Fluororubber fixed-weight cutting and packaging system and method based on YOLOX

Abstract

The invention discloses a fluororubber fixed-weight cutting and packaging system and method based on YOLOX. The system comprises a marking module, a cutting module and a packaging module, wherein the marking module is used for marking fluororubber by using a marking machine when the fluororubber reaches a set length; the visual identification module is used for carrying out real-time detection on the fluororubber image by utilizing a pre-trained YOLOX model so as to identify an image with a mark; the cutting module is used for cutting according to the mark position on the fluororubber by using a cutting knife; the weighing iteration module is used for weighing the cut fluororubber to obtain the weight of the cut fluororubber, calculating the cumulative weight of the fluororubber, and starting an iteration process to obtain the length needing to be cut when the difference value between the target weight and the cumulative weight is smaller than a preset weight threshold value; the automatic rejecting module is used for automatically rejecting the fluororubber which is unqualified after reinspection; and the packaging module is used for packaging qualified fluororubber. The fluororubber cutting device can accurately cut fluororubber with specified weight and automatically remove fluororubber which does not meet requirements, so that the labor cost is greatly reduced.

Description

Technical Field

[0001] This invention relates to the fields of machine vision and automated production technology, specifically to the field of fluororubber cutting technology, and more particularly to a YOLOX-based fluororubber fixed-weight cutting and packaging system and method. Background Technology

[0002] YOLOX is the latest variant of the YOLO (You Only Look Once) object detection model series. The YOLO series is renowned for its fast and accurate object detection capabilities, and YOLOX builds upon this foundation with numerous improvements and optimizations, aiming to provide a more flexible, scalable, and high-performance detection framework. YOLOX features a unified basic network structure, allowing models to share the same network architecture during training and inference phases, simplifying deployment and improving efficiency. Through modular design, YOLOX allows researchers to easily insert or replace different components, such as the backbone, neck, and head networks, to adapt to different task requirements or for algorithmic innovation.

[0003] YOLOX demonstrates superior performance on multiple standard datasets (such as COCO and PASCAL VOC), achieving significant improvements in accuracy while maintaining efficient inference speed, making it suitable for real-time object detection tasks in practical applications. Due to its outstanding performance, YOLOX is widely used in various fields, including but not limited to video surveillance, autonomous driving, drone inspection, intelligent security, and medical image analysis—scenarios requiring fast and accurate object detection.

[0004] Fluororubber is a synthetic polymer elastomer containing fluorine atoms on the carbon atoms of its main chain or side chains. The introduction of fluorine atoms endows the rubber with excellent heat resistance, oxidation resistance, oil resistance, corrosion resistance, and atmospheric aging resistance, leading to its widespread application in aerospace, aviation, automotive, petroleum, and household appliances. It is an irreplaceable key material in advanced defense industries. However, freshly produced fluororubber in factories is characterized by high temperature, high viscosity, and easy deformation. This means that cutting it to fixed lengths or at fixed times cannot meet the specified weight requirements, resulting in the inability to use automated cutting equipment and necessitating manual cutting and weighing.

[0005] Therefore, in actual production, computer vision technology is needed in conjunction with cutting, measuring, testing, and packaging equipment to obtain fluororubber of the specified weight on the production line, thus replacing manual production.

[0006] Currently, there are no articles directly applicable to fluororubber cutting. Some studies utilize pattern recognition and image processing technologies to determine the fixed length, and based on this length, use a closed-loop feedback device to ultimately achieve fixed-weight cutting. A camera remotely acquires process parameters from the production site, replacing operators in harsh environments. Operation and maintenance are convenient and reliable, enhancing the safety of personnel and equipment; it has strong anti-interference capabilities and accurate judgment; the system uses pattern recognition technology to effectively identify steel billets, thus effectively recognizing interference signals from external sources other than steel billets, preventing errors or missed judgments; it has high accuracy, small relative error in fixed-weight cutting, reducing production costs and increasing yield. This method converts the images captured by the camera into grayscale images, uses median filtering to reduce the influence of iron oxide scale on the image, then uses a high-pass filter to sharpen the image, uses region median filtering to replace noise, and finally uses a template to match the steel billet. Its drawback is that template matching requires different templates for different shaped steel billets, and the steel billet used as the template must be complete, without redundancy, and with complete information. For this project, the properties of fluororubber and steel billets differ significantly: fluororubber has irregular boundaries, its shape and size are easily variable, and it is semi-transparent, among other things. Therefore, a more advanced identification method is needed.

[0007] Some studies have proposed image processing-based precision cutting methods for metal materials. These methods include: acquiring grayscale and depth images of the metal to be cut, marked with a cutting path; extracting a skeleton from the grayscale image to obtain all cutting points along the cutting path; calculating the surface roughness of any cutting point in the depth image; calculating the orientation variation characteristics of any cutting point in the grayscale image; using the metal thickness, orientation variation characteristics, and surface roughness as cutting features of the cutting points; and querying the cutting dataset based on the cutting features of the cutting points, starting from a cutting point at one end of the cutting path, to sequentially determine the control parameters for each cutting point, and then cutting the metal. This method requires a flat surface on the material being cut and requires both top-view and cross-sectional images of the material. Due to the translucent nature of fluororubber and its softening at high temperatures, obtaining cross-sectional images is difficult, necessitating alternative methods to obtain the cutting parameters.

