Monitoring method, system and device in operation process of plastic mold and storage medium

Abstract

The invention discloses a monitoring method, system and device in the operation process of a plastic mold and a storage medium. The method comprises the following steps: obtaining mold video stream data through a visual system; performing image acquisition according to the mold video stream data to obtain a mold video frame; performing feature extraction processing according to the mold video frame to obtain feature point data; according to the feature point data, performing position offset calculation to obtain a position offset; according to the position offset and the feature point data, speed calculation is carried out, and the running speed of the mold is obtained; according to the mold operation speed, operation state judgment is conducted, and the mold operation state is obtained; and when it is judged that the mold operation state is abnormal, generating an early warning signal, and sending the early warning signal to a user side. According to the method, high-precision and high-stability monitoring of the operation process of the plastic mold can be realized.

Description

Technical Field

[0001] The present invention relates to the technical field of plastic molds, and in particular to a monitoring method, system, device and storage medium during the operation of a plastic mold. Background Art

[0002] With the rapid development of industry, the improvement of productivity and the continuous improvement of product quality requirements, the manufacturing industry has higher and higher requirements for the precision of production equipment, especially in the plastics industry. How to ensure the stability of the precision of injection molding machines and plastic molds during high-speed operation has become a key issue. Among them, the speed detection and control of plastic molds is the core link to ensure product quality and production efficiency. Traditional speed detection methods can no longer meet the needs of modern manufacturing for high precision, high reliability and low cost. Therefore, the development of a non-contact, high-precision speed monitoring technology has become an urgent need in the plastics industry.

[0003] At present, traditional speed detection technology mainly relies on contact measuring devices such as photoelectric encoders, incremental encoders, and Hall sensors. These devices collect and convert the speed signal of the plastic mold synchronously through AD, and transmit the data to the PLC controller, which controls the speed of the motor. In addition, speed measuring devices based on machine vision processing technology are also gradually being used. For example, a linear array camera is used to shoot the movement process of the plastic mold, and the real-time speed of the plastic mold is calculated through the camera grating ruler. These technologies can meet certain measurement requirements in low-speed motion scenarios and provide basic support for the speed control of plastic molds. However, during the high-speed motion of the plastic mold, the running speed deviation is caused by factors such as friction, which in turn affects the positioning accuracy. The faster the speed, the greater the deviation, which ultimately affects the product quality. In addition, the speed measuring device based on machine vision relies on the grating ruler, and the installation method is fixed and limited. It is only suitable for low-speed motion scenarios (usually less than 100m / min), the measurement range is limited, the accuracy is unstable, and the error control is poor.

[0004] In summary, the existing technology has the problems of low accuracy and instability in the monitoring and detection process of plastic mold operation. Summary of the invention

[0005] The present invention provides a monitoring method, system, device and storage medium during the operation of a plastic mold, so as to achieve high-precision and high-stability monitoring of the operation of the plastic mold.

[0006] In a first aspect, in order to solve the above technical problems, the present invention provides a monitoring method during the operation of a plastic mold, comprising: Acquire mold video stream data through the visual system; According to the mold video stream data, image acquisition is performed to obtain the mold video frame; Performing feature extraction processing according to the mold video frame to obtain feature point data; Calculating the position offset according to the feature point data to obtain the position offset; Calculate the speed according to the position offset and the feature point data to obtain the mold running speed; According to the mold running speed, the running state is determined to obtain the mold running state; When it is determined that the mold operation state is abnormal, an early warning signal is generated and sent to a user end.

[0007] As an optional implementation manner, performing image acquisition according to the mold video stream data to obtain the mold video frame includes: Extract video frames according to the mold video stream data to obtain initial video frames; Performing denoising processing on the initial video frame to obtain a denoised video frame; Image enhancement processing is performed according to the denoised video frame to obtain a mold video frame.

[0008] As an optional implementation manner, performing feature extraction processing according to the mold video frame to obtain feature point data includes: According to the mold video frame, a convolutional neural network is used to perform preliminary feature extraction to obtain a preliminary feature map; According to the preliminary feature map, a region recommendation network is used to process the map to obtain a candidate feature region; According to the candidate feature regions, a similarity measurement network is used to perform feature point matching to identify feature point data of the mold.

[0009] As an optional implementation manner, the calculating the position offset according to the feature point data to obtain the position offset includes: According to the feature point data, coordinate analysis is performed to obtain the reference position feature point coordinates and the target position feature point coordinates; Calculating pixel difference based on the reference position feature point coordinates and the target position feature point coordinates to obtain pixel difference; Calculate the position offset according to the pixel difference and the relationship between the preset pixel and the actual size to obtain the position offset; The calculation formula of the pixel difference is as follows: The calculation formula of the position offset is as follows: in, , The horizontal and vertical coordinates of the feature point coordinates of the reference position, , Represents the horizontal and vertical coordinates of the feature point coordinates of the target position, represents the horizontal axis pixel difference, represents the vertical pixel difference, Indicates the position offset, Indicates the conversion factor between the pixels on the horizontal axis and the actual size. Indicates the conversion factor between the pixels on the vertical axis and the actual size.

