Intelligent control system and method for cable processing

Abstract

The present application provides an intelligent control system and method for cable processing, which relates to the field of cable processing control. It uses artificial intelligence analysis technology based on deep learning, and extracts features of the motor winding speed and the image of the completed cable at multiple predetermined time points within a predetermined time period to understand the relationship between the cable winding quality and the motor winding speed, and then obtains the control direction of the current cable winding speed. In this way, the precise control of the cable winding speed is achieved, which is conducive to improving the quality of cable winding processing.

Description

Technical Field

[0001] The present application relates to the field of cable processing control, and more specifically, to an intelligent control system and method for cable processing. Background Art

[0002] As the era of electrification continues to deepen, the cable processing industry faces increasingly severe challenges. In the cable production process, the winding process of multiple single cables is crucial, which directly affects the quality and performance of the cable. The traditional winding method usually places multiple single cables side by side and winds them with winding equipment. During the cable winding process, when the cable winding speed is too high, the internal stress of the cable increases, which is prone to deformation or damage; while the speed is too low, the winding will be loose, affecting the overall performance and appearance quality of the cable.

[0003] Therefore, an optimized intelligent control solution for cable processing is needed. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present application provides an intelligent control system and method for cable processing.

[0005] An intelligent control system for cable processing, comprising:

[0006] The cable winding processing related data acquisition module is used to obtain the images of the completed wound cables and the winding speed values ​​of the motor at multiple predetermined time points within a predetermined time period;

[0007] A cable winding processing quality feature encoding module, used for performing feature encoding on the images of the completed cable winding at the plurality of predetermined time points to obtain a time-series enhanced feature vector of the cable winding surface;

[0008] A cable winding processing speed time series feature encoding module is used to extract the features of the motor winding speed values ​​at the plurality of predetermined time points to obtain a motor winding speed time series feature vector;

[0009] A cable winding responsiveness estimation module, used to calculate the responsiveness estimation of the cable winding surface timing enhancement feature vector relative to the motor winding speed timing feature vector to obtain a motor winding responsiveness control feature matrix;

[0010] The cable winding speed control result generation module is used to determine the winding speed control direction of the motor at the current time point based on the motor winding responsiveness control characteristic matrix.

[0011] An intelligent control method for cable processing, comprising:

[0012] Acquire images of the completed cable winding and winding speed values ​​of the motor at a plurality of predetermined time points within a predetermined time period;

[0013] Performing feature encoding on the images of the completed cable at the plurality of predetermined time points to obtain a time-series enhanced feature vector of the cable winding surface;

[0014] Performing feature extraction on the winding speed values ​​of the motor at the plurality of predetermined time points to obtain a motor winding speed time series feature vector;

[0015] Calculating the responsiveness estimation of the cable winding surface timing enhancement characteristic vector relative to the motor winding speed timing characteristic vector to obtain the motor winding responsiveness regulation characteristic matrix;

[0016] Based on the motor winding responsiveness control characteristic matrix, the winding speed control direction of the motor at the current time point is determined.

[0017] This application has significant technical effects due to the adoption of the above technical solutions:

[0018] The intelligent control system and method for cable processing provided by the present application adopts artificial intelligence analysis technology based on deep learning, and extracts features of the motor winding speed and the image of the completed cable at multiple predetermined time points within a predetermined time period to understand the relationship between the cable winding quality and the motor winding speed, and then obtains the control direction of the current cable winding speed. In this way, the precise control of the cable winding speed is achieved, which is conducive to improving the quality of cable winding processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0020] Figure 1 It is a system block diagram of an intelligent control system for cable processing according to an embodiment of the present application.

[0021] Figure 2 It is a block diagram of a cable winding processing quality feature encoding module in an intelligent control system for cable processing according to an embodiment of the present application.

[0022] Figure 3 It is a block diagram of a cable winding processing speed timing feature encoding module in an intelligent control system for cable processing according to an embodiment of the present application.

[0023] Figure 4 It is a block diagram of a cable winding speed control result generation module in an intelligent control system for cable processing according to an embodiment of the present application.

