A method for manufacturing a mandrel

Abstract

The present invention relates to a mandrel manufacturing method, which relates to the field of automobile parts, and comprises: obtaining a mandrel image of the mandrel; detecting whether there is a burr feature on the mandrel surface from a preset burr detection model according to the mandrel image; when there is a burr feature, matching the burr three-dimensional information from a preset burr three-dimensional recognition model based on the burr feature; removing the burr by a preset burr removal method according to the burr three-dimensional information. The present application has the effect of improving the qualified rate of the mandrel and reducing the burrs on the mandrel.

Description

Technical Field

[0001] The invention relates to the field of automobile parts, in particular to a core shaft manufacturing method. Background Art

[0002] A mandrel is a shaft used to support rotating parts and only bear bending moments but not transmit torque.

[0003] In the prior art, the mandrel is used in the mechanical structure of components such as the seat belts of automobiles. The mandrel is generally a metal part, and the mandrel is generally manufactured through the steps of rough blank preparation, rough machining, heat treatment, fine machining, surface treatment and quality inspection.

[0004] After the mandrel is manufactured, burrs are likely to appear, which may cause the mandrel to fail quality inspection. Summary of the invention

[0005] In order to improve the qualified rate of the mandrel and reduce the burrs on the mandrel, the present invention provides a mandrel manufacturing method.

[0006] In a first aspect, the present invention provides a method for manufacturing a mandrel, which adopts the following technical solution:

[0007] A method for manufacturing a mandrel, comprising:

[0008] acquiring a mandrel image of the mandrel;

[0009] Detecting whether there is a burr feature on the surface of the mandrel from a preset burr detection model according to the mandrel image;

[0010] When there are burr features, matching the burr three-dimensional information from a preset burr three-dimensional recognition model based on the burr features;

[0011] The burrs are removed by a preset burr removal method according to the 3D information of the burrs.

[0012] By adopting the above technical solution, when burrs are identified on the surface of the mandrel, the three-dimensional information of the burrs is matched through images, and the burrs are removed according to the three-dimensional information, thereby reducing the burrs on the mandrel and improving the qualified rate of the mandrel.

[0013] Optionally, the burr removal method includes:

[0014] Matching the grinding range from a preset grinding range recognition model according to the burr three-dimensional information;

[0015] Matching the burr position from a preset burr position recognition model based on the burr feature;

[0016] Matching the grinding stroke from a preset grinding stroke matching model according to the grinding range and the burr position;

[0017] The preset grinding device is controlled according to the grinding stroke to grind the burrs to remove the burrs.

[0018] By adopting the above technical solution, the grinding stroke of the grinding device is matched by the burr three-dimensional information and the position of the burr, so that the grinding device is controlled according to the grinding stroke to grind and remove the burrs on the surface of the mandrel, thereby reducing the burrs on the mandrel and improving the qualified rate of the mandrel.

[0019] Optionally, the burr removal method further includes:

[0020] Judging whether the length of the burr exceeds a preset burr length from a preset length judgment model according to the burr three-dimensional information;

[0021] When the length of the burr exceeds a preset burr length, a shearing position is matched from a preset shearing position matching model based on the burr feature;

[0022] Matching the shear stroke from a preset shear stroke matching model according to the shear position and the mandrel image;

[0023] The preset shearing device is controlled according to the shearing stroke to shear the burrs.

[0024] By adopting the above technical solution, the excessively long burrs are cut off by the shearing device, thereby reducing the situation where the burrs are too long and are easily bent by the grinding device during grinding, resulting in the grinding device being unable to grind and remove the burrs cleanly, thereby improving the convenience of burr removal.

[0025] Optionally, the burr removal method further includes:

[0026] According to the shearing position and the burr three-dimensional information, the burr diameter at the shearing position is matched from the preset diameter judgment model;

[0027] When the burr diameter falls within the preset grinding range, the grinding depth is matched from a preset grinding depth matching model according to the burr diameter;

[0028] Matching the grinding stroke from a preset grinding stroke matching model according to the grinding depth, shearing position and mandrel image;

[0029] The preset grinding device is controlled according to the grinding stroke to grind the burrs for easy shearing.

