Intelligent cutting control system and method for aluminum alloy profile

Abstract

The invention relates to the technical field of aluminum alloy profile cutting, and particularly discloses an intelligent cutting control system and method for an aluminum alloy profile. An image processing and analysis algorithm based on artificial intelligence is introduced at the rear end to analyze the state image of the cutting surface of the aluminum alloy section, so that semantic features of the state of the cutting surface of the aluminum alloy section expressed in the image are captured; therefore, based on the semantic representation of the cutting surface state of the aluminum alloy profile, the cutting surface state is recognized to determine whether burn exists or not, and a corresponding cutting optimization control instruction is automatically generated according to the recognition result, for example, the flow of cooling liquid is increased, so that surface defects are prevented and relieved. Thus, the cooling system parameters during cutting can be dynamically optimized and controlled according to the actual state of the cutting face of the aluminum alloy profile, the intelligent degree of the cutting control process of the aluminum alloy profile is improved, and the reliability and stability of the cutting process are further improved.

Description

Technical Field

[0001] The present application relates to the technical field of aluminum alloy profile cutting, and more specifically, to an intelligent cutting control system and method for aluminum alloy profiles. Background Art

[0002] In modern industrial production, aluminum alloy profiles are widely used in many fields such as construction, transportation, and electronics due to their advantages such as light weight, high strength, good conductivity and corrosion resistance. However, in the processing of aluminum alloy profiles, especially in the cutting process, a series of technical difficulties may arise, among which the most significant is the surface defects caused by overheating during the cutting process, such as burns. These surface defects not only affect the appearance quality of the product, but also may reduce the mechanical properties of the material, thereby limiting the application range of the product.

[0003] Traditionally, to avoid overheating during the cutting process, people usually rely on fixed parameter control methods or rely on the operator's experience to adjust the cutting parameters and cooling system settings. Fixed parameter control methods cannot adapt to changes in different material properties, environmental conditions or cutting requirements. For example, for some complex cutting tasks, fixed cooling parameters may not be sufficient to prevent overheating, thereby increasing the risk of surface defects. At the same time, setting overly conservative cooling parameters to avoid overheating may lead to excessive use of resources such as coolant. The manual control method of the operator will be affected by the differences in experience and technical level of different operators, resulting in inconsistent product quality. In addition, when some surface defects occur, the operator may not be able to detect them immediately or react quickly enough, resulting in an increase in defective products.

[0004] Therefore, an intelligent cutting control solution for aluminum alloy profiles is desired. Summary of the invention

[0005] The present application provides an intelligent cutting control system and method for aluminum alloy profiles, which can dynamically optimize and control the cooling system parameters during cutting according to the actual cutting surface state of the aluminum alloy profile, thereby improving the intelligence level of the aluminum alloy profile cutting control process and further improving the reliability and stability of the cutting process.

[0006] In a first aspect, a method for intelligent cutting control of an aluminum alloy profile is provided, comprising:

[0007] Acquire the cutting surface state image of the aluminum alloy profile captured by the camera;

[0008] Extracting features from the aluminum alloy profile cutting surface state image to obtain aluminum alloy profile cutting surface state features;

[0009] The aluminum alloy profile cutting surface state feature is subjected to pixel granularity-based explicit feature enhancement processing to obtain the aluminum alloy profile cutting surface state semantic enhancement feature, including: performing pixel granularity feature decoupling on the aluminum alloy profile cutting surface state feature to obtain a set of aluminum alloy profile cutting surface state pixel granularity decoupling features; calculating the significant feature weight of each aluminum alloy profile cutting surface state pixel granularity decoupling feature in the set of aluminum alloy profile cutting surface state pixel granularity decoupling features, and performing pixel granularity weighted enhancement on the aluminum alloy profile cutting surface state feature based on the significant feature weight of each aluminum alloy profile cutting surface state pixel granularity decoupling feature to obtain the aluminum alloy profile cutting surface state semantic enhancement feature;

[0010] Based on the semantic enhancement features of the cutting surface state of the aluminum alloy profile, cutting surface state recognition is performed to determine whether there is a burn. If the recognition result is that there is a burn, a cutting optimization control instruction is generated to indicate a cutting optimization control instruction for increasing the coolant flow rate.

[0011] In a second aspect, an intelligent cutting control system for aluminum alloy profiles is provided, comprising:

[0012] Aluminum alloy profile cutting image acquisition module, used to acquire the aluminum alloy profile cutting surface state image collected by the camera;

[0013] An aluminum alloy profile cutting surface state feature extraction module, used for extracting features from the aluminum alloy profile cutting surface state image to obtain aluminum alloy profile cutting surface state features;

[0014] An aluminum alloy profile cutting surface state feature enhancement processing module is used to perform pixel-granularity-based explicit feature enhancement processing on the aluminum alloy profile cutting surface state feature to obtain an aluminum alloy profile cutting surface state semantic enhancement feature, wherein the aluminum alloy profile cutting surface state feature enhancement processing module includes: a pixel granularity feature decoupling processing unit, used to perform pixel granularity feature decoupling on the aluminum alloy profile cutting surface state feature to obtain a set of aluminum alloy profile cutting surface state pixel granularity decoupling features; a weighted enhancement processing unit, used to calculate the significant feature weight of each aluminum alloy profile cutting surface state pixel granularity decoupling feature in the set of aluminum alloy profile cutting surface state pixel granularity decoupling features, and perform pixel granularity weighted enhancement on the aluminum alloy profile cutting surface state feature based on the significant feature weight of each aluminum alloy profile cutting surface state pixel granularity decoupling feature to obtain an aluminum alloy profile cutting surface state semantic enhancement feature;

[0015] A cutting surface state recognition module is used to perform cutting surface state recognition based on the semantic enhancement features of the cutting surface state of the aluminum alloy profile to determine whether there is a burn. If the recognition result is that there is a burn, a cutting optimization control instruction is generated to indicate a cutting optimization control instruction for increasing the coolant flow rate.

[0016] The application provides an intelligent cutting control system and method for aluminum alloy profiles, which collects the state image of the cutting surface of the aluminum alloy profile, and introduces an image processing and analysis algorithm based on artificial intelligence at the back end to analyze the state image of the cutting surface of the aluminum alloy profile, so as to capture the semantic features of the cutting surface state of the aluminum alloy profile shown in the image, thereby performing cutting surface state recognition based on the semantic representation of the cutting surface state of the aluminum alloy profile to determine whether there is a burn, and automatically generating corresponding cutting optimization control instructions based on the recognition results, such as increasing the flow rate of coolant, so as to prevent and reduce the occurrence of surface defects. In this way, the cooling system parameters during cutting can be dynamically optimized and controlled according to the actual cutting surface state of the aluminum alloy profile, thereby improving the intelligence of the aluminum alloy profile cutting control process and further improving the reliability and stability of the cutting process. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings of the embodiments of the present application are briefly introduced below. Obviously, the drawings described below only relate to some embodiments of the present application, and are not intended to limit the present application.

