Quartz tube production quality control method and system
By collecting and analyzing the striking pulse signals of quartz tubes, building feature vectors and using logistic regression models for detection, the intelligent automation and optical characteristics limitation of quartz tube production quality detection is solved, and more accurate detection results and rapid detection are achieved speed.
Patent Information
- Application Number
- CN202411320457.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-23
- Publication Date
- 2024-12-24
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
At this stage, there is a lack of methods and systems for intelligent automated detection of the production quality of quartz tubes, and the optical properties of quartz tubes make it difficult to use machine vision for quality detection.
By using the tapping head to collect the tapping pulse signals at both ends and the middle of the quartz tube, perform filtering processing and power spectrum analysis, build feature vectors, and detect them through the logistic regression quality detection model, and generate 16-dimensional high-order feature vectors for Quality inspection.
Intelligent automated quality detection of quartz tubes is realized, avoiding optical characteristics limitations, more accurate detection results, easy training and rapid detection of logistic regression models, and data set equality improves detection accuracy.
Smart Images
Figure CN119178801A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of intelligent control technology, and in particular to a quartz tube production quality control method and system. Background Art
[0002] The production of quartz tubes requires comprehensive consideration of multiple links including raw material preparation, melting, crystal growth, processing and quality inspection, which involve a variety of production processes. At present, with the development of science and technology, the production control method of quartz tubes is gradually moving towards intelligence and automation. By introducing advanced sensors, machine vision systems and big data analysis technology, real-time monitoring and precise control of process parameters in each link can be achieved. As the last line of defense in the production process, the importance of quality inspection is self-evident. However, at this stage, there is a lack of methods and systems for intelligent and automated detection of the production quality of quartz tubes. On the other hand, due to the optical characteristics of quartz tubes, it is difficult to use machine vision to detect production quality. Summary of the invention
[0003] In view of the above-mentioned technical deficiencies, the purpose of the present invention is to provide a quartz tube production quality control method and system, aiming to solve the problem that there is a lack of methods and systems for intelligent and automated detection of the production quality of quartz tubes at this stage, and due to the optical properties of quartz tubes, it is difficult to use machine vision to detect the production quality of quartz tubes.
[0004] In view of the above problems, the present application provides a quartz tube production quality control method and system.
[0005] The first aspect disclosed in the present application provides a quartz tube production quality control method, the method comprising: moving the produced quartz tube to a detection station through a transfer device; using a knocking head to knock at both ends of the quartz tube and in the middle of the quartz tube, respectively, and using a micro microphone to collect knocking sound pulse signals at these three points; filtering the knocking sound pulse signals at the three points, and then performing power spectrum analysis to construct a feature vector of the current quartz tube. , where t is the average time for the knocking sound pulse signal at three points to decay by 90%, and f is the minimum resonance peak frequency among the three points. is the maximum value of the resonance peak frequency difference among the three points ; The feature vector Perform feature mapping to generate a 16-dimensional high-order feature vector ,in, is a high-order eigenvector The amount will The trained logistic regression quality inspection model is input for quality inspection; the qualified quartz tubes are moved to the packaging station through the transfer device, and the unqualified quartz tubes are removed from the production queue.
[0006] Preferably, the logistic regression quality detection model is calculated using formula (1): (1) Among them, the vector , weight To complete the training of the logistic regression quality detection model parameters, if If it is greater than or equal to 0.5, it is a qualified quartz tube. If it is less than 0.5, it is an unqualified quartz tube.
[0007] Preferably, the training step of the logistic regression quality detection model includes: Collect the feature vectors of 1000 qualified quartz tubes and 1000 unqualified quartz tubes, randomly arrange and combine them, and generate a quartz tube feature vector data set with a capacity of 2000; Perform feature mapping on all feature vectors of the quartz tube feature vector data set, generate 16-dimensional high-order feature vectors and replace the original feature vectors; The quartz tube feature vector data set is input into the initialized logistic regression quality detection model, and the cost function of the logistic regression quality detection model is used for gradient descent, where the cost function is in the form of formula (2): (2) in, represents the high-order feature vector of the i-th quartz tube in the quartz tube feature vector dataset, represents the sample label of the i-th quartz tube in the quartz tube feature vector dataset, is the detection result of the logistic regression quality detection model on the i-th quartz tube in the quartz tube feature vector dataset, , in the current model training phase, the component The parameters of the logistic regression quality detection model to be trained by gradient descent are used to perform gradient descent using formula (3): (3) Among them, a is the learning rate, The cost function is The partial differential of , i=0,1,2,...,16, Indicates updating the assignment, and the gradient is iterated cyclically until the value of the cost function converges, completing the training.
