Tire pattern block automatic groove milling device control method based on binocular 3D vision

Abstract

The invention discloses a control method of an automatic groove milling device for a tire pattern block based on binocular 3D vision in the technical field of machine vision. The control method comprises the following steps: acquiring an image of a calibration plate by a camera, recording the position of a machine tool, and calibrating vision system parameters and the center of a rotating shaft; after the pattern blocks are fixed, the machine tool moves to a photographing position, a camera photographs and recognizes hole poses, and the machine tool moves a milling cutter to align the holes and records the positions; fixing and moving the pattern block again for photographing, positioning the center coordinate of the hole, carrying out binocular matching correction to obtain a 3D coordinate, and fitting and correcting the initial pose of the 3D coordinate; the automatic groove milling method comprises the steps of positioning a groove center line, constructing a point cloud, screening projection points, correcting the depth of a hole, calculating the position of the groove and finally calculating the machining position to machine the pattern block, and through calibration, teaching, positioning, chamfering and groove milling, the four-axis robot visual guidance automatic groove milling task with the overall precision requirement being 0.3 mm or above is finally completed.

Description

Technical Field

[0001] The invention relates to the technical field of machine vision, and in particular to a machine vision guided precision machining method. Background Art

[0002] The grooving of the back holes of the tire mold tread blocks is a crucial step in the tire manufacturing process. These back holes not only affect the bearing performance of the tire, but also have an important impact on the tire's balance, wear resistance and other performance indicators. Since the aluminum-magnesium material tread blocks have irregular shrinkage during the casting process and the position of the drilled holes also has errors, it is impossible to use CNC numerical control processing methods to groove the back hole air path according to the model. At present, the method of grooving the back hole air path of aluminum-magnesium material tread blocks is manual grooving. In order to solve the problem of low manual efficiency of grooving the back hole air path, a visual camera is used to measure the position of the back hole point, and guide the four-axis machine tool to groove the back hole air path. Summary of the invention

[0003] In response to the urgent need for automatic tire tread block milling equipment, the present invention provides a tire tread block automatic milling device and control method based on binocular 3D vision, which can ultimately complete the four-axis robot vision-guided automatic milling task with an overall accuracy requirement of more than 0.3mm through calibration, teaching, positioning, chamfering, and milling.

[0004] The object of the present invention is achieved as follows: A control method for an automatic groove milling device for a tire tread block based on binocular 3D vision includes two stages: calibration and teaching, and positioning and processing:

[0005] Phase 1) Calibration and teaching:

[0006] Step S1) First, fix the calibration plate on the clamping platform, control the movement of the four-axis machine tool, and use the camera to collect images of the calibration plate in different positions in the machine tool coordinate system, while recording the position of the machine tool;

[0007] Step S2) calibrating the visual system parameters and the rotation axis center;

[0008] Step S3) placing the pattern block on the clamping mechanism and fixing it, controlling the machine tool to move the pattern block to the photo taking position, recording the machine tool position at this time, and at the same time, taking photos with the camera to identify the posture of the hole closest to the center of the pattern block and recording it, controlling the machine tool to align the end of the milling cutter with the hole, and recording the machine tool position;

[0009] Phase 2) Positioning and processing phase:

[0010] Step S4) placing the pattern block on the clamping mechanism and fixing it, moving it to the photo taking position, and controlling the left and right cameras to collect images simultaneously;

[0011] Step S5) locating the coordinates of the centers of all holes in the two images, and performing binocular matching and correction on the hole coordinates in the two images to calculate the 3D coordinates of all holes;

[0012] Step S6) fitting the initial pose of the tread block according to the 3D coordinates of all holes and performing correction;

[0013] Step S7) locating the center line of the groove, constructing a point cloud of the back of the tread block in combination with the posture and radius information of the tread block, calculating the coordinates of the projection points of the point cloud in the image, selecting the projection points closest to the center line of the groove, and obtaining the corresponding point cloud;

[0014] Step S8) Correcting the depth information of the holes on the back of the tread block according to the point cloud, and calculating the positions of the grooves corresponding to all the holes;

[0015] Step S9) Calculate the position that the machine tool needs to reach during chamfering and slot milling, and process the pattern block.

[0016] Furthermore, step S1 is specifically as follows: the calibration plate is fixed on the clamping platform to ensure that the calibration plate and the platform are in a rigid connection state; the machine tool is controlled to drive the calibration plate to translate multiple times at the same angle within the camera field of view, and 20 sets of calibration plate images and corresponding machine tool positions are collected; the machine tool is controlled to drive the calibration plate to rotate multiple times at the same position within the camera field of view, and 20 sets of calibration plate images and corresponding machine tool positions are collected.