[0030] By adopting the above technical solution, the overly coarse burrs are ground down to be easier to shear by a grinding device, thereby reducing the situation where the shearing device cannot shear off the overly long burrs due to the burrs being too coarse, thereby improving the convenience of burr removal.

[0031] Optionally, the burr removal method further includes:

[0032] When the length of the burr exceeds the preset burr length, judging whether there is a shearing space for the preset shearing device to extend into from a preset shearing space judgment model based on the burr characteristics and the shearing position;

[0033] When there is no shearing space for the preset shearing device to extend into, matching the rotation position from a preset rotation position matching model based on the burr feature;

[0034] According to the burr diameter, the rotation force is matched from a preset rotation force matching model;

[0035] The preset rotating device is controlled according to the rotation position to grab the burr and rotate the burr according to the rotation force.

[0036] By adopting the above technical solution, when there is no shearing space for the shearing device to extend into, the burrs cannot be sheared by the shearing device. At this time, the burrs are rotated by the rotating device, so that the burrs are twisted and then broken, thereby improving the convenience of burr removal.

[0037] Optionally, a burr fixing method is also included, and the burr fixing method includes:

[0038] When the length of the burr exceeds the preset burr length, the lifting position is matched from the preset lifting position matching model according to the mandrel image and the shear stroke;

[0039] Control the preset lifting device to lift the burr according to the lifting position;

[0040] According to the lifting position, the preset clamping device is controlled to cooperate with the lifting device to clamp the burr;

[0041] matching a removal stroke from a preset removal stroke matching model according to the mandrel image and the shear stroke;

[0042] When the shearing is completed by the shearing device, the clamping device and the lifting device are controlled to move according to the removal stroke to remove the burrs.

[0043] By adopting the above technical solution, the burr is supported by a lifting device below the burr, thereby reducing the bending of the burr due to force when the burr is removed by a shearing device or a grinding device, and the burr is fixed from above by a clamping device in cooperation with the lifting device, reducing the slippage of the burr and facilitating the removal of the burr separated from the core shaft, thereby improving the convenience of burr removal.

[0044] Optionally, a burr bending method is also included, and the burr bending method includes:

[0045] After the burrs are ground by the grinding device, the bending position is matched from a preset bending position matching model according to the shearing position and the mandrel image;

[0046] Matching the bending stroke from a preset bending stroke matching model according to the lifting position and the bending position;

[0047] According to the bending stroke, the preset clamping device and the lifting device are controlled to move from the lifting position to the bending position and clamp the burr;

[0048] According to the preset bending parameters, the preset clamping device and the lifting device are controlled to cooperate with each other to reciprocate and bend the burrs for easy shearing.

[0049] By adopting the above technical scheme, the shearing position is easily deformed after being ground by the grinding device, and the burr is bent up and down by the clamping device and the lifting device, so that the shearing position of the burr is deformed, thereby facilitating the shearing device to separate the burr from the core shaft, thereby improving the convenience of burr removal.

[0050] Optionally, a debris cleaning method is also included, and the debris cleaning method includes:

[0051] After the burrs are removed by a preset burr removal method, a grinding image of the mandrel is obtained;

[0052] Matching a chip area from a preset chip area recognition model according to the polishing image;

[0053] Matching a cleaning stroke from a preset cleaning stroke matching model according to the debris area;

[0054] The preset cleaning device is controlled according to the cleaning stroke to remove the debris on the mandrel from the mandrel surface.

[0055] By adopting the above technical solution, a large amount of debris is easily generated when grinding burrs with a grinding device, and the debris is easily retained in the groove cavity on the surface of the core shaft. The debris is removed by a cleaning device to facilitate further processing or inspection of the core shaft.