[0024] Figure 5 The figure is a flow chart of an intelligent control method for cable processing according to an embodiment of the present application. DETAILED DESCRIPTION

[0025] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0026] As mentioned in the background technology above, currently, in the cable processing process, it is necessary to wind multiple single cables into one cable. The existing method is to place multiple single cables side by side, and the winding equipment winds the cables. In the winding process, the winding speed of the motor will directly affect the tightness of the cable winding. When the speed is too high, the cable will be wound too tightly, resulting in excessive internal stress in the cable, which is prone to deformation or damage. On the contrary, if the speed is too low, the cable will be wound too loosely, affecting the overall performance and appearance quality of the cable. Therefore, an optimized intelligent control solution for cable processing is expected.

[0027] In recent years, deep learning and neural networks have been widely used in computer vision, natural language processing, text signal processing and other fields. In addition, deep learning and neural networks have also shown a level close to or even beyond that of humans in image classification, object detection, semantic segmentation, text translation and other fields. The development of deep learning and neural networks provides new solutions and solutions for the intelligent control of cable processing.

[0028] Figure 1 FIG. 1 is a system block diagram of an intelligent control system for cable processing according to an embodiment of the present application. Figure 1As shown, in the intelligent control system 100 for cable processing, it includes: a cable winding processing related data acquisition module 110, which is used to obtain the cable images of the completed winding and the winding speed values ​​of the motor at multiple predetermined time points within a predetermined time period; a cable winding processing quality feature encoding module 120, which is used to feature encode the cable images of the completed winding at the multiple predetermined time points to obtain the cable winding surface timing enhancement feature vector; a cable winding processing speed timing feature encoding module 130, which is used to feature extract the winding speed values ​​of the motor at the multiple predetermined time points to obtain the motor winding speed timing feature vector; a cable winding responsiveness estimation module 140, which is used to calculate the responsiveness estimation of the cable winding surface timing enhancement feature vector relative to the motor winding speed timing feature vector to obtain the motor winding responsiveness control feature matrix; a cable winding speed control result generation module 150, which is used to determine the winding speed control direction of the motor at the current time point based on the motor winding responsiveness control feature matrix.

[0029] Specifically, in the technical solution of the present application, first, the images of the completed wound cables and the winding speed values ​​of the motors at multiple predetermined time points within a predetermined time period are obtained. It should be well understood that the cable image directly reflects the appearance of the cable after winding. The cable image can be used to understand the tightness, uniformity and whether there are defects in the cable winding. This information is crucial for judging the winding quality. By extracting features from the cable image, the surface features of the cable winding can be obtained. The winding speed of the motor is a key factor affecting the cable winding effect, so it is also necessary to analyze the winding speed of the motor. In reality, cable processing is a continuous process, and the cable winding effect will change over time. By obtaining the cable images and the speed values ​​of the motor at multiple time points, a time series analysis can be performed to establish a correlation model between the motor speed and the cable winding effect, thereby realizing intelligent control of the speed.

[0030] In an embodiment of the present application, an implementation method for obtaining images of completed wound cables and winding speed values ​​of motors at multiple predetermined time points within a predetermined time period may be: a. Acquisition of image data of completed wound cables: Install a high-resolution camera near the cable processing equipment to ensure that the camera can capture clearly. The camera can be fixed on a bracket to ensure a stable shooting angle. During the processing, the image of the completed wound cable is collected at a set recording frequency; b. Acquisition of winding speed value data of the motor: Install a speed sensor on the motor to measure the motor speed in real time, and record the motor winding speed value at each time point at a set recording frequency.

[0031] Figure 2FIG. 1 is a block diagram of a cable winding processing quality feature encoding module in an intelligent control system for cable processing according to an embodiment of the present application. Figure 2 As shown, the cable winding processing quality feature encoding module 120 includes: a cable image feature extraction unit 121, which is used to pass the completed cable images at the multiple predetermined time points through a cable surface feature extractor to obtain multiple cable winding surface feature matrices; a cable winding timing feature enhancement unit 122, which is used to arrange the multiple cable winding surface feature matrices into a cable winding surface timing feature tensor and then pass it through a cable surface feature enhancer to obtain a cable winding surface timing enhanced feature map; a feature map pooling unit 123, which is used to pool each feature matrix of the cable winding surface timing enhanced feature map along the channel dimension to obtain the cable winding surface timing enhanced feature vector.

