A machine vision-based defect recognition and analysis system for cord fabric production

Abstract

The present invention belongs to the field of image processing technology, and specifically is a cord fabric production defect recognition and analysis system based on machine vision, which includes a multispectral imaging module, an image preprocessing module, a target detection module, a result verification module, a weight correction module and a control response module. By collecting multispectral image sequences of each partition of the cord fabric forming surface, the preprocessing accuracy is improved by combining motion compensation, geometric calibration and noise filtering. Adaptive weight fusion is used to generate a composite spectral image based on considerations of image signal-to-noise ratio and texture clarity. Defects are detected based on dual-modal matching of frequency domain texture and morphological gradient parameters. The detection credibility is further verified by the enhancement of defect-sensitive spectral image features, the fusion weight is dynamically corrected and iteratively optimized, and finally the production control instruction is triggered according to the detection result to realize closed-loop management of high-precision defect detection of cord fabric and linkage of production control.

Description

Technical Field

[0001] The present invention belongs to the technical field of image processing, and in particular relates to a cord fabric production defect recognition and analysis system based on machine vision. Background Art

[0002] As the core structural material of tires, the quality of tire cord directly impacts tire strength, durability, and safety. As the global tire industry transitions toward lightweighting and high reliability, defect detection in tire cord production has become a key technology for improving quality.

[0003] Traditional manual inspection is limited by low efficiency, strong subjectivity, and a high rate of missed detections, making it difficult to adapt to the demands of modern industrial automation. A machine vision-based defect recognition and analysis system, integrating high-precision image acquisition, deep learning algorithms, and real-time data processing, provides an efficient and accurate intelligent quality control solution for cord fabric production.

[0004] There are already a large number of machine vision-based defect recognition solutions in the existing technology that can be directly applied to the cord fabric production process. For example, China Patent Publication No. CN117115147B is a machine vision-based textile detection method and system. It is based on textile images and uses convolutional neural networks to extract features to generate a preliminary recognition library. It combines recurrent neural networks, multispectral image acquisition, computer vision, blockchain and other technologies to form a detection system. It can detect flaws and hidden defects with high accuracy, realize real-time quality monitoring and adjustment, and build a reliable traceability system with the help of blockchain and the Internet of Things.

[0005] Another Chinese patent, CN115294097A, is a machine vision-based textile surface defect detection method that obtains the grayscale information of the textile image, determines the window shape and size based on the template image, determines the adaptive weighted denoising mean through the grayscale and pixel distance information, and calculates the denoised data to achieve accurate detection of textile surface defects.

[0006] The above solution shows that existing technologies can generate fused images through multispectral image acquisition and denoising to support defect detection, but there are still significant technical bottlenecks in the dynamic production scenario of cord fabric. First, during the high-speed continuous production of cord fabric, the operation of the loom and the movement of the conveyor belt cause image acquisition to face multi-source interference coupling, including spatial deformation caused by motion blur, registration deviation between multispectral imaging modules due to mechanical vibration or timing asynchrony, and superimposed pollution from loom vibration noise. Due to the lack of dynamic perception optimization capabilities, the image preprocessing in existing technologies is difficult to achieve spatiotemporal synchronization restoration of multimodal data under complex working conditions, resulting in structural distortion and signal-to-noise ratio degradation in the fused image, seriously affecting the baseline data quality of defect characteristics.

[0007] Second, existing technologies lack a mechanism for verifying the reliability of multispectral fusion results. Fusion weight assignment relies on static parameters calibrated based on historical experience, which can easily lead to the suppression of the significance of key defects in specific spectral dimensions by unbalanced fusion weights, resulting in attenuation of characteristic information about micro-defects. Furthermore, the lack of multispectral reverse verification and confidence assessment of fused detection results increases the misjudgment rate of false defects against complex textures in cord fabric, creating a systematic conflict between detection accuracy and generalizability. Summary of the Invention

[0008] In order to overcome the shortcomings of the background technology, an embodiment of the present invention provides a cord fabric production defect recognition and analysis system based on machine vision, which can effectively solve the problems involved in the above background technology.

[0009] The purpose of the present invention can be achieved through the following technical solutions: a cord fabric production defect recognition and analysis system based on machine vision, including: a multispectral imaging module, an image preprocessing module, a target detection module, a result verification module, a weight correction module and a control response module.