[0056] Optionally, also include:

[0057] When there is a burr feature, the burr posture is matched from a preset posture matching model according to the mandrel image;

[0058] Recognize the mandrel posture from a preset posture recognition model according to the mandrel picture;

[0059] Matching an adjustment step from a preset adjustment step matching model according to the mandrel posture and the burr posture;

[0060] Matching an adjustment stroke from a preset adjustment stroke matching model according to the adjustment steps;

[0061] The preset adjustment arm adjusts the posture of the burr on the mandrel according to the adjustment stroke control to facilitate the removal of the burr.

[0062] By adopting the above technical solution, the placement posture of the mandrel is adjusted by the mechanical arm, thereby adjusting the placement posture of the burrs on the mandrel, thereby facilitating the grinding device, shearing device and grinding device to operate the burrs, thereby improving the convenience of burr removal.

[0063] In a second aspect, the present application provides a mandrel, which adopts the following technical solution:

[0064] A core shaft comprises a body, one side of which is symmetrically provided with raised pins, and a weight-reducing cavity is evenly opened in the circumferential direction of the body.

[0065] By adopting the above technical solution, the overall mass of the mandrel is reduced through the weight-reducing cavity, and the mass distribution of the mandrel is made more uniform, thereby reducing the wear of the mandrel during use and improving the stability of the mandrel in use.

[0066] In summary, the present application includes at least one of the following beneficial technical effects:

[0067] 1. When burrs are identified on the surface of the mandrel, the 3D information of the burrs is matched through images, and the burrs are removed according to the 3D information, thereby reducing the burrs on the mandrel and improving the qualified rate of the mandrel;

[0068] 2. The grinding stroke of the grinding device is matched by the 3D information of the burr and the position of the burr, so as to control the grinding device to grind and remove the burrs on the surface of the mandrel according to the grinding stroke, thereby reducing the burrs on the mandrel and improving the qualified rate of the mandrel;

[0069] 3. The excessively long burrs are cut off by the shearing device, thereby reducing the situation where the burrs are too long and are easily bent by the grinding device during grinding, resulting in the grinding device being unable to grind and remove the burrs cleanly, thereby improving the convenience of burr removal. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 It is a schematic diagram of the structure of a mandrel.

[0071] Figure 2 A process of a mandrel manufacturing method Figure 1 .

[0072] Figure 3 The process of burr removal method Figure 1 .

[0073] Figure 4 The process of burr removal method Figure 2 .

[0074] Figure 5 The process of burr removal method Figure 3 .

[0075] Figure 6It is a flow chart of the burr fixing method.

[0076] Figure 7 It is a flow chart of the burr bending method.

[0077] Figure 8 is a flow chart of the debris cleaning method.

[0078] Fig. 9 A process of a mandrel manufacturing method Figure 2 .

[0079] The parts indicated by the numerical symbols in the above drawings are as follows: 1. main body; 2. raised pin; 3. weight-reducing cavity. DETAILED DESCRIPTION

[0080] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the following Figure 1-9 It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0081] Reference Figure 1 The embodiment of the present application discloses a core shaft, including a main body 1, on one side of which a raised pin 2 for installing the core shaft is integrally arranged, the raised pin 2 is symmetrically arranged about the axis of the main body 1, and a weight-reducing cavity 3 is evenly opened in the circumferential direction of the main body 1. The weight-reducing cavity 3 is used to reduce the mass of the main body and make the mass of the main body evenly distributed, thereby improving the stability of the core shaft structure.

[0082] Reference Figure 2 Based on the same inventive concept, an embodiment of the present invention provides a mandrel manufacturing method, comprising the following steps:

[0083] Step 100: Acquire a mandrel image of the mandrel.

[0084] The mandrel image refers to a picture of the mandrel, which displays the condition of the mandrel surface. The mandrel image can be obtained through a camera. The method for obtaining the mandrel image is selected by the staff according to the actual situation and will not be elaborated here.

[0085] Step 101: Detect whether there are burr features on the surface of the mandrel from a preset burr detection model according to the mandrel image.