[0010] As an optional implementation manner, the speed calculation is performed according to the position offset and the feature point data to obtain the mold running speed, including: According to the feature point data, the feature point data is parsed to obtain the time length corresponding to moving from the reference position to the target position; According to the position offset and the duration, a sliding window average method is used to calculate the speed to obtain the mold running speed; The speed calculation formula is as follows: in, represents the size of the sliding window, Indicates the window The calculated position offset, Indicates the mold running speed, Indicates the window The duration of the calculation.

[0011] As an optional implementation manner, the step of determining the operation state according to the mold operation speed to obtain the mold operation state includes: According to the historical running speed data, the isolation forest model is constructed to obtain the isolation forest model; Calculate the path length according to the isolation forest model and the running speed of the mold to obtain the path length; According to the path length, an average path length is calculated, and an anomaly score is calculated based on the average path length; Performing an abnormality determination according to the abnormality score, and if the abnormality score exceeds a preset threshold, determining that the mold operation state is abnormal; The calculation formula of the anomaly score is as follows: in, Indicates the current sample speed The anomaly score, Representation sample The average path length in the isolation forest, represents the normalization factor, Indicates the sample size.

[0012] As an optional implementation, it also includes: According to the speed and the position offset, a time series curve is drawn to obtain a parameter change trend; According to the parameter change trend, trend prediction is performed to obtain a prediction result; According to the prediction result, if a trend toward an abnormal state is detected, a weak warning signal is generated and sent to the user end.

[0013] In a second aspect, the present invention provides a monitoring system for a plastic mold during operation, comprising: A data acquisition module, used for acquiring mold video stream data through a visual system; An image acquisition module is used to acquire images according to the mold video stream data to obtain a mold video frame; A feature extraction module, used to perform feature extraction processing according to the mold video frame to obtain feature point data; A position calculation module, used to calculate the position offset according to the feature point data to obtain the position offset; A speed calculation module, used to perform speed calculation according to the position offset and the feature point data to obtain the mold running speed; An operation status module, used to determine the operation status according to the operation speed of the mold to obtain the operation status of the mold; The early warning module is used to generate an early warning signal when it is determined that the operating state of the mold is abnormal, and send the early warning signal to the user end.

[0014] In a third aspect, the present invention also provides a monitoring device for a plastic mold during operation, characterized in that it comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, it implements any one of the above-mentioned methods for monitoring a plastic mold during operation.

[0015] In a fourth aspect, the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the above-mentioned methods for monitoring the operation of a plastic mold.

[0016] Compared with the prior art, the present invention has the following beneficial effects: The present invention provides a monitoring method during the operation of a plastic mold, comprising: acquiring mold video stream data through a visual system; performing image acquisition according to the mold video stream data to obtain a mold video frame; performing feature extraction processing according to the mold video frame to obtain feature point data; performing position offset calculation according to the feature point data to obtain the position offset; performing speed calculation according to the position offset and the feature point data to obtain the mold operation speed; performing operation state determination according to the mold operation speed to obtain the mold operation state; when it is determined that the mold operation state is abnormal, generating an early warning signal, and sending the early warning signal to a user end.

[0017] In the present invention, the operation process of the plastic mold is monitored in real time through a visual system to obtain the mold video stream data, and then image acquisition is performed according to the mold video stream data to obtain a mold video frame; according to the mold video frame, feature extraction processing is performed through a deep learning algorithm to obtain feature point data; according to the feature point data, the position offset is calculated to obtain the position offset; according to the position offset and time information, the speed is calculated to obtain the mold operation speed; according to the mold operation speed, the operation state is determined by an isolated forest algorithm to obtain the mold operation state; when the mold operation state is determined to be abnormal, an early warning signal is generated and sent to the user end. Wherein, the early warning signal contains information such as the time, location, abnormal speed and duration of the abnormal state, which is convenient for users to respond and process quickly. Through the combination of deep learning algorithm and isolated forest algorithm, the accuracy and stability of plastic mold speed measurement and operation state determination are improved. In addition, by drawing time series curves for key parameters such as plastic molds, glue guns, and injection molding machines, and combining moving average and slope detection algorithms, it is possible to predict parameter change trends in advance and issue a weak warning when an abnormal trend is detected, reminding operation and maintenance personnel to pay attention in advance and avoid potential production accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flow chart of a monitoring method during the operation of a plastic mold provided by an embodiment of the present invention; Figure 2 It is a schematic diagram of the structure of a monitoring system during the operation of a plastic mold provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0019] With the rapid development of industry, the improvement of productivity and the continuous improvement of product quality requirements, the manufacturing industry has higher and higher requirements for the precision of production equipment, especially in the plastics industry. How to ensure the stability of the precision of injection molding machines and plastic molds during high-speed operation has become a key issue. Among them, the speed detection and control of plastic molds is the core link to ensure product quality and production efficiency. Traditional speed detection methods can no longer meet the needs of modern manufacturing for high precision, high reliability and low cost. Therefore, the development of a non-contact, high-precision speed monitoring technology has become an urgent need in the plastics industry.