[0017] Furthermore, step S2 is specifically as follows: calibrate the 40 sets of calibration plate images collected in step S1 to obtain the left camera intrinsic parameters, the right camera intrinsic parameters, and the pose relationship between the left camera coordinate system and the right camera coordinate system: ; Wherein the subscript C1 represents the left camera, and the superscript C2 represents the right camera; the calibration step S1 multiple translations of the collected 20 sets of images and the corresponding machine tool positions to obtain the position of the left camera coordinate system in the machine tool coordinate system , where the superscript B represents the machine tool coordinate system; the 20 sets of images collected by multiple rotations in step S1 are calibrated with the corresponding machine tool positions to obtain the position and posture of the rotation axis in the machine tool coordinate system, and the rotation center coordinate system is obtained , the origin and direction of the coordinate system are obtained by calibration.

[0018] Furthermore, step S5 is specifically as follows: first, the approximate area of ​​all holes is roughly located by a deep learning target detection method, and the accurate coordinates of all holes are accurately located in these areas by using a scalable shape template matching method; then, a hole closest to the center of the tread block is located, and a point cloud of the back of the entire tread block is fitted in combination with a known radius, and all 3D points are back-projected onto the image coordinate system, and the corresponding hole in the right camera image is found based on the accurate positioning result of the left camera image, and an NCC template is created based on the hole in the left image, and the hole is found at the corresponding position of the right image, and the coordinates of the right image are corrected, and finally the 3D coordinates of all holes in the binocular system are obtained. , where the superscript n is the hole number.

[0019] Further, step S6 is specifically as follows: firstly, a right cylinder is fitted according to the point cloud formed by all holes to obtain the pattern block coordinate system , with its origin at the highest point of the fitted cylinder side, the X-axis is parallel to the cylinder axis, and the Z-axis points to the geometric center of the cylinder; then locate the area where the entire pattern block is located, divide it into 60 rectangular areas, project the 3D coordinates of the holes in each area onto the bottom surface of the cylinder for circle fitting, find the highest point of the circle in the pattern block coordinate system, and project it back to the three-dimensional space, and finally fit the highest generatrix in the initial coordinate system of the pattern block according to the 60 highest points, and adjust the pattern block coordinate system according to the direction of the generatrix. Make corrections.

[0020] Furthermore, step S7 is specifically as follows: combining target detection with traditional image processing methods to obtain the center line of the groove, establishing a point cloud set for the back of the pattern block according to the pattern block coordinate system and known radius information obtained in step S6, calculating the projection coordinate set of the point cloud in the left image, finding the line connecting the projection points closest to the center line of the groove, and the point cloud coordinates corresponding to the projection points are the point cloud coordinates corresponding to the center line of the groove.

[0021] Furthermore, step S8 is specifically as follows: according to the point cloud model established in step S7, the Z-axis coordinates of the 3D coordinates of all the back holes obtained in step S5 are corrected: first, the coordinate closest to the back hole in the 2D projection coordinate set of the point cloud in the left image is calculated, and the Z-axis coordinate of the point cloud corresponding to the coordinate is used to replace the back hole The Z-axis coordinate of the hole; calculate the point on the groove centerline point cloud closest to the back hole 3D coordinate to obtain the position of the groove corresponding to all holes .

[0022] Further, step S9 is specifically as follows: first, according to The back hole coordinates in the camera coordinate system Convert to the machine tool coordinate system, that is: ,in It only includes the 3D coordinate information of the holes, and also calculates the rotation angle of the machine tool A axis when all holes rotate around the machine tool X axis and reach the highest point , this angle is the rotation axis coordinate system In the YZ plane, the angle between the line connecting the rotation center and the hole and the Z axis is calculated, and then the rotation of the hole around the rotation axis is calculated. The coordinates after the angle are obtained to obtain the position and posture of the rotated hole ; Similarly, the position of the groove corresponding to the rotated hole can be obtained ; According to the photographing posture recorded in step S3 , the position of the milling cutter aligned with the hole closest to the center of the pattern block , the position of the tread block closest to the center hole ; Calculate the position that the machine tool should reach when chamfering ,in:

[0023] ,

[0024] ,

[0025] By the same token, we can get the position that the machine tool needs to reach when milling the groove.

