Cutter detection method for numerical control machining

Abstract

The invention discloses a cutter detection method for numerical control machining, and relates to the technical field of cutter detection, and the method comprises the steps: obtaining image data of a cutter in real time through a sensor, and carrying out the preprocessing; constructing an image recognition model based on the image data and automatically extracting macroscopic and microscopic features of the cutting edge of the cutter; comprehensively evaluating the macroscopic shape recognition result of the cutting edge and the microcosmic wear and tipping detection result; generating a tool detection report according to the evaluation result and outputting the report to a control system of the numerical control machining equipment to dynamically adjust machining parameters; and retraining and optimizing the image recognition model by using the real-time cutter image data. According to the method, the macroscopic and microscopic characteristics of the cutting edge of the cutter can be comprehensively detected in real time, the machining parameters are adjusted in time according to the state of the cutter, and machining errors caused by cutter abrasion and deformation factors are effectively avoided, so that the overall quality of products is improved, the rejection rate is reduced, and the requirement for high-precision machining is met.

Description

Technical Field

[0001] The invention relates to the technical field of tool detection, in particular to a tool detection method for numerical control machining. Background Art

[0002] In the field of CNC machining, the performance and status of the tool directly affect the machining accuracy, efficiency and product quality. As the modern manufacturing industry continues to develop towards high precision, high efficiency and intelligence, extremely stringent requirements are placed on the detection of CNC machining tools. Traditional tool detection methods mainly rely on manual regular inspections or the use of some simple measuring tools. Manual detection methods have many disadvantages. They are highly subjective, and the detection results are easily affected by the experience, skill level and working status of the detection personnel. It is difficult to ensure the accuracy and consistency of the detection. In addition, manual detection is inefficient and cannot meet the rhythm of high-speed and continuous production of modern CNC machining. It is easy to cause production interruptions and increase production costs. Some simple Single measuring tools, such as calipers and micrometers, can only measure some of the macroscopic dimensions of the tool. It is difficult to conduct comprehensive and accurate detection of the microscopic wear, chipping and complex geometric shape changes of the tool edge. In recent years, with the continuous advancement of sensor technology and image processing technology, image-based tool detection methods have gradually emerged. Most of the existing image detection methods only focus on a certain type of feature detection of the tool edge, such as either only focusing on the macroscopic shape or only analyzing the microscopic wear. There is a lack of comprehensive consideration and effective integration of the macroscopic and microscopic features of the tool edge. This one-sided detection cannot comprehensively and accurately evaluate the true state of the tool, and it is difficult to achieve precise control of the CNC machining process.

[0003] However, the common solutions currently available have many shortcomings, including: existing technologies often make it difficult to simultaneously and accurately detect and comprehensively analyze the macro and micro features of the tool edge; many traditional methods can only focus on one aspect; this one-sided detection cannot provide a complete and accurate assessment of the tool status, resulting in the inability to fine-tune the machining parameters according to the actual status of the tool during CNC machining, thus affecting machining accuracy and product quality. Summary of the invention

[0004] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0005] In view of the above problems existing in the existing tool detection method for numerical control machining, the present invention is proposed.

[0006] Therefore, the purpose of the present invention is to provide a tool detection method for CNC machining, which is suitable for solving the problem that the prior art often finds it difficult to simultaneously perform accurate and comprehensive detection and comprehensive analysis of the macro and micro features of the tool cutting edge, resulting in the inability to perform fine-grained adjustment of machining parameters according to the actual state of the tool during CNC machining, thereby affecting machining accuracy and product quality.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In the first aspect, an embodiment of the present invention provides a tool detection method for CNC machining, which includes using a sensor to obtain image data of the tool in real time and perform preprocessing; constructing an image recognition model based on the image data and automatically extracting the macro and micro features of the tool edge; comprehensively evaluating the macro shape recognition results of the cutting edge, and the micro wear and chipping detection results; generating a tool detection report based on the evaluation results and outputting the report to the control system of the CNC machining equipment to dynamically adjust the machining parameters; and retraining and optimizing the image recognition model using real-time tool image data.