[0020] At present, traditional speed detection technology mainly relies on contact measuring devices such as photoelectric encoders, incremental encoders, and Hall sensors. These devices collect and convert the speed signal of the plastic mold synchronously through AD, and transmit the data to the PLC controller, which controls the speed of the motor. In addition, speed measuring devices based on machine vision processing technology are also gradually being used. For example, a linear array camera is used to shoot the movement process of the plastic mold, and the real-time speed of the plastic mold is calculated through the camera grating ruler. These technologies can meet certain measurement requirements in low-speed motion scenarios and provide basic support for the speed control of plastic molds. However, during the high-speed motion of the plastic mold, the running speed deviation is caused by factors such as friction, which in turn affects the positioning accuracy. The faster the speed, the greater the deviation, which ultimately affects the product quality. In addition, the speed measuring device based on machine vision relies on the grating ruler, and the installation method is fixed and limited. It is only suitable for low-speed motion scenarios (usually less than 100m / min), the measurement range is limited, the accuracy is unstable, and the error control is poor.

[0021] In summary, the existing technology has the problems of low accuracy and instability in the monitoring and detection process of plastic mold operation.

[0022] In view of the above problems existing in the prior art, the present invention provides a monitoring method, system, device and storage medium during the operation of a plastic mold, so as to achieve high-precision and high-stability monitoring of the operation of the plastic mold.

[0023] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] Reference Figure 1 The first embodiment of the present invention provides a method for monitoring a plastic mold during operation, comprising the following steps: S11, obtaining mold video stream data through a visual system; S12, performing image acquisition according to the mold video stream data to obtain a mold video frame; S13, performing feature extraction processing according to the mold video frame to obtain feature point data; S14, calculating the position offset according to the feature point data to obtain the position offset; S15, performing speed calculation according to the position offset and the feature point data to obtain the mold running speed; S16, determining the operation state of the mold according to the mold operation speed to obtain the mold operation state; S17, when it is determined that the mold operation state is abnormal, a warning signal is generated and the warning signal is sent to the user end.

[0025] It should be noted that the present invention realizes high-precision, non-contact monitoring of the operation process of plastic molds through the combination of visual systems, deep learning algorithms and edge computing technologies. Specifically, the visual system can obtain mold video stream data in real time, and the deep learning algorithm accurately measures the running speed of the mold through feature extraction and position offset calculation, and determines the running status in combination with the isolation forest algorithm. When an abnormal state is detected, the system can generate an early warning signal in time and send it to the user end to ensure that the operation and maintenance personnel can respond and process quickly. In addition, the present invention also uses the sliding window averaging method and the slope detection algorithm to predict the parameter change trend in advance, further improving the stability and reliability of the system.

[0026] In step S11, the mold video stream data is acquired through the visual system.

[0027] It should be noted that the visual system includes a high-frame rate linear array camera or other high-speed camera equipment, which can capture the dynamic image of the plastic mold during high-speed operation in real time. The visual system is installed around the injection molding machine to ensure that the entire motion range of the plastic mold can be covered. The mold video stream data obtained by the visual system has high resolution and high frame rate, and can provide high-quality input data for subsequent image processing and analysis. In addition, the visual system is connected to the data processing center through an edge computing device, which can transmit video stream data in real time to ensure the real-time and high efficiency of the monitoring process. The non-contact measurement method of the visual system avoids the loss of the mold by traditional contact measurement devices, while improving the reliability and stability of the system.

[0028] In step S12, the image acquisition is performed according to the mold video stream data to obtain the mold video frame, including: Extract video frames according to the mold video stream data to obtain initial video frames; Performing denoising processing on the initial video frame to obtain a denoised video frame; Image enhancement processing is performed according to the denoised video frame to obtain a mold video frame.

[0029] It should be noted that the image acquisition operation refers to obtaining the mold video stream data through the visual system and extracting high-quality mold video frames from it. This operation can provide clear image data for subsequent feature extraction and position offset calculation. In an embodiment of the present invention, the image acquisition operation is based on the mold video stream data and is implemented through steps such as video frame extraction, denoising and image enhancement. In the operation monitoring of plastic molds, the use of denoising and image enhancement technology can significantly improve the image quality and ensure the accuracy and stability of subsequent processing.