[0026] Compared with the prior art, the beneficial effects of the present invention are: the present invention can identify all the position information required for processing in one photo, the time consumption of visual processing is greatly shortened compared with laser scanning, and the automatic processing of molds with irregular shrinkage in the casting process can be realized, with high processing accuracy, fast speed and good consistency; before this, the milling task could only be performed manually, with low efficiency and poor consistency. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0028] Figure 1 The figure is a process flow chart of tire mold processing in the present invention.

[0029] Figure 2 It is the partial calibration plate image after translation and rotation in the present invention.

[0030] Figure 3 This is a binocular back projection effect diagram in the present invention.

[0031] Figure 4 Schematic diagram of the rotation center coordinate system in the present invention.

[0032] Figure 5 It is a schematic diagram of the point cloud projection effect in the present invention.

[0033] Figure 6 It is a schematic diagram of the local hole positioning and correction effect in the present invention.

[0034] Figure 7 It is a schematic diagram of the side surface of a cylinder fitted with point cloud in the present invention.

[0035] Figure 8 It is a schematic diagram of the rectangular area division in the present invention.

[0036] Fig. 9 This is a schematic diagram of the highest point fitting in the present invention.

[0037] Fig.10 It is a schematic diagram of busbar positioning in the present invention.

[0038] Fig.11 It is a schematic diagram of the groove positioning effect in the present invention.

[0039] Fig.12 Schematic diagram of the error before and after local point cloud correction in the present invention.

[0040] Fig.13 It is a schematic diagram corresponding to the through hole and the groove in the present invention.

[0041] Fig.14 This is a diagram of the on-site implementation and local chamfering effect of the present invention. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] like Figure 1 The device and control method for automatic milling grooves of tire tread blocks based on binocular 3D vision are shown, including two stages: calibration and teaching, positioning and processing (see Appendix Figure 1 ):

[0044] Calibration and teaching stage:

[0045] Step S1: First, fix the calibration plate on the clamping platform, control the movement of the four-axis machine tool, and use the camera to collect images of the calibration plate in different positions in the machine tool coordinate system, while recording the position of the machine tool; specifically:

[0046] The calibration plate is fixed on the clamping platform to ensure that the calibration plate and the platform are rigidly connected; the machine tool is controlled to drive the calibration plate to translate multiple times at the same angle within the camera field of view, and 20 sets of calibration plate images and corresponding machine tool positions are collected; the machine tool is controlled to drive the calibration plate to rotate multiple times at the same position within the camera field of view, and 20 sets of calibration plate images and corresponding machine tool positions are collected. The calibration plate images are referenced Figure 2 .

[0047] In this step, the translation and rotation processes are separately calibrated and image acquisition is performed, which helps to calibrate the translation and rotation separately and avoid motion coupling.

[0048] Step S2: Calibrate the visual system parameters and the center of the rotation axis; specifically:

[0049] The 40 sets of calibration plate images collected in the calibration step S1 are used to obtain the left camera intrinsic parameters, the right camera intrinsic parameters, and the pose relationship between the left camera coordinate system and the right camera coordinate system:

[0050] ,

[0051] The translation unit is m, the rotation unit is °, the subscript C1 represents the left camera, the superscript C2 represents the right camera, and the back-projection error is 0.133194 pixels. The back-projection effect is as follows: Figure 3 ; Calibration step S1 multiple translations of the collected 20 sets of images and the corresponding machine tool positions to obtain the position of the left camera coordinate system in the machine tool coordinate system:

[0052] ,

[0053] The superscript B represents the machine tool coordinate system, the average calibration error of the X-axis is 0.0131255 mm, the average calibration error of the Y-axis is 0.0165558 mm, and the average calibration error of the Z-axis is 0.0294939 mm;

[0054] Calibration step S1: Multiple rotations of 20 sets of images collected and the corresponding machine tool positions to obtain the position and posture of the rotation axis in the machine tool coordinate system and the rotation center coordinate system , the origin and direction of the coordinate system are obtained by calibration, such as Figure 4 , the conversion relationship between this coordinate system and the machine tool coordinate system is:

[0055] .

[0056] This step calibrates the XYZ axis of the machine tool based on the calibration plate image corresponding to the translation process and the motion coordinate values ​​of the machine tool, and calibrates the R axis of the machine tool based on the calibration plate image corresponding to the rotation process and the motion coordinate values ​​of the machine tool. This can greatly simplify the calibration process and provide a basis for subsequent high-precision calculations.