[0018] Figure 1 This is a schematic flow chart of the intelligent cutting control method for aluminum alloy profiles according to an embodiment of the present application.

[0019] Figure 2 This is a data flow diagram of the intelligent cutting control method for aluminum alloy profiles according to an embodiment of the present application.

[0020] Figure 3 This is a schematic flow chart of step S3 in the intelligent cutting control method for aluminum alloy profiles according to an embodiment of the present application.

[0021] Figure 4 This is a schematic flowchart of step S32 in the intelligent cutting control method for aluminum alloy profiles according to an embodiment of the present application.

[0022] Figure 5 This is a schematic flow chart of step S4 in the intelligent cutting control method for aluminum alloy profiles according to an embodiment of the present application.

[0023] Figure 6 This is a schematic block diagram of an intelligent cutting control system for aluminum alloy profiles according to an embodiment of the present application. DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work also fall within the scope of protection of the present application.

[0025] With the development of computer vision technology and machine learning algorithms, intelligent control methods have begun to be applied in the field of metal processing to achieve more accurate and automated cutting process control. Based on this, in the technical solution of this application, an intelligent cutting control method for aluminum alloy profiles is proposed, which can monitor and analyze the cutting process in real time through a camera to determine whether there is a burn, and dynamically adjust the cutting parameters according to the actual situation, thereby improving the reliability and stability of the aluminum alloy profile cutting process. In addition, this method reduces the reliance on manual intervention and improves the automation level of production control, thereby avoiding the defects brought about by traditional control methods.

[0026] Specifically, the technical concept of the present application is to collect the state image of the cutting surface of the aluminum alloy profile, and introduce an image processing and analysis algorithm based on artificial intelligence at the back end to analyze the state image of the cutting surface of the aluminum alloy profile, so as to capture the semantic features of the state of the cutting surface of the aluminum alloy profile shown in the image, and then identify the state of the cutting surface based on the semantic representation of the state of the cutting surface of the aluminum alloy profile to determine whether there is a burn, and automatically generate corresponding cutting optimization control instructions based on the recognition results, such as increasing the flow rate of coolant, so as to prevent and reduce the occurrence of surface defects. In this way, the cooling system parameters during cutting can be dynamically optimized and controlled according to the actual state of the cutting surface of the aluminum alloy profile, which improves the intelligence of the aluminum alloy profile cutting control process and further improves the reliability and stability of the cutting process.

[0027] Based on this, Figure 1 and Figure 2 As shown, the intelligent cutting control method of the aluminum alloy profile comprises: S1, acquiring a cutting surface state image of the aluminum alloy profile collected by a camera; S2, performing feature extraction on the cutting surface state image of the aluminum alloy profile to obtain a cutting surface state feature of the aluminum alloy profile; S3, performing explicit feature enhancement processing on the cutting surface state feature of the aluminum alloy profile based on pixel granularity to obtain a semantic enhancement feature of the cutting surface state of the aluminum alloy profile; S4, performing cutting surface state recognition based on the semantic enhancement feature of the cutting surface state of the aluminum alloy profile to determine whether there is a burn, and if the recognition result is that there is a burn, generating a cutting optimization control instruction for indicating a cutting optimization control instruction for increasing the coolant flow rate.

[0028] Exemplarily, in step S1, a state image of the cutting surface of the aluminum alloy profile captured by a camera is obtained. It should be understood that the purpose of obtaining the state image of the cutting surface of the aluminum alloy profile captured by the camera is to achieve intelligent monitoring and real-time feedback control of the cutting process. Specifically, this is to be able to dynamically analyze the state of the surface of the aluminum alloy profile during the cutting process, identify whether there are defects such as burns, and adjust the cutting parameters (such as coolant flow) in real time based on this information to prevent and reduce the occurrence of surface defects and ensure product quality. Traditional methods rely on fixed parameters or manual experience to adjust cutting conditions. This method is difficult to adapt to changing working environments and material properties, and is prone to product defects or waste of resources.

[0029] In one embodiment, obtaining the state image of the cutting surface of the aluminum alloy profile collected by the camera includes: installing one or more high-definition cameras at appropriate positions on the cutting equipment to ensure that the camera can clearly capture the key areas on the cutting surface. The selection of the camera should take into account the working environment (such as whether there is high temperature, dust, etc.) and the required resolution and frame rate. In order to ensure the accuracy of the image data, the camera can be calibrated, including optical distortion correction, focal length adjustment, etc., so that the collected image can truly reflect the actual cutting surface. The camera is connected to the cutting control system through a programming interface or other means to ensure time synchronization between the two, that is, the camera can start recording the image at the same time when each cutting operation occurs.

[0030] Exemplarily, in step S2, feature extraction is performed on the aluminum alloy profile cutting surface state image to obtain the aluminum alloy profile cutting surface state feature. It should be understood that the original image data usually has a very high dimension (for example, each pixel is a data point), and directly processing these high-dimensional data is a huge challenge for computing resources and algorithm efficiency. Through feature extraction, the original image can be converted into a set of more compact and representative feature vectors or feature maps, thereby greatly reducing the amount and complexity of subsequent processing. Feature extraction is not only a process of reducing the amount of data, but more importantly, it can capture high-level semantic information related to specific tasks in the image. In this case, feature extraction is performed on the cutting surface state image to focus on important visual patterns that can reflect the cutting quality (such as whether there is a burn), while ignoring insignificant background noise or other interference factors. Different cutting tasks may have different requirements and challenges. Through feature extraction, the system can be more flexible in responding to various changes, because it can extract common but differentiated features from different types of input images, and thus be applicable to a wider range of application scenarios.