[0086] The burr feature refers to the burr on the mandrel. The burr detection model refers to a neural network model obtained through sample training in advance. The burr detection model can identify whether there is a burr feature in the image.

[0087] Step 102: When burr features exist, burr three-dimensional information is matched from a preset burr three-dimensional recognition model based on the burr features.

[0088] The presence of burr features means that the burrs need to be processed to remove the burrs, thereby improving the qualified rate of the mandrel. The burr three-dimensional information refers to the three-dimensional information corresponding to the burr features, and the burr three-dimensional recognition model refers to the neural network model obtained by sample training in advance. The burr three-dimensional recognition model recognizes the three-dimensional information corresponding to the burr features from the image.

[0089] Step 103: removing burrs by a preset burr removal method according to the 3D burr information.

[0090] The burr removal method refers to a method of removing burrs according to the three-dimensional information of the burrs. When the burrs are too long or too thick, different means are selected to remove the burrs, thereby improving the convenience of burr removal.

[0091] Reference Figure 3 , the burr removal method includes the following steps:

[0092] Step 200: matching a grinding range from a preset grinding range recognition model according to the burr three-dimensional information.

[0093] The grinding range recognition model refers to a neural network model obtained through sample training in advance. The grinding range recognition model can match the area that needs to be polished and removed from the three-dimensional information of the burr. The grinding range is the result of the grinding range recognition model matching the three-dimensional information of the burr.

[0094] Step 201: matching the burr position from a preset burr position recognition model based on the burr feature.

[0095] The burr position recognition model refers to a neural network model obtained through sample training in advance. The burr position recognition model can identify the coordinate position of the burr from the image, where the coordinate position of the burr refers to the coordinate used to show the spatial position of the burr. The coordinate position can use the coordinate of the center point where the burr and the mandrel are connected. The coordinate position of the burr is selected by the staff according to the actual situation and will not be elaborated here. The burr position is the result obtained by the burr position recognition model based on the mandrel image recognition.

[0096] Step 202: Match the grinding stroke from a preset grinding stroke matching model according to the grinding range and the burr position.

[0097] The grinding device refers to a device used to grind burrs to remove burrs. The grinding device can be a grinding wheel, etc. The grinding device and its control method are selected by the staff according to the actual situation, and will not be described in detail here. The grinding stroke matching model refers to a neural network model obtained through sample training in advance. The grinding stroke matching model can match the path required for the grinding device to remove the burrs according to the area to be polished and the location of the burrs.

[0098] Step 203: Control a preset grinding device to grind the burrs according to the grinding stroke to remove the burrs.

[0099] The burrs on the surface of the mandrel are removed by grinding the burrs with a grinding device, thereby improving the qualified rate of the mandrel.

[0100] Reference Figure 4 The burr removal method further comprises the following steps:

[0101] Step 204: judging whether the burr length exceeds a preset burr length from a preset length judgment model according to the burr three-dimensional information.

[0102] The burr length refers to the upper limit of the length of a shorter burr. The length judgment model refers to a neural network model obtained by sample training in advance. The length judgment model can identify the length of the burr from the three-dimensional information of the burr and judge whether the length of the burr exceeds the burr length.

[0103] Step 205: When the length of the burr exceeds a preset burr length, a cutting position is matched from a preset cutting position matching model based on the burr feature.

[0104] The length of the burr exceeds the burr length, which means that the burr is too long. At this time, when the burr is directly ground by a grinding device, the burr is easy to bend, resulting in the inability to completely remove the burr. The shear position matching model refers to a neural network model obtained through sample training in advance. The shear position matching model can identify the coordinate position suitable for shearing the burr from the image. The shear position is the result of the shear position matching model matching the mandrel image.

[0105] Step 206: matching the burr diameter at the shearing position from a preset diameter judgment model according to the shearing position and the burr three-dimensional information.

[0106] The diameter judgment model refers to a neural network model obtained by sample training in advance. The diameter judgment model can identify the diameter value corresponding to the coordinate position in the three-dimensional information of the burr. The burr diameter is the diameter value of the shear position of the burr identified by the diameter judgment model.