[0057] Step S3: Place the pattern block on the clamping mechanism and fix it, control the machine tool to move the pattern block to the photo taking position, record the machine tool position at this time, and at the same time, use the camera to identify and record the posture of the hole closest to the center of the pattern block, control the machine tool to align the end of the milling cutter to the hole, and record the machine tool position.

[0058] Positioning and processing stage:

[0059] Step S4: placing the pattern block on the clamping mechanism and fixing it, moving it to the photo taking position, and controlling the left and right cameras to collect images simultaneously;

[0060] Step S5: locate the coordinates of the centers of all holes in the two images, perform binocular matching and correction on the hole coordinates in the two images, and calculate the 3D coordinates of all holes; specifically: firstly, roughly locate the approximate areas of all holes by using the deep learning target detection method, and accurately locate the exact coordinates of all holes in these areas by using the scalable shape template matching method; then locate a hole closest to the center of the tread block, fit the point cloud of the back of the entire tread block with the known radius, and back-project all 3D points onto the image coordinate system, such as Figure 5 , based on the precise positioning result of the left camera image, find the corresponding hole in the right camera image, and create the NCC template based on the hole in the left image, find the hole at the corresponding position in the right image, and correct the coordinates of the right image. The local positioning result is as follows Figure 6 Finally, the 3D coordinates of all holes in the binocular system are obtained , where the superscript n is the hole number.

[0061] All through holes are stably located through deep learning, and then the coordinates of the holes in the right camera are corrected through NCC matching to make them consistent with the coordinates of the hole center in the left camera, thereby improving the accuracy of the binocular system. This step not only achieves the stability of hole feature extraction by deep learning, but also achieves the accuracy of hole position extraction by traditional NCC matching.

[0062] Step S6: fitting the initial position of the pattern block according to the 3D coordinates of all holes and performing correction; specifically:

[0063] First, fit the right cylinder according to the point cloud formed by all the holes to obtain the pattern block coordinate system , its origin is at the highest point of the fitted cylinder side, the X axis is parallel to the cylinder axis, and the Z axis points to the geometric center of the cylinder, such as Figure 7 ; Then locate the area where the entire pattern block is located and divide it into 60 rectangular areas, such as Figure 8 , project the 3D coordinates of the holes in each area onto the bottom surface of the cylinder for circle fitting, find the highest point of the circle in the tread block coordinate system, and project it back to the three-dimensional space, such as Fig. 9 Finally, the highest generatrix in the initial coordinate system of the tread block is fitted according to the 60 highest points, as shown in Fig.10 , according to the direction of the generatrix, the coordinate system of the tread block Make corrections.

[0064] This step first obtains the pattern block coordinate system by fitting the point cloud, which is the basis for subsequent calculations and ensures the accuracy of subsequent correction and calculation steps. The highest points of the circle fitting in 60 areas are integrated for straight line fitting, which can ensure that the direction of the mother line is not affected by the center point of a single outlier on the extraction accuracy, achieving the best stability.

[0065] Step S7: locate the center line of the groove, build a point cloud on the back of the pattern block in combination with the posture and radius information of the pattern block, calculate the coordinates of the projection points of the point cloud in the image, screen out the projection points closest to the center line of the groove, and obtain the corresponding point cloud; specifically: combine target detection and traditional image processing methods to obtain the center line of the groove, establish a point cloud set for the back of the pattern block according to the pattern block coordinate system obtained in step S6 and the known radius information, calculate the projection coordinate set of the point cloud in the left image, find the line connecting the projection points closest to the center line of the groove, and the point cloud coordinates corresponding to the projection point are the point cloud coordinates corresponding to the center line of the groove. The groove positioning effect is as follows: Fig.11 .

[0066] This step combines the point cloud projection of the holes on the back of the overall pattern block extracted in step 5 with the groove center line obtained by image processing, and can efficiently find the groove center line closest to the force corresponding hole, and then calculate its connection line. It does not need to accurately segment the position of the groove on the image, which reduces the difficulty of image recognition and improves the stability of groove positioning.

[0067] Step S8: Correct the depth information of the back holes of the pattern block according to the point cloud, and calculate the positions of the grooves corresponding to all the holes; specifically, correct the Z-axis coordinates of the 3D coordinates of all the back holes obtained in step S5 according to the point cloud model established in step S7: first, find the coordinate closest to the back hole in the 2D projection coordinate set of the point cloud in the left image, and use the Z-axis coordinate of the point cloud corresponding to the coordinate to replace the back hole The Z-axis coordinates of the local point cloud before and after correction are as follows: Fig.12 Calculate the point on the groove centerline point cloud that is closest to the back hole 3D coordinates to obtain the positions of the grooves corresponding to all holes ,like Fig.13 .

