Automatic part identifying and machining system based on 3D point cloud

Abstract

The invention provides an automatic part identification and machining system based on 3D point cloud, and the system comprises a workbench which is provided with a magnetic plane for adsorbing a mechanical workpiece; an X-axis displacement table; the bottom ends of the first portal frame and the second portal frame are connected with the X-axis displacement table; the linear array laser point cloud camera is fixed on the first portal frame and is used for acquiring a 3D point cloud image of the mechanical workpiece; the Y-axis displacement table is arranged on the inner side face of the second portal frame; the Z-axis displacement table is arranged on the Y-axis displacement table; the machining tool is detachably mounted on the Z-axis displacement table and moves in the vertical direction under the driving of the Z-axis displacement table; and the upper computer is in communication connection with the X-axis displacement table, the linear array laser point cloud camera, the Y-axis displacement table and the Z-axis displacement table. According to the invention, through cooperation with point cloud information, identification and classified carrying of mechanical parts, automatic processing of mechanical workpieces, and defect detection and position marking of the surfaces of finished workpieces are realized respectively.

Description

Technical Field

[0001] The invention belongs to the field of automated processing, and in particular relates to a part automatic recognition and machining system based on 3D point cloud. Background Art

[0002] Mechanical processing refers to the industry that uses mechanical equipment to perform cutting, milling, drilling and other operations on metals, plastics, ceramics and other materials to produce parts or products that meet design requirements. With the rapid development of modern manufacturing, the machining industry has become an indispensable link. With the continuous development of automation technology, mechanical processing has become more efficient and the precision has become higher and higher. Production efficiency and product quality are constantly improving. However, there are still some problems, such as the need for manual intervention, manual intervention for different mechanical parts, and only one part can be processed at a time.

[0003] Current machining lathes require manual coding and path planning, and cannot process multiple parts at the same time. They need to be calibrated before processing the workpiece, and the tasks they perform are single. Summary of the invention

[0004] In view of the problems existing in the prior art, the present invention provides a 3D point cloud-based automatic parts recognition and machining system, which realizes the recognition and classified handling of mechanical parts, the automated machining of mechanical workpieces, and the defect detection and position marking of the surface of finished workpieces by coordinating point cloud information.

[0005] The technical solution of the present invention is as follows: 3D point cloud-based parts automatic recognition and machining system, including: A workbench, on which a magnetic surface is provided to absorb the mechanical workpiece; An X-axis translation stage is arranged at the edge positions of both sides of the magnetic attraction plane; A first gantry and a second gantry are arranged opposite to each other and the bottom ends of the gantry are connected to the X-axis translation table and are driven by the X-axis translation table to move along the length direction of the workbench; A linear array laser point cloud camera, which is fixed on the first gantry and collects a 3D point cloud image of the mechanical workpiece; A Y-axis translation stage, which is arranged on the inner side of the second gantry; A Z-axis translation stage, which is arranged on the Y-axis translation stage and moves along the width direction of the workbench under the drive of the Y-axis translation stage; A machining tool, which is detachably mounted on the Z-axis translation stage and moves in a vertical direction under the drive of the Z-axis translation stage; A host computer, which is communicatively connected with the X-axis translation stage, the linear array laser point cloud camera, the Y-axis translation stage and the Z-axis translation stage; in; The host computer has a mechanical workpiece shape database pre-stored therein; The host computer receives the 3D point cloud image sent by the linear array laser point cloud camera and fits the outline of the mechanical workpiece; The host computer compares the contour line with the mechanical workpiece shape in the mechanical workpiece shape database, thereby identifying and classifying the mechanical workpiece, and determining the three-dimensional coordinate point of each mechanical workpiece; The host computer uses the three-dimensional coordinate points to fit a machining path through a deep learning algorithm, and determines the working paths of the X-axis translation stage, the Y-axis translation stage, and the Z-axis translation stage according to the machining path to complete the machining operation.

[0006] Preferably, in the 3D point cloud-based automatic part recognition and machining system, the machining tool is a robotic arm and an electromagnetic chuck installed at the end of the robotic arm.

[0007] Preferably, in the 3D point cloud-based automatic part recognition and machining system, the machining tool is a fixed bracket and a cutting tool mounted on the fixed bracket.

[0008] Preferably, in the 3D point cloud-based automatic part recognition and machining system, the machining tool is a fixed bracket and an edge grinding tool mounted on the fixed bracket.

