Control method based on visual identification and automatic cutter sharpening system

Abstract

The invention discloses a control method based on visual identification and an automatic cutter sharpening system, the control method based on visual identification comprises a compensation step: an image sequence of a disc-shaped grinding wheel is acquired, and the image sequence comprises a plurality of first images; preprocessing a plurality of first images of the image sequence to obtain a plurality of first binary images; and performing image synthesis on the plurality of first binary images by adopting a preset algorithm to obtain a synthesized contour image. And identifying the radius variation of the disc-shaped grinding wheel based on the synthesized contour image and the standard contour image. And a compensation configuration matched with the radius variable quantity is inquired, and position compensation is conducted on the sharpening mechanism based on the compensation configuration. Adjustment can be accurately carried out in real time according to the abrasion condition of the grinding wheel based on the visual identification and position compensation mode, the situation that the tool grinding quality is reduced due to abrasion of the grinding wheel is effectively avoided, and stable and reliable continuous production is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer data processing, and particularly relates to a control method based on visual recognition and an automatic tool sharpening system. BACKGROUND

[0002] With the development of industrialization, mechanical processing is widely used in kitchen knife manufacturing. Through various mechanical equipment such as punch presses, grinding machines, etc., the forming, polishing and other processes of the tool can be well completed, which can meet the demand of large-scale market and is one of the main ways of kitchen knife processing at present.

[0003] The polishing process of the kitchen knife is a crucial link in the entire processing process, which directly affects the sharpness, durability, appearance quality and use performance of the kitchen knife. A Chinese invention patent with publication number CN119772669A discloses a data processing method, a data processing system and a full-automatic intelligent tool production equipment. The data processing method is realized based on the full-automatic intelligent tool production equipment, and the full-automatic tool production equipment includes a camera assembly, a mechanical hand and an edge opening grinder. The full-automatic intelligent tool production equipment is used for opening the edge of the tool. The data processing method includes: instructing the mechanical hand to grab the tool to be opened; acquiring first image data of the camera assembly, the image content of the first image data being the tool to be opened grabbed by the mechanical hand; identifying the tool type in the first image data; querying preset trajectory data matched with the tool type, the preset trajectory data including a first trajectory and a second trajectory; instructing the edge opening grinder to move to a target position according to the first trajectory; instructing the mechanical hand to grab the tool to be opened and move to an initial position, and opening the edge on the edge opening grinder according to the second trajectory.

[0004] The prior art reduces the control requirements of the mechanical hand by instructing the movement of the edge opening grinder to be compatible with the polishing and opening of different kitchen knives. However, the applicant finds that in the continuous operation process, the polishing quality of part of the kitchen knives cannot meet the index requirements, and it is difficult to ensure the stability and reliable continuous batch production of the kitchen knife product quality. SUMMARY

[0005] To solve the technical problems existing in the prior art, the purpose of the present application is to provide a control method based on visual recognition and an automatic tool sharpening system.

[0006] The purpose of the present application is achieved by the following technical solutions:

[0007] In a first aspect, the present application provides a control method based on visual recognition, which is implemented based on a tool automatic sharpening system, the tool automatic sharpening system comprising a camera assembly, a manipulator, a moving mechanism and a sharpening mechanism, the sharpening mechanism having a disc-shaped grinding wheel; wherein the control method comprises the following position compensation steps:

[0008] An image sequence of the disc-shaped grinding wheel is acquired, the image sequence comprising a plurality of first images, the plurality of first images being obtained by rotating sampling a local circumferential profile of the disc-shaped grinding wheel in a vertical direction by the camera assembly;

[0009] The plurality of first images of the image sequence are preprocessed to obtain a plurality of first binary images;

[0010] The plurality of first binary images are image-synthesized by using a preset algorithm to obtain a synthesized profile image;

[0011] Based on the synthesized profile image and a standard profile image, a radius variation of the disc-shaped grinding wheel is identified;

[0012] A compensation configuration matched with the radius variation is queried, and the sharpening mechanism is compensated in position based on the compensation configuration.

[0013] In combination with the first aspect, the present application provides a first specific implementation of the first aspect, specifically, the plurality of first binary images are image-synthesized by using a preset algorithm to obtain a synthesized profile image, specifically comprising:

[0014] The plurality of first images are globally threshold segmented;

[0015] The plurality of first images are anti-binary, the pixel gray value of the part of the disc-shaped grinding wheel in the first image is set to 1, and the pixel gray value of the part other than the disc-shaped grinding wheel is set to 0;

[0016] The plurality of first images are open operation morphological processed;

[0017] The plurality of first images are synthesized based on a profile frequency image generation algorithm to obtain a profile image.

[0018] In combination with the first aspect, the present application provides a second specific implementation of the first aspect, specifically, the compensation configuration matched with the radius variation is queried, specifically comprising:

[0019] When the radius variation ≤ a preset threshold S min , no position compensation is needed;

[0020] When the preset threshold S min < the radius variation < the preset threshold S n , a compensation configuration R nn is an integer greater than or equal to 0;

[0021] When the preset threshold S n When the radius change amount is less than the preset threshold S n+1 , a compensation configuration R n+1 is called.

[0022] When the preset threshold S max is less than the radius change amount, the disc-shaped grinding wheel is identified as being in a sharp wear state, and alert information for replacing or reconditioning the disc-shaped grinding wheel is generated.

[0023] In combination with the first aspect, the present application provides a third specific implementation of the first aspect, specifically, the control method further comprises:

[0024] The number of tools that have completed the sharpening process by the sharpening mechanism is accumulated, and a real-time tool sharpening value is output.

[0025] When the sharpening value meets a preset sharpening value, the position compensation step is re-executed.

[0026] In combination with the first aspect, the present application provides a fourth specific implementation of the first aspect, specifically, the control method, when executing the position compensation step, further comprises the following steps:

[0027] A second image of the disc-shaped grinding wheel is obtained, the second image being obtained by the camera assembly shooting the end face of the disc-shaped grinding wheel from the horizontal direction;

[0028] The second image is identified, and the outer circle contour of the disc-shaped grinding wheel is output.

[0029] The actual cross-sectional area of the disc-shaped grinding wheel is calculated according to the outer circle contour of the disc-shaped grinding wheel.

[0030] Based on the preset cross-sectional area and the actual cross-sectional area, the cross-sectional area loss of the disc-shaped grinding wheel is obtained.

[0031] When the cross-sectional area loss is greater than a preset threshold, the disc-shaped grinding wheel is identified as being in a sharp wear state, and alert information for replacing or reconditioning the disc-shaped grinding wheel is generated.