[0068] In this step, the corrected points are used to ensure the consistency of positioning accuracy of all through holes, reduce the positioning error of through holes at the edge of the mold caused by angles, and ensure the overall processing quality.

[0069] Step S9: Calculate the position that the machine tool needs to reach during chamfering and slot milling, and process the pattern block (see Appendix Fig.14 ); specifically:

[0070] First, according to The back hole coordinates in the camera coordinate system Convert to the machine tool coordinate system, that is: ,in It only includes the 3D coordinate information of the holes, and also calculates the rotation angle of the machine tool A axis when all holes rotate around the machine tool X axis and reach the highest point , this angle is the rotation axis coordinate system In the YZ plane, the angle between the line connecting the rotation center and the hole and the Z axis is calculated, and then the rotation of the hole around the rotation axis is calculated. The coordinates after the angle are obtained to obtain the position and posture of the rotated hole ; Similarly, the position of the groove corresponding to the rotated hole can be obtained According to the photographing posture recorded in step S3 , the position of the milling cutter aligned with the hole closest to the center of the pattern block , the position of the tread block closest to the center hole ; Calculate the position that the machine tool should reach when chamfering ,in:

[0071] ,

[0072] ,

[0073] By the same token, we can get the position that the machine tool needs to reach when milling the groove.

[0074] This calculation method only requires guiding the mold under the camera and taking a photo once to calculate the position information of all axes required for processing without changing the mechanical structure. The subsequent pattern block processing can be fully automatic, which greatly improves the processing efficiency.

[0075] The above embodiments are only used to help understand the method and core idea of ​​the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made to the present invention without departing from the principles of the present invention, and these improvements and modifications also fall within the scope of protection of the claims of the present invention.

Claims

1. A control method for an automatic tire tread block milling device based on binocular 3D vision, characterized in that: It includes two stages: calibration and teaching, positioning and processing: Phase 1) Calibration and teaching: Step S1) First, fix the calibration plate on the clamping platform, control the movement of the four-axis machine tool, and use the camera to collect images of the calibration plate in different positions in the machine tool coordinate system, while recording the position of the machine tool; Step S2) calibrating the visual system parameters and the rotation axis center; Step S3) placing the pattern block on the clamping mechanism and fixing it, controlling the machine tool to move the pattern block to the photo taking position, recording the machine tool position at this time, and at the same time, taking photos with the camera to identify the posture of the hole closest to the center of the pattern block and recording it, controlling the machine tool to align the end of the milling cutter with the hole, and recording the machine tool position; Phase 2) Positioning and processing phase: Step S4) placing the pattern block on the clamping mechanism and fixing it, moving it to the photo taking position, and controlling the left and right cameras to collect images simultaneously; Step S5) locating the coordinates of the centers of all holes in the two images, and performing binocular matching and correction on the hole coordinates in the two images to calculate the 3D coordinates of all holes; Step S6) fitting the initial pose of the tread block according to the 3D coordinates of all holes and performing correction; Step S7) locating the center line of the groove, constructing a point cloud of the back of the tread block in combination with the posture and radius information of the tread block, calculating the coordinates of the projection points of the point cloud in the image, selecting the projection points closest to the center line of the groove, and obtaining the corresponding point cloud; Step S8) Correcting the depth information of the holes on the back of the tread block according to the point cloud, and calculating the positions of the grooves corresponding to all the holes; Step S9) Calculate the position that the machine tool needs to reach during chamfering and slot milling, and process the pattern block.

2. The control method of a tire tread block automatic groove milling device based on binocular 3D vision according to claim 1, characterized in that: Step S1 is specifically as follows: the calibration plate is fixed on the clamping platform to ensure that the calibration plate and the platform are in a rigid connection state; the machine tool is controlled to drive the calibration plate to translate multiple times at the same angle within the camera field of view, and 20 sets of calibration plate images and corresponding machine tool positions are collected; the machine tool is controlled to drive the calibration plate to rotate multiple times at the same position within the camera field of view, and 20 sets of calibration plate images and corresponding machine tool positions are collected.