[0107] Step 207: When the burr diameter falls within the preset grinding range, the grinding depth is matched from a preset grinding depth matching model according to the burr diameter.

[0108] The grinding range refers to the diameter range of burrs that cannot be directly sheared. If the burr diameter falls into the grinding range, it means that the burr diameter is too large and cannot be directly sheared. The burr needs to be ground first to reduce the burr diameter.

[0109] The grinding depth matching model refers to a neural network model obtained through sample training in advance. The grinding depth matching model can match the required grinding depth value according to the burr diameter. The grinding depth is the result obtained by the grinding depth matching model according to the burr diameter matching.

[0110] Step 208: Match the grinding stroke from a preset grinding stroke matching model according to the grinding depth, the shearing position and the mandrel image.

[0111] The grinding device refers to a device that reduces the diameter of small burrs by grinding to facilitate shearing by the shearing device. The grinding device and its control method are selected by the staff according to the actual situation and will not be described in detail here. The grinding stroke matching model refers to a neural network model obtained through sample training in advance. The grinding stroke matching model can match the path of the grinding device according to the required grinding depth, position and the spatial relationship of the position required to be ground shown in the picture. The grinding stroke is the result obtained by the grinding stroke matching model.

[0112] Step 209: Control a preset grinding device according to the grinding stroke to grind the burrs for easy shearing.

[0113] The burrs are ground by a grinding device, thereby reducing the diameter of the shearing position, thereby facilitating shearing by the shearing device.

[0114] Step 210: Match the shearing stroke from a preset shearing stroke matching model according to the shearing position and the mandrel image.

[0115] The shearing device refers to a device used to shear burrs to separate the burrs from the mandrel. The shearing device and its control method are selected by the staff according to the actual situation and will not be described in detail here. The shearing stroke matching model refers to a neural network model obtained through sample training in advance. The shearing stroke matching model can match the path for shearing burrs according to the spatial relationship between the coordinate position to be sheared and the coordinate position shown in the picture. The shearing stroke is the result obtained by matching the shearing stroke matching model.

[0116] Step 211: Control a preset shearing device to shear burrs according to the shearing stroke.

[0117] The overlong burrs are cut off by the shearing device, and the remaining burrs are ground by the grinding device after the cutting, so as to remove the burrs on the mandrel. When the burr diameter does not fall within the preset grinding range, the overlong burrs are directly cut off by the shearing device.

[0118] Reference Figure 5 The burr removal method further comprises the following steps:

[0119] Step 212: When the length of the burr exceeds the preset burr length, it is determined from a preset shearing space determination model whether there is a shearing space for the preset shearing device to extend into based on the burr feature and the shearing position.

[0120] The length of the burr exceeding the burr length indicates that the burr needs to be sheared by a shearing device. The shearing space refers to the space for the shearing device to extend to shear off the burr. The shearing space judgment model refers to a neural network model obtained through sample training in advance. The shearing space judgment model can judge from the image whether the shearing device can extend into the mandrel to shear off the burr.

[0121] Step 213: When there is no shearing space for the preset shearing device to extend into, the rotation position is matched from the preset rotation position matching model based on the burr feature.

[0122] The absence of a shearing space for the preset shearing device to extend into means that the overlong burrs cannot be cut short by the shearing device. The rotational position matching model refers to a neural network model obtained through sample training in advance. The rotational position matching model can match the coordinate position on the burr that is suitable for grasping when the burr needs to be rotated from the image, where suitable for grasping means that it is not easy to detach after grasping and it is more labor-saving when rotating. The rotational position is the result obtained by matching the rotational position matching model.

[0123] Step 214: Match the rotation force from a preset rotation force matching model according to the burr diameter.

[0124] The rotation force refers to the force of rotating the burr at the rotation position. The rotation force matching model refers to a neural network model obtained by sample training in advance, and the rotation force matching model can match the rotation force corresponding to the diameter of the burr in order to separate the burr from the core shaft.

