A high-speed wire coil profile on-line measuring method

Abstract

The application discloses a kind of high-speed wire coil shape on-line measurement method, it is related to image recognition field, the method includes: based on self-organizing mapping and dynamic equilibrium coupling mechanism, the automatic adjustment and input-output ratio optimization of layout to infrared temperature sensor array, obtain the optimized infrared temperature sensor array;The three-dimensional model of coil is constructed, and the temperature field distribution of coil is acquired;The three-dimensional model of coil is associated with the temperature field distribution of coil Analysis, obtain the real-time temperature field model of coil;The deviation of real-time temperature field model and historical data is calculated, and by comparing with preset tolerance threshold, filter out overproof coil;The retest result of coil is acquired;When retest result is still overproof, then determine that coil is seriously overproof.The application improves the accuracy and efficiency of coil shape detection, reduces the influence of artificial factors on detection result, ensures the further improvement of product quality and production efficiency.

Description

Technical Field

[0001] This invention relates to the field of image recognition, and more specifically, to a method for online measurement of the shape of high-speed wire coils. Background Technology

[0002] With the rapid advancement of industrial automation and informatization, coil shape inspection has become crucial in the production process. Currently, traditional coil shape inspection mainly relies on manual operation and contact measuring equipment, such as mechanical calipers, which has several drawbacks. First, the inspection accuracy is insufficient, easily affected by coil surface deformation, leading to large dimensional measurement errors. This, in turn, causes jamming accidents due to dimensional deviations after warehousing, with a frequency of approximately 5%. Second, existing systems typically can only perform inspections upon warehousing after the coil has come off the production line. If the inspection results in deviations, manual intervention is required, resulting in lengthy rework times and significantly impacting the continuity of the production line. Furthermore, the lack of real-time monitoring and prediction capabilities for coil temperature changes makes it impossible to promptly identify the risk of dimensional deviations after cold-state warehousing.

[0003] These problems severely restrict production efficiency and product quality. Traditional inspection methods are not only time-consuming and labor-intensive, but also susceptible to human error, making it difficult to guarantee the accuracy and stability of inspection results. Furthermore, the lack of predictive capabilities prevents the timely detection of potential dimensional anomalies, increasing the difficulty and cost of subsequent processing. Therefore, a high-speed, accurate, and online inspection method is needed to improve the accuracy and efficiency of coil shape inspection and reduce the impact of human factors on inspection results.

[0004] To address these issues, the industry has been exploring new inspection technologies and methods. While some improvements have been proposed, such as the use of laser scanning, these methods still suffer from insufficient accuracy, susceptibility to interference and noise, and high measurement error rates. Furthermore, existing inspection equipment typically only performs two-dimensional measurements, making it difficult to simultaneously acquire three-dimensional information of the product, thus failing to meet the demands of inspecting complex shapes. Therefore, developing an intelligent early warning method that combines multi-dimensional data fusion, dynamic benchmark self-correction, and online closed-loop control is of great significance for improving the accuracy, efficiency, and reliability of coil shape inspection.

[0005] To address these issues, image recognition technology has offered new approaches to coil shape inspection in recent years. The development of machine vision technology has made image-based non-contact measurement possible, effectively avoiding the influence of coil surface deformation on traditional contact measurement devices and improving measurement accuracy. However, existing image measurement methods still have limitations, such as slow measurement speed and insufficient accuracy in detecting complex shapes. Furthermore, the lack of real-time monitoring and prediction capabilities for coil temperature changes makes it impossible to promptly detect the risk of dimensional deviations after cold storage, resulting in compromised accuracy and stability of the inspection results.

[0006] Existing patents, such as Chinese Publication No. CN113074633B, disclose an automatic detection system and method for the external dimensions of materials. This system utilizes sensors to monitor the presence or absence of materials, automatically completing the continuous reconstruction of a single material's three-dimensional system and automatically starting and stopping the reconstruction of subsequent materials. Simultaneously, it calculates basic dimensions, deformation dimensions, and smoothness parameters of the reconstructed model. However, in practical applications, this system still suffers from the problem of insufficient precision in aligning the center of the measuring device with the vertical geometric centerline of the material passing through its center.

[0007] For example, Chinese patent application CN118654573B proposes a method for measuring the shape and size of sweet cherries based on image recognition. This method acquires top and side views of the sweet cherry using a sweet cherry shape image acquisition device. The top and side views are preprocessed to remove useless information and noise interference, determining the calibration template of the sweet cherry in the top and side views. The transverse diameter, longitudinal diameter, and thickness of the sweet cherry are determined by combining pixel equivalents with the two points furthest apart in the horizontal, vertical, and center directions of the cherry outline in the top and side views. However, this method still needs improvement by incorporating more image processing and analysis techniques, such as shape analysis and surface reconstruction, to improve the accuracy and reliability of the size measurement.