[0030] Among them, the video frame extraction operation refers to extracting continuous initial video frames from the mold video stream data. Through a high frame rate linear array camera or other high-speed camera equipment, each frame of the plastic mold in high-speed operation can be captured. This step provides basic data for subsequent denoising and image enhancement processing. The denoising operation refers to denoising the initial video frame to remove noise interference in the image and obtain a denoised video frame. Common denoising algorithms include Gaussian filtering, median filtering, and Kalman filtering, which can effectively reduce random noise and salt and pepper noise in the image and improve the clarity of the image. In an embodiment of the present invention, the denoising algorithm uses the Kalman filtering algorithm. Of course, according to different application scenarios and user needs, the denoising algorithm can also use other methods such as Gaussian filtering and median filtering, which are not limited by the present invention. The image enhancement operation refers to performing image enhancement processing on the denoised video frame to further improve the contrast and detail performance of the image and obtain a mold video frame. Common image enhancement techniques include histogram equalization, contrast stretching, etc., which can significantly improve the visual effect of the image and provide high-quality input data for subsequent feature extraction and position offset calculation. In the embodiment of the present invention, the image enhancement technology adopts the histogram equalization technology. Of course, according to different application scenarios and user needs, the image enhancement technology can also adopt other technologies such as contrast stretching technology, and the present invention is not limited to this.

[0031] In step S13, the feature extraction process is performed according to the mold video frame to obtain feature point data, including: According to the mold video frame, a convolutional neural network is used to perform preliminary feature extraction to obtain a preliminary feature map; According to the preliminary feature map, a region recommendation network is used to process the map to obtain a candidate feature region; According to the candidate feature regions, a similarity measurement network is used to perform feature point matching to identify feature point data of the mold.

[0032] It should be noted that the feature extraction processing operation refers to extracting the feature point data of the mold from the mold video frame through a deep learning algorithm. This operation can provide key feature point information for subsequent position offset calculation and speed measurement. In an embodiment of the present invention, the feature extraction processing operation is based on the mold video frame and is implemented through steps such as a convolutional neural network, a region recommendation network, and a similarity measurement network. In the operation monitoring of plastic molds, the use of a deep learning algorithm can significantly improve the accuracy and stability of feature point recognition and ensure the accuracy of subsequent processing.

[0033] Among them, the preliminary feature extraction operation refers to the use of convolutional neural network (CNN) to process the mold video frame and extract the preliminary feature map. Convolutional Neural Network (CNN) is a deep learning model specifically used to process image data. It extracts local features of the image through convolution operations and gradually extracts higher-level features through a multi-layer network structure. The main components of the convolutional neural network include: convolution layer, which performs convolution operations on the input image through convolution kernels (Filter) to extract local features of the image. Each convolution kernel can learn different features, such as edges, textures, etc.; pooling layer, which reduces the size of the feature map through downsampling operations, reduces computational complexity, and enhances the robustness of the features. Commonly used pooling operations include maximum pooling and average pooling; activation function, which introduces nonlinearity and enhances the expression ability of the model; the fully connected layer integrates the features extracted by the convolution layer and the pooling layer, and outputs the final classification or regression results. Through multi-layer convolution and pooling operations, the convolutional neural network can effectively capture local and global features in the image and provide high-quality input data for the subsequent regional recommendation network. The region proposal network processing operation refers to the use of the region proposal network (RPN) to process the preliminary feature map and generate candidate feature regions. The region proposal network (RPN) is a neural network used to generate candidate target regions, which is usually used in target detection tasks. The main components of the region proposal network include: anchor mechanism, RPN generates candidate regions of different scales and aspect ratios by setting multiple anchors (Anchor) on the feature map. Each anchor corresponds to a candidate region. RPN determines whether the region contains the target through classification and regression operations, and adjusts the position and size of the region; the classification branch is used to determine whether the candidate region corresponding to each anchor contains the target region, and the output is a binary classification result (target / non-target); the regression branch is used to adjust the position and size of the candidate region so that it covers the target more accurately, and the output is the offset of the candidate region. The recommendation network can quickly locate the area in the image that may contain feature points through the sliding window and anchor mechanism, and provide candidate regions for subsequent feature point matching. The feature point matching operation refers to the use of a similarity measurement network to process the candidate feature region and identify the feature point data of the mold. Similarity Metric Network is a neural network model used to calculate the similarity between two inputs. It maps the input data into a feature space and calculates the similarity score between the inputs in this space. Similarity Metric Network is widely used in tasks such as image matching, target tracking, and face recognition.The goal of the similarity measurement network is to learn a function that can map input data (such as images, feature vectors, etc.) to a feature space and calculate the similarity between inputs in this space. Similarity is represented by distance metrics (such as Euclidean distance, cosine similarity) or similarity scores. The similarity measurement network can accurately match the feature points of the mold by calculating the similarity between the candidate area and the feature point template library, and output the coordinates and type information of the feature points. The specific steps include: feature extraction, inputting the candidate feature area and the template in the feature point template library into the feature extraction network respectively to obtain the corresponding feature vector; similarity calculation, calculating the similarity score between the feature vector of the candidate feature area and the feature vector of each template in the template library, feature point matching, selecting the template with the highest score as the matching result according to the similarity score, and outputting the coordinates and type information of the feature points.

