Intelligent dust removal system and dust removal method for cutting machine and electronic equipment

Abstract

The invention relates to the technical field of intelligent dust removal scheme design of cutting machines, in particular to an intelligent dust removal system and method for a cutting machine and electronic equipment. According to the method, the internal image of the cutting impurity containing box is obtained, and the YOLO model is used for accurate recognition, so that the actual condition of the cutting impurities can be accurately judged. When the impurities are recognized, the dust removal device is started in time, invalid operation of a traditional regular dust removal mode when the impurities are few is avoided, the dust removal efficiency is effectively improved, it is ensured that the cutting machine is in a good working environment all the time, the influence of impurity accumulation on equipment performance is reduced, and then the working efficiency and cutting quality of the cutting machine are improved. The running state of the dust removal device is intelligently controlled according to the judgment result of the YOLO model, unnecessary energy consumption is avoided through the intelligent control mode, and compared with a traditional dust removal system in a fixed mode, the energy cost is greatly reduced, and the development trend of energy conservation and environmental protection is met.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent dust removal scheme design for cutting machines, and particularly relates to an intelligent dust removal system and method for cutting machines, and an electronic device. Background Art

[0002] In many industrial fields, such as mechanical manufacturing, construction engineering, automobile manufacturing, etc., cutting machines are widely used as important processing equipment. When cutting machines process various materials (including metals, stones, woods, plastics, etc.), a large amount of cutting impurities will inevitably be generated. These impurities have various forms, mainly including the following: Metal debris: When using metal cutting equipment (such as plasma cutting machines, laser cutting machines, numerical control lathes, etc.) for metal processing, the high temperature generated by high-speed cutting will instantly melt the metal surface, and then cool and solidify to form tiny metal particles, which are the metal debris. Due to the high speed and high temperature of the cutting process, the metal debris often has a certain shape and hardness, and the particle sizes are different, some may be only a few micrometers, while some can reach several millimeters.

[0003] Dust agglomerates: When cutting or grinding some brittle materials (such as stones, ceramics, etc.) or fine-grained powder materials (such as carbon powder, coal powder, etc.), the materials will break and pulverize under the action of cutting force, forming a large number of dust particles. Due to the adsorption and agglomeration properties of these dust particles, they are prone to adsorb and aggregate with each other to form a lump-like mixture, that is, dust agglomerates.

[0004] Fiber filaments: During the processing of materials such as wood and plastics, when the materials are subjected to cutting force, the internal fibers or fibrous structures may be broken or torn, forming fibrous debris, that is, fiber filaments.

[0005] The above impurities will not only affect the normal operation of the cutting machine and reduce the cutting quality, but may also pose a hazard to the health of operators. Traditional dust removal methods are often carried out regularly, lacking real-time monitoring of the actual generation situation of cutting impurities, resulting in poor dust removal effect or excessive dust removal, wasting energy. Therefore, it is necessary to develop an intelligent dust removal system that can intelligently identify cutting impurities and provide real-time dust removal.

[0006] Therefore, the prior art still needs to be further developed. Summary of the Invention

[0007] The purpose of the present invention is to overcome the above technical deficiencies, and provide an intelligent dust removal system and method for cutting machines, and an electronic device, so as to solve the problems existing in the prior art.

[0008] To achieve the above technical objectives, according to the first aspect of the present invention, the present invention provides an intelligent dust removal system for a cutting machine, including: A cutting impurity retention box, which is arranged at the bottom of the cutting machine and is used for retaining the cutting impurities of the cutting machine; An acquisition module, the acquisition module includes a camera, and the camera is used for acquiring the internal image of the cutting impurity retention box to form a data set; A control module, which is used for training the YOLO model by using the data set. After the YOLO model is trained, it acquires the internal image of the cutting impurity retention box and inputs it into the trained YOLO model, and uses the trained YOLO model to judge whether an anchor box for indicating cutting impurities can be recognized, and controls the operating state of the dust removal device according to the judgment result; A dust removal device, including a blower unit connected to the cutting impurity retention box, and the blower unit is used for discharging the cutting impurities in the cutting impurity retention box to the outside of the cutting impurity retention box when it is operating.

[0009] Specifically, the blower unit includes a hair dryer and an exhaust fan. The cutting impurity retention box is provided with a blowing port and an exhaust port. The blowing port is connected to the hair dryer through a pipeline, and the exhaust port is connected to the exhaust fan through a pipeline.

[0010] According to the second aspect of the present invention, an intelligent dust removal method for a cutting machine is provided, including: S100. Acquire the internal image of the cutting impurity retention box to form a data set; S200. Train the YOLO model by using the data set. After the YOLO model is trained, acquire the internal image of the cutting impurity retention box and input it into the trained YOLO model, and use the trained YOLO model to judge whether an anchor box for indicating cutting impurities can be recognized; S300. Control the operating state of the dust removal device according to the judgment result.

[0011] Specifically, the controlling the operating state of the cutting machine according to the judgment result includes: If an anchor box for indicating cutting impurities can be recognized, control the dust removal device to operate; If an anchor box for indicating cutting impurities cannot be recognized, control the dust removal device to stop operating.