[0009] As a preferred solution of the tool detection method for CNC machining described in the present invention, wherein: the sensor includes an industrial camera, an optical microscope, an infrared thermal imaging sensor and a force sensor; the image data includes the original image of the tool edge, the edge images at different processing stages, the high-resolution edge microscopic image, the edge image with a mark or reference object, and the edge image under different lighting; the edge macroscopic shape recognition result includes the edge geometry category judgment, edge size change information, edge contour shape change and edge angle change data; the tool detection report includes tool identification information, image acquisition time and environment, edge macroscopic shape analysis, edge microscopic wear condition, chipping detection details, comprehensive tool wear rating, remaining service life prediction, detection method and model information, and suggestions and warning information.

[0010] As a preferred solution of the tool detection method for CNC machining described in the present invention, the specific steps of constructing the image recognition model are as follows: collect a large amount of image data of tool edges through a CNC machining environment and preprocess them; construct an image recognition model based on a recurrent neural network and extract the features of the tool edges; use a deep learning algorithm to further extract and analyze the tool edges; evaluate the model and adjust the model structure or parameters according to the evaluation results.

[0011] As a preferred solution of the tool detection method for CNC machining of the present invention, the specific formula of the comprehensive evaluation is as follows:

[0012]

[0013] Among them, S is a comprehensive evaluation index; I ij is the pixel value of the tool image after preprocessing; m is the number of rows of the tool image; n is the number of columns of the tool image; M is the macroscopic shape recognition result of the tool; N is the microscopic wear and chipping detection result of the tool edge.

[0014] As a preferred scheme of the tool detection method for CNC machining described in the present invention, the specific meaning of the comprehensive evaluation index is as follows: when the comprehensive evaluation index S is greater than 1, it indicates that the integrity of the tool edge is in a very good state, which means that the macro state recognition results and the micro wear and chipping detection results are very ideal, indicating that the image pixel features and the differences in macro and micro results have little negative impact on the overall evaluation; when the comprehensive evaluation index S is close to 0, it indicates that the integrity of the tool edge is very poor, indicating that the tool edge has serious problems in both macro shape and micro state, and the tool needs to be replaced or repaired on a large scale.

[0015] As a preferred solution of the tool detection method for CNC machining described in the present invention, the dynamic adjustment of machining parameters includes the following steps: combining the tool macro-shape recognition results and the micro-wear and chipping detection results to establish a comprehensive data set; formulating a comprehensive evaluation algorithm based on the comprehensive data set; generating a detailed tool detection report based on the comprehensive evaluation results, and dynamically adjusting the machining parameters; and outputting the tool detection report to the control system of the CNC machining equipment.

[0016] As a preferred solution of the tool detection method for numerical control machining of the present invention, the specific situation of the optimized image recognition model is as follows:

[0017] When it is found that the image recognition model has insufficient learning ability for complex edge features, the number of hidden layers of the recurrent neural network or the number of neurons in each hidden layer should be increased.

[0018] In the second aspect, in order to further solve the above-mentioned technical problems, the embodiment of the present invention provides a tool detection method for CNC machining, which includes: a data collection module, which is used to collect and preprocess the image data of the tool in real time; a model building module, which is used to build an image recognition model and automatically extract the macro and micro features of the tool edge; a comprehensive evaluation module, which is used to comprehensively evaluate the edge shape recognition results, micro wear and chipping detection results; a dynamic adjustment module, which is used to dynamically adjust the machining parameters according to the tool detection report; a model optimization module, which is used to retrain and optimize the image recognition model using real-time tool image data.

[0019] The beneficial effects of the present invention are as follows: the present invention can detect the macro and micro features of the tool edge in real time and comprehensively, and adjust the processing parameters in time according to the tool status, effectively avoiding processing errors caused by tool wear and deformation factors. In the processing of precision parts, the micro wear of the tool edge can be discovered in time and the parameters can be adjusted, which can significantly reduce the roughness of the workpiece surface and improve dimensional accuracy, thereby improving the overall quality of the product, reducing the scrap rate, and meeting the requirements for high-precision processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0021] Figure 1 This is a flow chart for implementing the present invention in Example 1.

[0022] Figure 2 This is a dynamic adjustment diagram of the processing parameters in Example 1. DETAILED DESCRIPTION

[0023] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0024] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0025] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0026] Example 1

[0027] Reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, provides a tool detection method for numerical control machining, comprising the following steps:

[0028] S1: Use the sensor to obtain the image data of the tool in real time and perform preprocessing.