[0031] In one embodiment, feature extraction is performed on the aluminum alloy profile cutting surface state image to obtain the aluminum alloy profile cutting surface state feature, including: inputting the aluminum alloy profile cutting surface state image into a cutting surface state feature extractor based on the Mobile-Former model to obtain an aluminum alloy profile cutting surface state feature map as the aluminum alloy profile cutting surface state feature. That is, the aluminum alloy profile cutting surface state image is input into the cutting surface state feature extractor based on the Mobile-Former model for feature mining to extract implicit feature information of the aluminum alloy profile cutting surface state, thereby obtaining an aluminum alloy profile cutting surface state feature map. It should be understood that Mobile-Former is a lightweight deep learning model that reduces computational complexity and memory usage while maintaining high accuracy, making real-time processing possible. This is very important for industrial application scenarios, because real-time is one of the key factors to ensure the smooth progress of the cutting process and reduce defective products. At the same time, Mobile-Former can effectively extract meaningful features from complex cutting surface images, providing a solid foundation for subsequent defect detection. Based on this, the use of the cutting surface state feature extractor based on the Mobile-Former model can provide powerful feature extraction capabilities while ensuring efficient calculation to capture the cutting surface state feature information of aluminum alloy profiles in different cutting scenarios.

[0032] Exemplarily, in step S3, the state feature of the aluminum alloy profile cutting surface is enhanced by an explicit feature based on pixel granularity to obtain the semantic enhancement feature of the state of the aluminum alloy profile cutting surface. It should be understood that, considering that the aluminum alloy profile cutting surface state feature map contains the state representation information of the aluminum alloy profile cutting surface, however, when cutting the aluminum alloy profile, the initial burns are relatively slight, and even very subtle changes or defects, such as slight burn marks, may have a significant impact on the quality of the final product, so it is necessary to timely discover and perform corresponding control optimization adjustments. In addition, it is also considered that the defects generated during the cutting process are often not isolated, but have a close relationship with the surrounding local semantics. Therefore, in order to focus on the fine-grained semantic feature information of the aluminum alloy profile cutting surface state related to the subsequent burn state recognition task, in the technical solution of the present application, the state feature of the aluminum alloy profile cutting surface is further enhanced by an explicit feature based on pixel granularity to obtain the semantic enhancement feature of the aluminum alloy profile cutting surface state. In this way, the model's sensitivity to these tiny burn details can be improved, which helps to more comprehensively and accurately distinguish normal cutting marks from defects (such as burns), and provide support for subsequent cutting optimization control. In addition, the process of enhancing the explicit features based on pixel granularity can decouple unnecessary connections between pixels, so that the features of each pixel granularity can be evaluated separately, thereby better capturing the local feature information of subtle burn defects that are easily overlooked but are critical to product quality.

[0033] In one embodiment, Figure 3 As shown, in step S3, the aluminum alloy profile cutting surface state feature is subjected to explicit feature enhancement processing based on pixel granularity to obtain the aluminum alloy profile cutting surface state semantic enhancement feature, including: S31, pixel granularity feature decoupling of the aluminum alloy profile cutting surface state feature to obtain a set of aluminum alloy profile cutting surface state pixel granularity decoupled features; S32, calculating the significant feature weight of each aluminum alloy profile cutting surface state pixel granularity decoupled feature in the set of aluminum alloy profile cutting surface state pixel granularity decoupled features, and based on the significant feature weights of each aluminum alloy profile cutting surface state pixel granularity decoupled feature, pixel granularity weighted enhancement of the aluminum alloy profile cutting surface state feature is performed to obtain the aluminum alloy profile cutting surface state semantic enhancement feature.

[0034] In one embodiment, in step S31, the pixel granularity feature decoupling of the aluminum alloy profile cutting surface state feature is performed to obtain a set of aluminum alloy profile cutting surface state pixel granularity decoupling features, including: performing pixel granularity feature decoupling of the aluminum alloy profile cutting surface state feature map to obtain a set of aluminum alloy profile cutting surface state pixel granularity decoupling feature vectors as the set of aluminum alloy profile cutting surface state pixel granularity decoupling features. Specifically, the process can be expressed by a formula as follows:

[0035] decouple(F i ) = {v1, v2, ..., v i , ..., v n}

[0036] Among them, F i is the characteristic diagram of the cutting surface state of the aluminum alloy profile, decouple(F i ) is for F i Decouple pixel-level features, v1, v2, v i and v n They are respectively the 1st, 2nd, i-th and n-th aluminum alloy profile cutting surface state pixel granularity decoupling feature vectors in the set of aluminum alloy profile cutting surface state pixel granularity decoupling feature vectors.

[0037] It should be understood that by analyzing each pixel-level feature independently, the fine-grained semantic feature information related to the burn recognition task can be separated, and irrelevant or redundant information, such as background noise or changes in non-critical areas, can be excluded. This processing method enables more acute perception of local features that are easily overlooked but critical to product quality, such as initial slight burn marks, which are not easy to detect at the macro level but show unique texture or brightness differences at the micro level. Through pixel-level feature decoupling, each pixel-level feature can be evaluated separately, which not only helps to better capture the semantic features closely related to the recognition of the burn state, but also enhances the intrinsic connection between the features, making the final generated feature map more semantically meaningful. Such feature representation is very important for distinguishing normal cutting marks from defects (such as burns), and provides support for subsequent cutting optimization control. In addition, this method allows the focus to be dynamically adjusted according to the actual material properties, ensuring that the state of the cutting surface can be accurately judged even on complex or irregularly shaped profiles, while flexibly responding to multiple variables that may exist in the production environment, such as changes in temperature, humidity and lighting conditions. The decoupled feature vector sets usually have better distribution characteristics, which means that they can provide better quality input data for the deep learning model in training, helping to speed up the model's learning process and improve the final performance.

[0038] In one embodiment, Figure 4 As shown, in step S32, the significant feature weights of the pixel granularity decoupling features of the cutting surface state of each aluminum alloy profile in the set of the pixel granularity decoupling features of the cutting surface state of the aluminum alloy profile are calculated, and the pixel granularity weighted enhancement of the cutting surface state features of the aluminum alloy profile is performed based on the significant feature weights of the pixel granularity decoupling features of the cutting surface state of each aluminum alloy profile to obtain the semantic enhancement features of the cutting surface state of the aluminum alloy profile, including: S321, calculating the significant features of the set of pixel granularity decoupling feature vectors of the cutting surface state of the aluminum alloy profile to obtain the structural semantic enhancement features of the cutting surface state of the aluminum alloy profile S322, performing masking processing on each eigenvalue in the aluminum alloy profile cutting surface state structure semantic significance distribution matrix based on feature explicit layering to obtain an aluminum alloy profile cutting surface state structure semantic significance distribution weight matrix composed of multiple aluminum alloy profile cutting surface state structure semantic weights; S323, performing weighted optimization on the aluminum alloy profile cutting surface state feature map based on the aluminum alloy profile cutting surface state structure semantic significance distribution weight matrix to obtain an aluminum alloy profile cutting surface state semantic enhancement feature map as the aluminum alloy profile cutting surface state semantic enhancement feature.