[0012] Specifically, the controlling the operating state of the cutting machine according to the judgment result specifically includes: If an anchor box for indicating cutting impurities can be recognized, control the hair dryer and the exhaust fan to operate; If an anchor box for indicating cutting impurities cannot be recognized, control the hair dryer and the exhaust fan to stop operating.

[0013] Specifically, training the YOLO model using the dataset includes: S1. Create a dataset; S2. Use Labelme to mark the prediction boxes for cutting impurities; S3. Train the YOLO model; S4. Save the optimal YOLO model.

[0014] Specifically, in step S1, it specifically includes: S11. Collect the original image data containing cutting impurities during the working process of the cutting machine. The original image data is from the working scenarios of the cutting machine under multiple different perspectives, different lighting conditions, and different cutting conditions; S12. Screen the collected original image data, and remove the images that are blurred, have poor image quality, and do not contain valid information of cutting impurities; S13. Perform preprocessing operations before annotation on the screened image data. The preprocessing operations include adjusting the image size to a unified size, grayscaling or color space conversion of the image, and normalizing the image pixel values, so that the image data meets the requirements of subsequent model training; S14. Divide the preprocessed image data into a training set, a validation set, and a test set according to a certain ratio to form a dataset for training the YOLO model.

[0015] Specifically, in step S2, it specifically includes: S21. Start the Labelme annotation tool and import the images in the dataset created in step S1 into the Labelme software interface in sequence; S22. For each image, carefully observe the specific shape, position, and distribution of the cutting impurities in the image, and determine different annotation categories according to the types of cutting impurities, where the types include metal debris, dust clumps, and fiber filaments; S23. Use the drawing function in the Labelme tool to accurately draw prediction boxes along the edges of the cutting impurities on the image, ensure that the prediction boxes completely cover the cutting impurity areas, and at the same time minimize the misselection of non-impurity areas. For cutting impurities with irregular shapes, the prediction boxes should fit their outer contours as much as possible; S24. Assign corresponding class labels to each drawn prediction box, and accurately fill in the determined cutting impurity class names in the annotation information column of Labelme to ensure that the class labels are consistent with the pre-set class system; S25. Conduct a preliminary inspection on the annotated images to check whether the prediction boxes are drawn accurately and whether the class labels are correct. For images with doubts or inaccurate annotations, return to observe and annotate again to ensure the annotation quality; S26. After all the images are labeled, export the data file with prediction box annotation information and class labels in the format specified by Labelme for subsequent training of the YOLO model.

[0016] Specifically, in the step S3, it specifically includes: S31. Set the initial training parameters of the YOLO model, and the initial training parameters include learning rate, momentum, weight decay coefficient, batch size, and maximum number of iterations; S32. Input the dataset with annotation information obtained after marking and cutting the impurity prediction box by Labelme in step S2 into the YOLO model, and start the training process according to the set initial training parameters; S33. During the training process, monitor the loss function value, accuracy, recall rate, and mAP evaluation metrics of the model on the validation set in real time; S34. Dynamically adjust the training parameters according to the monitored changes in the evaluation metrics. For example, when the loss function value no longer decreases or the evaluation metrics fluctuate, reduce the learning rate; S35. Continuously perform training and parameter adjustment until the evaluation metrics of the model on the validation set reach the preset optimal threshold, or the training iteration times reach the maximum number of iterations.

[0017] According to the third aspect of the present invention, there is provided an electronic device, including: a memory; and a processor, where computer-readable instructions are stored on the memory, and when the computer-readable instructions are executed by the processor, the above-mentioned intelligent dust removal method for the cutting machine is implemented.

[0018] Beneficial effects: By acquiring the internal image of the cutting impurity accommodation box and using the YOLO model for accurate identification, the present invention can accurately judge the actual situation of the cutting impurities. When impurities are identified, the dust removal device is started in a timely manner, avoiding the ineffective operation of the traditional timed dust removal method when there are few impurities, effectively improving the dust removal efficiency, ensuring that the cutting machine is always in a good working environment, reducing the impact of impurity accumulation on the equipment performance, and thus improving the working efficiency and cutting quality of the cutting machine. The present invention intelligently controls the operating state of the dust removal device according to the judgment result of the YOLO model, that is, when the cutting impurity anchor box is identified, the hair dryer and the exhaust fan are started for dust removal, and when it cannot be identified, the operation is stopped. This intelligent control method avoids unnecessary energy consumption, and compared with the traditional fixed-mode dust removal system, greatly reduces the energy cost, meeting the development trend of energy conservation and environmental protection. Description of the Drawings

[0019] Figure 1 It is a schematic diagram of the installation position of the camera provided in the specific embodiment of the present invention; Figure 2It is a schematic structural diagram of the intelligent dust removal system of the cutting machine provided in the specific embodiment of the present invention; Figure 3 It is a schematic flowchart of the intelligent dust removal method of the cutting machine provided in the specific embodiment of the present invention; Figure 4 It is a schematic diagram of the system composition of the intelligent dust removal system of the cutting machine provided in the specific implementation of the present invention; In the above-mentioned drawings, there are the following reference numerals: 1. Cutting machine; 2. Cutting impurity retention box; 3. Inclined plate; 4. Blowing port; 5. Exhaust port; 6. Branch blowing air pipe; 7. Main blowing air pipe; 8. Blower; 9. Suction air pipe; 10. Suction port; 11. Branch exhaust air pipe; 12. Main exhaust air pipe; 13. Exhaust fan; 14. Dust exhaust pipe; 15. Dust loading box; 16. Fan support; 17. Camera; 100. Acquisition module; 200. Control module; 300. Dust removal device. Detailed implementation manners

[0020] In order to enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Based on the embodiments in this application, other similar embodiments obtained by those of ordinary skill in the art without creative efforts shall all fall within the scope of protection of this application. In addition, the directional terms mentioned in the following embodiments, such as "upper", "lower", "left", "right", etc., are only references to the directions in the drawings. Therefore, the directional terms used are for illustration rather than limitation of the present invention.