[0032] During the cable processing, the cable winding state at each time point may be different, and it is necessary to perform feature extraction on the cable image at each time point, that is, the cable images of the completed winding at the plurality of predetermined time points are respectively passed through the cable surface feature extractor to obtain a plurality of cable winding surface feature matrices. By performing feature extraction on the cable image at each time point, we can obtain detailed information about the cable winding state at each time point. The structure of the cable surface feature extractor in this application is a deep separable convolutional neural network. It should be easy for those skilled in the art to understand that the deep separable convolution decomposes the traditional convolution operation into two independent steps of deep convolution and point-by-point convolution, which greatly reduces the number of parameters of the network. Due to the reduction in the number of parameters, the network is more lightweight, which is particularly important when processing a large amount of cable image data, and can significantly reduce the consumption of computing resources. Although the deep separable convolution reduces the parameters and the amount of calculation, it can still maintain a strong feature extraction capability. Because the point-by-point convolution layer can fuse the features of different channels and provide a richer feature representation, it helps to improve the accuracy and generalization ability of the model in identifying the cable winding state. And due to the lightweight and high efficiency of the deep separable convolutional neural network, it is easier to deploy on devices with limited computing resources. In addition, its smaller number of parameters also means that it is easier to converge during training, reducing the risk of overfitting.

[0033] Cable processing is a continuous process. The cable winding state at each time point is associated with other time points. In order to capture the changing trend of the cable winding state over time, so as to more comprehensively understand the dynamic changes in the cable processing process, it is necessary to first arrange the multiple cable winding surface feature matrices into a cable winding surface time series feature tensor. Through arrangement, the cable winding features at different time points can be effectively integrated. Then the arranged cable winding surface time series feature tensor is passed through the cable surface feature enhancer to obtain the cable winding surface time series enhanced feature map. In this application, the cable surface feature enhancer is a convolutional neural network model using a channel attention mechanism. Convolutional neural networks have powerful feature extraction and learning capabilities in the field of image processing, and the channel attention mechanism can further enhance the network's attention to important features. In the processing of cable winding surface features, the features of different channels may correspond to the cable winding quality features at different time points. By introducing the channel attention mechanism, the network can automatically learn the weight distribution between different channels, so as to pay more attention to the winding features of those important time points.

[0034] In an embodiment of the present application, after arranging the multiple cable winding surface feature matrices into a cable winding surface temporal feature tensor, a possible implementation method for obtaining a cable winding surface temporal enhanced feature map can be as follows: each layer of the cable surface feature enhancer performs the following on the input data in the forward transfer of the layer: convolution processing on the input data based on the convolution kernel to generate a convolution feature map; pooling processing on the convolution feature map to generate a pooling feature map; activation processing on the pooling feature map to generate an activation feature map; calculating the quotient of the eigenvalue mean of the feature matrix corresponding to each channel in the activation feature map and the sum of the eigenvalue mean of the feature matrix corresponding to all channels as the weighting coefficient of the feature matrix corresponding to each channel; weighting the feature matrix of each channel with the weighting coefficient of each channel in the activation feature map to generate a channel attention feature map; wherein the output of the last layer of the cable surface feature enhancer is the cable winding surface temporal enhanced feature map, and the input of the first layer of the cable surface feature enhancer is the cable winding surface temporal feature tensor.

[0035] Next, the cable winding surface timing enhancement feature map is pooled along the channel dimension to obtain the cable winding surface timing enhancement feature vector. The cable winding surface timing enhancement feature map extracted by the channel attention mechanism has a high dimension and contains a large amount of feature information. Directly processing these high-dimensional features will lead to a large increase in the amount of calculation, which not only reduces the processing speed, but also may introduce noise and redundant information. Through the pooling operation, the feature map can be compressed in the channel dimension to extract the most representative features, thereby simplifying the subsequent processing process. In addition, during the cable processing process, due to the interference of environmental noise, equipment vibration and other factors, the cable winding surface features may fluctuate and change to a certain extent. Through the pooling operation, these features can be smoothed to reduce the impact of noise, making the model's judgment of the cable winding state more stable and accurate.

[0036] Figure 3 FIG. 1 is a block diagram of a speed sequence feature encoding module for cable winding processing in an intelligent control system for cable processing according to an embodiment of the present application. Figure 3 As shown, the cable winding processing speed timing feature encoding module 130 includes: a motor winding speed data structuring unit 131, which is used to arrange the winding speed values ​​of the motor at the multiple time points according to the time dimension as a motor winding speed timing input vector; a winding speed timing feature generating unit 132, which is used to pass the motor winding speed timing input vector through a winding speed timing feature encoder to obtain the motor winding speed timing feature vector.