[0029] Preferably, Figure 1 The figure shows the implementation process of the present invention. First, the image data of the tool is acquired in real time by using a sensor and preprocessed. Then, an image recognition model is constructed based on the image data and the macro and micro features of the tool edge are automatically extracted. The macro shape recognition results of the edge and the micro wear and chipping detection results are comprehensively evaluated. Subsequently, a tool detection report is generated according to the evaluation results and the report is output to the control system of the CNC machining equipment to dynamically adjust the machining parameters. Finally, the image recognition model is retrained and optimized using the real-time tool image data.

[0030] Furthermore, the sensors include industrial cameras, optical microscopes, infrared thermal imaging sensors, and force sensors.

[0031] Furthermore, the image data includes original images of the tool edge, images of the edge at different processing stages, high-resolution microscopic images of the edge, images of the edge with markers or references, and images of the edge under different lighting conditions.

[0032] Specifically, preprocessing includes data cleaning, image enhancement, geometric transformation, and data normalization.

[0033] S2: Build an image recognition model based on image data and automatically extract the macro and micro features of the tool edge.

[0034] Furthermore, the specific steps of constructing the image recognition model are as follows: a large amount of image data of tool edges are collected through a CNC machining environment and pre-processed.

[0035] An image recognition model is constructed based on a recurrent neural network to extract the features of the tool edge.

[0036] Deep learning algorithms are used to further extract and analyze the tool edge.

[0037] Evaluate the model and adjust the model structure or parameters based on the evaluation results.

[0038] Preferably, the image recognition model constructed using the deep learning algorithm can automatically learn complex feature patterns from the tool edge image data without the need for manual tedious feature extraction. For macro features, it can accurately identify the type of cutting edge geometry, size changes, contour shape changes, and angle changes, and fully grasp the macro morphology of the tool edge. For micro features, it can effectively detect the type of micro wear, degree of wear, and specific details of chipping, and achieve early and accurate detection of subtle damage to the tool edge. Compared with traditional manual feature extraction methods, this automated feature extraction method not only greatly improves efficiency, but also avoids human errors, ensuring the objectivity, accuracy, and consistency of feature extraction.

[0039] S3: Comprehensively evaluate the cutting edge macro shape recognition results, micro wear and chipping detection results.

[0040] Furthermore, the macroscopic shape recognition result of the cutting edge includes the judgment of the cutting edge geometry category, the cutting edge size change information, the cutting edge contour shape change and the cutting edge angle change data.

[0041] Furthermore, the specific formula for comprehensive evaluation is as follows:

[0042]

[0043] Among them, S is a comprehensive evaluation index; I ij is the pixel value of the tool image after preprocessing; m is the number of rows of the tool image; n is the number of columns of the tool image; M is the macroscopic shape recognition result of the tool; N is the microscopic wear and chipping detection result of the tool edge.

[0044] Specifically, the specific meanings of the comprehensive evaluation index are as follows: When the comprehensive evaluation index S is greater than 1, it indicates that the integrity of the tool edge is in a very good state, which means that the macroscopic state recognition results and the microscopic wear and chipping detection results are ideal, indicating that the image pixel features and the differences in macroscopic and microscopic results have little negative impact on the overall evaluation.

[0045] When the comprehensive evaluation index S is close to 0, it means that the integrity of the tool edge is very poor, indicating that there are serious problems with the tool edge in both macroscopic shape and microscopic state, and the tool needs to be replaced or repaired on a large scale.

[0046] Preferably, the macro-shape recognition results of the cutting edge are comprehensively evaluated with the micro-wear and chipping detection results, fully considering the different levels of factors of the cutting edge state of the tool, avoiding the one-sided conclusion caused by evaluating the macro or micro characteristics alone, and by designing a reasonable comprehensive evaluation function or algorithm, the overall integrity of the cutting edge can be quantitatively evaluated, and the actual use status and remaining service life of the tool can be more accurately judged. Even if the macro-shape change of the cutting edge of the tool is relatively small, but the micro-wear is already serious, the comprehensive evaluation can timely discover potential problems of the tool. Conversely, when the macro-deformation is large and the micro-wear is still light, the overall condition of the tool can also be accurately evaluated, thereby achieving a comprehensive and accurate grasp of the tool state, improving the safety of tool use and the stability of processing quality.

[0047] S4: Generate a tool inspection report based on the evaluation results and output the report to the control system of the CNC machining equipment to dynamically adjust the machining parameters.