[0008] Some studies have developed automatic fixed-length cutting machines for all-steel radial tire pad rubber, solving problems such as cut shrinkage and deformation after fixed-length cutting in the calendering process, the need to apply adhesive to increase adhesion, and inconvenience in jointing. However, these machines only address the problem of cutting rubber materials and are not effective for high-temperature, high-viscosity rubber materials such as fluororubber. Furthermore, they do not utilize machine vision solutions for fixed-weight cutting of materials.

[0009] Some studies utilize computer vision measurement technology to explore online edge detection algorithms and dynamic tracking principles for continuously cast slab images. Multiple area-array CCD (charge-coupled device) cameras capture images of the billet's movement, which are then digitized by a high-speed image acquisition card and input into an industrial control computer. The computer preprocesses the digital images of the billet, detects the slab edges, and calculates the real-time slab length using a measurement model. When the set slab length is reached, a switch output card sends a signal to the cutting machine's PLC (Programmable Logic Controller) control system for fixed-length cutting. This method requires multiple cameras with their field of view axes perpendicular to the conveyor belt. Each camera is responsible for detecting a specific area, and each camera requires a reference line and calculations of the distance between the camera and the conveyor belt. Mathematical calculations are then used to obtain the continuous casting slab length. For fluororubber production lines, too many cameras can lead to program structure chaos. With current hardware and computer vision technology, a single camera can suffice. Furthermore, this method requires fixed cameras and production lines, making it difficult to apply a single solution to fluororubber production lines with varying operating conditions.

[0010] Therefore, to address the complexity of fluororubber product output during production and the challenge of controlling the output from extrusion machines, a mature solution for achieving automated and precise cutting to meet production standards is currently unavailable on the market. To address this challenge, this invention proposes an innovative system to solve the aforementioned problems. Summary of the Invention

[0011] The purpose of this invention is to address the shortcomings of existing technologies by providing a YOLOX-based fluororubber fixed-weight cutting and packaging system and method.

[0012] The objective of this invention is achieved through the following technical solution: The first aspect of this invention provides a YOLOX-based fluororubber fixed-weight cutting and packaging system, comprising:

[0013] The marking module is used to determine the optimal marking size and shape based on the Mooney properties of fluororubber products, and to mark the fluororubber products according to the optimal marking size and shape using a marking machine when the fluororubber products reach the set length.

[0014] The visual recognition module is used to capture images of fluororubber products through a camera and use a pre-trained YOLOX model to detect the images in real time to identify the marked images and generate a cutting signal to send to the cutting module.

[0015] The cutting module is used to pause the conveyor belt after receiving a cutting signal, use a cutting blade to cut the fluororubber product according to the marked position on the fluororubber product, and restart the conveyor belt after the cutting is completed.

[0016] The weighing iteration module is used to transport the cut fluororubber products to the online weighing area via a conveyor belt to obtain the weight data of the fluororubber products, record and save the data and calculate the cumulative weight of the fluororubber products. When the difference between the target weight and the cumulative weight is less than the preset weight threshold, the iteration process is started to perform iterative calculations to obtain the length that still needs to be cut.

[0017] The automatic rejection module is used to stack multiple fluororubber products that have reached the target weight using a stacking machine, and to re-inspect the stacked fluororubber products using a re-inspection scale. The fluororubber products that fail the re-inspection are sent to the waste bin, thus realizing the automatic rejection of unqualified products.

[0018] The packaging module is used to package qualified fluororubber products through the collaborative work of various automated devices.

[0019] Furthermore, a marking machine, a measuring wheel, a camera, a supplementary light, and a cutting knife are sequentially arranged behind the conveyor belt.

[0020] Furthermore, the marking machine is a carbon dioxide laser marking machine.

[0021] Furthermore, the visual recognition module specifically includes:

[0022] Under the illumination of the supplementary light, the camera captures images of fluororubber products. The pre-trained YOLOX model is used to perform real-time detection on the captured images to detect marks in the images in order to identify images with marks. Once the marks on the fluororubber products are identified, the vision recognition module generates a cutting signal and sends it to the cutting module.

[0023] Furthermore, the cutting module specifically includes:

[0024] When the cutting module receives the cutting signal, it pauses the conveyor belt; sets parameters such as power, speed, and stroke for the cutting blade to prepare it; the cutting blade advances and cuts the fluororubber product according to the marked position on the fluororubber product; after cutting, the blade retracts and is turned off; the conveyor belt is restarted after the cutting blade is turned off; if any abnormal situation occurs during the above cutting process, an alarm must be triggered for the abnormal situation.

[0025] Abnormal situations include incorrect parameter settings, parameters exceeding limits, inability to prepare the cutting blade, inability to advance the blade, failure to retract the blade to the specified position after cutting, and inability to turn off the cutting blade.