[0039] In one embodiment, in step S321, the significant features of the set of the aluminum alloy profile cutting surface state pixel granularity decoupling feature vectors are calculated to obtain the aluminum alloy profile cutting surface state structural semantic significance distribution matrix, including: inputting each aluminum alloy profile cutting surface state pixel granularity decoupling feature vector in the set of the aluminum alloy profile cutting surface state pixel granularity decoupling feature vectors into a hyperbolic space mapper to obtain a set of aluminum alloy profile cutting surface state hyperbolic space modulation pixel granularity decoupling feature vectors; calculating the fine-grained structural semantic significance measurement coefficient of each aluminum alloy profile cutting surface state hyperbolic space modulation pixel granularity decoupling feature vector in the set of the aluminum alloy profile cutting surface state hyperbolic space modulation pixel granularity decoupling feature vectors to obtain multiple aluminum alloy profile cutting surface state fine-grained structural semantic significance measurement coefficients; arranging the multiple aluminum alloy profile cutting surface state fine-grained structural semantic significance measurement coefficients into a matrix to obtain the aluminum alloy profile cutting surface state structural semantic significance distribution matrix. Specifically, the process can be expressed by the formula:

[0040] h i =W1v i W2

[0041]

[0042] Wherein, v is the i-th aluminum alloy profile cutting surface state pixel granularity decoupling feature vector in the set of aluminum alloy profile cutting surface state pixel granularity decoupling feature vectors, W1 and W2 are the first weight matrix and the second weight matrix respectively, and hj v i The corresponding hyperbolic spatial modulation pixel granularity decoupling feature vector of the aluminum alloy profile cutting surface state, ||·|| 2 To calculate the square of the vector norm, s i is the semantic saliency distribution matrix of the cutting surface state structure of aluminum alloy profile h i The corresponding semantic saliency coefficient of the fine-grained structure of the cutting surface state of the aluminum alloy profile.

[0043] It should be understood that the hyperbolic space has a natural hierarchical structure and exponential expansion characteristics, which is very suitable for representing the similarity and distance between data points with complex relationships. For the analysis of the cutting surface state, this mapping can enhance the fine-grained structural semantic information that may not be easily perceived in the original image space but is crucial for burn recognition. When the individual decoupled feature vectors are mapped to the hyperbolic space, they are transformed into the so-called "hyperbolic space modulated pixel granularity decoupled feature vectors". In this new representation, the subtle texture, brightness changes and other features that originally existed on the cutting surface are amplified or emphasized, especially those local abnormalities related to burns. Next, the fine-grained structural semantic saliency metric coefficients of these hyperbolic space feature vectors are calculated. This metric coefficient reflects the importance of the feature at each pixel position, that is, whether the feature at this position is more likely to be part of the burn mark, and its relative importance in the entire image. By quantifying all the correlation coefficients, a series of numerical values ​​can be obtained, which together constitute a distribution matrix describing the semantic saliency of the entire cutting surface state structure. This saliency distribution matrix is ​​essentially a two-dimensional array, in which each element corresponds to a pixel in the original image and carries information about whether there are potential signs of burns at that point and its surrounding area. A high value means that more obvious burn features are found at this location, while a low value indicates that this may be a normal cut mark or other less critical surface details. Such a matrix not only condenses a large amount of information extracted from the image, but also presents it in a structured way, which is convenient for direct use in subsequent processing steps.

[0044] In one embodiment, in step S322, each eigenvalue in the aluminum alloy profile cutting surface state structure semantic significance distribution matrix is ​​subjected to masking processing based on feature explicit hierarchical processing to obtain an aluminum alloy profile cutting surface state structure semantic significance distribution weight matrix composed of multiple aluminum alloy profile cutting surface state structure semantic weights, including: inputting the aluminum alloy profile cutting surface state structure semantic significance distribution matrix into a feature explicit hierarchical precipitation module based on a multi-layer masking function to obtain the aluminum alloy profile cutting surface state structure semantic significance distribution weight matrix. Specifically, the process can be expressed as follows:

[0045]

[0046] Among them, s i is the semantic saliency distribution matrix of the cutting surface state structure of aluminum alloy profile h i The corresponding semantic saliency coefficient of the fine-grained structure of the cutting surface state of the aluminum alloy profile, θ is the preset threshold, mask is the masking operation, s i ' is the weight matrix of semantic saliency distribution of the cutting surface state structure of aluminum alloy profile h i The corresponding semantic significance distribution weight value of the aluminum alloy profile cutting surface state structure.

[0047] It should be understood that through the feature explicit hierarchical precipitation module based on the multi-layer mask function, the features at different levels can be selectively amplified or suppressed, so as to more accurately capture the fine-grained semantic features most relevant to the burn recognition task and reduce the influence of noise and other irrelevant factors. The generated weight matrix of the semantic saliency distribution of the cutting surface state structure of the aluminum alloy profile not only retains the spatial layout of the original saliency distribution matrix, but also adds a deeper understanding and explanation on this basis. Each element now not only represents the saliency of the feature at that location, but also includes an assessment of its relative importance in the entire cutting surface state. A high weight value means that there is a high probability that there are obvious burn marks or other defects that need attention at that location; conversely, a low weight value indicates that this may be a normal cutting surface or unimportant background information. Such a weight matrix provides guidance for subsequent weighted optimization, so that more valuable features can be given more attention in the decision-making process, while ignoring less relevant information.

[0048] In one embodiment, in step S323, the aluminum alloy profile cutting surface state feature map is weighted optimized based on the aluminum alloy profile cutting surface state structure semantic saliency distribution weight matrix to obtain the aluminum alloy profile cutting surface state semantic enhancement feature map as the aluminum alloy profile cutting surface state semantic enhancement feature, including: calculating the aluminum alloy profile cutting surface state structure semantic saliency distribution weight matrix and the aluminum alloy profile cutting surface state feature map along the channel dimension The feature matrix is ​​multiplied by the position point to obtain the aluminum alloy profile cutting surface state semantic enhancement feature map. Specifically, the process can be expressed as:

[0049]

[0050] Among them, F i is the aluminum alloy profile cutting surface state feature map, S is the aluminum alloy profile cutting surface state structure semantic significance distribution weight matrix, In order to use the matrix to perform weighted multiplication on each feature matrix along the channel dimension of the feature map, F fIt is the semantically enhanced feature map of the cutting surface state of the aluminum alloy profile.