[0037] The motor winding speed is a key parameter in the cable processing process, which directly affects the cable winding quality and efficiency. The cable winding quality depends not only on the average speed of the motor, but also on dynamic factors such as speed fluctuations and change trends. In order to fully understand the dynamic behavior of the motor during the cable processing process, the motor winding speed values ​​at the multiple time points are arranged according to the time dimension as the motor winding speed timing input vector in this application. Through the arrangement, a complete motor speed change sequence can be obtained.

[0038] The original motor winding speed timing input vector contains rich speed information, but this information is difficult to be directly used in the subsequent motor control strategy. In order to extract the most representative and most conducive to the control decision-making key features of the motor winding speed from the original motor winding speed, it is necessary to pass the motor winding speed timing input vector through the winding speed timing feature encoder to obtain the motor winding speed timing feature vector. In this application, the winding speed timing feature encoder is a convolutional neural network model using a one-dimensional convolution layer and a fully connected layer. First, the one-dimensional convolution layer is particularly suitable for processing time series data. The one-dimensional convolution operation can slide along the time dimension to effectively capture the local patterns and changing trends in the motor speed timing data. This local perception feature enables the model to pay attention to the key details in the speed data and extract useful feature representations. Secondly, the fully connected layer can integrate the local features extracted by the one-dimensional convolution layer to form a global feature representation. The fully connected layer fuses and integrates the local features extracted by the convolution layer through weighted connections, thereby obtaining a more comprehensive and abstract feature vector. This structure helps the model capture the global patterns and laws in the motor speed timing data. In addition, the convolutional neural network model has a strong feature learning ability. Through the back propagation algorithm and gradient descent optimization during the training process, the model can automatically learn the mapping relationship from raw speed data to useful features. This enables the model to adaptively adjust its parameters to better extract the key features of cable speed related to cable processing control.

[0039] In an embodiment of the present application, one possible implementation method of passing the motor winding speed timing input vector through a winding speed timing feature encoder to obtain a motor winding speed timing feature vector may be: using the fully connected layer of the winding speed timing feature encoder to fully connect the motor winding speed timing input vector using the following fully connected encoding formula to extract high-dimensional implicit features of the characteristic values ​​at each position in the motor winding speed timing input vector, wherein the fully connected encoding formula is:

[0040]

[0041] Where X is the motor winding speed timing input vector, Y is the output vector, W is the weight matrix, B is the bias vector, Represents matrix multiplication; using the one-dimensional convolution layer of the winding speed timing feature encoder to perform one-dimensional convolution encoding on the motor winding speed timing input vector using the following one-dimensional convolution formula to extract the high-dimensional implicit correlation features of the correlation between the feature values ​​at each position in the motor winding speed timing input vector, wherein the one-dimensional convolution encoding formula is:

[0042]

[0043] Among them, a is the width of the first convolution kernel in the x direction, F(a) is the first convolution kernel parameter vector, G(xa) is the local vector matrix operated with the convolution kernel function, w is the size of the first convolution kernel, X represents the motor winding speed timing input vector, and Cov(X) represents one-dimensional convolution encoding of the input vector.

[0044] The cable winding surface timing enhancement feature vector is a key feature extracted from the cable winding surface image at multiple time points, which reflects the quality and state of the cable winding; the motor winding speed timing feature vector is an encoded representation of the speed information of the motor during the cable winding process; in order to analyze the influence of the motor speed on the cable winding surface, it is necessary to calculate the responsiveness estimate of the cable winding surface timing enhancement feature vector relative to the motor winding speed timing feature vector in this application to obtain the motor winding responsiveness control feature matrix.

[0045] In an embodiment of the present application, a possible implementation method of calculating the responsiveness estimation of the cable winding surface timing enhancement feature vector relative to the motor winding speed timing feature vector to obtain the motor winding responsiveness control feature matrix may be: calculating the responsiveness estimation of the cable winding surface timing enhancement feature vector relative to the motor winding speed timing feature vector using the following responsiveness estimation formula to obtain the motor winding responsiveness control feature matrix; wherein the responsiveness estimation formula is:

[0046]

[0047] Wherein V1 represents the cable winding surface timing enhancement feature vector, V2 represents the motor winding speed timing feature vector, M represents the motor winding responsiveness control feature matrix, Represents matrix multiplication.