[0021] The present invention will be further described below in conjunction with the drawings and preferred embodiments.

[0022] Please refer to Figures 1 - 4 , the present invention provides an intelligent dust removal system for a cutting machine, including: A cutting impurity retention box 2, which is arranged at the bottom of the cutting machine and is used to retain the cutting impurities of the cutting machine; An acquisition module 100, the acquisition module 100 includes a camera 17, and the camera 17 is used to acquire the internal image of the cutting impurity retention box 2 to form a data set, Specifically, the camera 17 is fixedly installed above the inside of the cutting impurity retention box 2 and is used to take the internal image of the cutting impurity retention box 2.

[0023] A control module 200, which is used to train the YOLO model by using the data set. After the YOLO model is trained, the internal image of the cutting impurity retention box 2 is acquired and input into the trained YOLO model, and the trained YOLO model is used to judge whether an anchor box for indicating cutting impurities can be recognized, and the operating state of the dust removal device 300 is controlled according to the judgment result; The dust removal device 300 includes a fan unit connected to the cutting impurity containing box 2. The fan unit is used to discharge the cutting impurities in the cutting impurity containing box 2 to the outside of the cutting impurity containing box 2 during operation.

[0024] Specifically, the fan unit includes a blower 8 and an exhaust fan 13. The cutting impurity containing box 2 is provided with a blowing port 4 and an exhaust port 5. The blowing port 4 is connected to the blower 8 through a pipeline, and the exhaust port 5 is connected to the exhaust fan 13 through a pipeline.

[0025] It includes a cutting machine 1. A plurality of cutting impurity containing boxes 2 are arranged in the middle of the cutting machine 1. An inclined plate 3 is arranged at the bottom of the cutting impurity containing box 2. Blowing ports 4 and exhaust ports 5 are symmetrically arranged on the two inner side plates of the cutting impurity containing box 2. The blowing port 4 is connected to a main blowing pipe 7 through a branch blowing pipe 6. One end of the main blowing pipe 7 is provided with a blower 8. The blower 8 is connected to a suction port 10 through a suction pipe 9. The exhaust port 5 is connected to a main exhaust pipe 12 through a branch exhaust pipe 11. One end of the main exhaust pipe 12 is provided with an exhaust fan 13. The exhaust fan 13 is connected to a dust loading box 15 through a dust exhaust pipe 14. The exhaust fan 13 and the blower 8 are fixedly installed on the side surface of the cutting machine 1 through a fan bracket 16.

[0026] Preferably, the number of the blowing ports 4 and the exhaust ports 5 is the same, and both are six. The shape of the exhaust port 5 is semi-circular, and its diameter is half of the side length of the cutting impurity containing box 2, which can improve the dust removal effect. The main blowing pipe 7 and the main exhaust pipe 12 are both made of random copolymer polypropylene pipes, which have the advantages of energy saving, material saving, environmental protection, light weight, high strength, corrosion resistance, smooth inner wall without scaling, simple construction and maintenance, and long service life.

[0027] By means of the main blower installed on the side of the main body of the cutting machine 1, air can be obtained from the environment under the action of the air suction port, and the air is conveyed into the main air blowing pipe through the air suction pipe. Then, the main air blowing pipe conveys the air into the branch air blowing pipes, and then the branch air blowing pipes convey the ambient air into the cutting impurity containment box 2 to accelerate the dust removal efficiency. Moreover, by means of the inclined plate 3 installed at the bottom of the cutting impurity containment box 2, it is possible to effectively prevent dust from being blown up, so as not to affect the working environment of the user, which is beneficial for people to use and operate for a long time. Additionally, by means of the air extraction port installed on the inner side surface of the cutting impurity containment box 2, the dust removal effect can be improved, and the dust is conveyed into the branch air extraction pipe 11 and the main air extraction pipe 12, and then the dust is conveyed into the dust loading box 15 by means of the air extraction fan 13. Among them, by setting the main air extraction pipe 12, the number of air extraction fans 13 used can be reduced, so as to play a role in reducing the waste of resources and lowering the cost. Moreover, one end of the dust discharge pipe 14 is provided with a dust loading box 15, which can facilitate people to handle the dust and also prevent the dust from polluting the environment.

[0028] It can be understood that by obtaining the internal image of the cutting impurity containment box 2 and using the YOLO model for accurate identification, the present invention can accurately judge the actual situation of the cutting impurities. When impurities are identified, the dust removal device 300 is started in a timely manner, avoiding the ineffective operation of the traditional timed dust removal method when there are few impurities, effectively improving the dust removal efficiency, ensuring that the cutting machine is always in a good working environment, reducing the impact of impurity accumulation on the equipment performance, and thus improving the working efficiency and cutting quality of the cutting machine. The present invention intelligently controls the operating state of the dust removal device 300 according to the judgment result of the YOLO model, that is, when the anchor box indicating the cutting impurities can be identified, the blower and the air extraction fan are started for dust removal, and when it cannot be identified, the operation is stopped. This intelligent control method avoids unnecessary energy consumption, and compared with the traditional fixed-mode dust removal system, greatly reduces the energy cost, meeting the development trend of energy conservation and environmental protection.