[0051] The positional point multiplication between the semantic saliency distribution weight matrix of the aluminum alloy profile cutting surface state structure and each feature matrix along the channel dimension in the cutting surface state feature map is calculated. This process aims to reapply the feature importance information (i.e., weight) obtained after processing by the multi-layer mask function back to the original feature map. Doing so can further strengthen those areas that are critical to burn recognition, while suppressing or ignoring less relevant parts, thereby generating a semantically enhanced feature map that is more focused on key features.

[0052] It should be understood that when these weight values ​​are dot-multiplied with the corresponding positions in the original feature map, the importance of the original features is actually adjusted: for those areas that are considered very important, their features will be amplified to make them more prominent in subsequent analysis; while for less important areas, the influence of the features is weakened, reducing their interference with the final result. The semantically enhanced feature map generated in this way not only retains the spatial structure and texture information of the original image, but also incorporates the feature importance evaluation after careful screening. This enhanced feature map can more accurately capture the subtle burn marks and other potential defects on the cutting surface, greatly improving the system's sensitivity to these subtle changes. For example, in the actual cutting process, even extremely slight overheating may cause the surface quality to deteriorate, but traditional detection methods may find it difficult to detect such early signs. However, with the help of semantically enhanced feature maps, intelligent control systems can quickly identify these risks and take timely measures (such as increasing the coolant flow) to prevent the problem from worsening and ensure product quality.

[0053] In summary, the process of explicit feature enhancement based on pixel granularity is to enhance the features by using explicit modeling and hierarchical mask modulation of the fine-grained features of the cutting surface state pixels of each aluminum alloy profile in the cutting surface state feature map of the aluminum alloy profile, which effectively improves the quality of the cutting surface state feature map of the aluminum alloy profile. Specifically, after the pixel granularity feature decoupling operation, the explicit feature enhancement based on pixel granularity is further performed with hyperbolic space mapping and the fine-grained structure semantic saliency coefficient is calculated, which can quantify the importance of each pixel-level fine-grained feature in the cutting surface state feature map of the aluminum alloy profile, and provide a scientific basis for subsequent feature selection. This helps to more accurately judge the state of the cutting surface, especially on profiles with complex or irregular shapes. Then, a multi-layer mask function is used to achieve selective amplification of features at different levels while suppressing noise or irrelevant features. This process is similar to the attention mechanism, which allows the dynamic adjustment of the degree of attention to the semantic features of different pixel granularities in the cutting surface state image of the aluminum alloy profile. This semantic level understanding is very important for distinguishing normal cutting marks from burn defects, and helps to improve the quality of the final feature representation. Finally, element-by-element multiplication is used to generate the final semantic enhancement feature map of the cutting surface state of the aluminum alloy profile as the semantic enhancement feature of the cutting surface state of the aluminum alloy profile, ensuring that the output semantic enhancement feature can highlight the most representative and distinctive pixel-level granularity semantic feature information of the cutting surface state of the aluminum alloy profile. These semantic features can focus on the subtle burn defect characteristics of the cutting surface state of the aluminum alloy profile and the feature parts that are crucial to the classification or recognition tasks, so that the model has higher sensitivity, accuracy and stronger generalization ability when detecting and evaluating the cutting surface state of the aluminum alloy profile, providing a solid foundation for subsequent cutting surface state recognition and burn defect detection.

[0054] In one embodiment, Figure 5As shown, in step S4, based on the semantic enhancement feature of the cutting surface state of the aluminum alloy profile, the cutting surface state is identified to determine whether there is a burn. If the identification result is that there is a burn, a cutting optimization control instruction is generated to indicate the cutting optimization control instruction of increasing the coolant flow rate, including: S41, inputting the semantic enhancement feature map of the cutting surface state of the aluminum alloy profile into the cutting surface state identifier based on the classifier to obtain a recognition result, and the recognition result is used to indicate whether there is a burn; S42, in response to the recognition result that there is a burn, generating the cutting optimization control instruction, and the cutting optimization control instruction is used to indicate the increase of the coolant flow rate. That is to say, the classification processing is carried out by using the enhanced semantics of the cutting surface state of the aluminum alloy profile to identify the cutting surface state to determine whether there is a burn. Furthermore, in response to the recognition result that there is a burn, a cutting optimization control instruction is generated to prevent and reduce the occurrence of surface defects, wherein the cutting optimization control instruction is used to indicate the increase of the coolant flow rate. In this way, the cooling system parameters during cutting can be dynamically optimized and controlled according to the actual cutting surface state of the aluminum alloy profile, which improves the intelligence of the aluminum alloy profile cutting control process and further improves the reliability and stability of the cutting process.

[0055] In a specific embodiment, the classifier-based cutting surface state identifier uses an SVM classifier, which is trained on a large number of labeled aluminum alloy profile cutting surface image data sets, one part of which represents a normal cutting surface, while the other part contains cases of burns of varying degrees. Through such a training process, the SVM learns how to distinguish between normal cutting and the presence of burns based on input features. When a new semantically enhanced feature map enters the SVM, the classifier maps it to a high-dimensional space, in which it searches for the optimal hyperplane to maximize the interval between different categories. For binary classification problems such as "burn" and "no burn" in this example, the SVM attempts to find a decision boundary that separates the two types of samples as much as possible. Once this boundary is determined, for a new test sample (i.e., the current semantically enhanced feature map), the SVM can determine which category it belongs to based on its relative position relative to the decision boundary. When the intelligent control system uses the SVM classifier to identify that there are signs of slight overheating on the cutting surface of the aluminum alloy profile (i.e., there may be a risk of burns), the system will immediately respond to this recognition result and automatically generate a clear cutting optimization control instruction. Based on the recognition results of the SVM, the intelligent control system immediately generates a control instruction containing specific parameters and sends the instruction to the PLC (programmable logic controller) responsible for regulating the cooling system through an industrial communication protocol (such as Modbus, Profibus or EtherCAT). For example, this instruction requires the coolant pumping rate to increase from the current 20 liters / minute to 30 liters / minute and activate the spare cooling nozzle group to ensure that more coolant is evenly distributed to cover the entire cutting area. At the same time, the system instructs to maintain this enhanced cooling setting for at least 10 seconds to ensure that the temperature is effectively controlled before evaluating whether to restore normal flow. After receiving the instruction, the PLC parses the specific parameters and prepares to perform the corresponding action. It sends a signal to the coolant pump, instructing it to increase the operating frequency or pressure so that the coolant flow rate is increased to 30 liters / minute. At the same time, the PLC triggers the solenoid valves of the spare cooling nozzles and opens these nozzles to provide additional cooling support. During this period, the system continues to monitor the status of the cutting surface in real time and collects data from sensors, such as temperature sensor readings, to verify the cooling effect. If the temperature drops successfully and there are no further signs of overheating, the system gradually returns to normal coolant flow after 10 seconds; if the problem persists, the system will continue to maintain or further increase cooling measures and may notify the operator for manual intervention.

