5), and use the processing area three-dimensional visual sensor (7) at the end to perform a three-dimensional scan on the front of the part to be processed to obtain original point cloud data; Step S7: Send the original point cloud data to the control module, filter out the background point cloud and interference point cloud through the point cloud preprocessing algorithm, and retain the point cloud of the part to be processed; Step S8: extracting the outer contour of the part to be processed based on the three-dimensional point cloud processing algorithm and performing point cloud analysis, and automatically distinguishing the area to be processed based on the part processing rules; Step S9: extracting the free edge to be processed, generating a free edge processing path, including path point coordinates and processing posture, and sending it to the processing robot (6); Step S10: the processing robot (6) performs processing according to the processing path provided by the system; Step S11: Process the other side to complete the processing.

10. The multifunctional processing method for non-standard parts based on three-dimensional vision drive according to claim 9 is characterized in that: The step S11 includes the following sub-steps: Step S11.1: The turning mechanism of the magnetic flippable processing platform (5) turns over and demagnetizes one side of the platform, turns over the part to be processed to the processing platform on the other side, controls the platform on one side to reset, and magnetizes the platform on the other side, so as to adsorb the part on the second magnetic processing station; Step S11.2: Move the processing robot (6) to the top of the magnetic reversible processing platform (5), and use the processing area three-dimensional visual sensor (7) at the end to perform a three-dimensional scan on the back of the part to be processed, and repeat steps S7- Step S10; Step S11.3: After the front and back sides of the part have been processed, the handling robot (2) is moved to the top of the magnetic reversible processing platform (5), and the three-dimensional visual sensor (7) in the end processing area is used to perform three-dimensional scanning and photographing of the finished part, and steps S2 to S3 are repeated to obtain the centroid coordinates of the finished part; Step S11.4: The handling robot (2) uses the magnetic handling device (4) to absorb the finished parts according to the coordinates calculated by the system, and places the parts into the finished product tray (9); Step S11.5: Move the transport robot (2) to the top of the loading and sorting tray (1), and repeat steps S1 to S11 until it is detected that there are no parts in the loading and sorting tray (1).

Images