[0032] In combination with the first aspect, the present application provides a fifth specific implementation of the first aspect, specifically, the second image is identified, and the outer circle contour of the disc-shaped grinding wheel is output, specifically comprising:

[0033] The second image is converted into a gray-scale image, and the gray-scale image is binarized to obtain a second binary image;

[0034] The second binary image is identified using a sub-pixel edge detection method to obtain a plurality of contour edge points of the disc-shaped grinding wheel.

[0035] The plurality of profile edge points are input into a MATLAB application program, and an outer circle profile of the disc-shaped grinding wheel is output.

[0036] In a second aspect, the present application also provides a tool automatic sharpening system, which comprises:

[0037] A rack is provided with a moving mechanism, a sharpening mechanism, a camera assembly, a light source assembly and a controller, the moving mechanism is fixedly connected to the rack, the moving mechanism is drivingly connected to the sharpening mechanism, the sharpening mechanism has a disc-shaped grinding wheel, and the camera assembly and the light source assembly are arranged on the sides of the sharpening mechanism, respectively;

[0038] A robot is located on the side of the rack, and the robot is used to grab the tool to be sharpened from the loading tool holder and move the tool to be sharpened to the sharpening mechanism for sharpening processing;

[0039] The controller is configured to perform a control method as follows:

[0040] An image sequence collected by the camera assembly is obtained, the image sequence comprises a plurality of first images, and the plurality of first images are obtained by rotating sampling the local circumferential profile of the disc-shaped grinding wheel in the vertical direction by the camera assembly;

[0041] The plurality of first images of the image sequence are preprocessed to obtain a plurality of first binary images;

[0042] A preset algorithm is used to perform image synthesis on the plurality of first binary images to obtain a synthesized profile image;

[0043] Based on the synthesized profile image and a standard profile image, a radius variation of the disc-shaped grinding wheel is identified;

[0044] A compensation configuration matched with the radius variation is queried, and the moving mechanism is controlled to perform position compensation on the sharpening mechanism based on the compensation configuration.

[0045] In combination with the second aspect, the present application provides a first specific implementation mode of the second aspect, specifically, the moving mechanism comprises:

[0046] A transverse moving screw pair is fixedly connected to the rack;

[0047] A longitudinal moving screw pair is connected to the transverse moving screw pair, and the longitudinal moving screw pair is connected to the sharpening mechanism;

[0048] The controller controls the longitudinal moving screw pair of the moving mechanism to perform position compensation on the sharpening mechanism based on the compensation configuration, so as to adjust the distance between the sharpening mechanism and the robot.

[0049] In combination with the second aspect, the present application provides a second specific implementation of the second aspect, in particular, the rack is further provided with a white background plate, the white background plate is arranged on one side of the disc-shaped grinding wheel of the sharpening mechanism, and the surface of the white background plate is arranged in parallel with the end surface of the disc-shaped grinding wheel.

[0050] In combination with the second aspect, the present application provides a third specific implementation of the second aspect, in particular, the sharpening mechanism has a disc-shaped grinding wheel, the outer circumference of the disc-shaped grinding wheel is used for sharpening, the disc-shaped grinding wheel is arranged horizontally and transversely, and the outer circumference of the disc-shaped grinding wheel faces the mechanical hand, and the outer circumference of the disc-shaped grinding wheel is used for sharpening.

[0051] Through the in-depth research of the applicant, it is found that the prior art opens the blade of the grinder by moving the command to be compatible with the polishing and sharpening of different kitchen knives, and reduces the control requirements of the mechanical hand. However, the prior art ignores the wear and tear of the grinding wheel of the sharpening grinder after a large number of continuous polishing, which directly affects the polishing quality of the product, causing the polishing quality of part of the kitchen knives to fail to meet the technical problem of the index requirement.

[0052] Compared with the prior art, the present application has at least the following beneficial effects:

[0053] The present application provides a control method based on visual recognition, which is realized based on a tool automatic sharpening system, the tool automatic sharpening system comprising a camera assembly, a mechanical hand, a moving mechanism and a sharpening mechanism, and the sharpening mechanism having a disc-shaped grinding wheel. The control method comprises the following position compensation steps: obtaining an image sequence of the disc-shaped grinding wheel, the image sequence comprising a plurality of first images, and the plurality of first images being obtained by rotating sampling the local circumferential profile of the disc-shaped grinding wheel in the vertical direction by the camera assembly. The plurality of first images of the image sequence are preprocessed to obtain a plurality of first binary images. The plurality of first binary images are synthesized by a preset algorithm to obtain a synthesized profile image. Based on the synthesized profile image and a standard profile image, the radius variation of the disc-shaped grinding wheel is identified. The compensation configuration matched with the radius variation is queried, and the position of the sharpening mechanism is compensated based on the compensation configuration.

[0054] The present application obtains and processes the image sequence of the disc-shaped grinding wheel, identifies the radius variation of the grinding wheel, and then compensates the position of the sharpening mechanism. This way of visual recognition and position compensation can adjust in real time and accurately according to the wear of the grinding wheel, effectively avoid the decrease of the tool polishing quality caused by the wear of the grinding wheel, thereby significantly improve the stability of the kitchen knife polishing quality, ensure that each kitchen knife can meet the expected quality index, and realize stable and reliable continuous batch production.

[0055] The control method based on visual recognition realizes automatic monitoring and compensation, avoids the problem of uncontrollable product quality caused by the error of the mill in the prior art, automatically identifies and calculates the change amount of the grinding wheel radius through a preset algorithm, and automatically compensates the position of the grinding edge mechanism according to the compensation configuration, thereby improving the automation degree and intelligent level of production.

[0056] The application also provides a tool automatic grinding edge system, and the device composition and the control method of the controller, which also achieve the technical effects. BRIEF DESCRIPTION OF DRAWINGS

[0057] Figure 1 A system composition diagram of a tool automatic grinding edge system of the application;

[0058] Figure 2 A rack structure schematic diagram of a tool automatic grinding edge system of the application;

[0059] Figure 3 A structure schematic diagram of a moving mechanism and a grinding edge mechanism of a tool automatic grinding edge system of the application;

[0060] Figure 4 A layout diagram of a first camera of a tool automatic grinding edge system of the application;

[0061] Figure 5 A flowchart of a control method based on visual recognition of the application;

[0062] In the drawings:

[0063] 100 - manipulator;

[0064] 200 - grinding edge mechanism, 210 - disc-shaped grinding wheel;

[0065] 300 - moving mechanism, 310 - transverse moving screw pair, 320 - longitudinal moving screw pair. DETAILED DESCRIPTION

[0066] For the convenience of understanding the present application, the technical solutions and advantages of the present application are further described in detail below in combination with the drawings and examples. The mechanisms or methods not described in the present application can refer to the prior art. The specific structure and characteristics of the present application are described below by way of example, which should not constitute any limitation on the present application. At the same time, any one of the technical features mentioned below (including implied or disclosed), as well as any one of the technical features directly shown or implied in the drawings, can continue to be combined or deleted between these technical features, thereby forming more other embodiments that can not be directly or indirectly mentioned in the present application. The preferred embodiments of the present application are shown in the drawings. However, the present application can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present application more thorough and comprehensive.