[0029] The present invention provides another embodiment. This embodiment provides a method for intelligent dust removal of a cutting machine, and the method for intelligent dust removal of the cutting machine includes: S100. Obtain the internal image of the cutting impurity containment box 2 to form a data set.

[0030] Specifically, S200. Use the data set to train the YOLO model. After the YOLO model is trained, obtain the internal image of the cutting impurity containment box 2 and input it into the trained YOLO model, and use the trained YOLO model to judge whether the anchor box for indicating the cutting impurities can be identified.

[0031] S300. Control the operating state of the dust removal device 300 according to the judgment result.

[0032] Specifically, controlling the operating state of the cutting machine according to the judgment result includes: If an anchor box for indicating cutting impurities can be recognized, control the dust removal device 300 to operate; If an anchor box for indicating cutting impurities cannot be recognized, control the dust removal device 300 to stop operating.

[0033] Specifically, controlling the operating state of the cutting machine according to the judgment result specifically includes: If an anchor box for indicating cutting impurities can be recognized, control the blower 8 and the exhaust fan 13 to operate; If an anchor box for indicating cutting impurities cannot be recognized, control the blower 8 and the exhaust fan 13 to stop operating.

[0034] Specifically, training the YOLO model using the data set includes: S1. Create a data set.

[0035] Specifically, in the step S1, it specifically includes: S11. Collect the original image data containing cutting impurities during the working process of the cutting machine. The original image data is from the working scenarios of the cutting machine under multiple different perspectives, different lighting conditions, and different cutting conditions.

[0036] S12. Screen the collected original image data, and remove the images that are blurred, have poor image quality, and do not contain effective information of cutting impurities.

[0037] S13. Perform preprocessing operations before annotation on the screened image data. The preprocessing operations include adjusting the image size to a unified size, graying or performing color space conversion on the image, and normalizing the image pixel values, so that the image data meets the requirements of subsequent model training; S14. Divide the preprocessed image data into a training set, a validation set, and a test set according to a certain ratio to form a data set for training the YOLO model.

[0038] S2. Use Labelme to mark the prediction boxes of cutting impurities.

[0039] Specifically, in the step S2, it specifically includes: S21. Start the Labelme annotation tool, and sequentially import the images in the data set created in step S1 into the Labelme software interface; S22. For each image, carefully observe the specific shape, position, and distribution of the cutting impurities in the image, and determine different annotation categories according to the types of cutting impurities. The types include metal debris, dust clumps, and fiber filaments; S23. Use the drawing function in the Labelme tool to accurately draw a prediction box along the edge of the cut impurity on the image, ensuring that the prediction box completely covers the cut impurity area while minimizing the misselection of non-impurity areas. For cut impurities with irregular shapes, the prediction box should fit its outer contour as closely as possible. S24. Assign a corresponding class label to each drawn prediction box, and accurately fill in the determined cut impurity class name in the annotation information column of Labelme to ensure that the class label is consistent with the pre-set class system. S25. Conduct a preliminary inspection on the annotated image to check whether the prediction box is drawn accurately and whether the class label is correct. For images with doubts or inaccurate annotations, return them for re-observation and annotation to ensure the annotation quality. S26. After all images are annotated, export the data file with prediction box annotation information and class labels in the format specified by Labelme for subsequent training of the YOLO model.

[0040] S3. Train the YOLO model.

[0041] Specifically, in the step S3, it specifically includes: S31. Set the initial training parameters of the YOLO model, and the initial training parameters include learning rate, momentum, weight decay coefficient, batch size, and maximum number of iterations. S32. Input the dataset with annotation information obtained after marking the cut impurity prediction box in step S2 into the YOLO model and start the training process according to the set initial training parameters. S33. During the training process, monitor the loss function value, accuracy, recall rate, and mAP evaluation metrics of the model on the validation set in real time. S34. Dynamically adjust the training parameters according to the monitored changes in the evaluation metrics. For example, when the loss function value no longer decreases or the evaluation metrics fluctuate, reduce the learning rate. S35. Continuously perform training and parameter adjustment until the evaluation metrics of the model on the validation set reach the preset optimal threshold or the training iteration number reaches the maximum number of iterations.

[0042] S4. Save the optimal YOLO model.

[0043] The following uses a specific example to illustrate the process of the present invention: 1. Obtain the internal image of the cut impurity accommodation box 2 and form a dataset (S100).

[0044] 1.1 Set the parameters of the image acquisition device.

[0045] Use a high-definition industrial camera with a resolution of not less than 1920×1080 as the image acquisition device, and set its frame rate to 30 frames per second to ensure that dynamic cutting impurity images during the working process of the cutting machine can be captured. The installation position of the camera should ensure a stable relative angle with the cutting impurity retention box 2, so that the captured images can completely cover the key areas inside the retention box, and the distance from the retention box should be maintained within the range of 30 - 50 cm to obtain the best image clarity and perspective.