3. The control method of a tire tread block automatic groove milling device based on binocular 3D vision according to claim 2 is characterized in that: Step S2 is specifically as follows: calibrate the 40 sets of calibration plate images collected in step S1 to obtain the left camera intrinsic parameters, the right camera intrinsic parameters, and the pose relationship between the left camera coordinate system and the right camera coordinate system: ; The subscript C1 represents the left camera, and the superscript C2 represents the right camera. The calibration step S1 is to translate the 20 sets of images collected multiple times and the corresponding machine tool positions to obtain the position of the left camera coordinate system in the machine tool coordinate system. , where the superscript B represents the machine tool coordinate system; the 20 sets of images collected by multiple rotations in step S1 are calibrated with the corresponding machine tool positions to obtain the position and posture of the rotation axis in the machine tool coordinate system, and the rotation center coordinate system is obtained , the origin and direction of the coordinate system are obtained by calibration.

4. The control method of the tire tread block automatic groove milling device based on binocular 3D vision according to claim 3 is characterized in that: Step S5 is specifically as follows: first, the approximate area of ​​all holes is roughly located by the deep learning target detection method, and the accurate coordinates of all holes are precisely located in these areas by using the scalable shape template matching method; then, a hole closest to the center of the pattern block is located, and the point cloud of the back of the entire pattern block is fitted based on the known radius, and all 3D points are back-projected onto the image coordinate system. Based on the precise positioning result of the left camera image, the corresponding hole in the right camera image is found, and the NCC template is created based on the hole in the left image, and the hole is found at the corresponding position in the right image, and the coordinates of the right image are corrected. Finally, the 3D coordinates of all holes in the binocular system are obtained. , where the superscript n is the hole number.

5. The control method of the tire tread block automatic groove milling device based on binocular 3D vision according to claim 4 is characterized in that: Step S6 is as follows: first, a right cylinder is fitted according to the point cloud formed by all holes to obtain the pattern block coordinate system , with its origin at the highest point of the fitted cylinder side, the X-axis is parallel to the cylinder axis, and the Z-axis points to the geometric center of the cylinder; then locate the area where the entire pattern block is located, divide it into 60 rectangular areas, project the 3D coordinates of the holes in each area onto the bottom surface of the cylinder for circle fitting, find the highest point of the circle in the pattern block coordinate system, and project it back to the three-dimensional space, and finally fit the highest generatrix in the initial coordinate system of the pattern block according to the 60 highest points, and adjust the pattern block coordinate system according to the direction of the generatrix. Make corrections.

6. The control method of a tire tread block automatic groove milling device based on binocular 3D vision according to claim 5, characterized in that: Step S7 is specifically as follows: combining target detection with traditional image processing methods to obtain the center line of the groove, establishing a point cloud set for the back of the pattern block according to the pattern block coordinate system and known radius information obtained in step S6, calculating the projection coordinate set of the point cloud in the left image, finding the line connecting the projection points closest to the center line of the groove, and the point cloud coordinates corresponding to the projection points are the point cloud coordinates corresponding to the center line of the groove.

7. The control method of a tire tread block automatic groove milling device based on binocular 3D vision according to claim 6, characterized in that: Step S8 is specifically as follows: according to the point cloud model established in step S7, the Z-axis coordinates of the 3D coordinates of all the back holes obtained in step S5 are corrected: first, the coordinate closest to the back hole in the 2D projection coordinate set of the point cloud in the left image is calculated, and the Z-axis coordinate of the point cloud corresponding to the coordinate is used to replace the back hole. The Z-axis coordinate of the hole; calculate the point on the groove centerline point cloud closest to the back hole 3D coordinate to obtain the position of the groove corresponding to all holes .

8. The control method of a tire tread block automatic groove milling device based on binocular 3D vision according to claim 6, characterized in that: Step S9 specifically includes: first, according to The back hole coordinates in the camera coordinate system Convert to the machine tool coordinate system, that is: ,in It only includes the 3D coordinate information of the holes, and also calculates the rotation angle of the machine tool A axis when all holes rotate around the machine tool X axis and reach the highest point , this angle is the rotation axis coordinate system In the YZ plane, the angle between the line connecting the rotation center and the hole and the Z axis is calculated, and then the rotation of the hole around the rotation axis is calculated. The coordinates after the angle are obtained to obtain the position and posture of the rotated hole ; Similarly, the position of the groove corresponding to the rotated hole can be obtained ; According to the photographing posture recorded in step S3 , the position of the milling cutter aligned with the hole closest to the center of the pattern block , the position of the tread block closest to the center hole ; Calculate the position the machine tool will reach during chamfering ,in: ， ， By the same token, we can get the position that the machine tool needs to reach when milling the groove.

Images