[0008] Preferably, the feature mapping method is: (4) in, Represents a mapping that generates a high-order feature vector.
[0009] The second aspect disclosed in the present application provides a quartz tube production quality control system, which is used for the above-mentioned quartz tube production quality control method, and the system includes: a moving module, which is used to move the produced quartz tube to the inspection station through a transfer device; a collection module, which is used to use a knocking head to knock at both ends of the quartz tube and the middle of the quartz tube, and use a micro microphone to respectively collect the knocking sound pulse signals of these three points; a power spectrum analysis module, which is used to filter the knocking sound pulse signals of the three points, and then perform power spectrum analysis to construct a feature vector of the current quartz tube. , where t is the average time for the knocking sound pulse signal at three points to decay by 90%, and f is the minimum resonance peak frequency among the three points. is the maximum value of the resonance peak frequency difference among the three points ; A quality detection module, the quality detection module is used to convert the feature vector Perform feature mapping to generate a 16-dimensional high-order feature vector ,in, is a high-order eigenvector The amount will The trained logistic regression quality inspection model is input for quality inspection; a sorting module is used to move the qualified quartz tubes to the packaging station through the transfer device, and move the unqualified quartz tubes out of the production queue.
[0010] The third aspect disclosed in the present application provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the following steps when executing the computer program: moving the produced quartz tube to the inspection station through a transfer device; using a knocking head to knock at both ends and the middle of the quartz tube respectively, and using a miniature microphone to collect knocking sound pulse signals at these three points respectively; filtering the knocking sound pulse signals at the three points, and then performing power spectrum analysis to construct a feature vector of the current quartz tube. , where t is the average time for the knocking sound pulse signal at three points to decay by 90%, and f is the minimum resonance peak frequency among the three points. is the maximum value of the resonance peak frequency difference among the three points ; The feature vector Perform feature mapping to generate a 16-dimensional high-order feature vector ,in, is a high-order eigenvector The amount will The trained logistic regression quality inspection model is input for quality inspection; the qualified quartz tubes are moved to the packaging station through the transfer device, and the unqualified quartz tubes are removed from the production queue.
[0011] The fourth aspect disclosed in the present application provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: the produced quartz tube is moved to the inspection station through a transfer device; a knocking head is used to knock at both ends and the middle of the quartz tube respectively, and a miniature microphone is used to collect knocking sound pulse signals at these three points respectively; the knocking sound pulse signals at the three points are filtered and processed, and then a power spectrum analysis is performed to construct a feature vector of the current quartz tube. , where t is the average time for the knocking sound pulse signal at three points to decay by 90%, and f is the minimum resonance peak frequency among the three points. is the maximum value of the resonance peak frequency difference among the three points ; The feature vector Perform feature mapping to generate a 16-dimensional high-order feature vector ,in, is a high-order eigenvector The amount will The trained logistic regression quality inspection model is input for quality inspection; the qualified quartz tubes are moved to the packaging station through the transfer device, and the unqualified quartz tubes are removed from the production queue.
[0012] The fifth aspect disclosed in the present application provides a computer program product, including a computer program or an instruction, which implements the following steps when executed by a processor: moving the produced quartz tube to a detection station through a transfer device; using a knocking head to knock at both ends and the middle of the quartz tube respectively, and using a miniature microphone to collect knocking sound pulse signals at these three points respectively; filtering the knocking sound pulse signals at the three points, and then performing power spectrum analysis to construct a feature vector of the current quartz tube. , where t is the average time for the knocking sound pulse signal at three points to decay by 90%, and f is the minimum resonance peak frequency among the three points. is the maximum value of the resonance peak frequency difference among the three points ; The feature vector Perform feature mapping to generate a 16-dimensional high-order feature vector ,in, is a high-order eigenvector The amount will The trained logistic regression quality inspection model is input for quality inspection; the qualified quartz tubes are moved to the packaging station through the transfer device, and the unqualified quartz tubes are removed from the production queue.
[0013] The beneficial effects of the present invention are: 1) It solves the current lack of methods and systems for intelligent and automated detection of the production quality of quartz tubes; 2) By analyzing the knocking sound of the quartz tube for production quality detection, the problem of difficulty in using machine vision to detect production quality due to the optical properties of the quartz tube is avoided. 3) By mapping the feature vector of the quartz tube to a higher dimension through feature mapping, the quality of the quartz tube can be detected from a higher dimensional vector space, making the detection result more accurate. 4) The proposed logistic regression quality detection model is easy to train, has a fast convergence speed, and a fast detection speed. 5) The constructed quartz tube feature vector data set satisfies the balance of positive and negative samples, which can make the detection results of the trained model more accurate. BRIEF DESCRIPTION OF THE DRAWINGS
[0014] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0015] Figure 1 This is the overall flow chart of the quartz tube production quality control method.