[0010] The multispectral imaging module is connected to the image preprocessing module, the image preprocessing module is connected to the target detection module, the target detection module is connected to the result verification module, the result verification module is respectively connected to the weight correction module and the control response module, and the weight correction module is connected to the target detection module.

[0011] The multispectral imaging module consists of a linear array of industrial cameras with integrated dichroic mirrors and adaptive fill light sources. Each camera in the array covers a rectangular monitoring area of ​​the cord fabric forming surface at the loom outlet according to a preset field of view mapping relationship and simultaneously collects a multispectral image sequence of the covered area.

[0012] The image preprocessing module performs preprocessing operations including motion compensation registration and noise filtering on the multispectral image sequences of each subarea.

[0013] The target detection module performs weighted fusion on the preprocessed multispectral image sequences of each partition to generate a composite spectral image. Based on the multi-scale texture features of the composite spectral image, it outputs the target detection results including the number, type and location coordinates of defects in each partition.

[0014] The result verification module associates the mapping relationship between the defect type and the corresponding sensitive spectral channel, and verifies the credibility of the composite spectral image target detection result based on the feature enhancement degree of each partition target defect type in the sensitive spectral channel image. If the verification is credible, the control response module is executed; otherwise, the weight correction module is executed.

[0015] The weight correction module corrects the fusion weight of the multispectral image sequence according to the feature enhancement deviation and feeds it back to the target detection module to regenerate the composite spectral image.

[0016] The control response module sends the target detection results to the production control terminal, and the terminal responds with the corresponding control instructions.

[0017] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects: (1) The present invention collects multispectral image sequences of each partition of the tire cord fabric forming surface, combines motion compensation, geometric calibration and noise filtering to improve preprocessing accuracy, generates a composite spectral image and performs defect target detection, verifies the detection credibility based on the target detection result by the feature enhancement degree of the sensitive spectral image where the defect type is located, and dynamically performs fusion weight correction, and finally triggers production control instructions according to the detection results, thereby realizing closed-loop management of high-precision defect detection and production control linkage.

[0018] (2) The preprocessing module of the present invention adopts motion compensation, sub-pixel geometric calibration and cross-channel noise filtering to repair the multi-source interference of the multispectral image caused by loom vibration and motion blur, ensuring the internal spatiotemporal alignment of the multispectral image sequence, thereby effectively avoiding the structural distortion and signal-to-noise ratio degradation of the subsequent composite spectral image.

[0019] (3) The present invention detects defects based on dual-modal feature matching of frequency domain texture and morphological gradient parameters of composite spectral images, covering both the frequency domain characteristics and spatial structure characteristics of defects, thereby effectively avoiding the limitations of a single feature mode and significantly enhancing the ability to identify micro-defects in a complex texture background, thereby reducing the missed detection rate.

[0020] (4) The present invention establishes a defect type-sensitive spectral channel mapping relationship and realizes reverse verification of the credibility of the composite spectral image detection result based on the feature enhancement degree of the sensitive spectral channel image where the defect type is located. If the verification is unreliable, a forward or reverse adjustment factor correction weight is further constructed based on the deviation ratio to perform closed-loop iterative optimization, suppress non-sensitive channel interference, reduce false defect misjudgment, and improve detection generalization. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The present invention is further described with reference to the accompanying drawings. However, the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative effort.

[0022] Figure 1 This is a schematic diagram of module connection of the present invention.

[0023] Figure 2 This is a schematic structural diagram of the multi-spectral dynamic imaging array for cord fabrics of the present invention.

[0024] Figure 3 A logic diagram is obtained for the target detection results in the target detection module of the present invention.

[0025] Reference numerals: 1. industrial camera; 2. adaptive fill light source; 3. loom; 4. cord fabric. DETAILED DESCRIPTION

[0026] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making creative efforts are within the scope of protection of the present invention.

[0027] Reference Figure 1 As shown, the present invention provides a cord fabric production defect recognition and analysis system based on machine vision, including: a multispectral imaging module, an image preprocessing module, a target detection module, a result verification module, a weight correction module and a control response module.

[0028] The multispectral imaging module is connected to the image preprocessing module, the image preprocessing module is connected to the target detection module, the target detection module is connected to the result verification module, the result verification module is respectively connected to the weight correction module and the control response module, and the weight correction module is connected to the target detection module.