[0067] At present, in the kitchen knife manufacturing field, with the acceleration of industrialization process, automatic batch processing has become the mainstream production mode. In order to better adapt to the production needs of different kitchen knives, the kitchen knife production equipment in the industry has been continuously optimized and improved. The prior art moves the opening blade grinder by command to be compatible with the polishing opening blade of different kitchen knives, and reduces the control requirements of the mechanical hand. However, the applicant found that in the continuous operation process, the polishing quality of part of the kitchen knives cannot meet the index requirements, and it is difficult to ensure the stability and reliable continuous batch production of the kitchen knife product quality.

[0068] Through the applicant's in-depth research, it is found that the prior art moves the opening blade grinder by command to be compatible with the polishing opening blade of different kitchen knives, and reduces the control requirements of the mechanical hand. However, the prior art ignores the wear and tear of the grinding wheel of the opening blade grinder after a large number of continuous polishing, which directly affects the polishing quality of the product, causing the polishing quality of part of the kitchen knives to not meet the index requirements.

[0069] Therefore, with reference to Figure 5 , the embodiment of the present application provides a flowchart of a control method based on visual recognition. The main method of the present application is to obtain an image sequence of a disc-shaped grinding wheel and process it to identify the radius variation of the grinding wheel, and then compensate the position of the blade grinding mechanism. This way based on visual recognition and position compensation can adjust in real time and accurately according to the wear of the grinding wheel, effectively avoid the decline of the tool polishing quality caused by the wear of the grinding wheel, thereby significantly improve the stability of the kitchen knife polishing quality, ensure that each kitchen knife can meet the expected quality index, and realize stable and reliable continuous batch production.

[0070] Embodiment one

[0071] As shown in Figures 1-4 , the preferred structure of the automatic blade grinding system of the present application.

[0072] In the present application, the automatic knife sharpening system is used for kitchen knife sharpening. The automatic knife sharpening system comprises a rack and a robot. The rack is provided with a moving mechanism, a sharpening mechanism, a camera assembly, a light source assembly and a controller. The moving mechanism is fixedly connected to the rack. The moving mechanism is drivingly connected to the sharpening mechanism. The sharpening mechanism has a disc-shaped grinding wheel. The camera assembly and the light source assembly are arranged on the sides of the sharpening mechanism. The robot is arranged on the side of the rack. The robot is used to grab the knife to be sharpened from the knife rack and move the knife to be sharpened to the sharpening mechanism for sharpening.

[0073] The automatic knife sharpening system of the present application realizes automatic monitoring and compensation, avoiding the problem of uncontrollable product quality caused by grinding machine errors in the prior art. The preset algorithm is used to automatically identify and calculate the grinding wheel radius variation, and the sharpening mechanism is automatically compensated in position according to the compensation configuration, thereby improving the automation and intelligence level of production.

[0074] In a specific implementation, the rack is a metal frame assembled from multiple square tubes and a metal shell. A workbench is arranged above the rack to deploy the moving mechanism, the sharpening mechanism, the camera assembly and the light source assembly, etc. The electrical elements such as the controller are installed in the electrical cabinet of the rack.

[0075] In some implementations, the moving mechanism is responsible for driving the sharpening mechanism to move along a preset path to adapt to the sharpening requirements at different positions, and the moving mechanism is also responsible for the position compensation of the sharpening mechanism. Specifically, the moving mechanism comprises:

[0076] A transverse moving screw pair is fixedly connected to the rack.

[0077] A longitudinal moving screw pair is connected to the transverse moving screw pair, and the longitudinal moving screw pair is connected to the sharpening mechanism.

[0078] The controller controls the longitudinal moving screw pair of the moving mechanism to compensate the position of the sharpening mechanism based on the compensation configuration, so as to adjust the distance between the sharpening mechanism and the robot.

[0079] In the automatic knife sharpening system, the design of the moving mechanism ensures that the sharpening mechanism can accurately move along a preset path to perform position compensation to ensure sharpening accuracy. For this purpose, the screw pair is a very suitable choice. The screw pair has high positioning accuracy and can provide very precise fine movement control, which is crucial for the wear compensation of the grinding wheel of the sharpening mechanism, as the position of the sharpening mechanism needs to be accurately controlled to ensure the sharpening quality. The screw pair provides excellent positioning accuracy, stability and repeatability.

[0080] It can be understood that the transverse moving screw pair is responsible for driving the sharpening mechanism to move in the horizontal direction (X-axis). The screw pair is fixedly connected to the rack to provide a stable transverse moving platform for the sharpening mechanism. The longitudinal moving screw pair is responsible for driving the sharpening mechanism to move in the longitudinal direction (Y-axis) perpendicular to the horizontal direction. The longitudinal moving screw pair is connected to the transverse moving screw pair to realize the position compensation relative to the position between the mechanical hands. The controller controls the longitudinal moving screw pair of the moving mechanism based on the compensation configuration to compensate the position of the sharpening mechanism.

[0081] In specific implementation, the transverse moving screw pair and the longitudinal moving screw pair are each provided with a guide rail pair to realize guiding and provide stability through the guide rail and the sliding block.

[0082] In specific implementation, as shown in Figure 3 The sharpening mechanism has a disc-shaped grinding wheel, the outer circumference of the disc-shaped grinding wheel is used for sharpening, the disc-shaped grinding wheel is horizontally arranged, the outer circumference of the disc-shaped grinding wheel faces the mechanical hand, and the outer circumference of the disc-shaped grinding wheel is used for sharpening. The main function of the sharpening mechanism is to grind the cutting tool through the high-speed rotating grinding wheel to complete the sharpening work of the cutting edge. In the automatic system, the sharpening mechanism cooperates with the mechanical hand and the controller to ensure the accuracy and efficiency of the cutting tool sharpening.

[0083] The sharpening mechanism is composed of a base, a motor, a transmission assembly, a main shaft and a disc-shaped grinding wheel. The base is connected to the transverse moving screw pair. The outer circumference of the disc-shaped grinding wheel is the part directly contacting the cutting edge, which is responsible for grinding and trimming the cutting edge to achieve the required sharpness and shape. The outer circumference of the grinding wheel faces the mechanical hand, which can ensure that when the mechanical hand grabs the cutting tool and moves to the sharpening mechanism, the cutting edge can directly contact the abrasive of the grinding wheel, facilitating the sharpening operation.