[0016] Figure 2 This is the overall structure diagram of the quartz tube production quality control system. DETAILED DESCRIPTION
[0017] The embodiments of the present application provide a quartz tube production quality control method and system, thereby solving the problem that there is currently a lack of methods and systems for intelligent and automated detection of the production quality of quartz tubes, and that due to the optical properties of quartz tubes, it is difficult to use machine vision to detect the production quality of quartz tubes.
[0018] 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.
[0019] like Figure 1 As shown, the embodiment of the present application provides a quartz tube production quality control method, the method comprising: Move the produced quartz tube to the inspection station through the transfer device; Use a knocking head to knock at both ends and the middle of the quartz tube respectively, and use a micro microphone to collect the knocking sound pulse signals at these three points respectively; Specifically, the knocking device consists of an epoxy resin knocking head and an electromagnet driver, the sound collection consists of a miniature microphone and a sound card, and the collected knocking sound pulse signal is processed by a computer.
[0020] The knocking sound pulse signals at the three points are filtered and then subjected to power spectrum analysis to construct the characteristic vector of the current quartz tube. , where t is the average time for the knocking sound pulse signal at three points to decay by 90%, and f is the minimum resonance peak frequency among the three points. is the maximum value of the resonance peak frequency difference among the three points ; Specifically, MATLAB is used to execute the percussion sound collection and processing program, and the andlogin put() function is used to collect analog sound signals. The collection frequency is 22050Hz and the number of collection points is 512. After the data is collected, it is first filtered and then subjected to power spectrum analysis.
[0021] The feature vector Perform feature mapping to generate a 16-dimensional high-order feature vector ,in, is a high-order eigenvector The amount will Input the trained logistic regression quality detection model for quality detection; Specifically, by mapping the feature vector of the quartz tube to a higher dimension through feature mapping, the quality of the quartz tube can be detected from a higher dimensional vector space, making the detection result more accurate.
[0022] The qualified quartz tubes are moved to the packaging station through the transfer device, and the unqualified quartz tubes are removed from the production queue.
[0023] Furthermore, the logistic regression quality detection model is calculated using formula (1): (1) Among them, the vector , weight To complete the training of the logistic regression quality detection model parameters, the function Then change x from Projected into the interval [0,1], and the function graph is symmetric about the point (0,0.5), if If it is greater than or equal to 0.5, it is a qualified quartz tube. If it is less than 0.5, it is an unqualified quartz tube.
[0024] Furthermore, the training steps of the logistic regression quality detection model include: Collect the feature vectors of 1000 qualified quartz tubes and 1000 unqualified quartz tubes, randomly arrange and combine them, and generate a quartz tube feature vector data set with a capacity of 2000; Perform feature mapping on all feature vectors of the quartz tube feature vector data set, generate 16-dimensional high-order feature vectors and replace the original feature vectors; The quartz tube feature vector data set is input into the initialized logistic regression quality detection model, and the cost function of the logistic regression quality detection model is used for gradient descent, where the cost function is in the form of formula (2): (2) in, represents the high-order feature vector of the i-th quartz tube in the quartz tube feature vector dataset, represents the sample label of the i-th quartz tube in the quartz tube feature vector dataset, represents true, represents false, and is used to represent qualified quartz tubes and unqualified quartz tubes respectively. is the detection result of the logistic regression quality detection model on the i-th quartz tube in the quartz tube feature vector dataset, , in the current model training phase, the component is the parameter of the logistic regression quality detection model to be trained by gradient descent. The purpose of training is to make the detection result of the logistic regression quality detection model for the quartz tube in the quartz tube feature vector data set as close as possible to the corresponding sample label. Gradient descent is performed using formula (3): (3) Among them, a is the learning rate, The cost function is The partial differential of , i=0,1,2,...,16, Indicates updating the assignment, and the gradient is iterated cyclically until the value of the cost function converges, completing the training.
[0025] Specifically, define , combining equation (1) and equation (2), equation (2) can be simplified to the form of equation (4): (4) At this time, equation (4) is used to By calculating the partial differential of , we can get the form of formula (5): (5) Where j = [0, 1, 2, ... , 16], substituting equation (5) into equation (3), we can get the update formula of gradient descent: (6) When the logistic regression quality detection model starts training, it is randomly sampled from a Gaussian distribution N(0,0.01) with a standard deviation of 0.01 and a mean of 0, and the parameters are Initialize and set the learning rate a to 0.0001, use formula (6) to perform gradient descent, and iteratively update the parameters until the value of the cost function converges.