[0029] Reference Figure 2 As shown, the multispectral imaging module consists of a linear array of industrial cameras with integrated dichroic mirrors and adaptive fill light sources. Each camera in the array covers a rectangular monitoring partition of the cord fabric forming surface at the loom outlet according to a preset field of view mapping relationship and synchronously collects a multispectral image sequence of the covered partition.

[0030] The image preprocessing module performs preprocessing operations including motion compensation registration and noise filtering on the multispectral image sequences of each subarea.

[0031] In a preferred embodiment of the present invention, the preprocessing operation refers to the following execution process: (a) based on the fixed transmission speed of the tire cord fabric and the exposure response time of the camera, the displacement of the tire cord fabric at the camera sampling moment is determined, and the multispectral image sequence of each partition is reversely translated to eliminate motion blur.

[0032] It should be added that the displacement of the cord fabric at the above-mentioned camera sampling moment is the product of the fixed transmission speed of the cord fabric and the exposure response time of the camera.

[0033] (b) Performing intra-group geometric calibration on the multispectral image sequence of the same partition, including: establishing an affine transformation matrix between spectral channels based on the optical properties of the dichroic mirror, performing global geometric correction on the images of each spectral channel in the same partition, converting the corrected images of each spectral channel to the frequency domain, obtaining the sub-pixel shift between each spectral channel through the phase correlation method, and implementing sub-pixel shift compensation for each spectral channel image using an interpolation algorithm.

[0034] It should be added that the spectroscopic characteristics of the dichroic mirror can lead to differences in the optical paths of different spectral channels, which manifest as translation, rotation, or scaling between images. Using the pre-set optical design parameters of the dichroic mirror, an affine transformation matrix is ​​established between the spectral channels and applied to each spectral channel image one by one to correct the geometric distortion caused by the optical path differences. The affine transformation matrix includes a scaling factor to eliminate differences in optical path magnification, a shear factor to eliminate optical axis tilt, and a translation to eliminate optical path offset.

[0035] The above-mentioned method of obtaining the sub-pixel shift between each spectral channel through the phase correlation method specifically refers to: after each spectral channel image is converted into the frequency domain, the spectrum information is obtained by Fourier transform, the cross-power spectrum between the two spectral channel images is calculated based on the spectrum information, the cross-power spectrum is inverse Fourier transformed to obtain the impulse response function, the peak position of the impulse response function is retrieved and Gaussian fitting is performed simultaneously to quantify the sub-pixel shift between the two spectral channel images, thereby obtaining the sub-pixel shift between each spectral channel.

[0036] The above interpolation algorithm specifically refers to a bicubic interpolation algorithm.

[0037] (c) The grayscale distribution of the background area of ​​each spectral channel image in the same partition is extracted, the noise distribution characteristics of the multispectral channel are analyzed and the noise probability density function is fitted. A dynamic threshold is set to distinguish the true features of the image from abnormal noise points, and then isolated noise points caused by the mechanical vibration of the loom are filtered out based on the cross-channel signal correlation.

[0038] The preprocessing module of the present invention adopts motion compensation, sub-pixel geometric calibration and cross-channel noise filtering to repair the multi-source interference of multispectral images caused by loom vibration and motion blur, ensure the internal spatiotemporal alignment of the multispectral image sequence, and thus effectively avoid the structural distortion and signal-to-noise ratio degradation of the subsequent composite spectral images.

[0039] The target detection module performs weighted fusion on the preprocessed multispectral image sequences of each partition to generate a composite spectral image, and outputs target detection results including the number, type and location coordinates of defects in each partition based on the multi-scale texture features of the composite spectral image.

[0040] In a preferred embodiment of the present invention, the weighted fusion processing of the multispectral image sequence adopts an adaptive weight allocation method, and the fusion weight coefficient corresponding to each spectral channel image uses its predefined benchmark weight as the initial parameter, and is dynamically adjusted in combination with the image signal-to-noise ratio and texture clarity obtained in real time, wherein the predefined benchmark weight is determined through an offline calibration experiment based on the intrinsic spectral reflectance characteristics of the cord fabric material.

[0041] It should be added that the specific process of obtaining the above-mentioned image signal-to-noise ratio and texture clarity is as follows: select a preset pixel size window in the texture area of ​​the spectral channel image and calculate the grayscale mean of the window, select a texture-free background area of ​​the spectral channel image and calculate the grayscale standard deviation of the area, and substitute the grayscale mean and the grayscale standard deviation into the standard SNR formula to obtain the image signal-to-noise ratio of the spectral channel image.