[0084] In a specific implementation, the sharpening mechanism of the present application is a mobile design. Because the length, width, thickness and other dimensions of each cutting tool are different, the shape and cutting angle of the cutting edge also differ. Combined with position compensation, the sharpening mechanism can be adjusted with high precision to ensure the accuracy of the contact position and angle between the disc-shaped grinding wheel and the cutting tool, thereby obtaining ideal sharpening quality. Different types of cutting tools may require different sharpening processes, and the automatic adjustment of the sharpening mechanism combined with the disc-shaped grinding wheel wear detection can flexibly cope with these differences. For cutting tools of the same type and small specification difference, the processing path of the mechanical hand does not need to be changed or slightly adjusted to meet the processing needs of different cutting tools. Moreover, combined with the visual recognition and position compensation based method, the sharpening mechanism can be adjusted in real time and accurately according to the wear condition of the disc-shaped grinding wheel, effectively avoiding the decline of the cutting tool polishing quality caused by the wear of the disc-shaped grinding wheel, thereby significantly improving the stability of the kitchen knife polishing quality, ensuring that each kitchen knife can meet the expected quality indicators, and realizing stable and reliable continuous mass production.

[0085] It can be seen that the automatic adjustment of the position of the sharpening mechanism and the position compensation of the grinding wheel wear can realize accurate adjustment according to the specific shape, size, angle and the like of each tool, and realize high-quality batch continuous production operation. The high-precision position-adjustable design of the sharpening mechanism is particularly effective for tools with small shape and size differences, and can flexibly meet the processing needs of various tools without complex programming and path adjustment of the mechanical hand. In production, especially in small-batch tool production, this way is more flexible and economical. Because compared with the position change of the sharpening mechanism, it is simple and easier to implement. And for some tools, it may be faster to adjust the position of the sharpening mechanism with high precision than to adjust the path of the mechanical hand, so that the adjustment can be quickly made without waiting for the design of the complex path of the mechanical hand.

[0086] As shown in FIG. 4, the camera assembly includes a first industrial camera and a second industrial camera. The first industrial camera is installed above the disc-shaped grinding wheel to capture a sequence of images of the grinding wheel, i.e., a plurality of first images. The second industrial camera is fixedly installed on the base of the sharpening mechanism through a support and moves synchronously with the sharpening mechanism. The optical axis of the second industrial camera and the axis of the disc-shaped grinding wheel are coaxial and at the same level. The second industrial camera is used to acquire second images.

[0087] In combination with the implementation of the camera assembly, the light source assembly is arranged directly below the first camera, and the disc-shaped grinding wheel is located between the first camera and the light source assembly. The light source assembly is fixed on the workbench, and the light source assembly adopts a telecentric parallel light source. In a preferred embodiment, a blue LED lamp bead with a wavelength of 460 nm is used, which has a lower reflectivity than a white LED. Such a choice can suppress the scattering and reflection of the circumferential surface of the grinding wheel, obtain an image with a sharper profile edge, and meet the subsequent profile analysis.

[0088] In a preferred embodiment, the first industrial camera is equipped with a double-telecentric lens, and a telecentric parallel light source is used to build an image acquisition system to acquire high-quality grinding wheel profile images. In a preferred embodiment, the first industrial camera uses a MV-CH120-10TM type black and white gray industrial camera, and uses a SONY IMX253 CMOS chip with a pixel size of 3.45 μm.

[0089] In a preferred embodiment, the front side of the disc-shaped grinding wheel is a circumferential surface facing the mechanical hand, which is used for sharpening the tool to be sharpened. The rear side of the disc-shaped grinding wheel is the other side of the principle mechanical hand. Specifically, the first industrial camera and the light source assembly are both arranged on the rear side of the disc-shaped grinding wheel. In this way, the pollution of the first industrial camera and the light source assembly by debris during processing can be avoided, such as being attached to the optical assembly of the double-telecentric lens or the light-transmitting glass of the light source assembly, which affects the image acquisition.

[0090] In a preferred embodiment, the rack is further provided with a white background plate, which is arranged on one side of the disc-shaped grinding wheel of the sharpening mechanism, and the surface of the white background plate is arranged in parallel with the end surface of the disc-shaped grinding wheel. The white background plate and the second industrial camera are arranged opposite to each other, and the disc-shaped grinding wheel is located between the white background plate and the second industrial camera, which aims to improve the accuracy of the contour recognition of the contour of the disc-shaped grinding wheel in the second image. The white background plate provides a uniform and high-contrast background, which makes the contour edge more sharp, and also enhances the accuracy of image processing, which is convenient for subsequent image processing and analysis.

[0091] In a specific embodiment, the manipulator is arranged beside the rack, and the execution end of the manipulator is provided with a tool clamp, which is used to grab the tool to be sharpened. Specifically, the manipulator grabs the back of the tool to be sharpened, which is the position of the non-sharp edge and the handle.

[0092] Specifically, the manipulator moves the tool to be sharpened from one position to another position, such as from the loading rack to the sharpening mechanism, or moves the tool that has completed the sharpening to the corresponding designated position, such as the quality inspection position or the unloading rack position, through precise control of the movement. The controller generates control instructions according to the data processing results and transmits them to the manipulator. For example, the controller instructs the manipulator to grab and position according to the tool type and target trajectory, and instructs the moving mechanism to drive the sharpening mechanism to perform the sharpening operation of the tool according to the preset trajectory. Specifically, the manipulator is a device commonly used in the art, and will not be described here.

[0093] For other technical implementations of the tool automatic sharpening system, please refer to the disclosure of the invention patent with publication number CN119772669A.

[0094] In a preferred embodiment, the tool automatic sharpening system further comprises a loading conveyor belt and an unloading conveyor belt, and the loading conveyor belt and the unloading conveyor belt are both provided with a plurality of placement racks for placing tools. The loading conveyor belt transports a plurality of tools to be sharpened to the side of the manipulator through the plurality of placement racks, and the manipulator grabs the tool to be sharpened from the placement rack on the loading conveyor belt. After the sharpening is completed, the manipulator places the sharpened tool on the placement rack of the unloading conveyor belt.

[0095] Embodiment two

[0096] Reference Figure 5The embodiment two of the present application provides a flowchart of a control method based on visual recognition. The present application provides a control technology of an automatic tool production device. The control method is based on the cooperative work of a camera assembly, a mechanical hand, a moving mechanism and a sharpening mechanism, can be compatible with different tool types, and can realize stable quality and batch continuous production of automatic tool sharpening operation. The control method obtains an image sequence of a disc-shaped grinding wheel, processes the image sequence, recognizes a radius variation of the disc-shaped grinding wheel, and then performs position compensation on the sharpening mechanism. This way based on visual recognition and position compensation can adjust in real time and accurately according to the wear of the grinding wheel, effectively avoid the quality decline of the tool polishing caused by the wear of the grinding wheel, and significantly improve the stability of the kitchen knife polishing quality, so that each kitchen knife can reach the expected quality index, and stable and reliable continuous batch production is realized.