[0026] Furthermore, the feature mapping method is: (7) in, Represents a mapping that generates a high-order feature vector.
[0027] In summary, the quartz tube production quality control method provided in the embodiment of the present application has the following technical effects: 1) It solves the current lack of intelligent and automated detection methods and systems for the production quality of quartz tubes; 2) Production quality inspection is performed by analyzing the knocking sound of the quartz tube, avoiding the problem of using machine vision to perform production quality inspection due to the optical characteristics of the quartz tube.
[0028] 3) By mapping the feature vector of the quartz tube to a higher dimension through feature mapping, the quality of the quartz tube can be detected from a higher dimensional vector space, making the detection result more accurate.
[0029] 4) The proposed logistic regression quality detection model is easy to train, converges quickly, and has a fast detection speed.
[0030] 5) The constructed quartz tube feature vector dataset satisfies the balance of positive and negative samples, which can make the detection results of the trained model more accurate.
[0031] Based on the same inventive concept as the quartz tube production quality control method in the above-mentioned embodiment, Figure 2 As shown, the present application provides a quartz tube production quality control system, the system comprising: A moving module, which is used to move the produced quartz tube to the inspection station through a transfer device; A collection module, wherein the collection module is used to use a knocking head to knock at both ends of the quartz tube and the middle of the quartz tube respectively, and use a micro microphone to collect knocking sound pulse signals at these three points respectively; The power spectrum analysis module is used to filter the knocking sound pulse signals at the three points, and then perform power spectrum analysis to construct the characteristic vector of the current quartz tube. , where t is the average time for the knocking sound pulse signal at three points to decay by 90%, and f is the minimum resonance peak frequency among the three points. is the maximum value of the resonance peak frequency difference among the three points ; A quality detection module, the quality detection module is used to convert the feature vector Perform feature mapping to generate a 16-dimensional high-order feature vector ,in, is a high-order eigenvector The amount will Input the trained logistic regression quality detection model for quality detection; The sorting module is used to move the qualified quartz tubes to the packaging station through the transfer device, and move the unqualified quartz tubes out of the production queue.
[0032] Through the above detailed description of a quartz tube production quality control method in this specification, those skilled in the art can clearly understand a quartz tube production quality control system in this embodiment. Since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.
[0033] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the following steps when executing the computer program: moving the produced quartz tube to the inspection station through a transfer device; using a knocking head to knock at both ends and the middle of the quartz tube respectively, and using a miniature microphone to collect knocking sound pulse signals at these three points respectively; filtering the knocking sound pulse signals at the three points, and then performing power spectrum analysis to construct a feature vector of the current quartz tube , where t is the average time for the knocking sound pulse signal at three points to decay by 90%, and f is the minimum resonance peak frequency among the three points. is the maximum value of the resonance peak frequency difference among the three points ; The feature vector Perform feature mapping to generate a 16-dimensional high-order feature vector ,in, is a high-order eigenvector The amount will The trained logistic regression quality inspection model is input for quality inspection; the qualified quartz tubes are moved to the packaging station through the transfer device, and the unqualified quartz tubes are removed from the production queue.
[0034] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: moving the produced quartz tube to the inspection station through a transfer device; using a knocking head to knock at both ends of the quartz tube and in the middle of the quartz tube, respectively, and using a miniature microphone to collect knocking sound pulse signals at these three points; filtering the knocking sound pulse signals at the three points, and then performing power spectrum analysis to construct a feature vector of the current quartz tube. , where t is the average time for the knocking sound pulse signal at three points to decay by 90%, and f is the minimum resonance peak frequency among the three points. is the maximum value of the resonance peak frequency difference among the three points ; The feature vector Perform feature mapping to generate a 16-dimensional high-order feature vector ,in, is a high-order eigenvector The amount will The trained logistic regression quality inspection model is input for quality inspection; the qualified quartz tubes are moved to the packaging station through the transfer device, and the unqualified quartz tubes are removed from the production queue.