[0026] Furthermore, the cutting blade is an ultrasonic cutting blade.

[0027] Furthermore, in the weighing iteration module, the iteration process is specifically implemented through the following steps:

[0028] The additional weight is determined by the difference between the target weight and the cumulative weight. The ratio between length and weight is calculated through iterative operations. Based on this ratio, the length that needs to be cut corresponding to the additional weight is calculated.

[0029] Furthermore, the fluororubber products that fail the re-inspection specifically refer to those whose re-inspection weight after stacking exceeds the weight required for qualified products.

[0030] Furthermore, the automated equipment includes automatic sealing machines, automatic packaging equipment, palletizers, and automatic box folding machines.

[0031] A second aspect of this invention provides a method for a YOLOX-based fluororubber fixed-weight cutting and packaging system, specifically including the following steps:

[0032] (1) Use the marking module to determine the optimal marking size and shape, use the measuring wheel to measure the length of the fluororubber product, and use the marking machine to mark the fluororubber product according to the optimal marking size and shape when the fluororubber product reaches the set length;

[0033] (2) After being conveyed by the conveyor belt, the image of the fluororubber product is captured by the camera. The pre-trained YOLOX model in the visual recognition module is used to detect the image in real time to identify the marked image and generate a cutting signal to be sent to the cutting module.

[0034] (3) After receiving the cutting signal, the cutting module pauses the conveyor belt, uses the cutting blade to cut the fluororubber product according to the marked position, and restarts the conveyor belt after the cutting is completed;

[0035] (4) The fluororubber products after cutting are transported to the online weighing area by the conveyor belt using the weighing iteration module. The weight of the fluororubber products is obtained by online weighing, recorded and saved and the cumulative weight of the fluororubber products is calculated. It is determined whether the difference between the target weight and the cumulative weight is less than the preset weight threshold. If the difference is less than the preset weight threshold, the iteration process is started to perform iterative calculation to obtain the length that still needs to be cut; otherwise, return to step (1).

[0036] (5) Take the remaining length to be cut as the new set length, and repeat steps (1)-(4) until the preset target weight is reached.

[0037] (6) Using an automatic rejection module, multiple fluororubber products that have reached the target weight are transferred to a stacking machine via a cooling conveyor belt. After being stacked by the stacking machine, they are re-inspected by a re-inspection scale. Fluororubber products that fail the re-inspection are sent to a waste bin to automatically reject unqualified fluororubber products.

[0038] (7) Using various automated equipment in the packaging module, qualified fluororubber products are packaged accordingly.

[0039] The beneficial effects of this invention are as follows: This invention innovatively proposes a fluororubber fixed-weight cutting and packaging system. This system integrates a marking module, a visual recognition module, and a weighing iteration module. Through the collaborative work of these modules, automatic and precise cutting of fluororubber can be achieved even in production environments with varying product characteristics. This invention uses the YOLOX model to construct the visual recognition module, which can achieve accurate identification and cutting of markings. After the cut products are re-inspected and weighed, they enter the weighing iteration module, which uses a built-in algorithm for adaptive dynamic adjustment to ensure the accuracy of the measuring wheel distance. Under complex product characteristics, this system can dynamically adjust the cutting strategy to achieve precise cutting. This invention can precisely cut fluororubber to a specified weight, automatically rejecting substandard fluororubber, and can operate in all weather conditions and at high temperatures. The YOLOX model provides accurate recognition, rapid cutting, and clean cuts. The cutting length can be adjusted according to different packaging requirements while maintaining the specified weight. No manual intervention is required from extrusion to packaging and stacking, significantly reducing labor costs and associated personnel safety risks. Furthermore, this invention includes a waste rejection mechanism to achieve the recycling of defective products. In summary, this invention overcomes the technical challenge of achieving automated and precise cutting of fluororubber products during extrusion and realizes a system capable of adaptively and dynamically adjusting the cutting strategy. Attached Figure Description

[0040] Figure 1 This is an overall workflow diagram of the YOLOX-based fluororubber fixed-weight cutting and packaging system of the present invention;

[0041] Figure 2 This is a production line deployment diagram of the YOLOX-based fluororubber fixed-weight cutting and packaging system of the present invention;

[0042] Figure 3 This is a flowchart of the cutting module of the present invention. Detailed Implementation

[0043] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the invention as detailed in the appended claims.

[0044] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The singular forms “a,” “the,” and “the” used in this invention and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0045] It should be understood that although the terms first, second, third, etc., may be used in this invention to describe various information, this information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, first information may also be referred to as second information without departing from the scope of this invention, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to a determination."

[0046] The present invention will now be described in detail with reference to the accompanying drawings. Unless otherwise specified, the features of the following embodiments and implementations can be combined with each other.