[0009] Preferably, in the 3D point cloud-based parts automatic recognition and machining system, The first gantry is also provided with a visible light wide-angle camera, the visible light camera collects a visible light picture of the mechanical workpiece and sends it to the host computer, and the host computer analyzes the visible light picture and the 3D point cloud image to determine the three-dimensional coordinates of the defect of the mechanical workpiece; The machining tool is a marking pen fixing bracket and a marking pen mounted on the marking pen fixing bracket; The host computer uses the three-dimensional coordinates of the defect to fit a machining path through a deep learning algorithm, and determines the working paths of the X-axis translation stage, the Y-axis translation stage and the Z-axis translation stage according to the machining path to complete the defect marking operation.

[0010] The present invention has the following beneficial effects: By using a linear array laser point cloud camera to obtain high-precision three-dimensional coordinates of mechanical parts, the processing path is fitted after processing the point cloud image, and the mechanical parts on the worktable can be accurately identified and classified, and automatically processed and polished. In addition, surface defects can be detected. This further improves the efficiency of traditional machine tools for processing parts of different mechanical structures. It can be widely used in mechanical processing, mechanical parts polishing, parts identification and classification, defect detection and other fields.

[0011] Other advantages, objectives and features of the present invention will be embodied in part through the following description, and in part will be understood by those skilled in the art through study and practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 A schematic structural diagram of an embodiment of a 3D point cloud-based parts automatic recognition and machining system provided by the present invention; Figure 2 for Figure 1 Workflow diagram of Figure 3 A schematic structural diagram of an embodiment of a 3D point cloud-based parts automatic recognition and machining system provided by the present invention; Figure 4 for Figure 3 Workflow diagram of Figure 5 A schematic structural diagram of an embodiment of a 3D point cloud-based parts automatic recognition and machining system provided by the present invention; Figure 6 for Figure 5 Workflow diagram of Figure 7 A schematic structural diagram of an embodiment of a 3D point cloud-based parts automatic recognition and machining system provided by the present invention; Figure 8 for Figure 7 Workflow diagram of . DETAILED DESCRIPTION

[0013] The present invention is further described in detail below in conjunction with the accompanying drawings so that those skilled in the art can implement the invention with reference to the description.

[0014] The present invention belongs to the field of automated processing technology, specifically machine vision, algorithm processing and automatic control methods, and can also be extended to other industrial detection fields.

[0015] The present invention aims to design an integrated machining system, which is an integrated system integrating part identification, path planning and automatic machining.

[0016] Embodiment 1 like Figure 1and Figure 2 As shown, the present invention provides a 3D point cloud-based parts automatic recognition and machining system, which includes: A workbench 102 on which a magnetic attraction plane 103 is arranged to attract the mechanical workpiece 101; An X-axis translation stage 106 is disposed at the edge positions of both sides of the magnetic attraction plane 103; The first gantry 105 and the second gantry 107 are arranged opposite to each other and the bottom ends of the gantry 105 are connected to the X-axis translation table 106 and are driven by the X-axis translation table 106 to move along the length direction of the workbench 102; A linear array laser point cloud camera 104, which is fixed on the first gantry 105 and collects a 3D point cloud image of the mechanical workpiece 101; A Y-axis translation stage 108, which is disposed on the inner side of the second gantry 107; A Z-axis translation stage 109, which is disposed on the Y-axis translation stage 108 and moves along the width direction of the workbench 102 under the drive of the Y-axis translation stage 108; A machining tool, which is detachably mounted on the Z-axis translation stage 109 and moves in a vertical direction under the drive of the Z-axis translation stage 109; A host computer, which is communicatively connected with the X-axis translation stage, the linear array laser point cloud camera, the Y-axis translation stage and the Z-axis translation stage; in; The host computer has a mechanical workpiece shape database pre-stored therein; The host computer receives the 3D point cloud image sent by the linear array laser point cloud camera and fits the outline of the mechanical workpiece; The host computer compares the contour line with the mechanical workpiece shape in the mechanical workpiece shape database, thereby identifying and classifying the mechanical workpiece, and determining the three-dimensional coordinate point of each mechanical workpiece; The host computer uses the three-dimensional coordinate points to fit a machining path through a deep learning algorithm, and determines the working paths of the X-axis translation stage, the Y-axis translation stage, and the Z-axis translation stage according to the machining path to complete the machining operation.

[0017] The machining tool is a robot arm 110 and an electromagnetic chuck installed at the end of the robot arm 110 .