[0034] In step S14, the position offset is calculated based on the feature point data to obtain the position offset, including: According to the feature point data, coordinate analysis is performed to obtain the reference position feature point coordinates and the target position feature point coordinates; Calculating pixel difference based on the reference position feature point coordinates and the target position feature point coordinates to obtain pixel difference; Calculate the position offset according to the pixel difference and the relationship between the preset pixel and the actual size to obtain the position offset; The calculation formula of the pixel difference is as follows: The calculation formula of the position offset is as follows: in, , The horizontal and vertical coordinates of the feature point coordinates of the reference position, , Represents the horizontal and vertical coordinates of the feature point coordinates of the target position, represents the horizontal axis pixel difference, represents the vertical pixel difference, Indicates the position offset, Indicates the conversion factor between the pixels on the horizontal axis and the actual size. Indicates the conversion factor between the pixels on the vertical axis and the actual size.

[0035] It should be noted that the position offset calculation operation refers to calculating the position change of the plastic mold during operation by analyzing the feature point data. This operation can provide key position change information for subsequent speed calculation and operation status determination. In an embodiment of the present invention, the position offset calculation operation is based on the feature point data and is implemented through steps such as coordinate analysis, pixel difference calculation, and position offset calculation. In the operation monitoring of the plastic mold, the relationship between pixels and actual sizes can be used to convert pixel changes in the image into actual physical size changes, ensuring the accuracy and practicality of the position offset calculation.

[0036] Among them, the coordinate parsing operation refers to extracting the coordinates of the reference position feature point and the target position feature point from the feature point data. These coordinates represent the position changes of the feature points during the operation of the plastic mold, and provide basic data for the subsequent pixel difference calculation. The pixel difference calculation operation refers to calculating the horizontal coordinate pixel difference and the vertical coordinate pixel difference according to the reference position feature point coordinates and the target position feature point coordinates. The position offset calculation operation refers to calculating the actual position offset based on the pixel difference and the preset relationship between the pixel and the actual size. The pixel difference calculation formula is used to calculate the pixel change between the reference position feature point coordinates and the target position feature point coordinates, thereby quantifying the position change of the feature point in the image. The position offset calculation formula is used to convert the pixel difference into the position offset of the actual physical size to ensure that the calculation result can accurately reflect the operating status of the plastic mold.

[0037] In step S15, the speed calculation is performed according to the position offset and the feature point data to obtain the mold running speed, including: According to the feature point data, the feature point data is parsed to obtain the time length corresponding to moving from the reference position to the target position; According to the position offset and the duration, a sliding window average method is used to calculate the speed to obtain the mold running speed; The speed calculation formula is as follows: in, represents the size of the sliding window, Indicates the window The calculated position offset, Indicates the mold running speed, Indicates the window The duration of the calculation.

[0038] It should be noted that the speed calculation operation refers to calculating the running speed of the plastic mold by analyzing the position offset and feature point data. This operation can provide key speed data for determining the running state in the operation monitoring of the plastic mold. In an embodiment of the present invention, the speed calculation operation is based on the position offset and feature point data, and is implemented through steps such as feature point data analysis and sliding window averaging. In the operation monitoring of the plastic mold, the use of the sliding window averaging method can effectively reduce the fluctuation of the speed data and improve the accuracy and stability of the speed measurement.

[0039] Among them, the feature point data analysis operation refers to extracting the time corresponding to moving from the reference position to the target position from the feature point data. These time data represent the movement time of the feature points during the operation of the plastic mold, providing basic data for subsequent speed calculations. The sliding window average method operation refers to calculating the mold operation speed using the sliding window average method based on the position offset and duration. The calculation formula is: in, represents the size of the sliding window, Indicates the window The calculated position offset, Indicates the mold running speed, Indicates the window The duration of the calculation.

[0040] It should be noted that the Sliding Window Average Method is a common technique for smoothing time series data. It averages the data in a fixed-size window to reduce the volatility of the data, thereby improving the stability and reliability of the data. The Sliding Window Average Method is widely used in signal processing, time series analysis, and real-time monitoring. Indicates the number of data points used in each average calculation. It can be adjusted according to specific needs. Larger windows can provide smoother data, but may delay the response to changes; smaller windows can respond to changes faster, but the smoothing effect is poor. The window slides on the time series, sliding one data point each time, and the data points in the window are constantly updated. For each window, the average value of all data points in the window is calculated as the smoothing value at the current moment. As the window slides, new data points enter the window, old data points move out of the window, and the average value is updated accordingly. The sliding window averaging method can effectively reduce the impact of random noise on the data and improve the smoothness of the data by averaging the data in the window. It is suitable for real-time data processing and can dynamically update the average value as new data arrives. The sliding window averaging method can effectively reduce noise and fluctuations and improve the stability and reliability of the data by averaging the data in the window. In the operation monitoring of plastic molds, the sliding window averaging method is used to calculate the mold operation speed, ensure the accuracy and real-time performance of speed measurement, and provide reliable data support for operation status determination and abnormal warning.