[0046] 1.2 Lighting condition setting.

[0047] To ensure the quality of the captured images, arrange a ring-shaped LED light source evenly around the cutting impurity retention box 2. Select white light for the light source color, and set the light intensity to 500 - 1000 lux. Use auxiliary devices such as light shields to reduce the interference of ambient light, and ensure that the captured images have good contrast, which is convenient for subsequent image processing and feature extraction.

[0048] 1.3 Control of data acquisition process.

[0049] During the normal working process of the cutting machine, continuously collect the image data inside the cutting impurity retention box 2, and set the acquisition duration to 24 hours (which can be adjusted appropriately according to the actual cutting conditions) to ensure that various cutting impurity situations generated during the working process of the cutting machine at different time periods are covered, thereby enriching the diversity of the data set.

[0050] 2. Train the YOLO model (S200) using the data set.

[0051] 2.1 Create a data set (S1).

[0052] 2.1.1 Collect original image data (S11).

[0053] Collect the image data of the cutting machine working under different cutting scenarios. During the collection process, ensure that the images are taken from at least 10 different perspectives (such as top view, bottom view, side view, etc.), covering at least 5 different lighting conditions (such as natural bright light, low - illumination dim light, direct light, scattered light, etc.) and 3 different cutting conditions (such as high - speed cutting, low - speed high - precision cutting, irregular - shape cutting, etc.). A total of no less than 10,000 original image data are collected to provide a rich data basis for subsequent model training.

[0054] 2.1.2 Screen the original image data (S12).

[0055] Establish screening rules to remove images with a low degree of clarity (evaluated by the clarity evaluation index of the image, such as the variance of the Laplacian operator, and a variance below 10 is set as a blurred image), poor image quality (evaluated by indicators such as the contrast and brightness of the image, and a contrast below 20 or a brightness range exceeding 10% of the 0 - 255 range is set as a low-quality image), and images that do not contain effective information on cutting impurities (screened by preliminary manual inspection). Retain valid images with good quality to ensure that the number of screened images is not less than 80% of the original images, that is, not less than 8000 images.

[0056] 2.1.3 Preprocessing operation (S13).

[0057] Image size adjustment: Uniformly adjust the size of the screened images to 416×416 pixels. This is a commonly used and appropriate data input size in YOLO model training, which helps the model learn image features.

[0058] Color space conversion and normalization: Convert to grayscale images (if the original images are color images, this operation can be selected according to the actual situation. Grayscale images can reduce the amount of calculation). At the same time, normalize the pixel values of the images so that their range is between 0 and 1 to improve the stability and convergence speed of model training.

[0059] 2.1.4 Dataset division (S14).

[0060] Divide the preprocessed image data into a training set, a validation set, and a test set according to a ratio of 7:2:1, that is, about 5600 images in the training set, about 1600 images in the validation set, and about 800 images in the test set, to form a dataset for training the YOLO model.

[0061] 2.2 Label the prediction boxes of cutting impurities with Labelme (S2).

[0062] 2.2.1 Start the annotation tool and import data (S21).

[0063] Start the Labelme annotation tool and import all the images in the dataset created in step S1 into the Labelme software interface in sequence to ensure that each image can be correctly loaded and processed.

[0064] 2.2.2 Determine the annotation categories (S22).

[0065] For each image, carefully observe the specific shape, location, and distribution of the cutting impurities. According to the types of cutting impurities, clearly divide them into the following 3 annotation categories: Metal debris: Usually in the form of irregular small particles or polyhedrons, with obvious metallic luster.

[0066] Dust agglomerate: Overall, it appears as a relatively loose but mutually adherent particle agglomerate with an irregular shape and a relatively low density.

[0067] Fiber filament: It is slender and has a certain flexibility, presenting a strip or filamentous form.

[0068] 2.2.3 Draw prediction boxes (S23).

[0069] Use the drawing function in the Labelme tool to accurately draw prediction boxes along the edges of the cut impurities. For regularly shaped cut impurities, the maximum gap between the edges of the prediction box and the impurity edges does not exceed 2 pixels; for irregularly shaped cut impurities, the prediction box should fit as closely as possible to its outer contour, with the similarity between the covered area and the actual impurity area reaching more than 90%. At the same time, ensure that the prediction box completely covers the cut impurity area, and the number of pixel points in the misselected non-impurity area does not exceed 5% of the total number of pixels in the selected area.

[0070] 2.2.4 Assign class labels (S24).

[0071] Assign the corresponding class labels to each drawn prediction box, and accurately fill in the determined cut impurity class name in the annotation information column of Labelme to ensure that the labels are consistent with the pre-set class system of "metal debris", "dust agglomerate", and "fiber filament", and avoid label confusion or errors.

[0072] 2.2.5 Check and correct (S25).

[0073] Conduct a 100% comprehensive inspection of the labeled images. Adopt a two-person cross-check method to ensure that the prediction boxes are drawn accurately and the class labels are correct. For images with doubts or inaccurate annotations (such as inconsistent inspection results between two people, the proportion of the prediction box covering the impurity being less than 90%, or the misselection proportion being higher than 5%, etc.), return them for re-observation and annotation until the annotation quality of all images meets the standards.

[0074] 2.2.6 Export data files (S26).