[0056] Preferably, inputting the aluminum alloy profile cutting surface state semantic enhancement feature map into a cutting surface state identifier based on a classifier to obtain a recognition result comprises:

[0057] The absolute value sum and the square root of the square sum of all eigenvalues ​​of the semantic enhancement feature graph of the aluminum alloy profile cutting surface state are calculated to obtain the first aluminum alloy profile cutting surface state semantic enhancement spatial structure value and the second aluminum alloy profile cutting surface state semantic enhancement spatial structure value, that is:

[0058] w1=∑ i |f i |

[0059]

[0060] Among them, f i represents the i-th eigenvalue of the semantically enhanced feature graph of the cutting surface state of the aluminum alloy profile, w1 represents the semantically enhanced spatial structure value of the cutting surface state of the first aluminum alloy profile, and w2 represents the semantically enhanced spatial structure value of the cutting surface state of the second aluminum alloy profile;

[0061] Determine the total number n of eigenvalues ​​of all eigenvalues ​​of the set of semantically enhanced feature graphs of the cutting surface state of the aluminum alloy profile;

[0062] For each eigenvalue of the semantic enhancement feature graph of the cutting surface state of the aluminum alloy profile, the first aluminum alloy profile cutting surface state semantic enhancement spatial structure value minus the product of the eigenvalue and the total number of eigenvalues ​​is calculated to obtain the first aluminum alloy profile cutting surface state semantic enhancement long-range dependency value x i =w1-f i ×n, where w1 represents the semantically enhanced spatial structure value of the cutting surface state of the first aluminum alloy profile, f j represents the i-th eigenvalue of the semantic enhancement feature graph of the aluminum alloy profile cutting surface state, n represents the number of eigenvalues ​​of all eigenvalues ​​of the semantic enhancement feature graph of the aluminum alloy profile cutting surface state, x i The semantically enhanced long-range dependency value of the cutting surface state of the first aluminum alloy profile is represented; and the semantically enhanced long-range dependency value of the cutting surface state of the second aluminum alloy profile is obtained by multiplying the square root of the total number of eigenvalues ​​by the product of the eigenvalues ​​minus the semantically enhanced spatial structure value of the cutting surface state of the second aluminum alloy profile. Among them, w2 represents the semantic enhancement spatial structure value of the cutting surface state of the second aluminum alloy profile, f i represents the i-th eigenvalue of the semantic enhancement feature graph of the aluminum alloy profile cutting surface state, n represents the number of eigenvalues ​​of all eigenvalues ​​of the semantic enhancement feature graph of the aluminum alloy profile cutting surface state, y i Indicates the semantically enhanced long-range dependency value of the cutting surface state of the second aluminum alloy profile;

[0063] The index value calculated by taking the semantically enhanced long-range dependency value of the cutting surface state of the first aluminum alloy profile as the exponent of the natural constant and the inverse of the semantically enhanced long-range dependency value of the cutting surface state of the second aluminum alloy profile are weighted summed to obtain the optimized eigenvalue. Among them, x i represents the semantically enhanced long-range dependency value of the cutting surface state of the first aluminum alloy profile, y i represents the semantically enhanced long-range dependency value of the cutting surface state of the second aluminum alloy profile, e represents a natural constant, α and β represent weighted hyperparameters, and f i represents the i-th optimized eigenvalue;

[0064] The optimized aluminum alloy profile cutting surface state semantic enhancement feature map composed of the optimized feature values ​​is input into a cutting surface state identifier based on a classifier to obtain a recognition result.

[0065] Here, in the case where the aluminum alloy profile cutting surface state feature map represents the image semantic features of the aluminum alloy profile cutting surface, after performing feature enhancement based on feature explicit hierarchical precipitation, the aluminum alloy profile cutting surface state semantic enhancement feature map will also have significant feature explicit spatial structure differences due to the differences in image semantic feature hierarchical precipitation, affecting the convergence consistency of the classifier and thus affecting the consistency of the recognition result.

[0066] Based on this, the spatial structure consistency loss of the feature set of the semantic enhancement feature graph of the cutting surface state of the aluminum alloy profile in the high-dimensional space may cause the weight matrix of the classifier to implicitly infer the spatial structure information based on the feature, resulting in inconsistent convergence. By establishing a long-distance feature dependency relationship based on the overall feature scale of the semantic enhancement feature graph of the cutting surface state of the aluminum alloy profile relative to the spatial structure representation of the semantic enhancement feature graph of the cutting surface state of the aluminum alloy profile, the local connectivity of the features of the semantic enhancement feature graph of the cutting surface state of the aluminum alloy profile is established, and the spatial ambiguity information of the feature value of the object is captured by predicting the unstructured feature value points of the semantic enhancement feature graph of the cutting surface state of the aluminum alloy profile, thereby improving the spatial inductive deviation perception ability of the feature set of the semantic enhancement feature graph of the cutting surface state of the aluminum alloy profile, improving the convergence consistency of the classifier, and improving the consistency of the recognition results obtained by the cutting surface state identifier based on the classifier. In this way, the cooling system parameters during cutting can be dynamically optimized and controlled according to the actual cutting surface state of the aluminum alloy profile, which improves the intelligence of the aluminum alloy profile cutting control process and further improves the reliability and stability of the cutting process.

[0067] In summary, the intelligent cutting control method of the aluminum alloy profile according to the embodiment of the present application is explained, which collects the state image of the cutting surface of the aluminum alloy profile, and introduces an image processing and analysis algorithm based on artificial intelligence at the back end to analyze the state image of the cutting surface of the aluminum alloy profile, so as to capture the semantic features of the cutting surface state of the aluminum alloy profile shown in the image, and then perform cutting surface state recognition based on the semantic representation of the cutting surface state of the aluminum alloy profile to determine whether there is a burn, and automatically generate corresponding cutting optimization control instructions based on the recognition results, such as increasing the flow rate of coolant, so as to prevent and reduce the occurrence of surface defects. In this way, the cooling system parameters during cutting can be dynamically optimized and controlled according to the actual cutting surface state of the aluminum alloy profile, which improves the intelligence of the aluminum alloy profile cutting control process and further improves the reliability and stability of the cutting process.