[0041] In step S16, the operation state is determined according to the mold operation speed to obtain the mold operation state, including: According to the historical running speed data, the isolation forest model is constructed to obtain the isolation forest model; Calculate the path length according to the isolation forest model and the running speed of the mold to obtain the path length; According to the path length, an average path length is calculated, and an anomaly score is calculated based on the average path length; Performing an abnormality determination according to the abnormality score, and if the abnormality score exceeds a preset threshold, determining that the mold operation state is abnormal; The calculation formula of the anomaly score is as follows: in, Indicates the current sample speed The anomaly score, Representation sample The average path length in the isolation forest, represents the normalization factor, Indicates the sample size.

[0042] It should be noted that the operation state determination operation refers to the abnormal detection of the mold operation speed through the isolation forest algorithm to determine whether the mold operation state is abnormal. This operation can provide key support for abnormal warning. In an embodiment of the present invention, the operation state determination operation is based on the isolation forest model and is implemented through the steps of path length calculation, abnormal score calculation and abnormal determination. In the operation monitoring of plastic molds, the use of the isolation forest algorithm can effectively detect the abnormal state of the operation speed and ensure the accuracy and reliability of the operation state determination.

[0043] Among them, the isolation forest model construction operation refers to building an isolation forest model based on historical running speed data. The isolation forest model divides the sample space into multiple subspaces by randomly selecting features and segmentation values ​​until each subspace contains only one sample or reaches the maximum depth of the tree. The isolation forest model is formed by constructing multiple isolated trees. The path length calculation operation refers to calculating the path length of the sample in each isolated tree based on the isolation forest model and the current mold running speed. The path length indicates the number of edges that the sample passes through from the root node to the leaf node, and abnormal samples usually have a shorter path length. The anomaly score calculation operation refers to calculating the anomaly score based on the average path length of the sample and the normalization factor. The calculation formula is: Among them, the normalization factor The calculation formula is: is the harmonic number, and the calculation formula is: in, It is a summation variable in the harmonic number calculation formula, which is used to traverse each item in the summation formula. The abnormality judgment operation refers to abnormality judgment based on the abnormality score. If the abnormality score exceeds the preset threshold, the mold operation state is judged to be abnormal; otherwise, the mold operation state is judged to be normal. In an embodiment of the present invention, the preset threshold of the abnormality score is a key parameter used by the isolation forest algorithm to determine whether the mold operation state is abnormal. The preset threshold of the abnormality score directly affects the sensitivity and accuracy of abnormality detection. Exemplarily, the abnormality threshold is set to 0.6, which is the 95% quantile of the abnormality score distribution of the historical operating speed data. Of course, the abnormality threshold can also be set to other quantiles such as 90%, 99%, etc., and the present invention is not limited to this.

[0044] In step S17, when it is determined that the mold operation state is abnormal, an early warning signal is generated and sent to the user end.

[0045] It should be noted that the generation and sending of early warning signals is intended to promptly notify users of abnormal conditions in the mold operation state so that appropriate measures can be taken to avoid production accidents or equipment damage. Early warning signals usually contain the following information: abnormal time, the time point when the abnormality occurs; abnormal location, the specific location or equipment number where the abnormality occurs; abnormal speed, the abnormal value of the mold operation speed; abnormal duration, the length of time the abnormal state lasts; abnormal type, the specific type of abnormality (such as too high speed, too low speed, etc.).

[0046] In an embodiment of the present invention, the method further includes: According to the speed and the position offset, a time series curve is drawn to obtain a parameter change trend; According to the parameter change trend, trend prediction is performed to obtain a prediction result; According to the prediction result, if a trend toward an abnormal state is detected, a weak warning signal is generated and sent to the user end.

[0047] It should be noted that the time series curve drawing and trend prediction operation refers to predicting the future change trend of the mold operation status by analyzing the historical data of speed and position offset, and generating a weak warning signal when a potential abnormal trend is detected. This operation can help users discover potential problems in advance and avoid production accidents and equipment damage. In an embodiment of the present invention, the time series curve drawing and trend prediction operation are based on speed and position offset data, and are implemented through steps such as time series analysis, trend prediction and weak warning signal generation. In the operation monitoring of plastic molds, the use of time series analysis and trend prediction technology can significantly improve the initiative and preventiveness of the monitoring system and ensure the stability and reliability of the production process.