[0125] Step 215: Control a preset rotating device to grab the burr according to the rotation position and rotate the burr according to the rotation force.

[0126] The rotating device is a device used to rotate the burr to separate the burr from the mandrel. The rotating device and its control method are selected by the staff according to the actual situation and are not described in detail here. When the burr cannot be sheared by the shearing device, the burr is rotated by the rotating device to twist the burr, thereby separating the burr from the mandrel.

[0127] Reference Figure 6 , the burr fixing method includes the following steps:

[0128] Step 300: When the length of the burr exceeds a preset burr length, a lifting position is matched from a preset lifting position matching model according to the mandrel image and the shearing stroke.

[0129] When the length of the burr exceeds the burr length, it means that the burr is too long, and the burr is easy to bend. The lifting position matching model refers to a neural network model obtained through sample training in advance. The lifting position matching model can match the coordinate position of the supporting burr to reduce the bending of the burr during shearing according to the spatial relationship of the burr shown in the picture and the shearing direction shown by the shearing stroke. The lifting position is the result obtained by matching the lifting position matching model.

[0130] Step 301: Control a preset lifting device to lift the burr according to the lifting position.

[0131] The lifting device refers to a device used to lift the burr. The lifting device and its control method are selected by the staff according to the actual situation, and will not be described here. When shearing and grinding burrs, the burr is supported by the lifting device, thereby reducing the bending of the burr and improving the convenience of burr removal.

[0132] Step 302: Control a preset clamping device to cooperate with the lifting device to clamp the burr according to the lifting position.

[0133] The clamping device refers to a device used to clamp burrs in cooperation with the lifting device. The clamping device and its control method are selected by the staff according to actual conditions and will not be elaborated here.

[0134] Step 303: matching a removal stroke from a preset removal stroke matching model according to the mandrel image and the shear stroke.

[0135] The removal stroke matching model refers to a neural network model obtained through sample training in advance. The removal stroke matching model can match the path of the lifting device and clamping device of the non-interfering shearing device to clamp the burr and remove it from the mandrel according to the spatial relationship and shear stroke shown in the picture. The removal stroke is the result obtained by matching the removal stroke matching model.

[0136] Step 304: After the shearing is completed by the shearing device, the clamping device and the lifting device are controlled to move according to the removal stroke to remove the burrs.

[0137] The clamping device and the lifting device are separated from the mandrel and clamp the burr at the same time, so that the burr is extracted to facilitate the cleaning of the burr debris, thereby improving the convenience of burr removal.

[0138] Reference Figure 7 The burr bending method includes the following steps:

[0139] Step 400: After the burrs are ground by the grinding device, the bending position is matched from a preset bending position matching model according to the shearing position and the mandrel image.

[0140] After the burr is ground by the grinding device, the diameter of the shear position is reduced, which makes it easy for the burr to bend back and forth with the shear position as the axis. The bending position matching model refers to a neural network model obtained through sample training in advance. The bending position matching model can match the force position that makes the burr easy to bend with the shear position as the axis according to the shear position and the burr shown in the picture. The bending position is the result obtained by matching the bending position matching model.

[0141] Step 401: matching a bending stroke from a preset bending stroke matching model according to the lifting position and the bending position.

[0142] The bending stroke matching model refers to a neural network model obtained through sample training in advance. The bending stroke matching model can match the path that causes the clamping device and the lifting device to first release the lifting position, then move to the bending position, and then clamp the burr at the bending position. The bending stroke is the result obtained by the bending stroke matching model.

[0143] Step 402: Control the preset clamping device and lifting device to move from the lifting position to the bending position and clamp the burr according to the bending stroke.

[0144] The clamping device and the lifting device are controlled to move to the bending position and clamp the burr at the bending position to facilitate bending the burr.

[0145] Step 403: Control the preset clamping device and the lifting device to cooperate with each other to reciprocately bend the burrs according to the preset bending parameters to facilitate shearing.