[0047] The present invention relates to a YOLOX-based fluororubber fixed-weight cutting and packaging system, specifically comprising a marking module, a vision recognition module, a cutting module, a weighing iteration module, an automatic rejection module, and a packaging module. The marking module determines the optimal marking size and shape based on the Mooney properties of the fluororubber product, and marks the product with a marking machine according to the optimal size and shape when the product reaches a set length. The vision recognition module captures images of the fluororubber product using a camera and performs real-time detection on the images using a pre-trained YOLOX model to identify marked images and generate a cutting signal, which is then sent to the cutting module. The cutting module pauses the conveyor belt upon receiving the cutting signal, cuts the fluororubber product using a cutting blade according to the marked positions, and restarts the conveyor belt after cutting. The weighing iteration module transports the cut fluororubber product to an online weighing area via the conveyor belt to obtain the weight data of the fluororubber product online, records and saves the data, and calculates the cumulative weight. When the difference between the target weight and the cumulative weight is less than a preset weight threshold, an iterative process is initiated to calculate the remaining cutting length. The automatic rejection module is used to stack multiple fluororubber products that have reached the target weight using a stacking machine, and then re-inspect the stacked fluororubber products using a re-inspection scale. Fluororubber products that fail the re-inspection are sent to a waste bin, thus automatically rejecting unqualified products. The packaging module is used to package the qualified fluororubber products through the collaborative work of various automated devices.

[0048] In this embodiment, the potential of laser marking technology was thoroughly explored and fully utilized during the design and implementation of the marking module. In particular, a meticulous study and analysis of the Mooney properties of fluororubber products was conducted to determine the most suitable marking size and shape for optimal marking results. For example, when processing A1 type fluororubber materials with Mooney values ​​between 20 and 30, after precise calculations and experiments, a square with a side length of 3 mm was ultimately chosen as the marking shape. This choice was based on a deep understanding of the material's physical properties and ensured that the marking clarity met the requirements of the visual model in the visual recognition module for accurate recognition during subsequent cutting and processing. It should be noted that the lower the Mooney value, the greater the deformation after conveyor belt transport. If not adjusted in real time, this would affect the subsequent cutting process and render the marking meaningless. Therefore, it is necessary to determine the optimal marking size and shape based on the Mooney properties of the fluororubber products. Specifically, this can be achieved by manually testing the relationship between the Mooney value and deformation of the fluororubber products experimentally, determining the relationship between the two, and then using this relationship to determine the direction for adjusting the marking size and shape. To ensure the quality of fluororubber products remains unaffected during production, a carbon dioxide laser marking machine was specifically selected for the marking process. This equipment produces fine black carbonized lines at the extrusion cutting point, which are clearly visible without causing excessive thermal damage to the material. Furthermore, to further minimize any potential interference during subsequent processing, the marking machine was positioned next to the cutting blade. This layout ensures that marking and cutting are performed simultaneously, thereby improving production efficiency while reducing potential impacts on product quality.

[0049] Preferably, a carbon dioxide laser marking machine is selected as the marking machine.

[0050] In this embodiment, on the fluororubber product industrial production line, a vision recognition module is carefully designed and placed behind the conveyor belt. It utilizes a pre-trained YOLOX model to monitor each fluororubber product passing through the conveyor belt in real time, especially those with specific markings. When these marked fluororubber products pass the vision recognition module located above the conveyor belt, a high-resolution camera captures an image of the product under supplementary lighting. The pre-trained YOLOX model then performs real-time detection on the image, identifying the markings. Once the markings are identified, the vision recognition module quickly generates a precise cutting signal and sends it to the cutting module. This allows the cutting module to cut at the precise marked location on the fluororubber product, ensuring cutting accuracy and product quality, thus significantly improving production efficiency and product quality. This vision recognition module guarantees product quality by capturing high-resolution images and utilizing the high efficiency of the YOLOX model, enabling rapid image data processing and real-time monitoring of the fluororubber product's status.

[0051] To verify the effectiveness of the YOLOX model proposed in this invention, the research data came from the actual production process of a real company. This invention used an industrial area scan camera to completely record the entire process of fluororubber slicing. The acquired monitoring videos were randomly divided into training and test sets in a 7:3 ratio. To ensure data reliability, the acquired video format was single-channel AVI, with a resolution of 2592*1944 and a capture rate of 30 frames per second. To ensure the continuous and stable operation of the visual inspection system for fluororubber cutting markers in industrial scenarios, supplementary lighting was used to enhance the video image labels. Ultimately, 2207 labeled images and 3814 unlabeled images were obtained, with the YOLOX model achieving a validation accuracy of 90%. To improve the processing speed of the YOLOX model to meet the needs of real-world scenarios, the video images were first adjusted to a standard size of 224*224. Furthermore, various data augmentation methods were employed, such as random cropping, random rotation, random noise, and random erasing, to expand the effectiveness and trainability of the training data. On the other hand, to enable the YOLOX model to adapt to complex environments under different lighting conditions, random brightness and random contrast were introduced to simulate real-world scenes. Finally, to alleviate the long-tail problem caused by the imbalance of the number of samples from different states, oversampling was performed on the fewer samples of the class to balance the number of samples from different states in the training queue. This helps reduce the impact of data imbalance on the YOLOX model training, ensuring that the YOLOX model can effectively learn various states. Based on this, various data augmentation methods were also employed, such as random pruning, random rotation, random noise, and random erasing, to expand the effectiveness and trainability of the training data. The YOLOX model is pre-trained using the aforementioned training data. Its structure and training process are publicly available and commonly used, and will not be elaborated upon here. Afterwards, the pre-trained YOLOX model can be used in the system to detect markers in images of fluororubber products.