[0042] The Sobel operator is used pixel by pixel on the spectral channel image to obtain the horizontal gradient and vertical gradient of each pixel and calculate the gradient amplitude based on them. The texture clarity is defined as the sum of the squares of the gradient amplitudes of all pixels in the spectral channel image.

[0043] The specific process of obtaining the corresponding fusion weight coefficients of the above spectral channel images is as follows: normalize the image signal-to-noise ratio and texture clarity of each spectral channel image so that the value is Within the interval, the image signal-to-noise ratio and the texture clarity corresponding adjustment factors are set to control the contribution ratio of the image signal-to-noise ratio and the texture clarity, the image signal-to-noise ratio and its adjustment factor are multiplied, and the product of the texture clarity and its adjustment factor is added, and the product of the calculation result and the predefined reference weight of the spectral channel image is used as the corresponding fusion weight coefficient of the spectral channel image. Finally, the corresponding fusion weight coefficient of each spectral channel image still needs to be normalized, wherein the setting of the image signal-to-noise ratio and texture clarity corresponding adjustment factors is based on the weighted fusion focus target of the multispectral image sequence. If the emphasis is on noise suppression, the adjustment factor assigned to the image signal-to-noise ratio is greater than the adjustment factor of the texture clarity, which can be exemplified as 0.7 or 0.3. If the emphasis is on detail retention, the adjustment factor assigned to the image signal-to-noise ratio is less than the adjustment factor of the texture clarity, which can be exemplified as 0.3 or 0.7. The sum of the adjustment factors of the image signal-to-noise ratio and the texture clarity is always required to be 1.

[0044] Reference Figure 3 As shown, in a preferred embodiment of the present invention, the target detection result refers to the following acquisition process: S1. Divide the composite spectral image into several image units, extract the frequency domain texture parameter set and morphological gradient parameter set of each image unit, and organize them into a texture parameter matrix and a morphological parameter matrix respectively.

[0045] It should be added that the above-mentioned frequency domain texture parameter set includes at least high-frequency sub-band energy, low-frequency sub-band energy, sub-band entropy, directional response mean and directional response variance, and the morphological gradient features include at least the multi-scale gradient mean, gradient extreme value density, directional gradient discreteness and longitudinal / latitudinal gradient concentration of each image unit.

[0046] S2. Perform similarity matching between the texture parameter matrix of each image unit and the pre-stored standard texture parameter interval matrix of each defect type of the cord fabric to generate a texture similarity score, and simultaneously perform similarity matching between the morphological parameter matrix and the standard morphological parameter interval matrix to generate a morphological similarity score.

[0047] S3. Calculate the product value of the bimodal similarity score between each image unit and each defect type. If the product value between an image unit and a specific defect type exceeds a preset value, mark the image unit as the corresponding defect type label.

[0048] S4. Perform a connected domain search on adjacent image units with the same defect type label, merge them to form the same defect area, and use the combination of the vertex coordinates and the centroid coordinates of the minimum circumscribed rectangle of the defect area as the defect position coordinates, so as to output the defect number, defect type and position coordinates of the corresponding partition of the composite spectral image.

[0049] In a preferred embodiment of the present invention, the similarity matching in step S3 refers to the following process: assigning a preset benchmark matching score to each element in the parameter matrix.

[0050] The numerical value of each element in the parameter matrix is ​​compared with the numerical interval of the element at the same position in the corresponding standard parameter interval matrix. If the element value is within the corresponding numerical interval, the benchmark matching score of the element is retained. If it exceeds the numerical interval, the score correction coefficient is calculated based on the deviation between the element value and the interval boundary, and the benchmark matching score is reduced according to the correction coefficient.

[0051] It should be noted that the above-mentioned calculation of the score correction coefficient based on the deviation between the element value and the interval boundary includes the following process: if the element value is less than the lower limit of the interval, the absolute difference between the element value and the lower limit of the interval and the ratio of the same to the lower limit of the interval are used as the deviation; if the element value is greater than the lower limit of the interval, the difference between the element value and the upper limit of the interval and the ratio of the same to the upper limit of the interval are used as the deviation, and the deviation is substituted into the standard exponential attenuation function with the control attenuation factor set to 1 to obtain the score correction coefficient.