[0068] Figure 6 Schematic block diagram of the intelligent cutting control system for aluminum alloy profiles according to the embodiment of the present application. Figure 6 As shown, the intelligent cutting control system 1 of the aluminum alloy profile comprises: an aluminum alloy profile cutting image acquisition module 10, which is used to acquire the cutting surface state image of the aluminum alloy profile collected by the camera; an aluminum alloy profile cutting surface state feature extraction module 20, which is used to extract features from the aluminum alloy profile cutting surface state image to obtain the cutting surface state features of the aluminum alloy profile; an aluminum alloy profile cutting surface state feature enhancement processing module 30, which is used to perform explicit feature enhancement processing on the cutting surface state features of the aluminum alloy profile based on pixel granularity to obtain the cutting surface state semantic enhancement features of the aluminum alloy profile; a cutting surface state recognition module 40, which is used to perform cutting surface state recognition based on the aluminum alloy profile cutting surface state semantic enhancement features to determine whether there is a burn, and if the recognition result is that there is a burn, generate a cutting optimization control instruction for indicating a cutting optimization control instruction for increasing the coolant flow rate.

[0069] In one embodiment, the aluminum alloy profile cutting surface state feature extraction module is used to: input the aluminum alloy profile cutting surface state image into a cutting surface state feature extractor based on the Mobile-Former model to obtain an aluminum alloy profile cutting surface state feature map as the aluminum alloy profile cutting surface state feature.

[0070] In one embodiment, the aluminum alloy profile cutting surface state feature enhancement processing module includes: a pixel granularity feature decoupling processing unit, which is used to perform pixel granularity feature decoupling on the aluminum alloy profile cutting surface state feature to obtain a set of aluminum alloy profile cutting surface state pixel granularity decoupling features; a weighted enhancement processing unit, which is used to calculate the significance feature weight of each aluminum alloy profile cutting surface state pixel granularity decoupling feature in the set of aluminum alloy profile cutting surface state pixel granularity decoupling features, and perform pixel granularity weighted enhancement on the aluminum alloy profile cutting surface state feature based on the significance feature weight of each aluminum alloy profile cutting surface state pixel granularity decoupling feature to obtain an aluminum alloy profile cutting surface state semantic enhancement feature.

[0071] Here, those skilled in the art will appreciate that the specific operations of the various modules and units in the above-mentioned intelligent cutting control system for aluminum alloy profiles have been described in detail above. Figures 1 to 5 The description of the intelligent cutting control method of aluminum alloy profiles has been introduced in detail, and therefore, its repeated description will be omitted.

[0072] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0073] It should be understood that the specific examples in this article are only intended to help those skilled in the art better understand the embodiments of the present application, rather than to limit the scope of the embodiments of the present application.

[0074] It should also be understood that in the various embodiments of the present application, the size of the serial number of each process does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0075] It should also be understood that the various implementation methods described in this specification can be implemented individually or in combination, and the embodiments of the present application are not limited to this.

[0076] Unless otherwise stated, all technical and scientific terms used in the embodiments of the present application are the same as the meanings generally understood by those skilled in the art of the technical field of the present application. The terms used in the present application are only for the purpose of describing specific embodiments and are not intended to limit the scope of the present application. The term "and / or" used in the present application includes any and all combinations of one or more related listed items. The singular forms of "a kind of", "above" and "the" used in the embodiments of the present application and the appended claims are also intended to include majority forms, unless the context clearly indicates other meanings. In addition, the terms "first", "second" etc. are only used for description purposes and cannot be understood as indicating or suggesting relative importance.

[0077] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0078] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0079] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0080] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0081] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. An intelligent cutting control method for aluminum alloy profiles, characterized in that: include: Acquire the cutting surface state image of the aluminum alloy profile captured by the camera; Extracting features from the aluminum alloy profile cutting surface state image to obtain aluminum alloy profile cutting surface state features; The aluminum alloy profile cutting surface state feature is subjected to pixel granularity-based explicit feature enhancement processing to obtain the aluminum alloy profile cutting surface state semantic enhancement feature, including: performing pixel granularity feature decoupling on the aluminum alloy profile cutting surface state feature to obtain a set of aluminum alloy profile cutting surface state pixel granularity decoupling features; calculating the significant feature weight of each aluminum alloy profile cutting surface state pixel granularity decoupling feature in the set of aluminum alloy profile cutting surface state pixel granularity decoupling features, and performing pixel granularity weighted enhancement on the aluminum alloy profile cutting surface state feature based on the significant feature weight of each aluminum alloy profile cutting surface state pixel granularity decoupling feature to obtain the aluminum alloy profile cutting surface state semantic enhancement feature; Based on the semantic enhancement features of the cutting surface state of the aluminum alloy profile, cutting surface state recognition is performed to determine whether there is a burn. If the recognition result is that there is a burn, a cutting optimization control instruction is generated to indicate a cutting optimization control instruction for increasing the coolant flow rate.

2. The intelligent cutting control method for aluminum alloy profiles according to claim 1, characterized in that: The cutting surface state image of the aluminum alloy profile is subjected to feature extraction to obtain the cutting surface state feature of the aluminum alloy profile, comprising: inputting the cutting surface state image of the aluminum alloy profile into a cutting surface state feature extractor based on a Mobile-Former model to obtain a cutting surface state feature map of the aluminum alloy profile as the cutting surface state feature of the aluminum alloy profile.

3. The intelligent cutting control method for aluminum alloy profiles according to claim 2 is characterized in that: The pixel granularity feature decoupling of the aluminum alloy profile cutting surface state feature is performed to obtain a set of aluminum alloy profile cutting surface state pixel granularity decoupling features, including: performing pixel granularity feature decoupling of the aluminum alloy profile cutting surface state feature map to obtain a set of aluminum alloy profile cutting surface state pixel granularity decoupling feature vectors as the set of aluminum alloy profile cutting surface state pixel granularity decoupling features.