[0052] Furthermore, the performance of the proposed fluororubber visual inspection system was thoroughly validated using accuracy and average accuracy metrics to ensure reliable precision and generalization ability. Accuracy is a simple and intuitive performance metric, commonly used in classification problems, representing the proportion of correctly predicted samples out of the total number of samples. However, accuracy is easily affected by class imbalance. Accuracy is generally low when a certain state has a large number of samples and low accuracy, while other states have high accuracy. Conversely, accuracy may be high when a certain state has a large number of samples and high accuracy, while other states have low accuracy. This indicates that accuracy alone cannot fairly evaluate the model's performance under imbalanced data. In contrast, average accuracy is an important performance metric, particularly suitable for evaluating the overall performance of a model in object detection or information retrieval tasks. Because it is a relative rather than an absolute metric, it still provides a good evaluation of model performance even under imbalanced data conditions.

[0053] Furthermore, this visual recognition module ensures the accuracy of weighing data. After the product is identified by the visual recognition module, if the weight cut exceeds 3kg, the system will automatically record the weight information of the product that needs to be replenished to 2kg and synchronize it with the cutting signal to ensure accurate data support for subsequent weighing and packaging processes. This accurate data support not only improves production efficiency but also ensures product quality, providing strong technical support for the company's production management.

[0054] In this embodiment, the measuring wheel in the cutting module transmits the product's length information through the system, ensuring that the length data of each product segment is accurately recorded and processed. When the product length reaches the preset standard length of 50cm, the marking module marks it. After being transported by the conveyor belt, the marking is detected by the visual recognition module. When a marked fluororubber product is detected, a cutting signal is generated and sent to the cutting module. Upon receiving the corresponding cutting signal instruction, the cutting module pauses the conveyor belt and uses a cutting blade to cut the fluororubber product according to the marked position. After the cutting action is completed, the length information is cleared to zero so that the length calculation and tracking of the next fluororubber product segment can begin again. At the same time, the weighing platform continuously monitors and accumulates the product's weight data, which is stored in the system's internal database. The system compares this accumulated weight data with the final weight of the target product (5kg) and calculates the difference between the two. Once this difference is less than 2kg, the control system activates the iterative process of the weighing iteration module, starting to execute a more refined calculation and control process. The weighing iteration module works by recording the weight information of each 50cm segment of the product being cut and determining the ratio between length and weight through a series of calculations. The module accumulates the current weight data and further calculates the final segment length needed to reach the target total weight of 5kg. Once the calculation is complete, this length information is input into the system. When the measuring wheel detects that the product length has reached the length corresponding to this accumulated weight, the system instructs the cutting blade to perform the cutting operation, ensuring that the final weight of the product accurately reaches 5kg.

[0055] Furthermore, the cutting signal of the cutting module mainly originates from two points: ① a cutting signal sent based on the default length input by the measuring wheel within 3 kg. ② a cutting signal sent when the measuring wheel reaches the remaining length calculated by the weighing iteration module.

[0056] Furthermore, such as Figure 3 As shown, the specific workflow of the cutting module includes: when the cutting module receives a cutting signal, it pauses the conveyor belt; it sets parameters such as power, speed, and stroke for the cutting blade to prepare it; the cutting blade advances and cuts the fluororubber product according to the marked position on the product; after cutting, it retracts the blade and shuts it off; after the blade shuts off, the conveyor belt restarts; if any abnormal situation occurs during the above cutting process, an alarm must be triggered. Abnormal situations include incorrect parameter settings, parameters exceeding limits, failure to prepare the cutting blade, inability to advance the blade, failure to retract the blade to the designated position after cutting, and inability to shut off the cutting blade.

[0057] It should be noted that due to the inherent physical properties of fluororubber—namely, its high fluoride ion content and the significant decrease in its Mooney coefficient during the heating and slicing process—ordinary cutting blades are not well-suited for this situation. In traditional production methods, ordinary mechanical cutting blades frequently experience blade sticking, posing a significant safety hazard in the factory. Furthermore, the adhesion of fluororubber products at the cutting blade location can damage the blade itself. Damage to the cutting blade, in turn, increases production costs.