[0097] The control method of the present application can be executed by a control system which can be realized in the form of hardware and / or software, and the control system can be configured in an electronic device of the automatic tool sharpening system, such as a controller. The control method of the present application is realized based on the automatic tool sharpening system, and specifically, the automatic tool sharpening system is completely consistent with that described in embodiment 1. As shown in the figure, the control method based on visual recognition includes the following position compensation steps: Figure 5

[0098] Step 501: obtaining an image sequence of the disc-shaped grinding wheel, the image sequence including a plurality of first images, and the plurality of first images being obtained by rotating sampling the local circumferential profile of the disc-shaped grinding wheel in the vertical direction by the camera assembly.

[0099] In a possible implementation, the image sequence is obtained by the first industrial camera of the camera assembly in embodiment 1, and the image sequence is combined with the image sequence obtained by the second industrial camera of the camera assembly in embodiment 1. Figure 4 As can be seen, for example, the controller starts the sharpening mechanism, controls the first camera to continuously and randomly capture a plurality of photos, obtains a plurality of first images of the disc-shaped grinding wheel in the circumferential direction, and forms an image sequence.

[0100] The visual detection based on a single cross-sectional profile cannot accurately reflect the actual wear profile of the disc-shaped grinding wheel. Therefore, in the present application, the detection is performed by synthesizing the circumferential direction multi-section profile, the circumferential profile of the disc-shaped grinding wheel is rotated and sampled, the circumferential profile information of the disc-shaped grinding wheel is obtained, and the radius variation is detected according to the synthesized profile frequency image, for position compensation.

[0101] Step 502: pre-processing a plurality of first images of the image sequence to obtain a plurality of first binary images.

[0102] In a possible implementation, the pre-processing includes image filtering and image enhancement. The implementation process of this step S502 includes: ​

[0103] Step 5021: filtering the first image by using guided filtering.

[0104] For example, guided filtering assumes that there is a local linear relationship within the image, and considers that the image is a non-analytic two-dimensional function, but can be expressed by an analytic linear function locally:

[0105]

[0106] In the formula, q i represents the filtered output image, I i is a guided image (the first image), Q k is a local window, a k and b k are constant coefficients in the k local window.

[0107] In order to reduce distortion, the mean square error is used for optimization:

[0108]

[0109] In the formula, ε is a compensation coefficient, p i is an original image (the first image).

[0110] Specifically, the least square method can be used to calculate the values of a k and b k . The effect of guided filtering is slightly better, which can remove a large number of noise points while retaining the contour edge details of the disc-shaped grinding wheel.

[0111] Step 5022: performing gray scale transformation on the first image after filtering.

[0112] Gray scale transformation refers to mapping the pixel gray scale values of an image according to a certain relationship, which can be divided into linear gray scale transformation and nonlinear gray scale transformation. In the present application, piecewise linear transformation is used for image enhancement.

[0113] Step 5023: converting the first image after gray scale transformation into a binary image, and setting a threshold to divide the pixels in the image into foreground (grinding wheel contour) and background. Common binarization methods include global thresholding method and adaptive thresholding method.

[0114] Step 503: performing image synthesis on a plurality of first binary images by using a preset algorithm to obtain a synthesized contour image.

[0115] In the present application, the single section in the traditional visual detection cannot reflect the global contour information of the grinding wheel, and the detection precision is poor. The present application synthesizes the multi-section contour images of the disc-shaped grinding wheel for subsequent analysis. In one possible implementation manner, the implementation process of the step 503 includes:

[0116] Step 5031: Global threshold segmentation is performed on the plurality of first images.

[0117] In step 5031, the disc-shaped grinding wheel part in the first image is separated from the background, providing a basis for subsequent binaryzation processing. For example, a global threshold is selected, which is used to divide the pixels in the image into foreground (disc-shaped grinding wheel part) and background (non-disc-shaped grinding wheel part). The global threshold can be automatically selected by histogram analysis, Otsu method, etc., or manually set. The gray value of each pixel in the first image is compared with the global threshold, and the pixel greater than the threshold is classified as foreground, and the pixel less than or equal to the threshold is classified as background.

[0118] It can be understood that the gray difference between the background of the preprocessed contour image and the grinding wheel contour in the image is large, so the contour edge can be determined by threshold segmentation. Threshold segmentation is one of the most commonly used image segmentation methods, which divides the pixels into different subsets according to the gray value of the image pixels, and each subset corresponds to different features.

[0119] Step 5032: Inverse binaryzation is performed on the plurality of first images, and the pixel gray value of the disc-shaped grinding wheel part in the first image is set to 1, and the pixel gray value of the non-disc-shaped grinding wheel part is set to 0.

[0120] In step 5032, the purpose is to generate a clear binary image, highlighting the contour of the disc-shaped grinding wheel, to facilitate subsequent morphological processing and contour extraction. For example, the segmented image is subjected to inverse binaryzation processing, i.e. the pixel gray value of the disc-shaped grinding wheel part is set to 1, and the pixel gray value of the non-disc-shaped grinding wheel part is set to 0. In this way, the disc-shaped grinding wheel can be represented in white (pixel value 1) and the background in black (pixel value 0) in the binary image.

[0121] Step 5033: Opening morphological processing is performed on the plurality of first images.

[0122] In this application, due to the environmental factors (dust, etc.) existing in the production workshop, which will interfere with the subsequent detection process, therefore, it is necessary to reduce the interference points in the contour image by image morphological operation.

[0123] In one possible implementation, the use of opening morphological operation can achieve the removal of discrete noise points and edge burrs. Opening is a mathematical morphological image processing operation used to remove small objects, discrete noise points and edge burrs in the image while preserving the main structure of the image. Opening consists of two basic operations: erosion and dilation. Specifically, opening first performs erosion operation on the image, and then performs dilation operation on the eroded image.

[0124] For example, the erosion operation is a structure element based image processing method for reducing the size of foreground objects in an image. The purpose of the erosion operation is to remove small objects and edge spurs in an image:

[0125] h(x, y) = min (x′，y')：elem(x',y')≠0 g(x + x', y + y');

[0126] where elem represents a structure element, (x, y) is the position of the anchor point of the structure element, (x', y') is the position of an element with value 1 in the structure element relative to its anchor point, h(x, y) is the output image, and g(x, y) is the input binary image.

[0127] For example, the dilation operation is a structure element based image processing method for increasing the size of foreground objects in an image. The purpose of the dilation operation is to fill in small holes in an image and connect disconnected parts.

[0128] h(x, y) = max (x′，y′)：elem(x',y′)≠0 g(x + x', y + y');

[0129] where elem represents a structure element, (x, y) is the position of the anchor point of the structure element, (x', y') is the position of an element with value 1 in the structure element relative to its anchor point, h(x, y) is the output image, and g(x, y) is the input binary image.