[0048] Specifically, dynamically adjusting the machining parameters includes the following steps: combining the tool macro shape recognition results and the micro wear and chipping detection results to establish a comprehensive data set.

[0049] Develop a comprehensive evaluation algorithm based on a comprehensive dataset.

[0050] Generate detailed tool inspection reports based on comprehensive evaluation results and dynamically adjust machining parameters.

[0051] Output the tool inspection report to the control system of the CNC machining equipment.

[0052] Preferably, based on the tool detection report generated according to the comprehensive evaluation results, the control system of the CNC machining equipment can dynamically adjust the machining parameters in real time. When the tool is in good condition, the control system can appropriately increase the machining parameters, such as increasing the cutting speed and feed rate, thereby improving the machining efficiency. When the tool is worn or other abnormal conditions occur, the machining parameters are timely reduced, such as reducing the cutting force related parameters. This can not only ensure the machining quality and reduce the scrapping of workpieces due to tool problems, but also extend the tool life and reduce the frequency and cost of tool replacement. This dynamic adjustment mechanism realizes the real-time adaptation of the machining process and the tool status, and improves the intelligence level of CNC machining and the overall machining efficiency.

[0053] S5: Retrain and optimize the image recognition model using real-time tool image data.

[0054] Specifically, the specific circumstances for optimizing the image recognition model are as follows: when it is found that the image recognition model has insufficient learning ability for complex blade edge features, the number of hidden layers of the recurrent neural network or the number of neurons in each hidden layer should be increased.

[0055] Preferably, the image recognition model is retrained and optimized using real-time tool image data, so that the model can continuously adapt to various changes in the machining process, such as tool type replacement, workpiece material change, and machining process adjustment. As the machining progresses, new data is continuously added to the model training, and the model can learn more laws and characteristics of tool edge state changes, thereby improving the model's accuracy and adaptability to tool edge integrity detection. When a new type of tool is introduced, the model can quickly learn its characteristic pattern through new data to ensure the effectiveness of the detection. When the machining process optimization causes the tool wear pattern to change, the model can also adjust the detection strategy in time, always providing reliable support for tool detection.

[0056] This embodiment also provides a tool detection method for CNC machining, including: a data collection module, used to collect image data of the tool in real time and perform preprocessing; a model building module, used to build an image recognition model and automatically extract the macro and micro features of the tool edge; a comprehensive evaluation module, used to comprehensively evaluate the edge shape recognition results, micro wear and chipping detection results; a dynamic adjustment module, used to dynamically adjust the machining parameters according to the tool detection report; a model optimization module, used to retrain and optimize the image recognition model using real-time tool image data.

[0057] In summary, the present invention can detect the macro and micro features of the tool edge in real time and comprehensively, and adjust the processing parameters in time according to the tool status, effectively avoiding processing errors caused by tool wear and deformation factors. In the processing of precision parts, timely discovering the micro wear of the tool edge and adjusting the parameters can significantly reduce the roughness of the workpiece surface and improve dimensional accuracy, thereby improving the overall quality of the product, reducing the scrap rate, and meeting the requirements for high-precision processing.

[0058] Example 2

[0059] Referring to Tables 1 to 3, the second embodiment of the present invention is shown. This embodiment is different from the first embodiment in that, in order to verify its beneficial effects, operating data and related instructions of the present invention in an actual environment are provided.

[0060] Tables 1 and 2 show the comparison between the technology of the present invention and the prior art in tool edge feature detection data, processing parameter adjustment and processing effect in this example.

[0061] Table 1 Tool edge feature detection data comparison table

[0062]

[0063] Table 2 Comparison of processing parameter adjustment and processing effect

[0064] Detection Methods Tool replacement frequency (times / month) Processing efficiency improvement ratio Processing scrap rate Average tool life Prior art 10 0％ 5 100 The technology of the present invention 3 20％ 1 200

[0065] Table 3 shows the training and optimization data table of the image recognition model according to the present invention.

[0066] Table 3 Model training and optimization data table

[0067] Model training times Amount of training data Model update time (minutes) 1 4200 － 2 4600 10 3 5000 15 4 5400 20 5 5800 25

[0068] It can be seen from the above table that by continuously incorporating new data for training, the model of the present invention can better learn the characteristic patterns of the tool edge, thereby improving the accuracy of tool edge integrity detection. This ensures that the model can adapt to changes in the tool state and processing conditions in a timely manner, and provides continuous and reliable support for tool detection in the CNC machining process. It reflects the advantages of the present invention in model optimization, and can effectively improve the accuracy and reliability of tool detection, thereby improving the intelligence level and processing efficiency of the entire CNC machining process.