[0058] Given the aforementioned factors, flexible materials (such as rubber and plastic films) are widely used in various fields in modern industry, including electronic devices, automotive manufacturing, and medical devices. However, how to efficiently and accurately cut these flexible materials has always been a technical challenge. Traditional cutting techniques often exhibit poor performance on flexible materials, such as easily causing material deformation and unsightly edges. Therefore, ultrasonic cutting technology has gradually gained attention, demonstrating significant advantages in processing flexible materials. Adjusting the frequency and amplitude of ultrasonic vibrations can significantly improve cutting quality. For example, higher frequencies typically produce smoother cutting edges and reduce cutting forces, thereby minimizing material deformation. Through precise parameter control, the cutting effect of flexible materials can be effectively improved, especially when processing complex structures or thin film materials. Cutting flexible materials using ultrasonic cutting technology faces significant challenges due to their inherent characteristics (such as high elasticity and sensitivity to vibration). However, these challenges can be effectively addressed by optimizing ultrasonic parameters, utilizing specially designed cutting tools, support materials, and employing hybrid techniques. Compared to traditional mechanical cutting, ultrasonic cutting offers significant improvements in precision, cutting quality, and material deformation control. Continuous advancements in ultrasonic cutting tool design and process optimization are paving the way for more reliable and precise cutting of flexible materials, and this technology has already been validated in food industry applications.

[0059] Therefore, this invention innovatively improves upon traditional fluororubber cutting tools. Specifically, an ultrasonic cutting tool is preferably used. When processing sticky materials, ultrasonic cutting tools exhibit the following advantages: Through high-frequency vibration, ultrasonic cutting significantly reduces friction and heat generation, effectively preventing burrs, tearing, and edge warping during the cutting process. Ultrasonic vibration-assisted cutting technology reduces the elastic recovery effect of the material, thereby improving the accuracy of the cutting dimensions and ensuring consistency in the cutting path and edge quality. Furthermore, ultrasonic cutting does not require excessive pressure or high temperatures, which helps maintain smooth and neat cutting edges while reducing material loss during the cutting process and the need for subsequent processing. Ultrasonic cutting is easily integrated into automated production lines, improving production efficiency and the consistency of cutting quality. These advantages of ultrasonic cutting technology make it particularly effective when processing sticky materials such as fluororubber. Since fluororubber easily adheres to the cutting tool during processing, the high-frequency vibration of the ultrasonic cutting tool effectively reduces this adhesion, thereby reducing safety risks during production. Simultaneously, the precise control capability of ultrasonic cutting technology allows for neater and more consistent edges when cutting fluororubber, which is crucial for improving the overall quality of the product. Furthermore, the superior material loss reduction performance of ultrasonic cutting technology means that material waste can be minimized when cutting fluororubber, thereby improving material utilization and production efficiency. In conclusion, ultrasonic cutting technology not only improves the efficiency and safety of cutting fluororubber but also enhances product quality, making it an ideal choice for processing flexible materials in modern industry.

[0060] In this embodiment, the weighing iteration module primarily employs online weighing technology, conveying the cut fluororubber products to the online weighing area via a conveyor belt. In the online weighing area, the fluororubber products are precisely weighed online, acquiring their weight data, which is then transmitted to the system's internal database via an interface. Given the inherent uncertainty in the output of fluororubber products, the core objective of this invention is to ensure a total product weight of 5kg. However, because fluororubber is prone to deformation at the extrusion sheet location, a 5kg product often results in a very long length, leading to various problems. This causes more internal changes in the fluororubber, potentially further impacting the overall product weight. To address this issue and achieve more accurate iterative weight calculation, this invention does not perform any iterative operations within the first 3kg weight range, only performing fixed-length cutting to accumulate weight. In the final 2kg weight range, an iterative process is initiated to accurately calculate the required additional length, ensuring a total product weight of 5kg. Through continuous iterative calculations, this invention ensures the stability of the product weight.

[0061] Furthermore, the iterative process is implemented as follows: The remaining weight is determined based on the difference between the target weight and the accumulated weight. The proportional relationship between length and weight is calculated through iterative operations to determine the remaining cutting length corresponding to the additional weight. Specifically, a long strip of fluororubber is cut into multiple pieces, and the weight of each piece is obtained. The weighing iteration module calculates the remaining cutting length based on the weight of each piece. In other words, the weighing iteration module accumulates the current weight data and further calculates how much weight the final section needs to be added to achieve the target total weight of 5kg. The remaining length of the final section can be determined based on the determined proportional relationship between length and weight.

[0062] In this embodiment, for products exceeding the preset weight range, the present invention incorporates an automatic rejection module. This module automatically rejects stacked, non-conforming products (target weight ±0.02 kg) and returns them to the waste bin for reprocessing in subsequent processes. This reduces losses and resource waste during production line operation and ensures product quality. Specifically, non-conforming fluororubber products are defined as those whose re-inspection weight exceeds the required weight for a qualified product after stacking. It should be understood that a product weight within ±0.02 kg of the target weight after re-inspection is considered a qualified product. Alternatively, a weight within ±0.03 kg of the target weight can also be defined as a qualified product; the specific setting can be adjusted according to actual needs.

[0063] The system described in this invention employs advanced algorithms to monitor and adjust the cutting process in real time, ensuring that the weight and length of each product meet predetermined standards. Furthermore, the system has a self-diagnostic function, capable of identifying and reporting any anomalies that may affect production efficiency and product quality. Through these measures, this research module significantly improves production efficiency while reducing the need for manual intervention, optimizing the automated production process. The control system can also adjust production parameters based on real-time data to adapt to different production conditions and requirements. This intelligent adjustment mechanism not only improves product quality consistency but also further reduces production costs. In addition, the system integrates a data analysis module, enabling in-depth analysis of data during the production process, providing a scientific basis for production decisions, thereby further enhancing the intelligence level of the production process. For example, it can calculate the output per minute based on the planned output and use this output per minute to estimate the weight; or it can predict the stability of the extruded fluororubber based on the material level in the upper fluororubber silo, and so on.