[0048] Figure 4 FIG. 1 is a block diagram of a cable winding speed control result generation module in an intelligent control system for cable processing according to an embodiment of the present application. Figure 4 As shown, the cable winding speed control result generation module 150 includes: a feature optimization unit 151, which is used to adjust the motor winding responsiveness control feature matrix based on the probability sensitivity of feature activation to obtain an optimized motor winding responsiveness control feature matrix; an optimized feature matrix classification unit 152, which is used to pass the optimized motor winding responsiveness control feature matrix through a classifier to obtain a classification result, and the classification result is used to indicate whether the winding speed value of the motor at the current time point should be increased or decreased.

[0049] In particular, in the technical solution of the present application, it is considered that the cable image contains rich visual information, such as the texture, shape, winding density, etc. of the cable. When this information is converted into a feature matrix, a large number of feature dimensions will be generated, resulting in the high dimensionality of the feature space. The cable winding process is dynamically changing, and the surface features of the cable at different time points may have significant differences. Arranging these time series features into tensors and enhancing them will further increase the dimension and complexity of the features. When the cable winding surface time series enhancement feature vector is fused with the motor winding speed time series feature vector, information loss or conversion may occur, resulting in a more divergent distribution of the fused feature matrix in the high-dimensional space. When calculating the motor winding responsiveness control feature matrix, it is necessary to estimate the responsiveness between the cable surface features and the motor speed. This estimate may be uncertain, resulting in a divergent distribution in the feature matrix. In order for the classifier to have good generalization ability, it is necessary to capture enough information in the feature space to distinguish different control requirements. This usually means that the model needs to find a suitable decision boundary in the high-dimensional space, and the determination of this boundary may be affected by the divergence of the distribution. Therefore, in the technical solution of the present application, in view of the distribution divergence in the high-dimensional feature semantic space constituted by the motor winding responsiveness control feature matrix, in the technical solution of the present application, the motor winding responsiveness control feature matrix is ​​adjusted for probability sensitivity based on feature activation.

[0050] Specifically, the motor winding responsiveness control feature matrix is ​​subjected to probability sensitivity adjustment based on feature activation to obtain an optimized motor winding responsiveness control feature matrix, including: summing the eigenvalues ​​of each position of the motor winding responsiveness control feature matrix to obtain the sum of the eigenvalues ​​of the motor winding responsiveness control feature matrix; adding one to the sum of the eigenvalues ​​of the motor winding responsiveness control feature matrix and dividing it by the feature distribution scale of the motor winding responsiveness control feature matrix to obtain a first probability value, wherein the feature distribution scale of the motor winding responsiveness control feature matrix is ​​the width of the motor winding responsiveness control feature matrix multiplied by the width of the motor winding responsiveness control feature matrix. The height of the eigenvalue matrix; using the eigenvalue of the predetermined position of the motor winding responsiveness control feature matrix as the exponent of the natural constant to calculate an exponential function with the natural constant as the base to obtain a first exponential function value; subtracting one from the first exponential function value and then calculating the absolute value of the eigenvalue of the predetermined position of the motor winding responsiveness control feature matrix to obtain a first absolute value; activating the first absolute value through a softmax function to obtain a second probability value; multiplying the first probability value and the second probability value by the eigenvalue of the predetermined position of the motor winding responsiveness control feature matrix respectively and then calculating the weighted sum to obtain the eigenvalue of the corresponding position of the optimized motor winding responsiveness control feature matrix.

[0051] In an embodiment of the present application, another possible implementation method of performing probability sensitivity adjustment based on feature activation on the motor winding responsiveness control characteristic matrix to obtain an optimized motor winding responsiveness control characteristic matrix may be: performing probability sensitivity adjustment based on feature activation on the motor winding responsiveness control characteristic matrix using the following optimization formula to obtain the optimized motor winding responsiveness control characteristic matrix; wherein the optimization formula is:

[0052]

[0053] Among them, m i,j is the eigenvalue of the (i, j)th position of the motor winding responsiveness control feature matrix, W and H are the width and height of the motor winding responsiveness control feature matrix, α and β represent weighting coefficients, softmax represents a normalized exponential function, and m i,j ′ is the eigenvalue of the (i, j)th position of the optimized motor winding responsiveness control characteristic matrix.