4. The intelligent cutting control method for aluminum alloy profiles according to claim 3 is characterized in that: The significant feature weight of each of the aluminum alloy profile cutting surface state pixel granularity decoupling features in the set of the aluminum alloy profile cutting surface state pixel granularity decoupling features is calculated, and the aluminum alloy profile cutting surface state features are weightedly enhanced at the pixel granularity based on the significant feature weight of each of the aluminum alloy profile cutting surface state pixel granularity decoupling features to obtain the aluminum alloy profile cutting surface state semantic enhancement feature, including: Calculating the significant features of the set of pixel granularity decoupling feature vectors of the aluminum alloy profile cutting surface state to obtain the aluminum alloy profile cutting surface state structure semantic significance distribution matrix; Performing a masking process based on feature explicit layering on each eigenvalue in the aluminum alloy profile cutting surface state structure semantic significance distribution matrix to obtain an aluminum alloy profile cutting surface state structure semantic significance distribution weight matrix composed of a plurality of aluminum alloy profile cutting surface state structure semantic weights; Based on the semantic significance distribution weight matrix of the aluminum alloy profile cutting surface state structure, the aluminum alloy profile cutting surface state feature map is weighted optimized to obtain the aluminum alloy profile cutting surface state semantic enhancement feature map as the aluminum alloy profile cutting surface state semantic enhancement feature.

5. The intelligent cutting control method for aluminum alloy profiles according to claim 4, characterized in that: Calculating the significant features of the set of pixel granularity decoupling feature vectors of the aluminum alloy profile cutting surface state to obtain the aluminum alloy profile cutting surface state structure semantic significance distribution matrix, including: Inputting each of the aluminum alloy profile cutting surface state pixel granularity decoupling feature vectors in the set of aluminum alloy profile cutting surface state pixel granularity decoupling feature vectors into a hyperbolic space mapper to obtain a set of aluminum alloy profile cutting surface state hyperbolic space modulation pixel granularity decoupling feature vectors; Calculating the fine-grained structural semantic saliency metric coefficient of each aluminum alloy profile cutting surface state hyperbolic space modulation pixel granularity decoupling feature vector in the set of aluminum alloy profile cutting surface state hyperbolic space modulation pixel granularity decoupling feature vectors to obtain a plurality of aluminum alloy profile cutting surface state fine-grained structural semantic saliency metric coefficients; The multiple aluminum alloy profile cutting surface state fine-grained structural semantic significance measurement coefficients are arranged into a matrix to obtain the aluminum alloy profile cutting surface state structural semantic significance distribution matrix.

6. The intelligent cutting control method for aluminum alloy profiles according to claim 5, characterized in that: Each eigenvalue in the aluminum alloy profile cutting surface state structure semantic significance distribution matrix is ​​subjected to masking processing based on feature explicit hierarchical processing to obtain an aluminum alloy profile cutting surface state structure semantic significance distribution weight matrix composed of multiple aluminum alloy profile cutting surface state structure semantic weights, including: inputting the aluminum alloy profile cutting surface state structure semantic significance distribution matrix into a feature explicit hierarchical precipitation module based on a multi-layer masking function to obtain the aluminum alloy profile cutting surface state structure semantic significance distribution weight matrix.

7. The intelligent cutting control method for aluminum alloy profiles according to claim 6, characterized in that: Based on the semantic significance distribution weight matrix of the aluminum alloy profile cutting surface state structure, the aluminum alloy profile cutting surface state feature map is weighted optimized to obtain the aluminum alloy profile cutting surface state semantic enhancement feature map as the aluminum alloy profile cutting surface state semantic enhancement feature, including: calculating the position point multiplication between the aluminum alloy profile cutting surface state structure semantic significance distribution weight matrix and each feature matrix along the channel dimension in the aluminum alloy profile cutting surface state feature map to obtain the aluminum alloy profile cutting surface state semantic enhancement feature map.

8. The intelligent cutting control method for aluminum alloy profiles according to claim 7, characterized in that: Based on the semantic enhancement feature of the cutting surface state of the aluminum alloy profile, the cutting surface state is identified to determine whether there is a burn. If the identification result is that there is a burn, a cutting optimization control instruction is generated to indicate a cutting optimization control instruction for increasing the coolant flow rate, including: Inputting the semantic enhancement feature map of the cutting surface state of the aluminum alloy profile into a cutting surface state identifier based on a classifier to obtain a recognition result, wherein the recognition result is used to indicate whether there is a burn; In response to the identification result that burns exist, the cutting optimization control instruction is generated, and the cutting optimization control instruction is used to indicate increasing the flow rate of coolant.

9. An intelligent cutting control system for aluminum alloy profiles, characterized in that: include: Aluminum alloy profile cutting image acquisition module, used to acquire the aluminum alloy profile cutting surface state image collected by the camera; An aluminum alloy profile cutting surface state feature extraction module, used for extracting features from the aluminum alloy profile cutting surface state image to obtain aluminum alloy profile cutting surface state features; An aluminum alloy profile cutting surface state feature enhancement processing module is used to perform pixel-granularity-based explicit feature enhancement processing on the aluminum alloy profile cutting surface state feature to obtain an aluminum alloy profile cutting surface state semantic enhancement feature, wherein the aluminum alloy profile cutting surface state feature enhancement processing module includes: a pixel granularity feature decoupling processing unit, used to perform pixel granularity feature decoupling on the aluminum alloy profile cutting surface state feature to obtain a set of aluminum alloy profile cutting surface state pixel granularity decoupling features; a weighted enhancement processing unit, used to calculate the significant feature weight of each aluminum alloy profile cutting surface state pixel granularity decoupling feature in the set of aluminum alloy profile cutting surface state pixel granularity decoupling features, and perform pixel granularity weighted enhancement on the aluminum alloy profile cutting surface state feature based on the significant feature weight of each aluminum alloy profile cutting surface state pixel granularity decoupling feature to obtain an aluminum alloy profile cutting surface state semantic enhancement feature; A cutting surface state recognition module is used to perform cutting surface state recognition based on the semantic enhancement features of the cutting surface state of the aluminum alloy profile to determine whether there is a burn. If the recognition result is that there is a burn, a cutting optimization control instruction is generated to indicate a cutting optimization control instruction for increasing the coolant flow rate.

10. The intelligent cutting control system for aluminum alloy profiles according to claim 9, characterized in that: The aluminum alloy profile cutting surface state feature extraction module is used to: input the aluminum alloy profile cutting surface state image into the cutting surface state feature extractor based on the Mobile-Former model to obtain the aluminum alloy profile cutting surface state feature map as the aluminum alloy profile cutting surface state feature.

Images