[0064] In this embodiment, the basic equipment of the packaging module includes various automated devices, such as automatic sealing machines, automatic packaging equipment, palletizers, and automatic carton folding machines, etc. Figure 2As shown, the coordinated operation of these automated devices ensures the efficiency and smoothness of the entire process from initial packaging to final packing. Products that pass re-inspection and weighing are then sent to a vacuum packaging machine for final packing. Packed products are then vacuum-sealed to ensure safety and stability during transfer. Subsequently, these products are loaded into an automatic box-folding machine, where they are precisely folded and arranged into neat boxes. Next, the products are secured with cable ties to ensure the boxes' stability during transport. Finally, the packed products are placed on pallets and stacked by a palletizer, preparing them for subsequent logistics and distribution. In the packaging process, the automatic stacking machine plays a crucial role, neatly stacking the packaged products to form stable pallet stacks. This process not only improves space utilization but also ensures the safety of the products during storage and transportation. The palletizer, according to a preset pattern, places these pallet stacks onto transport vehicles in a predetermined order and position. This automation greatly improves the efficiency and accuracy of logistics and distribution. The entire packaging module is designed to automate the product packaging process. By reducing manual operations, it not only improves production efficiency but also ensures the consistency and standardization of product packaging. Furthermore, the modular design makes future upgrades and maintenance more convenient, thereby ensuring the long-term stable operation of the production line and providing solid technical support for the company's sustainable development.

[0065] It is worth mentioning that the embodiments of the present invention also provide a method for fixed-weight cutting and packaging of fluororubber based on YOLOX, which is implemented based on the YOLOX-based fluororubber fixed-weight cutting and packaging system in the above embodiments.

[0066] like Figure 1 and Figure 2 As shown, the method specifically includes the following steps:

[0067] (1) Use the marking module to determine the optimal marking size and shape, use the measuring wheel to measure the length of the fluororubber product, and use the marking machine to mark the fluororubber product according to the optimal marking size and shape when the fluororubber product reaches the set length.

[0068] (2) After being conveyed by the conveyor belt, the image of the fluororubber product is captured by the camera. The image is detected in real time using the pre-trained YOLOX model in the visual recognition module to identify the marked image and generate a cutting signal to be sent to the cutting module.

[0069] (3) After receiving the cutting signal, the cutting module pauses the conveyor belt, uses the cutting blade to cut the fluororubber product according to the marked position, and restarts the conveyor belt after the cutting is completed.

[0070] (4) The fluororubber products after cutting are transported to the online weighing area by the conveyor belt using the weighing iteration module. The weight of the fluororubber products is obtained by online weighing, recorded and saved, and the cumulative weight of the fluororubber products is calculated. It is determined whether the difference between the target weight and the cumulative weight is less than the preset weight threshold. If the difference is less than the preset weight threshold, the iteration process is started to perform iterative calculation to obtain the length that still needs to be cut; otherwise, return to step (1).

[0071] (5) Take the remaining length to be cut as the new set length, and repeat steps (1)-(4) until the preset target weight is reached.

[0072] (6) Using an automatic rejection module, multiple fluororubber products that have reached the target weight are transferred to a stacking machine via a cooling conveyor belt. After being stacked by the stacking machine, they are re-inspected by a re-inspection scale. Fluororubber products that fail the re-inspection are sent to a waste bin to automatically reject unqualified fluororubber products.

[0073] (7) Using various automated equipment in the packaging module, qualified fluororubber products are packaged accordingly.

[0074] Furthermore, a marking machine, a measuring wheel, a camera, a supplementary light, and a cutting knife are sequentially installed behind the conveyor belt.

[0075] It should be noted that placing the marking machine in front of the measuring wheel is equivalent to allowing the measuring wheel to detect the length "in advance". When the material moves to the set length, the marking machine has already completed the marking in the correct position (compensating for the delay time), solving the problem of synchronization between measurement and marking, and avoiding errors caused by material elasticity, marking delay, or mechanical response time.