[0130] Step 5034: Synthesize the plurality of first images to obtain a contour image based on a contour frequency image generation algorithm.

[0131] Through the contour frequency image generation algorithm, the information of multiple images is integrated to generate a complete and accurate synthesized contour image, which provides high-quality contour data for subsequent radius change amount recognition. For example, for each pixel position, the frequency of the position being foreground (pixel value being 1) in multiple first binary images is counted. The frequency value can be normalized to the interval [0, 1] to represent the probability of the position belonging to the contour of the disc-shaped grinding wheel.

[0132] In one possible implementation, a contour frequency image is generated based on the contour frequency value, and a region with a high frequency value represents a contour of the disc-shaped grinding wheel more likely. According to the contour frequency image, a region with a frequency value higher than a certain threshold (such as 0.5) is determined as part of the synthesized contour image. In this way, the information of multiple first images can be integrated to obtain a more complete and more accurate contour of the disc-shaped grinding wheel.

[0133] In one possible implementation, the contour frequency image generation algorithm can simply sum the first images in the image sequence at the same pixel position. The mathematical expression of the contour frequency image generation algorithm is as follows:

[0134]

[0135] In the formula, N is the input binary image sequence after morphological processing, h(x, y) is the pixel value of the (x, y) point of the k-th image in the input sequence after binary inversion and morphological processing, and freq(x, y) is the value of the (x, y) point of the synthesized frequency image.

[0136] Step 504: Based on the synthesized contour image and the standard contour image, the radius variation of the disc-shaped grinding wheel is identified.

[0137] In a possible implementation, the implementation process of step 504 includes:

[0138] Step 5041: The actual contour is extracted from the synthesized contour image.

[0139] For example, the edge detection algorithm (such as the Canny algorithm) is used to extract the contour line of the disc-shaped grinding wheel from the synthesized contour image. The Canny algorithm can effectively detect the edges in the image and generate a clear contour line.

[0140] Step 5042: The long side of the synthesized contour image is identified.

[0141] For example, the long side is identified from the extracted contour of the synthesized contour image. It can be understood that the synthesized contour image is actually a cross-sectional contour image of the disc-shaped grinding wheel, which is like a rectangle. The four vertices of the contour can be determined by calculating the polygon approximation of the contour (such as using the cv2.approxPolyDP function of OpenCV), and then the long side is identified.

[0142] Step 5043: The contour pixel data of the standard contour image is called.

[0143] For example, the standard contour image has pre-associated stored contour pixel data obtained by S5041 and S5042, that is, the long side of the standard contour. In a possible implementation, the contour pixel data can be directly the pixel number value of the long side of the standard contour.

[0144] Step 5044: The pixel difference value between the long side of the synthesized contour image and the long side of the standard contour image is calculated pixel by pixel.

[0145] For example, the contour pixel data of the standard contour image is the pixel number value of the long side of the standard contour, and the pixel number value of the long side of the synthesized contour image is calculated pixel by pixel. The pixel number value of the long side of the standard contour minus the pixel number value of the long side of the synthesized contour image is used to obtain the pixel difference value. That is, the pixel difference value = the pixel number value of the long side of the standard contour image - the pixel number value of the long side of the synthesized contour image.

[0146] Step 5045: The radius variation is output.

[0147] Further, in actual position compensation, the radius wear value of the disc-shaped grinding wheel can be derived from the radius variation, i.e., radius variation = pixel difference value x pixel size, and the pixel size is the actual physical size of each pixel in the image. The pixel size can be obtained through the calibration process of the camera or can be found from the parameter specification of the camera.

[0148] For example, the pixel difference value is 10 pixels, and the pixel size of the camera is 0.05 mm / pixel. Then, the calculation process of the radius variation is as follows: radius variation = 10 pixels x 0.05 mm / pixel = 0.5 mm.

[0149] Step 505: querying a compensation configuration matched with the radius variation, and performing position compensation on the grinding mechanism based on the compensation configuration.

[0150] In this application, the wear of the disc-shaped grinding wheel is the main cause of the failure of the grinding quality of the tool. Therefore, position compensation is performed on the grinding mechanism based on the compensation configuration. Specifically, a corresponding control signal is generated according to the compensation configuration, and the control signal is sent to the stepping motor of the longitudinal movement mechanism of the movement mechanism to perform a corresponding movement operation. For example, the stepping motor accurately moves the grinding mechanism to the specified position coordinates according to the pulse number and direction signal of the control signal. For example, the feed amount of the longitudinal screw pair is set according to the radius variation, and the coordinate value required by the grinding mechanism for compensation is directly changed to achieve the purpose of compensation. The compensation principle is not complex, and the difficulty lies in how to identify the radius variation.

[0151] For example, a compensation configuration matched with the identified radius variation of the disc-shaped grinding wheel is queried. The compensation configuration stores the compensation measures to be taken for different radius variations, including the movement direction and movement distance.

[0152] In a possible implementation, in step 505, the compensation configuration matched with the radius variation is queried, and specifically includes:

[0153] Step 505A: when the radius variation is less than or equal to a preset threshold S min , position compensation is not required.

[0154] It can be understood that the radius variation is less than or equal to S min , which indicates that the wear of the grinding wheel is slight, and position compensation is not required. For the slightly worn grinding wheel, unnecessary compensation operation is avoided, and in addition, a certain visual recognition error can be compatible, so as to reduce system intervention and improve efficiency.

[0155] Step 505B: when the radius variation is greater than the preset threshold S min<The radius change <Preset threshold S n At that time, the compensation configuration R is invoked. n , where n is an integer greater than or equal to 0.

[0156] In this application, when the radius change meets certain conditions, the compensation configuration Rn is invoked. The compensation configuration Rn is pre-set to handle wear conditions within this range. By invoking the corresponding compensation configuration, the position of the grinding mechanism is adjusted to ensure the accuracy of the machining process.

[0157] Step 505C: When the preset threshold S n <The radius change <Preset threshold S n+1 At that time, the compensation configuration R is invoked. n+1 .

[0158] Understandably, when the radius change meets certain conditions, the compensation configuration Rn is invoked. The compensation configuration Rn is pre-set to handle wear conditions within that range. By using multiple layers of different compensation configurations, it adapts to different wear levels, allowing for better adjustments and ensuring the accuracy of the machining process.

[0159] Step 505D: When the preset threshold S max When the radius change is less than or equal to the specified amount, the disc grinding wheel is identified as being in a state of rapid wear, and an alarm message is generated indicating that the disc grinding wheel needs to be replaced or re-corrected.