[0146] Bending parameters refer to the control parameters used to control the reciprocating bending burrs of the clamping device and the lifting device. The bending parameters are set by the staff according to actual conditions and will not be elaborated here.

[0147] Reference Figure 8 , the debris cleaning method includes the following steps:

[0148] Step 500: After the burrs are removed by a preset burr removal method, a grinding image of the mandrel is obtained.

[0149] The grinding image refers to a picture of the mandrel after being ground by the grinding device. The grinding image can be obtained through a camera. The method for obtaining the grinding image is selected by the staff according to the actual situation and will not be elaborated here.

[0150] Step 501: matching a debris area from a preset debris area recognition model according to a polishing image.

[0151] The debris region recognition model refers to a neural network model obtained by sample training in advance. The debris region recognition model can recognize the region where debris exists in the polishing image based on the polishing image. The debris region is the result obtained by matching the debris region recognition model.

[0152] Step 502: Matching a cleaning stroke from a preset cleaning stroke matching model according to the debris area.

[0153] The cleaning device refers to a device used to clean grinding debris. The cleaning device can use a brush or a vacuum fan. The cleaning device and its control method are selected by the staff according to the actual situation and will not be described in detail here. The cleaning stroke matching model refers to a neural network model obtained through sample training in advance. The cleaning stroke matching model can match the path of cleaning the debris area through the cleaning device.

[0154] Step 503: Control a preset cleaning device according to the cleaning stroke to remove debris on the mandrel from the surface of the mandrel.

[0155] The grinding debris on the mandrel is removed by the cleaning device, thereby facilitating further inspection or processing of the mandrel.

[0156] Reference Fig. 9 , a mandrel manufacturing method further comprises the following steps:

[0157] Step 600: When a burr feature exists, the burr posture is matched from a preset posture matching model according to the mandrel image.

[0158] The posture matching model refers to a neural network model obtained through sample training in advance. The posture matching model can identify the direction of the burr from the image. The burr posture is the result obtained by the posture matching model.

[0159] Step 601: Identify the spindle posture from a preset posture recognition model according to the spindle image.

[0160] The posture recognition model refers to a neural network model obtained through sample training in advance. The posture recognition model can identify the placement state of the mandrel from the picture. The mandrel posture is the result obtained by matching the posture recognition model.

[0161] Step 602: Match an adjustment step from a preset adjustment step matching model according to the spindle posture and the burr posture.

[0162] The adjustment step matching model refers to a neural network model obtained through sample training in advance. The adjustment step matching model can match the adjustment steps for adjusting the placement of the mandrel according to the mandrel posture and the burr posture. The placement of the mandrel is adjusted through the adjustment step to adjust the burr to a direction suitable for processing, wherein the direction suitable for processing refers to the direction toward the grinding device, and the burr is generally adjusted to a vertically upward direction. The adjustment step is the result obtained by matching the adjustment step matching model.

[0163] Step 603: matching the adjustment stroke from a preset adjustment stroke matching model according to the adjustment step.

[0164] The adjustment robot arm refers to a robot arm used to adjust the placement of the mandrel. The control method of the adjustment robot arm is selected by the staff according to the actual situation and will not be described here. The adjustment stroke matching model refers to a neural network model obtained through sample training in advance. The adjustment stroke matching model can match the path that the adjustment robot arm takes when adjusting the mandrel according to the adjustment steps. The adjustment stroke is the result obtained by matching the adjustment stroke matching model.

[0165] Step 604: adjusting the posture of the burr on the mandrel according to the preset adjustment stroke control of the adjustment robot arm to facilitate burr removal.

[0166] The orientation of the burr is adjusted by adjusting the robot arm, thereby adjusting the orientation of the burr to facilitate the removal of the burr by a burr removal method.