[0052] According to the predefined defect type weight rule, the weight coefficient of each defect type of the cord fabric for each element in the parameter matrix is ​​obtained, and the matching score of each element in the parameter matrix is ​​weighted and summed according to the weight to generate the similarity score.

[0053] The parameter matrix includes a texture parameter matrix or a morphological parameter matrix, and the standard parameter interval matrix includes a corresponding standard texture parameter interval matrix or a standard morphological parameter interval matrix.

[0054] The embodiment of the present invention detects defects based on dual-modal feature matching of frequency domain texture and morphological gradient parameters of composite spectral images, covering the frequency domain characteristics and spatial structure characteristics of defects, thereby effectively avoiding the limitations of a single feature mode and significantly enhancing the ability to recognize micro-defects in a complex texture background, thereby reducing the missed detection rate.

[0055] The result verification module associates the mapping relationship between the defect type and the corresponding sensitive spectral channel, verifies the credibility of the composite spectral image target detection result based on the feature enhancement degree of each partition target defect type in the sensitive spectral channel image, and executes the control response module if the verification is credible; otherwise, the weight correction module is executed.

[0056] In a preferred embodiment of the present invention, the feature enhancement degree of each partitioned target defect type in the sensitive spectral channel image refers to the following acquisition process: based on the weight coefficient of each defect type of the cord fabric to each element in the parameter matrix, the weight coefficient of the target defect type to each element in the parameter matrix is ​​extracted, and the elements whose weight coefficient exceeds the average element weight coefficient of the parameter matrix are screened as the enhancement elements of the target defect type in the parameter matrix, so as to retrieve the enhancement elements of each partitioned target defect type in the texture parameter matrix and the morphological parameter matrix.

[0057] It should be added that the various strengthening elements of the above-mentioned target defect types in the texture parameter matrix and the morphology parameter matrix can be exemplified by referring to Table 1.

[0058] Table 1 Mapping relationship between strengthening elements and defect types

[0059]

[0060] The texture parameter matrix and morphological parameter matrix of the target defect position area in each spectral channel image of each partition are obtained, the values ​​of each strengthening element in the matrix are retrieved and the degree of defect feature manifestation in the strengthening direction is calculated, and the average is taken to obtain the characteristic manifestation degree of the target defect position area in each spectral channel image of each partition.

[0061] The above retrieval matrix calculates the degree of defect feature manifestation in the reinforcement direction, which can be explained by taking the high-frequency sub-band energy of the strong twist defect in the texture parameter matrix in the above table as an example: when a strong twist defect occurs in the cord fabric, its spiral texture will lead to an enhancement of the high-frequency energy, and the direction of reinforcement of this value should be positive increase. The relative increase ratio of the high-frequency sub-band energy value in the matrix to the preset high-frequency sub-band energy standard value under the normal state of the cord fabric is used as the degree of defect feature manifestation in the corresponding reinforcement direction.

[0062] If we take the low-frequency sub-band energy of the hole defect in the texture parameter matrix in the above table as an example to explain: when a hole defect appears in the cord fabric, the hole causes the low-frequency energy of the light-transmitting area to decrease, then the direction of numerical enhancement should be the opposite decrease, and the relative reduction ratio of the low-frequency sub-band energy value in the matrix to the preset low-frequency sub-band energy standard value under the normal state of the cord fabric is used as the degree of defect feature manifestation in the corresponding enhancement direction.

[0063] The preset high- and low-frequency sub-band energy standard values ​​under normal conditions of the cord fabric can be based on the testing requirements for the physical properties of the cord fabric in industry standards and specifications, or can be based on experimental testing and calibration of defect-free cord fabric samples in the early stages of system development.

[0064] The characteristic appearance degree of each subarea target defect type in the sensitive spectral channel image is calculated, and the minimum improvement rate relative to the characteristic appearance degree in other spectral channel images is multiplied by the characteristic appearance degree in the sensitive spectral channel image to obtain the characteristic enhancement degree of each subarea target defect type in the sensitive spectral channel image.

[0065] In a preferred embodiment of the present invention, the condition for verifying the credibility of the composite spectral image target detection result is that the feature enhancement degree of each partitioned target defect type in the sensitive spectral channel image is greater than or equal to a preset feature enhancement degree threshold.