[0160] It is understandable that when the preset threshold S max When the radius change is less than or equal to the stated amount, it indicates that the disc grinding wheel has entered the third stage of wear, reaching a state of rapid wear. For severely worn grinding wheels, an alarm should be issued promptly to remind operators to replace or readjust the grinding wheel to avoid machining quality problems caused by grinding wheel wear.

[0161] The multi-layer compensation configuration is designed to adapt to grinding wheels at different wear stages, ensuring that the system maintains stable machining accuracy throughout the entire service life of the grinding wheel. Considering the wear pattern of the grinding wheel, this layered approach ensures accurate compensation, avoiding machining errors caused by under-compensation or over-compensation.

[0162] In some preferred embodiments, the control method further includes:

[0163] Step 506: Accumulate the number of tools that have completed the sharpening process by the sharpening mechanism and output the tool sharpening value in real time.

[0164] Understandably, the counter automatically increments by 1 each time the sharpening mechanism completes the sharpening process for a tool. For example, after the robotic arm picks up a tool to be sharpened and completes the sharpening process, the automatic tool sharpening system automatically starts counting.

[0165] Step 507: When the number of sharpening satisfies the preset number of sharpening, re-perform the position compensation step.

[0166] A preset number of sharpening is set in the system, which can be determined according to the wear law of the sharpening mechanism and production requirements. For example, after processing a certain number of cutters, the wear of the grinding wheel may reach a certain degree, and the system automatically calls the position compensation step (steps 501-505) to perform new position compensation for the sharpening mechanism. This includes reacquiring the image sequence of the grinding wheel, image processing, calculating the radius change, and adjusting the position of the sharpening mechanism according to the new wear condition. By periodically re-executing the position compensation step, it ensures that the sharpening mechanism always maintains stable processing quality during long-term operation. This helps to reduce processing errors caused by grinding wheel wear, improve product quality and production efficiency.

[0167] In metal material grinding, the wear process of the disc-shaped grinding wheel can be divided into three stages: primary wear stage (first stage), wear wear stage (second stage) and rapid wear stage (third stage). In the third stage, due to the passivation of abrasive grain cutting edge exceeding a certain limit, the force acting on the abrasive grain increases sharply, causing large chunks of abrasive grain to break, binder to break, and the entire abrasive grain to fall off. Therefore, the wear of the grinding wheel increases rapidly, the curve rises very steeply, and the disc-shaped grinding wheel cannot work normally, so the disc-shaped grinding wheel needs to be trimmed or replaced.

[0168] Therefore, the present application determines the wear state of the disc-shaped grinding wheel by two methods, the first method is step 505D in step 505, which determines whether the disc-shaped grinding wheel is in the rapid wear stage by the identified radius change. The second method is that when the controller executes the position compensation step, it also includes the following steps:

[0169] Step 601: Acquire the second image of the disc-shaped grinding wheel, the second image is obtained by the camera assembly shooting the end face of the disc-shaped grinding wheel from the horizontal direction.

[0170] In one possible implementation, the second image is obtained by the second industrial camera of the camera assembly in embodiment one. For example, the controller controls the sharpening mechanism to stop, and controls the second camera to collect the second image.

[0171] Step 602: Identify the second image and output the outer circle contour of the disc-shaped grinding wheel.

[0172] In the present application, the outer circle contour of the disc-shaped grinding wheel is extracted to provide contour data for subsequent cross-sectional area calculation.

[0173] In a specific implementation, the second image needs to be pre-processed by image denoising. Since the outer contour image has high requirements for edge integrity, the median filter is used to process the noise in the second image. In one possible implementation, in step 602, the second image is identified, and the outer contour of the disc-shaped grinding wheel is output, specifically including:

[0174] Step 6021: Convert the second image to a grayscale image, binarize the grayscale image, and obtain a second binary image.

[0175] In this application, the image processing process is simplified, and the profile of the grinding wheel is highlighted to provide a clear image basis for subsequent edge detection. In one possible implementation, the second image in color is converted to a grayscale image. Grayscale conversion can be achieved by converting an RGB image to a grayscale image using a weighted average method.

[0176] For example, the grayscale image is binarized to divide the pixels in the image into foreground (grinding wheel part) and background (non-grinding wheel part). Common binarization methods include global thresholding methods (such as the Otsu method) and adaptive thresholding methods. The global thresholding method determines a global threshold by calculating the histogram of the image, while the adaptive thresholding method dynamically adjusts the threshold based on the local region of the image.

[0177] Step 6022: Use a sub-pixel edge detection method to identify the second binary image to obtain a plurality of contour edge points of the disc-shaped grinding wheel.

[0178] In one possible implementation, the sub-pixel edge detection method is used to process the binarized image. The sub-pixel edge detection method can improve the accuracy of edge detection, including the following steps:

[0179] (1) Coarse edge detection: use an edge detection algorithm (Canny algorithm) to preliminarily extract the contour edges in the second image.

[0180] (2) Sub-pixel level positioning: sub-pixel level positioning is performed on the coarse edges to improve the accuracy of edge detection. Sub-pixel level edge detection algorithms can be used, such as image interpolation-based methods or curve fitting-based methods. The contour edge points of the grinding wheel are extracted from the detected edges. These edge points will be used for subsequent contour fitting.

[0181] Step 6023: Input the plurality of contour edge points into a MATLAB application program, and output the outer contour of the disc-shaped grinding wheel.

[0182] In a possible implementation, the extracted plurality of profile edge points are imported into a MATLAB application. Specifically, a cftool curve fitting function of the MATLAB program is used. The extracted edge points are fitted by using the cftool fitting tool. In the tool window, x and y variables are selected, and a polynomial fitting mode is selected.

[0183] Step 603: According to the outer circle profile of the disc-shaped grinding wheel, the actual cross-sectional area of the disc-shaped grinding wheel is calculated

[0184] In a specific implementation, the outer circle profile of the disc-shaped grinding wheel is a curve obtained by polynomial fitting, and the cross-sectional area is calculated according to the parameters of the fitted curve and an integral method. Specifically, the numerical integral method is used to calculate the area.

[0185] Step 604: Based on the preset cross-sectional area and the actual cross-sectional area, the cross-sectional area loss of the disc-shaped grinding wheel is obtained.

[0186] In a possible implementation, the difference between the preset cross-sectional area and the actual cross-sectional area is calculated to obtain the cross-sectional area loss. The calculation formula is: cross-sectional area loss = preset cross-sectional area - actual cross-sectional area. By calculating the cross-sectional area loss, the degree of wear of the grinding wheel is quantified.

[0187] Step 605: When the cross-sectional area loss is greater than a preset threshold, the disc-shaped grinding wheel is identified as being in a sharp wear state, and alert information for replacing or reconditioning the disc-shaped grinding wheel is generated.