[0167] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A method for manufacturing a mandrel, characterized in that: include: acquiring a mandrel image of the mandrel; Detecting whether there is a burr feature on the surface of the mandrel from a preset burr detection model according to the mandrel image; When there are burr features, matching the burr three-dimensional information from a preset burr three-dimensional recognition model based on the burr features; Remove burrs by using a preset burr removal method according to the 3D information of the burrs; The burr removal method comprises: Matching the grinding range from a preset grinding range recognition model according to the burr three-dimensional information; Matching the burr position from a preset burr position recognition model based on the burr feature; Matching the grinding stroke from a preset grinding stroke matching model according to the grinding range and the burr position; Controlling a preset grinding device to grind the burrs according to the grinding stroke to remove the burrs; The burr removal method further comprises: Judging whether the burr length exceeds a preset burr length from a preset length judgment model according to the burr three-dimensional information; When the length of the burr exceeds a preset burr length, a shearing position is matched from a preset shearing position matching model based on the burr feature; Matching the shear stroke from a preset shear stroke matching model according to the shear position and the mandrel image; Control the preset shearing device to shear burrs according to the shearing stroke; The burr removal method further comprises: According to the shearing position and the burr three-dimensional information, the burr diameter at the shearing position is matched from the preset diameter judgment model; When the burr diameter falls within the preset grinding range, the grinding depth is matched from a preset grinding depth matching model according to the burr diameter; Matching the grinding stroke from a preset grinding stroke matching model according to the grinding depth, shearing position and mandrel image; The preset grinding device is controlled according to the grinding stroke to grind the burrs for easy shearing; The burr removal method further comprises: When the length of the burr exceeds the preset burr length, judging whether there is a shearing space for the preset shearing device to extend into from a preset shearing space judgment model based on the burr characteristics and the shearing position; When there is no shearing space for the preset shearing device to extend into, matching the rotation position from a preset rotation position matching model based on the burr feature; According to the burr diameter, the rotation force is matched from a preset rotation force matching model; Control the preset rotating device to grab the burr according to the rotating position and rotate the burr according to the rotating force; Also included is a burr fixing method, the burr fixing method comprising: When the length of the burr exceeds the preset burr length, the lifting position is matched from the preset lifting position matching model according to the mandrel image and the shear stroke; Control the preset lifting device to lift the burr according to the lifting position; According to the lifting position, the preset clamping device is controlled to cooperate with the lifting device to clamp the burr; matching a removal stroke from a preset removal stroke matching model according to the mandrel image and the shear stroke; After the shearing is completed by the shearing device, the clamping device and the lifting device are controlled to move according to the removal stroke to remove the burrs; Also included is a burr bending method, the burr bending method comprising: After the burrs are ground by the grinding device, the bending position is matched from a preset bending position matching model according to the shearing position and the mandrel image; Matching the bending stroke from a preset bending stroke matching model according to the lifting position and the bending position; According to the bending stroke, the preset clamping device and the lifting device are controlled to move from the lifting position to the bending position and clamp the burr; According to the preset bending parameters, the preset clamping device and the lifting device are controlled to cooperate with each other to reciprocate and bend the burrs for easy shearing.

2. A method for manufacturing a mandrel according to claim 1, characterized in that: Also included is a debris cleaning method, the debris cleaning method comprising: After the burrs are removed by a preset burr removal method, a grinding image of the mandrel is obtained; Matching a chip area from a preset chip area recognition model according to the polishing image; Matching a cleaning stroke from a preset cleaning stroke matching model according to the debris area; The preset cleaning device is controlled according to the cleaning stroke to remove the debris on the mandrel from the mandrel surface.

3. A method for manufacturing a mandrel according to claim 1, characterized in that: Also includes: When there is a burr feature, the burr posture is matched from a preset posture matching model according to the mandrel image; Recognize the mandrel posture from a preset posture recognition model according to the mandrel picture; Matching an adjustment step from a preset adjustment step matching model according to the mandrel posture and the burr posture; Matching an adjustment stroke from a preset adjustment stroke matching model according to the adjustment steps; The preset adjustment arm adjusts the posture of the burr on the mandrel according to the adjustment stroke control to facilitate the removal of the burr.

Images