[0066] The weight correction module corrects the multispectral image sequence fusion weight according to the feature enhancement degree deviation and feeds it back to the target detection module to regenerate the composite spectral image.

[0067] In a preferred embodiment of the present invention, the method of correcting the fusion weight of the multispectral image sequence based on the feature enhancement deviation includes: extracting the feature enhancement degree of the untrustworthy partition target defect type in the sensitive spectral channel image, constructing a positive adjustment factor based on the deviation ratio between the feature enhancement degree and a preset feature enhancement degree threshold, and upwardly correcting the corresponding fusion weight coefficient of the sensitive spectral channel image.

[0068] It should be noted that the above-mentioned positive adjustment factor can be constructed as a linear adjustment factor, that is, a combination of the normalized deviation ratio and the preset gain coefficient, to upward correct the corresponding fusion weight coefficient of the sensitive spectral channel image, that is, the product of the original fusion weight coefficient and the positive adjustment factor is used as the upward correction amount, which is increased on the basis of the original fusion weight coefficient.

[0069] The reverse adjustment factor is constructed based on the deviation ratio of the characteristic appearance degree of the untrustworthy partition target defect type in the sensitive spectral channel image relative to the maximum characteristic appearance degree in other spectral channel images, and the corresponding fusion weight coefficients of other spectral channel images are corrected downward.

[0070] It should be noted that the above-mentioned reverse adjustment factor can be constructed as an exponential attenuation factor, that is, the minimum deviation ratio is substituted into the exponential attenuation function under the preset suppression intensity coefficient, and the corresponding fusion weight coefficients of other spectral channel images are downwardly corrected, that is, the product of the original fusion weight coefficient and the reverse adjustment factor is used as the downward correction amount, which is reduced on the basis of the original fusion weight coefficient.

[0071] The corrected fusion weight coefficients corresponding to each spectral channel image are normalized to regenerate the fusion weights of the multispectral image sequence.

[0072] In a preferred embodiment of the present invention, the weight correction module also includes an iterative verification and rollback mechanism: setting a maximum number of iterations for multispectral image sequence fusion weight correction; if the verification is still unreliable after reaching the maximum number of iterations, the predefined baseline weight configuration of each spectral channel image is automatically rolled back, triggering an alarm signal and generating an error log, which contains the weight adjustment trajectory during the iteration process and the credibility verification data change curve.

[0073] The embodiment of the present invention establishes a defect type-sensitive spectral channel mapping relationship, and realizes reverse verification of the credibility of the composite spectral image detection result based on the feature enhancement degree of the sensitive spectral channel image where the defect type is located. If the verification is unreliable, a forward or reverse adjustment factor correction weight is further constructed based on the deviation ratio to perform closed-loop iterative optimization, suppress non-sensitive channel interference, reduce false defect misjudgment, and improve detection generalization.

[0074] The control response module sends the target detection results to the production control terminal, and the terminal responds to the corresponding control instructions.

[0075] In a preferred embodiment of the present invention, the control instruction includes the following operations: when no defects are detected in the target detection result, sending a maintenance operation instruction to the production control terminal to maintain the current production status.

[0076] When the target detection results contain the number, type and location coordinates of partitioned defects, a stop command is sent to the production control terminal to terminate the cord fabric production process. Based on the mapping relationship between the composite spectral image and the actual physical coordinates of the cord fabric, the defect location coordinates in the image are converted into the actual defect location coordinates of the cord fabric molding surface. A notification window containing the actual defect location coordinates, defect type and quantity is generated and fed back to the operator terminal to trigger manual intervention.

[0077] It should be noted that the system's preset parameters or rules are all stored in a cloud database, allowing for direct access when needed. The data sources for all preset parameters are pre-set during the system development phase to ensure the standardization and reliability of parameter access.

[0078] The embodiment of the present invention collects multispectral image sequences of each partition of the tire cord fabric forming surface, combines motion compensation, geometric calibration and noise filtering to improve preprocessing accuracy, generates a composite spectral image and performs defect target detection, verifies the detection credibility based on the target detection result using the feature enhancement degree of the sensitive spectral image where the defect type is located, and dynamically performs fusion weight correction. Finally, production control instructions are triggered according to the detection results, realizing closed-loop management of high-precision defect detection and production control linkage.

[0079] The above content is merely an example and explanation of the structure of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the present invention, they should all fall within the scope of protection of the present invention.