[0075] After all the images are labeled, export the data files with prediction box annotation information and class labels in the JSON format specified by Labelme and save them to a dedicated folder for subsequent use in training the YOLO model.

[0076] 2.3 Train the YOLO model (S3).

[0077] 2.3.1 Set initial training parameters (S31).

[0078] Set the initial training parameters of the YOLO model to ensure the stability and efficiency of the training process. The specific parameter settings are as follows: Learning rate: The initial learning rate is set to 0.001, which is a commonly used range of initial learning rates in deep learning training and helps the model quickly converge to near the optimal solution.

[0079] Momentum: The momentum parameter is set to 0.9. Momentum allows the model to accumulate the exponentially weighted average of previous gradients during training, helping to accelerate convergence and avoid getting stuck in local optima.

[0080] Weight decay coefficient: The weight decay coefficient is set to 0.0005, which is used to prevent overfitting of the model. By imposing a regularization penalty on the model weights, it restricts the complexity of the model.

[0081] Batch size: The batch size (batchsize) is set to 64. This parameter determines the number of samples input to the model during each training. An appropriate batch size can achieve a balance between computational efficiency and training effect.

[0082] Maximum number of iterations: The maximum number of iterations is set to 500 to ensure that the model has enough training epochs to converge to the optimal state.

[0083] 2.3.2 Start training (S32).

[0084] Input the dataset with annotation information obtained after cutting and predicting impurity bounding boxes using Labelme in step S2 into the YOLO model, and start the training process according to the initial training parameters set in step S31, recording the training loss value and validation set metrics for each iteration.

[0085] 2.3.3 Real-time monitoring of evaluation metrics (S33).

[0086] During the training process, real-time monitor various evaluation metrics of the model on the validation set, including but not limited to: Loss function value (Loss): Measures the difference between the model's prediction results and the true labels. The lower the loss function value, the better the model's fitting effect.

[0087] Accuracy: The proportion of correctly predicted samples to the total number of predicted samples, used to evaluate the overall prediction correctness of the model.

[0088] Recall: Among all samples that are actually positive samples, the proportion of samples correctly predicted as positive samples, reflecting the model's ability to capture positive samples.

[0089] mAP (mean Average Precision): Considers the average precision of the model at different IoU (Intersection over Union) thresholds and is an important indicator to measure the comprehensive performance of the object detection model.

[0090] 2.3.4 Dynamically adjust training parameters (S34).

[0091] Dynamically adjust the training parameters according to the monitored changes in evaluation metrics to ensure the effectiveness and stability of model training. The specific adjustment strategies are as follows: When the loss function value no longer decreases in 10 consecutive iterations (i.e., the decrease amplitude is less than 0.001) or the evaluation metrics (such as accuracy, recall, mAP) fluctuate by more than 5%, reduce the learning rate by 10% to more finely adjust the model parameters and avoid missing the optimal solution.

[0092] If, after adjusting the learning rate, the loss function value still does not decrease significantly (the decrease amplitude is less than 0.001) in 20 consecutive iterations and the optimization of the evaluation metrics is not obvious, appropriately increase the weight decay coefficient (increase by 0.0001 each time) to further prevent model overfitting.

[0093] 2.3.5 Continuously train and adjust until convergence (S35).

[0094] Continuously perform training and parameter adjustment until the evaluation metrics of the model on the validation set reach the preset optimal threshold, as follows: The loss function value is lower than 0.01; The accuracy is not less than 90%; The recall is not less than 85%; The mAP is not less than 0.8.

[0095] Or until the maximum number of training iterations (500 times) is reached. Use the model obtained when the optimal evaluation metrics are reached or the maximum number of iterations is reached as the final trained YOLO model.

[0096] 2.4 Save the optimal YOLO model (S4).

[0097] After determining that the optimal evaluation metrics are reached or the maximum number of iterations is reached, save the trained YOLO model in a standard model format (such as the.pt format) and store it in a dedicated model storage folder for subsequent invocation and loading in actual applications.

[0098] 3. Control the operating state of the dust removal device 300 according to the judgment result (S300).

[0099] 3.1 Obtain a new image and input it into the model for judgment (S200 and subsequent).

[0100] During the actual operation, continuously obtain the internal image of the cutting impurity retention box (2), preprocess it (such as size adjustment, normalization, etc.) in the same way as the training model in step S200, and then input it into the trained YOLO model to determine whether an anchor box indicating cutting impurities can be recognized by the model.

[0101] 3.2 Operation control of the dust removal device 300 (specific control of S300).

[0102] 3.2.1 When an anchor box can be recognized (S300 - situation where it can be recognized).

[0103] If an anchor box indicating cutting impurities can be recognized, it indicates that there are cutting impurities in the cutting impurity retention box 2. At this time, control the operation of the dust removal device 300. The specific control parameters are as follows: Hair dryer 8: Set the rotation speed of the hair dryer to 1500 - 2000 revolutions per minute and the air pressure to 800 - 1000 Pa to ensure sufficient suction to pick up the impurities.

[0104] Exhaust fan 13: Set the rotation speed of the exhaust fan to 2000 - 2500 revolutions per minute and the maximum air volume to 1000 - 1500 cubic meters per hour to efficiently discharge the cutting impurities in the retention box to the outside of the retention box.