[0069] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention can be modified or replaced without departing from the spirit and scope of the technical solutions of the present invention, which should be included in the scope of the claims of the present invention.

Claims

1. A tool detection method for CNC machining, characterized in that: include: Use sensors to obtain image data of cutting tools in real time and perform preprocessing; Build an image recognition model based on image data and automatically extract macro and micro features of the tool edge; Comprehensively evaluate the cutting edge macro shape recognition results, micro wear and chipping detection results; Generate tool inspection report based on evaluation results and output the report to the control system of CNC machining equipment to dynamically adjust machining parameters; Retrain and optimize image recognition models using real-time tool image data.

2. The tool detection method for CNC machining according to claim 1, characterized in that: The sensors include industrial cameras, optical microscopes, infrared thermal imaging sensors, and force sensors; The image data includes original images of the cutting edge of the tool, images of the cutting edge at different processing stages, high-resolution microscopic images of the cutting edge, images of the cutting edge with marks or reference objects, and images of the cutting edge under different lighting conditions; The cutting edge macroscopic shape recognition result includes cutting edge geometry category judgment, cutting edge size change information, cutting edge contour shape change situation and cutting edge angle change data; The tool inspection report includes tool identification information, image acquisition time and environment, cutting edge macro-shape analysis, cutting edge micro-wear condition, chipping detection details, comprehensive tool wear rating, remaining service life prediction, inspection method and model information, and suggestions and warning information.

3. The tool detection method for CNC machining according to claim 2, characterized in that: The specific steps of constructing the image recognition model are as follows: Collect a large amount of tool edge image data through CNC machining environment and perform preprocessing; Build an image recognition model based on recurrent neural network and extract the features of tool edge; Use deep learning algorithms to further extract and analyze tool edges; Evaluate the model and adjust the model structure or parameters based on the evaluation results.

4. The tool detection method for CNC machining according to claim 3, characterized in that: The specific formula for the comprehensive evaluation is as follows: Among them, S is a comprehensive evaluation index; I ij is the pixel value of the tool image after preprocessing; m is the number of rows of the tool image; n is the number of columns of the tool image; M is the macroscopic shape recognition result of the tool; N is the microscopic wear and chipping detection result of the tool edge.

5. The tool detection method for CNC machining according to claim 4, characterized in that: The specific meanings of the comprehensive evaluation indicators are as follows: When the comprehensive evaluation index S is greater than 1, it indicates that the integrity of the tool edge is in a very good state, which means that the macroscopic state recognition results and the microscopic wear and chipping detection results are very ideal, indicating that the image pixel features and the differences in macroscopic and microscopic results have little negative impact on the overall evaluation; When the comprehensive evaluation index S is close to 0, it means that the integrity of the tool edge is very poor, indicating that there are serious problems with the tool edge in both macroscopic shape and microscopic state, and the tool needs to be replaced or repaired on a large scale.

6. The tool detection method for CNC machining according to claim 5, characterized in that: The dynamic adjustment of processing parameters comprises the following steps: Combine tool macroscopic shape recognition results with microscopic wear and chipping detection results to create a comprehensive data set; Develop comprehensive evaluation algorithms based on comprehensive data sets; Generate detailed tool inspection reports based on comprehensive evaluation results and dynamically adjust machining parameters; Output the tool inspection report to the control system of the CNC machining equipment.

7. The tool detection method for CNC machining according to claim 3, characterized in that: The specific situation of the optimized image recognition model is as follows: When it is found that the image recognition model has insufficient learning ability for complex edge features, the number of hidden layers of the recurrent neural network or the number of neurons in each hidden layer should be increased.

8. A tool detection method for CNC machining according to any one of claims 1 to 7, characterized in that: include, A data collection module is used to collect image data of the tool in real time and perform preprocessing; A model building module, which is used to build an image recognition model and automatically extract macro and micro features of the tool edge; Comprehensive evaluation module, used to comprehensively evaluate the cutting edge shape recognition results, micro wear and chipping detection results; Dynamic adjustment module, used to dynamically adjust processing parameters according to tool detection reports; Model optimization module, used to retrain and optimize the image recognition model using real-time tool image data.

Images