[0048] Among them, the time series curve drawing operation refers to drawing the historical data of speed and position offset into a curve graph in chronological order, so as to intuitively observe the change trend of the parameters. The curve graph can be used to analyze the fluctuation of speed and position offset and identify the early signal of abnormal trend. The trend prediction operation refers to predicting the future change trend of speed and position offset by analyzing the time series curve. Commonly used prediction methods include moving average method, slope detection method and machine learning model (such as ARIMA, LSTM, etc.). In the embodiment of the present invention, the prediction method adopted is the slope detection method. Of course, according to different user needs and application scenarios, other prediction methods such as moving average method can also be adopted, and the present invention is not limited to this. Through these methods, it can be judged whether the speed and position offset show an upward, downward or stable trend, and predict the abnormal state that may occur in the future. The weak warning signal generation operation refers to generating a weak warning signal according to the trend prediction result if it is detected that the speed and position offset have a trend of developing towards an abnormal state. The weak warning signal contains information such as prediction time, prediction trend, prediction abnormal type and recommended measures, aiming to remind users to pay attention in advance and take preventive measures.

[0049] Reference Figure 2 The second embodiment of the present invention provides a monitoring system for a plastic mold during operation, comprising: A data acquisition module, used for acquiring mold video stream data through a visual system; An image acquisition module is used to acquire images according to the mold video stream data to obtain a mold video frame; A feature extraction module, used to perform feature extraction processing according to the mold video frame to obtain feature point data; A position calculation module, used to calculate the position offset according to the feature point data to obtain the position offset; A speed calculation module, used to perform speed calculation according to the position offset and the feature point data to obtain the mold running speed; An operation status module, used to determine the operation status according to the operation speed of the mold to obtain the operation status of the mold; The early warning module is used to generate an early warning signal when it is determined that the operating state of the mold is abnormal, and send the early warning signal to the user end.

[0050] It should be noted that a monitoring system for a plastic mold during operation provided by an embodiment of the present invention is used to execute all process steps of a monitoring method for a plastic mold during operation of the above embodiment. The working principles and beneficial effects of the two correspond one to one, and thus will not be elaborated on.

[0051] In summary, in the present invention, the operation process of the plastic mold is monitored in real time through a visual system to obtain the mold video stream data, and then image acquisition is performed according to the mold video stream data to obtain a mold video frame; according to the mold video frame, feature extraction processing is performed through a deep learning algorithm to obtain feature point data; according to the feature point data, the position offset is calculated to obtain the position offset; according to the position offset and time information, the speed is calculated to obtain the mold operation speed; according to the mold operation speed, the operation state is determined by the isolation forest algorithm to obtain the mold operation state; when the mold operation state is determined to be abnormal, an early warning signal is generated and sent to the user end. Wherein, the early warning signal contains information such as the time, location, abnormal speed and duration of the abnormal state, which is convenient for users to respond and process quickly. Through the combination of deep learning algorithm and isolation forest algorithm, the accuracy and stability of plastic mold speed measurement and operation state determination are improved. In addition, by drawing time series curves for key parameters such as plastic molds, glue guns, and injection molding machines, and combining moving average and slope detection algorithms, it is possible to predict parameter change trends in advance and issue a weak warning when an abnormal trend is detected, reminding operation and maintenance personnel to pay attention in advance and avoid potential production accidents.

[0052] The embodiment of the present invention also provides a monitoring device for the operation of a plastic mold. The monitoring device includes: a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a monitoring program for the operation of a plastic mold. When the processor executes the computer program, the steps of the above-mentioned monitoring method embodiments for the operation of a plastic mold are implemented, such as Figure 1 Alternatively, the processor implements the functions of each module / unit in the above-mentioned system embodiments when executing the computer program.

[0053] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, which are used to describe the execution process of the computer program in the terminal device.

[0054] The terminal device may be a computing device such as a desktop computer, a notebook, a PDA, and a smart tablet. The terminal device may include, but is not limited to, a processor and a memory. Those skilled in the art will appreciate that the above components are merely examples of terminal devices and do not constitute a limitation on the terminal device. The terminal device may include more or fewer components than the above components, or may combine certain components, or different components. For example, the terminal device may also include input and output devices, network access devices, buses, etc.

[0055] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the terminal device, and uses various interfaces and lines to connect various parts of the entire terminal device.

[0056] The memory can be used to store the computer program and / or module, and the processor realizes various functions of the terminal device by running or executing the computer program and / or module stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0057] Wherein, if the module / unit integrated in the terminal device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or system capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0058] It should be noted that the system embodiment described above is merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the drawings of the system embodiment provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art may understand and implement it without paying any creative effort.