[0054] Here, in view of the distribution divergence in the high-dimensional feature semantic space constituted in the motor winding responsiveness regulation feature matrix, in the technical solution of the present application, the motor winding responsiveness regulation feature matrix is ​​subjected to probability sensitivity adjustment based on feature activation, which adjusts the label probability response of each feature in the motor winding responsiveness regulation feature matrix so as to better capture and reflect the intrinsic structure and category distribution of the data. Specifically, it uses a statistical analysis method to evaluate the response probability of each feature in the motor winding responsiveness regulation feature matrix to different labels, and then adjusts the feature weights according to these probabilities, so that the model can more sensitively reflect the correlation between the feature distribution and the label, which can effectively reduce the classification uncertainty caused by the divergence of the feature distribution, so that the feature distribution of the motor winding responsiveness regulation feature matrix is ​​transferred to a range with a stable and structured boundary under the target domain, thereby improving the stability of the inductive iteration of the classification solution.

[0055] Finally, the optimized motor winding responsiveness control feature matrix is ​​passed through a classifier to obtain a classification result, and the classification result is used to indicate whether the winding speed value of the motor at the current time point should be increased or decreased. It should be well understood that the optimized motor winding responsiveness control feature matrix incorporates the information of the cable winding surface timing enhancement feature vector and the motor winding speed timing feature vector, and reflects the responsiveness relationship between them, but it is not directly used for control decision-making. A classifier is a machine learning model that can classify or label the input data. It learns the mapping relationship from features to classification results through a training process, and this mapping relationship is obtained based on a large amount of training data. According to the classification result, the control system can adjust the speed of the motor in real time to adapt to the changes in the cable winding surface. If the classification result shows that the speed should be increased, the control system will increase the speed of the motor accordingly to improve the winding quality of the cable; conversely, if the classification result shows that the speed should be reduced, the control system will reduce the speed of the motor to avoid excessive winding or damage to the cable.

[0056] In summary, the intelligent control system 100 for cable processing based on the embodiment of the present application is explained, which adopts artificial intelligence analysis technology based on deep learning, and extracts features of the motor winding speed and the cable image that has been wound at multiple predetermined time points within a predetermined time period to understand the relationship between the cable winding quality and the motor winding speed, and then obtains the control direction of the current cable winding speed. In this way, the precise control of the cable winding speed is achieved, which is conducive to improving the quality of cable winding processing.

[0057] As described above, the intelligent control system 100 for cable processing according to the embodiment of the present application can be implemented in various terminal devices, such as a server for intelligent control of cable processing. In one example, the intelligent control system 100 for cable processing according to the embodiment of the present application can be integrated into the terminal device as a software module and / or a hardware module. For example, the intelligent control system 100 for cable processing can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the intelligent control system 100 for cable processing can also be one of the many hardware modules of the terminal device.

[0058] Alternatively, in another example, the intelligent control system 100 for cable processing and the terminal device may also be separate devices, and the intelligent control system 100 for cable processing may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0059] Figure 5FIG. 1 is a flow chart of an intelligent control method for cable processing according to an embodiment of the present application. Figure 5 As shown, in the intelligent control method for cable processing, it includes: S110, obtaining the cable images of the completed winding and the winding speed values ​​of the motor at multiple predetermined time points within a predetermined time period; S120, feature encoding the cable images of the completed winding at the multiple predetermined time points to obtain the cable winding surface timing enhancement feature vector; S130, feature extraction of the motor winding speed values ​​at the multiple predetermined time points to obtain the motor winding speed timing feature vector; S140, calculating the responsiveness estimate of the cable winding surface timing enhancement feature vector relative to the motor winding speed timing feature vector to obtain the motor winding responsiveness control feature matrix; S150, based on the motor winding responsiveness control feature matrix, judging the winding speed control direction of the motor at the current time point.

[0060] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned intelligent control method for cable processing have been described in the above reference. Figures 1 to 4 The invention has been described in detail in the description of the intelligent control system for cable processing, and therefore, its repeated description will be omitted.

[0061] In summary, the intelligent control method for cable processing based on the embodiment of the present application is explained, which adopts artificial intelligence analysis technology based on deep learning, and extracts features of the motor winding speed and the cable image that has been wound at multiple predetermined time points within a predetermined time period to understand the relationship between the cable winding quality and the motor winding speed, and then obtains the control direction of the current cable winding speed. In this way, the precise control of the cable winding speed is achieved, which is conducive to improving the quality of cable winding processing.