Claims

1. A machine vision-based system for identifying and analyzing defects in tire cord fabric production, characterized in that: include: The multispectral imaging module consists of a linear array of industrial cameras with integrated dichroic mirrors and adaptive fill light sources. Each camera in the array covers a rectangular monitoring area of ​​the tire cord fabric forming surface at the loom outlet according to a preset field of view mapping relationship and simultaneously collects a multispectral image sequence of the covered area. Image preprocessing module, which performs motion compensation registration and noise filtering preprocessing operations on multispectral image sequences; The target detection module performs weighted fusion of the pre-processed multispectral image sequence to generate a composite spectral image. Based on the texture features of the composite spectral image, it outputs the target detection results of the number of defects, defect types, and location coordinates of each partition. The result verification module associates the mapping relationship between the defect type and the corresponding sensitive spectral channel, verifies the credibility of the composite spectral image target detection result based on the feature enhancement degree of the target defect type in the sensitive spectral channel image, and executes the control response module if it is credible, otherwise executes the weight correction module; The weight correction module corrects the fusion weight of the multispectral image sequence according to the feature enhancement deviation and regenerates the composite spectral image; The control response module sends the target detection results to the control terminal to execute the control instructions.

2. The machine vision-based cord fabric production defect recognition and analysis system according to claim 1, characterized in that: The preprocessing operation is as follows: (a) based on the fixed transmission speed of the tire cord fabric and the exposure response time of the camera, the displacement of the tire cord fabric at the camera sampling moment is determined, and the multispectral image sequence of each partition is reversely shifted to eliminate motion blur; (b) performing intra-group geometric calibration on the multispectral image sequence of the same partition, including: establishing an affine transformation matrix between spectral channels based on the optical properties of the dichroic mirror, performing global geometric correction on the images of each spectral channel in the same partition, converting the corrected images of each spectral channel to the frequency domain, obtaining the sub-pixel shift between each spectral channel through the phase correlation method, and performing sub-pixel shift compensation on each spectral channel image using an interpolation algorithm; (c) The grayscale distribution of the background area of ​​each spectral channel image in the same partition is extracted, the noise distribution characteristics of the multispectral channel are analyzed and the noise probability density function is fitted. A dynamic threshold is set to distinguish the true features of the image from abnormal noise points, and then isolated noise points caused by the mechanical vibration of the loom are filtered out based on the cross-channel signal correlation.

3. The machine vision-based cord fabric production defect recognition and analysis system according to claim 1, characterized in that: The weighted fusion processing of the multispectral image sequence adopts an adaptive weight allocation method. The fusion weight coefficient corresponding to each spectral channel image uses its predefined benchmark weight as the initial parameter and is dynamically adjusted in combination with the image signal-to-noise ratio and texture clarity obtained in real time. The predefined benchmark weight is determined through an offline calibration experiment based on the intrinsic spectral reflectance characteristics of the tire cord material.

4. The machine vision-based cord fabric production defect recognition and analysis system according to claim 1, characterized in that: The target detection results are obtained by referring to the following acquisition process: S1. Dividing the composite spectral image into several image units, extracting the frequency domain texture parameter set and morphological gradient parameter set of each image unit, and organizing them into a texture parameter matrix and a morphological parameter matrix respectively; S2. The texture parameter matrix of each image unit is similarly matched with the pre-stored standard texture parameter interval matrix of each defect type of the cord fabric to generate a texture similarity score, and the morphological parameter matrix is ​​simultaneously similarly matched with the standard morphological parameter interval matrix to generate a morphological similarity score; S3. Calculate the product value of the bimodal similarity score of each image unit and each defect type. If the product value of an image unit and a specific defect type exceeds a preset value, mark the image unit as the corresponding defect type label; S4. Perform a connected domain search on adjacent image units with the same defect type label, merge them to form the same defect area, and use the combination of the vertex coordinates and the centroid coordinates of the minimum circumscribed rectangle of the defect area as the defect position coordinates, so as to output the defect number, defect type and position coordinates of the corresponding partition of the composite spectral image.