[0188] In a possible implementation, according to the wear law of the disc-shaped grinding wheel, a preset threshold of the cross-sectional area loss is set to determine whether the disc-shaped grinding wheel enters a sharp wear state. For example, by collecting cross-sectional area data of the disc-shaped grinding wheel at different use stages, wear law analysis is performed to determine the preset threshold. When the cross-sectional area loss is greater than the preset threshold, the grinding wheel is identified as being in a sharp wear state.

[0189] In the embodiments of the present application, the terms "first", "second" and "third" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance. The term "at least one" means one or more, and the term "multiple" means two or more, unless otherwise explicitly limited.

[0190] In the present application, the term "and / or" is only used to describe the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B, which means that there are three cases of A alone, A and B together, and B alone. In addition, the character " / " in this paper generally represents that the front and rear associated objects are a "or" relationship.

[0191] Finally, it should be noted that the above examples are only used to illustrate the technical solutions of the present application, and are not intended to limit the same; although the present application has been described in detail with reference to the foregoing examples, those of ordinary skill in the art should understand that the technical solutions recorded in the foregoing examples can still be modified, or some of the technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the protection scope of the technical solutions of the embodiments of the present application.

Claims

1. A control method based on visual recognition, characterized in that, The control method is based on an automatic tool sharpening system, which includes a camera assembly, a robotic arm, a moving mechanism, and a sharpening mechanism, wherein the sharpening mechanism has a disc-shaped grinding wheel; the control method includes the following position compensation steps: An image sequence of the disc-shaped grinding wheel is acquired, the image sequence including multiple first images, the multiple first images being obtained by the camera component rotating and sampling a local circumferential contour of the disc-shaped grinding wheel in the vertical direction; The image sequence is preprocessed to obtain multiple first binary images; A pre-defined algorithm is used to synthesize multiple first binary images to obtain a synthesized contour image. Based on the synthesized contour image and the standard contour image, the radius change of the disc grinding wheel is identified; Query the compensation configuration that matches the radius change, and perform position compensation on the grinding mechanism based on the compensation configuration.

2. The control method based on visual recognition according to claim 1, characterized in that, A preset algorithm is used to synthesize multiple first binary images to obtain a synthesized contour image, specifically including: Perform global thresholding segmentation on multiple first images; Multiple first images are debinarized, and the pixel gray values ​​of the disc-shaped grinding wheel portion in the first image are set to 1, while the pixel gray values ​​of the non-disc-shaped grinding wheel portion are set to 0. Opening morphological processing is performed on multiple first images; Based on the contour frequency image generation algorithm, multiple first images are synthesized to obtain a contour image.

3. The control method based on visual recognition according to claim 1, characterized in that, Query the compensation configuration that matches the radius change, specifically including: When the radius change is less than or equal to a preset threshold S min At this time, no position compensation is required; When the preset threshold S min <The radius change <Preset threshold S n At that time, the compensation configuration R is invoked. n n is an integer greater than or equal to 0; When the preset threshold S n <The radius change <Preset threshold S n+1 At that time, the compensation configuration R is invoked. n+1 ; When the preset threshold S max When the radius change is less than or equal to the specified amount, the disc grinding wheel is identified as being in a state of rapid wear, and an alarm message is generated indicating that the disc grinding wheel needs to be replaced or re-corrected.

4. The control method based on visual recognition according to claim 1, characterized in that, The control method further includes: The number of tools that have completed the sharpening process by the sharpening mechanism is accumulated, and the sharpening value of the tools is output in real time. When the sharpening value meets the preset sharpening value, the position compensation step is executed again.

5. The control method based on visual recognition according to claim 1, characterized in that, The control method further includes the following steps when performing the position compensation step: A second image of the disc-shaped grinding wheel is obtained, which is a horizontal image of the end face of the disc-shaped grinding wheel captured by the camera assembly. Identify the second image and output the outer circumference contour of the disc-shaped grinding wheel; Calculate the actual cross-sectional area of ​​the disc-shaped grinding wheel based on its outer circular profile. Based on the preset cross-sectional area and the actual cross-sectional area, the cross-sectional area loss of the disc grinding wheel is obtained; When the cross-sectional area loss exceeds a preset threshold, the disc grinding wheel is identified as being in a state of rapid wear, and an alarm message is generated to replace or re-correct the disc grinding wheel.

6. The control method based on visual recognition according to claim 5, characterized in that, Recognizing the second image and outputting the outer circumference contour of the disc-shaped grinding wheel specifically includes: The second image is converted into a grayscale image, and the grayscale image is binarized to obtain a second binary image; The second binary image is identified using a subpixel edge detection method to obtain multiple contour edge points of the disc grinding wheel; Input multiple contour edge points into the MATLAB application to output the outer circle contour of the disc grinding wheel.

7. An automatic knife sharpening system, characterized in that, The automatic tool sharpening system includes: A frame is provided with a moving mechanism, a grinding mechanism, a camera assembly, a light source assembly, and a controller. The moving mechanism is fixedly connected to the frame and drives the grinding mechanism. The grinding mechanism has a disc-shaped grinding wheel. The camera assembly and the light source assembly are respectively located on the side of the grinding mechanism. A robotic arm, located on the side of the frame, is used to grab the cutting tool to be sharpened from the feeding tool holder and move the cutting tool to the sharpening mechanism for sharpening. The controller is configured to execute the following control method: The image sequence acquired by the camera component is obtained, the image sequence including multiple first images, the multiple first images being obtained by the camera component rotating and sampling a local circumferential contour of the disc grinding wheel in the vertical direction; The image sequence is preprocessed to obtain multiple first binary images; A pre-defined algorithm is used to synthesize multiple first binary images to obtain a synthesized contour image. Based on the synthesized contour image and the standard contour image, the radius change of the disc grinding wheel is identified; The system queries a compensation configuration that matches the radius change and controls the moving mechanism to perform position compensation on the grinding mechanism based on the compensation configuration.

8. The automatic tool sharpening system according to claim 7, characterized in that, The mobile mechanism includes: A transverse moving screw pair, which is fixedly connected to the frame; A longitudinal moving lead screw pair is connected to the transverse moving lead screw pair, and the longitudinal moving lead screw pair is connected to the grinding mechanism; The controller controls the longitudinal movement screw pair of the moving mechanism to perform position compensation on the grinding mechanism based on the compensation configuration, so as to adjust the distance between the grinding mechanism and the robot arm.

9. The automatic tool sharpening system according to claim 7, characterized in that: The frame is also provided with a white background plate, which is located on one side of the disc grinding wheel of the grinding mechanism. The surface of the white background plate is parallel to the end face of the disc grinding wheel.

10. An automatic tool sharpening system according to claim 7, characterized in that: The grinding mechanism has a disc-shaped grinding wheel, the outer circumference of which is used for sharpening. The disc-shaped grinding wheel is horizontally arranged, and the outer circumference of which faces the robot arm is used for sharpening.

Images