[0105] 3.2.2 When an anchor box cannot be recognized (S300 - situation where it cannot be recognized).

[0106] If an anchor box indicating cutting impurities cannot be recognized, it indicates that there may be no cutting impurities temporarily in the cutting impurity retention box 2, or the amount of impurities is extremely small. At this time, control the dust removal device 300 to stop running to avoid unnecessary energy consumption.

[0107] It should be noted here that the present invention can accurately judge the actual situation of cutting impurities by obtaining the internal image of the cutting impurity retention box and using the YOLO model for precise recognition. When impurities are recognized, the dust removal device is started in a timely manner, avoiding the ineffective operation of the traditional timed dust removal method when there are few impurities, effectively improving the dust removal efficiency, ensuring that the cutting machine is always in a good working environment, reducing the impact of impurity accumulation on the equipment performance, and thus improving the working efficiency and cutting quality of the cutting machine. The present invention intelligently controls the operation state of the dust removal device according to the judgment result of the YOLO model, that is, when the cutting impurity anchor box can be recognized, the hair dryer and the exhaust fan are started for dust removal, and when it cannot be recognized, the operation is stopped. This intelligent control method avoids unnecessary energy consumption and greatly reduces the energy cost compared with the traditional fixed-mode dust removal system, meeting the development trend of energy conservation and environmental protection.

[0108] In a preferred embodiment, the present application also provides an electronic device, and the electronic device includes: A memory; and a processor, wherein computer-readable instructions are stored on the memory, and when the computer-readable instructions are executed by the processor, the intelligent dust removal method for a cutting machine is implemented. The computer device can be broadly a server, a terminal, or any other electronic device having necessary computing and / or processing capabilities. In one embodiment, the computer device may include a processor, a memory, a network interface, a communication interface, etc. connected through a system bus. The processor of the computer device can be used to provide necessary computing, processing, and / or control capabilities. The memory of the computer device may include a non-volatile storage medium and an internal memory. An operating system, a computer program, etc. may be stored in or on the non-volatile storage medium. The internal memory can provide an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface and the communication interface of the computer device can be used to connect and communicate with external devices through a network. When the computer program is executed by the processor, the steps of the method of the present invention are executed.

[0109] The present invention can be implemented as a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method of the embodiments of the present invention are caused to be executed. In one embodiment, the computer program is distributed on a plurality of network-coupled computer devices or processors, so that the computer program is stored, accessed, and executed in a distributed manner by one or more computer devices or processors. A single method step / operation, or two or more method steps / operations, can be executed by a single computer device or processor or by two or more computer devices or processors. One or more method steps / operations can be executed by one or more computer devices or processors, and one or more other method steps / operations can be executed by one or more other computer devices or processors. One or more computer devices or processors can execute a single method step / operation, or execute two or more method steps / operations.

[0110] Those of ordinary skill in the art can understand that the method steps of the present invention can be completed by a computer program instructing relevant hardware such as a computer device or a processor, and the computer program can be stored in a non-transitory computer-readable storage medium, and when the computer program is executed, the steps of the present invention are caused to be executed. Depending on the situation, any reference to a memory, storage, database, or other medium herein may include non-volatile and / or volatile memories. Examples of non-volatile memories include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, magnetic tape, floppy disk, magneto-optical data storage device, optical data storage device, hard disk, solid state disk, etc. Examples of volatile memories include random access memory (RAM), external cache memory, etc.

[0111] It is understandable that by acquiring the internal image of the cutting impurity accommodation box and using the YOLO model for accurate identification, the present invention can accurately judge the actual situation of the cutting impurities. When impurities are identified, the dust removal device is started in a timely manner, avoiding the ineffective operation of the traditional timed dust removal method when there are few impurities, effectively improving the dust removal efficiency, ensuring that the cutting machine is always in a good working environment, reducing the impact of impurity accumulation on the equipment performance, and thus improving the working efficiency and cutting quality of the cutting machine. The present invention intelligently controls the operating state of the dust removal device according to the judgment result of the YOLO model, that is, when the cutting impurity anchor box is recognized, the hair dryer and the exhaust fan are started for dust removal, and when it cannot be recognized, the operation is stopped. This intelligent control method avoids unnecessary energy consumption and greatly reduces the energy cost compared with the traditional fixed-mode dust removal system, meeting the development trend of energy conservation and environmental protection.

[0112] The technical features described above can be combined arbitrarily. Although not all possible combinations of these technical features are described, any combination of these technical features should be considered to be covered by this specification as long as such a combination does not exist in contradiction.

[0113] The specific implementation manners of the present invention described above do not constitute a limitation to the protection scope of the present invention. Any other corresponding changes and deformations made according to the technical concept of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. An intelligent dust removal system for a cutting machine, characterized in that: The system comprises: A cutting impurity retaining box (2), the cutting impurity retaining box (2) being arranged at the bottom of the cutting machine and being used for retaining cutting impurities of the cutting machine; An acquisition module (100), the acquisition module (100) comprising a camera (17), the camera (17) being used to acquire an internal image of the cut impurity retaining box (2) to form a data set; A control module (200) is used to train a YOLO model using a data set. After the YOLO model is trained, an internal image of the cutting impurity holding box (2) is obtained and input into the trained YOLO model. The trained YOLO model is used to determine whether an anchor frame indicating cutting impurities can be identified, and the operating state of the dust removal device (300) is controlled according to the determination result. The dust removal device (300) comprises a fan unit connected to a cutting impurity retaining box (2), wherein the fan unit is used to discharge the cutting impurities in the cutting impurity retaining box (2) to the outside of the cutting impurity retaining box (2) when in operation.