[0018] Various mechanical workpieces 101 are manually placed on the machine tool workbench 102, and the workbench fixes the workpieces through the magnetic suction plane 103. The linear array laser point cloud camera 104 is fixed on the machine tool workbench through the gantry 105, and moves in the X-axis direction through the translation table 106 to collect the workpiece data on the machine tool workbench. Then, through comparison and identification, the model of the workpiece is distinguished, and the three-dimensional coordinate points are given for the mark position of each workpiece. The contour edges are identified through the deep learning algorithm, and the parts are identified and classified. In addition, a gantry 107 is configured on the machine tool workbench, and the gantry is configured with a Y-axis translation table 108 and a Z translation table 109, and a robotic arm 110 is installed on the translation table. By coordinating the point cloud information and the three-dimensional coordinate points, the electromagnetic suction cup at the end of the robotic arm is used to realize the handling and classification of unused mechanical parts.

[0019] Embodiment 2 like Figure 3 and Figure 4 As shown, the present invention provides a 3D point cloud-based parts automatic recognition and machining system, which includes: A workbench 102, on which a magnetic suction plane 103 is arranged to suck the machine blank 201; An X-axis translation stage 106 is disposed at the edge positions of both sides of the magnetic attraction plane 103; The first gantry 105 and the second gantry 107 are arranged opposite to each other and the bottom ends of the gantry 105 are connected to the X-axis translation table 106 and are driven by the X-axis translation table 106 to move along the length direction of the workbench 102; A linear array laser point cloud camera 104, which is fixed on the first gantry 105 and collects a 3D point cloud image of the mechanical workpiece 101; A Y-axis translation stage 108, which is disposed on the inner side of the second gantry 107; A Z-axis translation stage 109, which is disposed on the Y-axis translation stage 108 and moves along the width direction of the workbench 102 under the drive of the Y-axis translation stage 108; A machining tool, which is detachably mounted on the Z-axis translation stage 109 and moves in a vertical direction under the drive of the Z-axis translation stage 109; A host computer, which is communicatively connected with the X-axis translation stage, the linear array laser point cloud camera, the Y-axis translation stage and the Z-axis translation stage; in; The host computer has a mechanical workpiece shape database pre-stored therein; The host computer receives the 3D point cloud image sent by the linear array laser point cloud camera and fits the outline of the mechanical workpiece; The host computer compares the contour line with the mechanical workpiece shape in the mechanical workpiece shape database, thereby identifying and classifying the mechanical workpiece, and determining the three-dimensional coordinate point of each mechanical workpiece; The host computer uses the three-dimensional coordinate points to fit a machining path through a deep learning algorithm, and determines the working paths of the X-axis translation stage, the Y-axis translation stage, and the Z-axis translation stage according to the machining path to complete the machining operation.

[0020] The machining tool is a fixing bracket 210 and a cutting tool 211 installed on the fixing bracket 210 .

[0021] Various blanks 201 are manually placed on the machine tool workbench 102, and the workbench fixes the blanks through the magnetic plane 103. The linear array laser point cloud camera 104 is fixed on the machine tool workbench through the gantry 105, and moves in the X-axis direction through the translation table 106 to collect the workpiece data on the machine tool workbench. Then the 3D point cloud information of the workpiece to be processed is imported and matched with the 3D point cloud model of the blank. The processing path is planned through the deep learning algorithm. In addition, a gantry 107 is configured on the machine tool workbench, and the gantry is configured with a Y-axis translation table 108 and a Z translation table 109. A tool fixing bracket 210 is installed on the translation table, and a suitable tool 211 is installed. Automatic processing of mechanical parts is achieved by coordinating the three-dimensional coordinate information and the processing path map route.