[0076] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A YOLOX-based fluororubber fixed-weight cutting and packaging system, characterized in that, include: The marking module is used to determine the optimal marking size and shape based on the Mooney properties of fluororubber products, and to mark the fluororubber products according to the optimal marking size and shape using a marking machine when the fluororubber products reach the set length. The visual recognition module is used to capture images of fluororubber products through a camera and use a pre-trained YOLOX model to detect the images in real time to identify the marked images and generate a cutting signal to send to the cutting module. The cutting module is used to pause the conveyor belt after receiving a cutting signal, use a cutting blade to cut the fluororubber product according to the marked position on the fluororubber product, and restart the conveyor belt after the cutting is completed. The weighing iteration module is used to transport the cut fluororubber products to the online weighing area via a conveyor belt to obtain the weight data of the fluororubber products, record and save the cumulative weight of the fluororubber products, and start the iteration process to perform iterative calculations to obtain the length that still needs to be cut when the difference between the target weight and the cumulative weight is less than the preset weight threshold. The automatic rejection module is used to stack multiple fluororubber products that have reached the target weight using a stacking machine, and to re-inspect the stacked fluororubber products using a re-inspection scale. The fluororubber products that fail the re-inspection are sent to the waste bin, thus realizing the automatic rejection of unqualified products. The packaging module is used to package qualified fluororubber products through the collaborative work of various automated devices.

2. The YOLOX-based fluororubber fixed-weight cutting and packaging system according to claim 1, characterized in that, Behind the conveyor belt are arranged in sequence a marking machine, a measuring wheel, a camera, a supplementary light, and a cutting knife.

3. The YOLOX-based fluororubber fixed-weight cutting and packaging system according to claim 1, characterized in that, The marking machine used is a carbon dioxide laser marking machine.

4. The YOLOX-based fluororubber fixed-weight cutting and packaging system according to claim 1, characterized in that, The visual recognition module specifically includes: Under the illumination of the supplementary light, the camera captures images of fluororubber products. The pre-trained YOLOX model is used to perform real-time detection on the captured images to detect marks in the images in order to identify images with marks. Once the marks on the fluororubber products are identified, the vision recognition module generates a cutting signal and sends it to the cutting module.

5. The YOLOX-based fluororubber fixed-weight cutting and packaging system according to claim 1, characterized in that, The cutting module specifically includes: When the cutting module receives the cutting signal, it pauses the conveyor belt; sets parameters such as power, speed, and stroke for the cutting blade to prepare it; the cutting blade advances and cuts the fluororubber product according to the marked position on the fluororubber product; after cutting, the blade retracts and is turned off; the conveyor belt is restarted after the cutting blade is turned off; if any abnormal situation occurs during the above cutting process, an alarm must be triggered for the abnormal situation. Abnormal situations include incorrect parameter settings, parameters exceeding limits, inability to prepare the cutting blade, inability to advance the blade, failure to retract the blade to the specified position after cutting, and inability to turn off the cutting blade.

6. The YOLOX-based fluororubber fixed-weight cutting and packaging system according to claim 1 or 5, characterized in that, The cutting blade is an ultrasonic cutting blade.

7. The YOLOX-based fluororubber fixed-weight cutting and packaging system according to claim 1, characterized in that, In the weighing iteration module, the iteration process is specifically implemented through the following steps: The additional weight is determined by the difference between the target weight and the cumulative weight. The ratio between length and weight is calculated through iterative operations. Based on this ratio, the length that needs to be cut corresponding to the additional weight is calculated.

8. The YOLOX-based fluororubber fixed-weight cutting and packaging system according to claim 1, characterized in that, The fluororubber products that failed the re-inspection specifically refer to those whose re-inspection weight after stacking exceeded the weight required for qualified products.

9. The YOLOX-based fluororubber fixed-weight cutting and packaging system according to claim 1, characterized in that, The automated equipment includes automatic sealing machines, automatic packaging equipment, palletizers, and automatic box folding machines.

10. A method for a YOLOX-based fluororubber fixed-weight cutting and packaging system according to any one of claims 1-9, characterized in that, Specifically, the steps include the following: (1) Use the marking module to determine the optimal marking size and shape, use the measuring wheel to measure the length of the fluororubber product, and use the marking machine to mark the fluororubber product according to the optimal marking size and shape when the fluororubber product reaches the set length; (2) After being conveyed by the conveyor belt, the image of the fluororubber product is captured by the camera. The pre-trained YOLOX model in the visual recognition module is used to detect the image in real time to identify the marked image and generate a cutting signal to be sent to the cutting module. (3) After receiving the cutting signal, the cutting module pauses the conveyor belt, uses the cutting blade to cut the fluororubber product according to the marked position, and restarts the conveyor belt after the cutting is completed; (4) The fluororubber products after cutting are transported to the online weighing area by the conveyor belt using the weighing iteration module. The weight of the fluororubber products is obtained by online weighing, recorded and saved and the cumulative weight of the fluororubber products is calculated. It is determined whether the difference between the target weight and the cumulative weight is less than the preset weight threshold. If the difference is less than the preset weight threshold, the iteration process is started to perform iterative calculation to obtain the length that still needs to be cut; otherwise, return to step (1). (5) Take the remaining length to be cut as the new set length, and repeat steps (1)-(4) until the preset target weight is reached. (6) Using an automatic rejection module, multiple fluororubber products that have reached the target weight are transferred to a stacking machine via a cooling conveyor belt. After being stacked by the stacking machine, they are re-inspected by a re-inspection scale. Fluororubber products that fail the re-inspection are sent to a waste bin to automatically reject unqualified fluororubber products. (7) Using various automated equipment in the packaging module, qualified fluororubber products are packaged accordingly.

Images