2. The intelligent dust removal system for cutting machines according to claim 1, characterized in that: The fan unit comprises a blower (8) and an exhaust fan (13); the cutting impurity retaining box (2) is provided with a blowing port (4) and an exhaust port (5); the blowing port (4) is connected to the blower (8) via a pipeline; and the exhaust port (5) is connected to the exhaust fan (13) via a pipeline.

3. A cutting machine intelligent dust removal method, characterized in that: The method comprises: S100, obtaining an internal image of the cut impurity retaining box (2) to form a data set; S200, using the data set to train a YOLO model. After the YOLO model is trained, an internal image of the cut impurity retaining box (2) is obtained and input into the trained YOLO model. The trained YOLO model is used to determine whether an anchor frame for indicating the cut impurities can be identified; S300, controlling the operating state of the dust removal device (300) according to the judgment result.

4. The intelligent dust removal method for a cutting machine according to claim 3, characterized in that: The step of controlling the operating state of the cutting machine according to the judgment result includes: If an anchor frame for indicating cutting impurities can be identified, controlling the dust removal device (300) to operate; If the anchor frame for indicating the cutting impurities cannot be identified, the dust removal device (300) is controlled to stop operating.

5. The intelligent dust removal method for a cutting machine according to claim 4, characterized in that: The controlling the operating state of the cutting machine according to the judgment result specifically includes: If the anchor frame for indicating the cutting impurities can be identified, the blower (8) and the exhaust fan (13) are controlled to operate; If the anchor frame for indicating the cut impurities cannot be identified, the blower (8) and the exhaust fan (13) are controlled to stop running.

6. The intelligent dust removal method for a cutting machine according to claim 1, characterized in that: The YOLO model is trained using the data set, including: S1. Create a dataset. S2, Labelme marks the predicted box of the cutting impurity; S3, train the YOLO model; S4. Save the optimal YOLO model.

7. The intelligent dust removal method for a cutting machine according to claim 6, characterized in that: In the step S1, it specifically includes: S11, collecting original image data containing cutting impurities during the working process of the cutting machine, wherein the original image data is derived from the working scenes of the cutting machine under multiple different viewing angles, different lighting conditions and different cutting conditions; S12, screening the collected raw image data to remove blurred images, images of poor quality, and images that do not contain effective information of cutting impurities; S13, performing preprocessing operations on the screened image data before labeling, wherein the preprocessing operations include adjusting the image size to a uniform size, graying or color space conversion of the image, and normalizing the image pixel values, so that the image data meets the requirements of subsequent model training; S14, dividing the preprocessed image data into a training set, a validation set, and a test set according to a certain ratio to form a data set for training the YOLO model.

8. The intelligent dust removal method for a cutting machine according to claim 6, characterized in that: In the step S2, it specifically includes: S21, start the Labelme annotation tool, and import the images in the data set created in step S1 into the Labelme software interface in sequence; S22, for each image, carefully observing the specific shape, position and distribution of the cutting impurities in the image, and determining different labeling categories according to the types of the cutting impurities, including metal debris, dust lumps, and fiber filaments; S23. Use the drawing function in the Labelme tool to accurately draw a prediction box along the edge of the cut impurities on the image to ensure that the prediction box completely covers the cut impurity area and minimizes the misselection of non-impurity areas. For irregularly shaped cut impurities, the prediction box should fit its outer contour as closely as possible. S24, assigning a corresponding category label to each drawn prediction box, accurately filling in the determined cutting impurity category name in the label information column of Labelme, and ensuring that the category label is consistent with the pre-set category system; S25. Perform a preliminary check on the annotated images to check whether the prediction boxes are drawn accurately and the category labels are correct. For images that are questionable or inaccurately annotated, return to observe and annotate them again to ensure the annotation quality. S26. After all images are labeled, export the data file with the prediction box annotation information and category label in the format specified by Labelme for subsequent training of the YOLO model.

9. The intelligent dust removal method for a cutting machine according to claim 6, characterized in that: In the step S3, it specifically includes: S31, setting the initial training parameters of the YOLO model, wherein the initial training parameters include learning rate, momentum, weight decay coefficient, batch size, and maximum number of iterations; S32, inputting the data set with labeled information obtained after the impurity prediction frame is cut by Labelme in step S2 into the YOLO model, and starting the training process according to the set initial training parameters; S33. During the training process, the loss function value, accuracy, recall rate and mAP evaluation index of the model on the validation set are monitored in real time; S34. Dynamically adjust the training parameters according to the monitored changes in the evaluation indicators, for example, when the loss function value no longer decreases or the evaluation indicators fluctuate, reduce the learning rate; S35. Continue training and parameter adjustment until the evaluation index of the model on the validation set reaches a preset optimal threshold, or the number of training iterations reaches the maximum number of iterations.

10. An electronic device, characterized in that: include: Memory; and a processor, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the intelligent dust removal method for a cutting machine according to any one of claims 1 to 8 is implemented.

Images