[0022] Embodiment 3 like Figure 5 and Figure 6 As shown, the present invention provides a 3D point cloud-based parts automatic recognition and machining system, which includes: A workbench 102 on which a magnetic attraction plane 103 is arranged to attract the mechanical workpiece 101; An X-axis translation stage 106 is disposed at the edge positions of both sides of the magnetic attraction plane 103; The first gantry 105 and the second gantry 107 are arranged opposite to each other and the bottom ends of the gantry 105 are connected to the X-axis translation table 106 and are driven by the X-axis translation table 106 to move along the length direction of the workbench 102; A linear array laser point cloud camera 104, which is fixed on the first gantry 105 and collects a 3D point cloud image of the mechanical workpiece 101; A Y-axis translation stage 108, which is disposed on the inner side of the second gantry 107; A Z-axis translation stage 109, which is disposed on the Y-axis translation stage 108 and moves along the width direction of the workbench 102 under the drive of the Y-axis translation stage 108; A machining tool, which is detachably mounted on the Z-axis translation stage 109 and moves in a vertical direction under the drive of the Z-axis translation stage 109; A host computer, which is communicatively connected with the X-axis translation stage, the linear array laser point cloud camera, the Y-axis translation stage and the Z-axis translation stage; in; The host computer has a mechanical workpiece shape database pre-stored therein; The host computer receives the 3D point cloud image sent by the linear array laser point cloud camera and fits the outline of the mechanical workpiece; The host computer compares the contour line with the mechanical workpiece shape in the mechanical workpiece shape database, thereby identifying and classifying the mechanical workpiece, and determining the three-dimensional coordinate point of each mechanical workpiece; The host computer uses the three-dimensional coordinate points to fit a machining path through a deep learning algorithm, and determines the working paths of the X-axis translation stage, the Y-axis translation stage, and the Z-axis translation stage according to the machining path to complete the machining operation.

[0023] The machining tool is a fixed bracket 210 and an edge grinding tool 211 installed on the fixed bracket 210 .

[0024] Various mechanical workpieces 101 are manually placed on the machine tool workbench 102, and the workbench fixes the workpieces through the magnetic suction plane 103. The linear array laser point cloud camera 104 is fixed on the machine tool workbench through the gantry 105, and moves in the X-axis direction through the translation table 106 to collect the workpiece data on the machine tool workbench. Then, through comparison and identification, the model of the workpiece is distinguished, and the three-dimensional coordinate point is given for the mark position of each workpiece. The contour edge is identified through the deep learning algorithm, and the path planning is performed for the contour edge of the part. In addition, a gantry 107 is configured on the machine tool workbench, and the gantry is equipped with a Y-axis translation table 108 and a Z translation table 109, and a tool fixing bracket 210 and a corresponding edge grinding tool 211 are installed on the translation table. By coordinating the three-dimensional coordinate information and the edge grinding path, the automatic edge grinding and chamfering of mechanical parts can be realized.

[0025] Embodiment 4 like Figure 7 and Figure 8 As shown, the present invention provides a 3D point cloud-based parts automatic recognition and machining system, which includes: A workbench 102 on which a magnetic attraction plane 103 is arranged to attract the mechanical workpiece 101; An X-axis translation stage 106 is disposed at the edge positions of both sides of the magnetic attraction plane 103; The first gantry 105 and the second gantry 107 are arranged opposite to each other and the bottom ends of the gantry 105 are connected to the X-axis translation table 106 and are driven by the X-axis translation table 106 to move along the length direction of the workbench 102; A linear array laser point cloud camera 104, which is fixed on the first gantry 105 and collects a 3D point cloud image of the mechanical workpiece 101; A Y-axis translation stage 108, which is disposed on the inner side of the second gantry 107; A Z-axis translation stage 109, which is disposed on the Y-axis translation stage 108 and moves along the width direction of the workbench 102 under the drive of the Y-axis translation stage 108; A machining tool, which is detachably mounted on the Z-axis translation stage 109 and moves in a vertical direction under the drive of the Z-axis translation stage 109; A host computer, which is communicatively connected with the X-axis translation stage, the linear array laser point cloud camera, the Y-axis translation stage and the Z-axis translation stage; in; The host computer has a mechanical workpiece shape database pre-stored therein; The host computer receives the 3D point cloud image sent by the linear array laser point cloud camera and fits the outline of the mechanical workpiece; The host computer compares the contour line with the mechanical workpiece shape in the mechanical workpiece shape database, thereby identifying and classifying the mechanical workpiece, and determining the three-dimensional coordinate point of each mechanical workpiece; The host computer uses the three-dimensional coordinate points to fit a machining path through a deep learning algorithm, and determines the working paths of the X-axis translation stage, the Y-axis translation stage, and the Z-axis translation stage according to the machining path to complete the machining operation.

[0026] The first gantry is also provided with a visible light wide-angle camera, the visible light camera collects a visible light picture of the mechanical workpiece and sends it to the host computer, and the host computer analyzes the visible light picture and the 3D point cloud image to determine the three-dimensional coordinates of the defect of the mechanical workpiece; The machining tool is a marking pen fixing bracket 310 and a marking pen 311 mounted on the marking pen fixing bracket 310; The host computer uses the three-dimensional coordinates of the defect to fit a machining path through a deep learning algorithm, and determines the working paths of the X-axis translation stage, the Y-axis translation stage and the Z-axis translation stage according to the machining path to complete the defect marking operation.