5. The machine vision-based system for identifying and analyzing defects in tire cord fabric production according to claim 4, characterized in that: The similarity matching in step S3 refers to the following process: Assign a preset benchmark matching score to each element in the parameter matrix; Compare the value of each element in the parameter matrix with the value interval of the element at the same position in the corresponding standard parameter interval matrix. If the element value is within the corresponding value interval, the benchmark matching score of the element is retained. If it exceeds the value interval, the score correction coefficient is calculated based on the deviation between the element value and the interval boundary, and the benchmark matching score is reduced according to the correction coefficient. According to the predefined defect type weight rule, the weight coefficient of each defect type of the cord fabric for each element in the parameter matrix is ​​obtained, and the matching score of each element in the parameter matrix is ​​weighted and summed according to the weight to generate the similarity score; The parameter matrix includes a texture parameter matrix or a morphological parameter matrix, and the standard parameter interval matrix includes a corresponding standard texture parameter interval matrix or a standard morphological parameter interval matrix.

6. The machine vision-based system for identifying and analyzing defects in tire cord fabric production according to claim 5, characterized in that: The characteristic enhancement degree of each subarea target defect type in the sensitive spectrum channel image is obtained by referring to the following acquisition process: Based on the weight coefficients of each defect type of the cord fabric for each element in the parameter matrix, the weight coefficients of the target defect type for each element in the parameter matrix are extracted, and the elements whose weight coefficients exceed the average element weight coefficient of the parameter matrix are selected as the strengthening elements of the target defect type in the parameter matrix. In this way, the strengthening elements of the target defect type in each partition in the texture parameter matrix and the morphology parameter matrix are retrieved; Obtain the texture parameter matrix and morphological parameter matrix of the target defect location area in each spectral channel image of each partition, retrieve the value of each strengthening element in the matrix and calculate the degree of defect feature manifestation in the strengthening direction, and take the average to obtain the characteristic manifestation degree of the target defect location area in each spectral channel image of each partition; The characteristic appearance degree of each subarea target defect type in the sensitive spectral channel image is calculated, and the minimum improvement rate relative to the characteristic appearance degree in other spectral channel images is multiplied by the characteristic appearance degree in the sensitive spectral channel image to obtain the characteristic enhancement degree of each subarea target defect type in the sensitive spectral channel image.

7. The machine vision-based system for identifying and analyzing defects in tire cord fabric production according to claim 1, characterized in that: The condition for verifying the credibility of the composite spectral image target detection result is that the feature enhancement degree of each partition target defect type in the sensitive spectral channel image is greater than or equal to a preset feature enhancement degree reaching threshold.

8. The machine vision-based cord fabric production defect recognition and analysis system according to claim 1, characterized in that: The method of correcting the multispectral image sequence fusion weight according to the feature enhancement deviation includes: extracting the feature enhancement degree of the untrustworthy partition target defect type in the sensitive spectral channel image, constructing a positive adjustment factor based on the deviation ratio between the feature enhancement degree and a preset feature enhancement degree threshold, and upwardly correcting the corresponding fusion weight coefficient of the sensitive spectral channel image; A reverse adjustment factor is constructed based on the deviation ratio of the characteristic appearance degree of the untrustworthy partition target defect type in the sensitive spectral channel image relative to the maximum characteristic appearance degree in the other spectral channel images, and the corresponding fusion weight coefficients of the other spectral channel images are corrected downward; The corrected fusion weight coefficients corresponding to each spectral channel image are normalized to regenerate the fusion weights of the multispectral image sequence.

9. The machine vision-based system for identifying and analyzing defects in tire cord fabric production according to claim 1, characterized in that: The weight correction module also includes an iterative verification and rollback mechanism: setting a maximum number of iterations for multispectral image sequence fusion weight correction. If the verification is still unreliable after reaching the maximum number of iterations, the predefined baseline weight configuration of each spectral channel image is automatically rolled back, triggering an alarm signal and generating an error log. The error log contains the weight adjustment trajectory during the iteration process and the credibility verification data change curve.

10. The machine vision-based cord fabric production defect recognition and analysis system according to claim 1, characterized in that: The control instructions include the following operations: When no defects are detected in the target detection results, a maintenance instruction is sent to the production control terminal to maintain the current production status; When the target detection results contain the number, type and location coordinates of partitioned defects, a stop command is sent to the production control terminal to terminate the cord fabric production process. Based on the mapping relationship between the composite spectral image and the actual physical coordinates of the cord fabric, the defect location coordinates in the image are converted into the actual defect location coordinates of the cord fabric molding surface. A notification window containing the actual defect location coordinates, defect type and quantity is generated and fed back to the operator terminal to trigger manual intervention.

Images