[0062] The present application also relates to a computer-readable storage medium for implementing an intelligent control method for cable processing according to one or more embodiments of the present application. The computer-readable storage medium mentioned herein includes various types of computer storage media, which can be any available medium that can be accessed by a general or special-purpose computer. For example, the computer-readable storage medium may include RAM, ROM, EPROM, E2PROM, registers, hard disks, removable disks, CD-ROMs or other optical disk storage, disk storage or other magnetic storage devices, or any other temporary or non-temporary medium that can be used to carry or store desired program code units in the form of instructions or data structures and can be accessed by a general or special-purpose computer, or a general or special-purpose processor. The description of the computer-readable storage medium according to the present application can refer to the explanation of the intelligent control method for cable processing according to the present application, and no further description is given.

[0063] The terms used herein are for the purpose of describing specific embodiments only and are not intended to limit the present application. As used herein, the singular forms "a", "an", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprise and / or comprise" when used in this specification specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, element components, and / or groups thereof.

[0064] Those skilled in the art will appreciate that various example embodiments are shown and described herein, each example embodiment having certain features in a particular embodiment, but the present application is not limited thereto. On the contrary, the present application can be modified to incorporate any number of variants, changes, substitutions, combinations, sub-combinations, or equivalent arrangements not previously described but commensurate with the scope of the present application. In addition, although various embodiments of the present application have been described, it will be understood that aspects of the present application may include only some of the described embodiments. Therefore, the present application will not be considered to be limited by the foregoing description, but only by the scope of the appended claims.

Claims

1. An intelligent control system for cable processing, characterized in that: include: The cable winding processing related data acquisition module is used to obtain the images of the completed wound cables and the winding speed values ​​of the motor at multiple predetermined time points within a predetermined time period; A cable winding processing quality feature encoding module, used for performing feature encoding on the images of the completed cable winding at the plurality of predetermined time points to obtain a time-series enhanced feature vector of the cable winding surface; A cable winding processing speed time series feature encoding module is used to extract the features of the motor winding speed values ​​at the plurality of predetermined time points to obtain a motor winding speed time series feature vector; A cable winding responsiveness estimation module, used to calculate the responsiveness estimation of the cable winding surface timing enhancement feature vector relative to the motor winding speed timing feature vector to obtain a motor winding responsiveness control feature matrix; The cable winding speed control result generation module is used to determine the winding speed control direction of the motor at the current time point based on the motor winding responsiveness control characteristic matrix.

2. The intelligent control system for cable processing according to claim 1 is characterized in that: The cable winding processing quality feature encoding module comprises: A cable image feature extraction unit, used for passing the cable images of the plurality of predetermined time points that have been wound through a cable surface feature extractor to obtain a plurality of cable winding surface feature matrices; A cable winding timing feature enhancement unit, used for arranging the plurality of cable winding surface feature matrices into a cable winding surface timing feature tensor and then passing the tensor through a cable surface feature enhancer to obtain a cable winding surface timing enhancement feature map; The feature map pooling unit is used to perform pooling of each feature matrix along the channel dimension on the cable winding surface timing enhancement feature map to obtain the cable winding surface timing enhancement feature vector.

3. The intelligent control system for cable processing according to claim 2 is characterized in that: The cable winding processing speed timing feature encoding module includes: A motor winding speed data structuring unit, used to arrange the motor winding speed values ​​at the plurality of predetermined time points into a motor winding speed timing input vector according to a time dimension; The winding speed timing characteristic generating unit is used for passing the motor winding speed timing input vector through a winding speed timing characteristic encoder to obtain the motor winding speed timing characteristic vector.

4. The intelligent control system for cable processing according to claim 3 is characterized in that: The cable surface feature extractor is a deep separable convolutional neural network model, the cable surface feature enhancer is a first convolutional neural network model using a channel attention mechanism, and the winding speed timing feature encoder is a second convolutional neural network model using a one-dimensional convolutional layer and a fully connected layer.