[0027] For some machined parts, their surfaces may have defects such as scratches and pits due to various reasons. Therefore, it is necessary to detect the surface and mark the defect position after the processing is completed. This system can also achieve this function after a small modification. Various mechanical workpieces 101 are manually placed on the machine tool workbench 102, and the workbench fixes the workpieces through the magnetic suction plane 103. The linear array laser point cloud camera 104 and the visible light wide-angle camera 112 are fixed on the machine tool workbench through the gantry 105, and move in the X-axis direction through the translation table 106 to collect the workpiece data on the machine tool workbench. Then the visible light camera image and 3D point cloud information are combined. The surface defects of the workpiece and the size of the workpiece are analyzed, and the three-dimensional coordinates of the defects are obtained. In addition, a gantry 107 is configured on the machine tool workbench, and the Y-axis translation table 108 and the Z translation table 109 are configured on the gantry. The marking pen fixing bracket 310 and the marking pen 311 are installed on the translation table. The automatic detection and marking of the surface defects of mechanical parts are realized by coordinating the three-dimensional coordinate information.

[0028] Although the embodiments of the present invention have been disclosed as above, they are not limited to the applications listed in the specification and the implementation modes. They can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to specific details.

Claims

1. The automatic part recognition and machining system based on 3D point cloud is characterized by: include: A workbench, on which a magnetic surface is provided to absorb the mechanical workpiece; An X-axis translation stage is arranged at the edge positions of both sides of the magnetic attraction plane; A first gantry and a second gantry are arranged opposite to each other and the bottom ends of the gantry are connected to the X-axis translation table and are driven by the X-axis translation table to move along the length direction of the workbench; A linear array laser point cloud camera, which is fixed on the first gantry and collects a 3D point cloud image of the mechanical workpiece; A Y-axis translation stage, which is arranged on the inner side of the second gantry; A Z-axis translation stage, which is arranged on the Y-axis translation stage and moves along the width direction of the workbench under the drive of the Y-axis translation stage; A machining tool, which is detachably mounted on the Z-axis translation stage and moves in a vertical direction under the drive of the Z-axis translation stage; A host computer, which is communicatively connected with the X-axis translation stage, the linear array laser point cloud camera, the Y-axis translation stage and the Z-axis translation stage; in; The host computer has a mechanical workpiece shape database pre-stored therein; The host computer receives the 3D point cloud image sent by the linear array laser point cloud camera and fits the outline of the mechanical workpiece; The host computer compares the contour line with the mechanical workpiece shape in the mechanical workpiece shape database, thereby identifying and classifying the mechanical workpiece, and determining the three-dimensional coordinate point of each mechanical workpiece; The host computer uses the three-dimensional coordinate points to fit a machining path through a deep learning algorithm, and determines the working paths of the X-axis translation stage, the Y-axis translation stage, and the Z-axis translation stage according to the machining path to complete the machining operation.

2. The automatic part recognition and machining system based on 3D point cloud according to claim 1, characterized in that: The machining tool is a mechanical arm and an electromagnetic chuck installed at the end of the mechanical arm.

3. The automatic part recognition and machining system based on 3D point cloud according to claim 2, characterized in that: The machining tool is a fixed bracket and a cutting tool mounted on the fixed bracket.

4. The automatic part recognition and machining system based on 3D point cloud according to claim 3, characterized in that: The machining tool is a fixed bracket and an edge grinding tool installed on the fixed bracket.

5. The automatic part recognition and machining system based on 3D point cloud according to claim 4, characterized in that: The first gantry is also provided with a visible light wide-angle camera, the visible light camera collects a visible light picture of the mechanical workpiece and sends it to the host computer, and the host computer analyzes the visible light picture and the 3D point cloud image to determine the three-dimensional coordinates of the defect of the mechanical workpiece; The machining tool is a marking pen fixing bracket and a marking pen mounted on the marking pen fixing bracket; The host computer uses the three-dimensional coordinates of the defect to fit a machining path through a deep learning algorithm, and determines the working paths of the X-axis translation stage, the Y-axis translation stage and the Z-axis translation stage according to the machining path to complete the defect marking operation.

Images