[0059] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for monitoring a plastic mold during operation, characterized in that: include: Acquire mold video stream data through the visual system; According to the mold video stream data, image acquisition is performed to obtain the mold video frame; Performing feature extraction processing according to the mold video frame to obtain feature point data; Calculating the position offset according to the feature point data to obtain the position offset; Calculate the speed according to the position offset and the feature point data to obtain the mold running speed; According to the mold running speed, the running state is determined to obtain the mold running state; When it is determined that the mold operation state is abnormal, an early warning signal is generated and sent to a user end.

2. The method for monitoring the operation of a plastic mold according to claim 1, characterized in that: The method of performing image acquisition according to the mold video stream data to obtain the mold video frame includes: Extract video frames according to the mold video stream data to obtain initial video frames; Performing denoising processing on the initial video frame to obtain a denoised video frame; Image enhancement processing is performed according to the denoised video frame to obtain a mold video frame.

3. The method for monitoring the operation of a plastic mold according to claim 1, characterized in that: The step of performing feature extraction processing according to the mold video frame to obtain feature point data includes: According to the mold video frame, a convolutional neural network is used to perform preliminary feature extraction to obtain a preliminary feature map; According to the preliminary feature map, a region recommendation network is used to process the map to obtain a candidate feature region; According to the candidate feature regions, a similarity measurement network is used to perform feature point matching to identify feature point data of the mold.

4. The method for monitoring the operation of a plastic mold according to claim 1, characterized in that: The calculating the position offset according to the feature point data to obtain the position offset includes: According to the feature point data, coordinate analysis is performed to obtain the reference position feature point coordinates and the target position feature point coordinates; Calculating pixel difference based on the reference position feature point coordinates and the target position feature point coordinates to obtain pixel difference; Calculate the position offset according to the pixel difference and the relationship between the preset pixel and the actual size to obtain the position offset; The calculation formula of the pixel difference is as follows: The calculation formula of the position offset is as follows: in, , The horizontal and vertical coordinates of the feature point coordinates of the reference position, , Represents the horizontal and vertical coordinates of the feature point coordinates of the target position, represents the horizontal axis pixel difference, represents the vertical pixel difference, Indicates the position offset, Indicates the conversion factor between the pixels on the horizontal axis and the actual size. Indicates the conversion factor between the pixels on the vertical axis and the actual size.

5. The method for monitoring the operation of a plastic mold according to claim 1, characterized in that: The method of performing speed calculation according to the position offset and the feature point data to obtain the mold running speed includes: According to the feature point data, the feature point data is parsed to obtain the time length corresponding to moving from the reference position to the target position; According to the position offset and the duration, a sliding window average method is used to calculate the speed to obtain the mold running speed; The speed calculation formula is as follows: in, represents the size of the sliding window, Indicates the window The calculated position offset, Indicates the mold running speed, Indicates the window The duration of the calculation.

6. The method for monitoring the operation of a plastic mold according to claim 1, characterized in that: The step of determining the operation state according to the operation speed of the mold to obtain the operation state of the mold includes: According to the historical running speed data, the isolation forest model is constructed to obtain the isolation forest model; Calculate the path length according to the isolation forest model and the running speed of the mold to obtain the path length; According to the path length, an average path length is calculated, and an anomaly score is calculated based on the average path length; Performing an abnormality determination according to the abnormality score, and if the abnormality score exceeds a preset threshold, determining that the mold operation state is abnormal; The calculation formula of the anomaly score is as follows: in, Indicates the current sample speed The anomaly score, Representation sample The average path length in the isolation forest, represents the normalization factor, Indicates the sample size.

7. The method for monitoring the operation of a plastic mold according to claim 1, characterized in that: The method further comprises: According to the speed and the position offset, a time series curve is drawn to obtain a parameter change trend; According to the parameter change trend, trend prediction is performed to obtain a prediction result; According to the prediction result, if a trend toward an abnormal state is detected, a weak warning signal is generated and sent to the user end.

8. A monitoring system for a plastic mold during operation, characterized in that: include: A data acquisition module, used for acquiring mold video stream data through a visual system; An image acquisition module is used to acquire images according to the mold video stream data to obtain a mold video frame; A feature extraction module, used to perform feature extraction processing according to the mold video frame to obtain feature point data; A position calculation module, used to calculate the position offset according to the feature point data to obtain the position offset; A speed calculation module, used to perform speed calculation according to the position offset and the feature point data to obtain the mold running speed; An operation status module, used to determine the operation status according to the operation speed of the mold to obtain the operation status of the mold; The early warning module is used to generate an early warning signal when it is determined that the operating state of the mold is abnormal, and send the early warning signal to the user end.

9. A monitoring device for a plastic mold during operation, characterized in that: It comprises a processor, a memory and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, it implements a monitoring method during the operation of a plastic mold as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the monitoring method during the operation of the plastic mold as described in any one of claims 1 to 7.

Images