5. The intelligent control system for cable processing according to claim 4, characterized in that: The cable winding timing feature enhancement unit is used for: each layer of the cable surface feature enhancer performs the following operations on the input data in the forward transmission of the layer: Performing convolution processing on the input data based on a convolution kernel to generate a convolution feature map; Performing pooling processing on the convolutional feature map to generate a pooled feature map; Performing activation processing on the pooled feature map to generate an activation feature map; Calculate the quotient of the eigenvalue mean of the feature matrix corresponding to each channel in the activation feature map and the sum of the eigenvalue means of the feature matrices corresponding to all channels as the weighting coefficient of the feature matrix corresponding to each channel; Weighting the feature matrix of each channel by the weight coefficient of each channel in the activation feature map to generate a channel attention feature map; Among them, the output of the last layer of the cable surface feature enhancer is the cable winding surface temporal enhancement feature map, and the input of the first layer of the cable surface feature enhancer is the cable winding surface temporal feature tensor.

6. The intelligent control system for cable processing according to claim 5, characterized in that: The cable winding speed control result generating module comprises: A feature optimization unit, used for performing probability sensitivity adjustment on the motor winding responsiveness control feature matrix based on feature activation to obtain an optimized motor winding responsiveness control feature matrix; The optimized characteristic matrix classification unit is used to pass the optimized motor winding responsiveness control characteristic matrix through a classifier to obtain a classification result, and the classification result is used to indicate whether the winding speed value of the motor at the current time point should be increased or decreased.

7. The intelligent control system for cable processing according to claim 6, characterized in that: The feature optimization unit is used to: Summing the eigenvalues ​​of each position of the motor winding responsiveness control characteristic matrix to obtain the sum of the eigenvalues ​​of the motor winding responsiveness control characteristic matrix; Adding one to the sum of the eigenvalues ​​of the motor winding responsiveness control feature matrix and then dividing it by the feature distribution scale of the motor winding responsiveness control feature matrix to obtain a first probability value, wherein the feature distribution scale of the motor winding responsiveness control feature matrix is ​​the width of the motor winding responsiveness control feature matrix multiplied by the height of the motor winding responsiveness control feature matrix; Using the eigenvalue of the predetermined position of the motor winding responsiveness control characteristic matrix as the exponent of the natural constant, an exponential function with the natural constant as the base is calculated to obtain a first exponential function value; Subtracting one from the first exponential function value and dividing it by the eigenvalue of the predetermined position of the motor winding responsiveness control characteristic matrix and then calculating the absolute value to obtain a first absolute value; Activating the first absolute value through a softmax function to obtain a second probability value; The first probability value and the second probability value are respectively multiplied by the eigenvalues ​​of predetermined positions of the motor winding responsiveness control feature matrix, and then the weighted sum is calculated to obtain the eigenvalues ​​of corresponding positions of the optimized motor winding responsiveness control feature matrix.

8. An intelligent control method for cable processing, characterized in that: include: Acquire images of the completed cable winding and winding speed values ​​of the motor at a plurality of predetermined time points within a predetermined time period; Performing feature encoding on the images of the completed cable at the plurality of predetermined time points to obtain a time-series enhanced feature vector of the cable winding surface; Performing feature extraction on the winding speed values ​​of the motor at the plurality of predetermined time points to obtain a motor winding speed time series feature vector; Calculating the responsiveness estimation of the cable winding surface timing enhancement characteristic vector relative to the motor winding speed timing characteristic vector to obtain the motor winding responsiveness regulation characteristic matrix; Based on the motor winding responsiveness control characteristic matrix, the winding speed control direction of the motor at the current time point is determined.

9. The intelligent control method for cable processing according to claim 8, characterized in that: Performing feature encoding on the cable images of the plurality of predetermined time points to obtain a cable winding surface time sequence enhanced feature vector, comprising: Passing the images of the completed winding cables at the plurality of predetermined time points through a cable surface feature extractor to obtain a plurality of cable winding surface feature matrices; Arranging the plurality of cable winding surface feature matrices into a cable winding surface time series feature tensor and then passing it through a cable surface feature enhancer to obtain a cable winding surface time series enhanced feature map; The cable winding surface timing enhancement feature map is pooled along the channel dimension to obtain the cable winding surface timing enhancement feature vector.

10. The intelligent control method for cable processing according to claim 9, characterized in that: Based on the motor winding responsiveness control characteristic matrix, determining the winding speed control direction of the motor at the current time point includes: Performing probability sensitivity adjustment based on feature activation on the motor winding responsiveness regulation feature matrix to obtain an optimized motor winding responsiveness regulation feature matrix; The optimized motor winding responsiveness control characteristic matrix is ​​passed through a classifier to obtain a classification result, and the classification result is used to indicate whether the winding speed value of the motor at the current time point should be increased or decreased.

Images