[0035] In one embodiment, a computer program product is provided, including a computer program or an instruction, which implements the following steps when executed by a processor: moving the produced quartz tube to a detection station through a transfer device; using a knocking head to knock at both ends of the quartz tube and in the middle of the quartz tube, respectively, and using a miniature microphone to collect knocking sound pulse signals at these three points; filtering the knocking sound pulse signals at the three points, and then performing power spectrum analysis to construct a feature vector of the current quartz tube. , where t is the average time for the knocking sound pulse signal at three points to decay by 90%, and f is the minimum resonance peak frequency among the three points. is the maximum value of the resonance peak frequency difference among the three points ; The feature vector Perform feature mapping to generate a 16-dimensional high-order feature vector ,in, is a high-order eigenvector The amount will The trained logistic regression quality inspection model is input for quality inspection; the qualified quartz tubes are moved to the packaging station through the transfer device, and the unqualified quartz tubes are removed from the production queue.
[0036] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0037] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A quartz tube production quality control method, characterized in that: The method comprises: Move the produced quartz tube to the inspection station through the transfer device; Use a knocking head to knock at both ends and the middle of the quartz tube respectively, and use a micro microphone to collect the knocking sound pulse signals at these three points respectively; The knocking sound pulse signals at the three points are filtered and then subjected to power spectrum analysis to construct the characteristic vector of the current quartz tube. , where t is the average time for the knocking sound pulse signal at three points to decay by 90%, and f is the minimum resonance peak frequency among the three points. is the maximum value of the resonance peak frequency difference among the three points ; The feature vector Perform feature mapping to generate a 16-dimensional high-order feature vector ,in, is a high-order eigenvector The amount will Input the trained logistic regression quality detection model for quality detection; The qualified quartz tubes are moved to the packaging station through the transfer device, and the unqualified quartz tubes are removed from the production queue.
2. A quartz tube production quality control method as claimed in claim 1, characterized in that: The logistic regression quality detection model is calculated using formula (1): (1) Among them, the vector , weight To complete the training of the logistic regression quality detection model parameters, if If it is greater than or equal to 0.5, it is a qualified quartz tube. If it is less than 0.5, it is an unqualified quartz tube.
3. A quartz tube production quality control method as claimed in claim 1, characterized in that: The training steps of the logistic regression quality detection model include: Collect the feature vectors of 1000 qualified quartz tubes and 1000 unqualified quartz tubes, randomly arrange and combine them, and generate a quartz tube feature vector data set with a capacity of 2000; Perform feature mapping on all feature vectors of the quartz tube feature vector data set, generate 16-dimensional high-order feature vectors and replace the original feature vectors; The quartz tube feature vector data set is input into the initialized logistic regression quality detection model, and the cost function of the logistic regression quality detection model is used for gradient descent, where the cost function is in the form of formula (2): (2) in, represents the high-order feature vector of the i-th quartz tube in the quartz tube feature vector dataset, represents the sample label of the i-th quartz tube in the quartz tube feature vector dataset, is the detection result of the logistic regression quality detection model on the i-th quartz tube in the quartz tube feature vector dataset, , in the current model training phase, the component The parameters of the logistic regression quality detection model to be trained by gradient descent are used to perform gradient descent using formula (3): (3) Among them, a is the learning rate, The cost function is The partial differential of , i=0,1,2,...,16, Indicates updating the assignment, and the gradient is iterated cyclically until the value of the cost function converges, completing the training.
4. A quartz tube production quality control method as claimed in claim 1 or 3, characterized in that: The feature mapping method is: (4) in, Represents a mapping that generates a high-order feature vector.
5. A quartz tube production quality control system, characterized in that: include: A moving module, which is used to move the produced quartz tube to the inspection station through a transfer device; A collection module, wherein the collection module is used to use a knocking head to knock at both ends of the quartz tube and the middle of the quartz tube respectively, and use a micro microphone to collect knocking sound pulse signals at these three points respectively; The power spectrum analysis module is used to filter the knocking sound pulse signals at the three points, and then perform power spectrum analysis to construct the characteristic vector of the current quartz tube. , where t is the average time for the knocking sound pulse signal at three points to decay by 90%, and f is the minimum resonance peak frequency among the three points. is the maximum value of the resonance peak frequency difference among the three points ; A quality detection module, the quality detection module is used to convert the feature vector Perform feature mapping to generate a 16-dimensional high-order feature vector ,in, is a high-order eigenvector The amount will Input the trained logistic regression quality detection model for quality detection; The sorting module is used to move the qualified quartz tubes to the packaging station through the transfer device, and move the unqualified quartz tubes out of the production queue.
6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of a quartz tube production quality control method according to any one of claims 1 to 4 are implemented.
7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a quartz tube production quality control method according to any one of claims 1 to 4 are implemented.
8. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the steps of a quartz tube production quality control method described in any one of claims 1 to 4 are implemented.