[0008] In general, the existing technology has the following shortcomings:

[0009] (1) Traditional contact measuring devices are affected by the deformation of the coil, resulting in large dimensional detection errors and a jamming accident rate of about 5%. This error not only affects the quality control of the product, but may also cause the production line to stop and delay, thus affecting the overall production efficiency. In addition, traditional devices can usually only perform inspections when the product is put into storage after the lower limit, which requires manual return of non-conforming products, which is time-consuming and seriously affects the continuity and automation level of the production line.

[0010] (2) The lack of real-time monitoring and prediction of coil temperature changes makes it impossible to detect the risk of dimensional deviations after cold storage, which affects the accuracy and stability of the test results. Since temperature changes have a significant impact on the elasticity and size of materials, the lack of a real-time monitoring mechanism makes it impossible to effectively control temperature fluctuations during the production process, thus affecting the final test results.

[0011] (3) Existing equipment is mostly two-dimensional measurement, which makes it difficult to acquire three-dimensional information of the product at the same time. This makes it impossible to meet the needs of complex shape inspection and limits the comprehensiveness and accuracy of the inspection. This limitation makes it difficult for manufacturers to fully evaluate the geometric features and structural integrity of the product, which in turn affects the product design and manufacturing process.

[0012] (4) The lack of intelligent early warning and decision-making functions makes it impossible to automatically adjust the production process based on the detection results, resulting in low production efficiency and difficulty in adapting to the needs of modern industrial automation. The lack of an intelligent early warning system means that the production line cannot optimize and adjust itself based on real-time data, and cannot cope with anomalies and changes in the production process, leading to increased uncertainty in the production process.

[0013] (5) Existing technologies are insufficient in data processing and analysis, lacking multi-dimensional data fusion and real-time analysis capabilities, and are unable to fully utilize detection data to optimize production processes and improve product quality. The limitations of data not only restrict a comprehensive understanding of the production process, but also hinder data-driven production decisions and process improvements.

[0014] (6) The arrangement of infrared temperature sensor arrays mostly adopts manual experience or static optimization mode, which cannot adjust the position and density of sensors in a timely manner according to the dynamic changes and spatial complexity of the field environment. This leads to monitoring blind spots, insufficient coverage, node redundancy, high energy consumption and increased maintenance costs. In addition, the existing technology lacks a dynamic calculation mechanism for quantifying environmental complexity and matching monitoring supply and demand, making it difficult to balance monitoring performance and economy.

[0015] No effective solutions have yet been proposed to address the problems in the relevant technologies. Summary of the Invention

[0016] Existing technologies suffer from problems such as low accuracy in coil shape detection, low process efficiency, lack of predictive capabilities, and inability to achieve multi-dimensional data fusion. Furthermore, existing equipment struggles to provide sufficient accuracy and stability when handling complex shape detection, affecting the accuracy and reliability of the results. Therefore, this invention provides a high-speed online measurement method for the shape of wire coils to overcome the aforementioned technical problems in related existing technologies.

[0017] Therefore, the specific technical solution adopted by the present invention is as follows:

[0018] According to a first aspect of the present invention, a method for online measurement of the shape of high-speed wire coils is provided, comprising:

[0019] Based on self-organizing mapping and dynamic equilibrium coupling mechanism, the layout of the infrared temperature sensor array is automatically adjusted and the input-output ratio is optimized to obtain the optimized infrared temperature sensor array.

[0020] The image outline of the coil is obtained, and a three-dimensional model of the coil is constructed by combining image processing algorithms; the temperature field distribution of the coil is obtained by using an optimized infrared temperature sensor array.

[0021] By correlating the three-dimensional model of the coil with the temperature field distribution of the coil, a real-time temperature field model of the coil is obtained.

[0022] The deviation between the real-time temperature field model and historical data is calculated, and the non-compliant coils are screened out by comparing them with the preset tolerance threshold.

[0023] The real-time temperature field model of the coils exceeding the standard is regenerated, and the deviation between the real-time temperature field model and historical data is recalculated to obtain the retest results of the coils; if the retest results are still exceeding the standard, the coils are judged to be seriously exceeding the standard.

[0024] Furthermore, based on self-organizing mapping and dynamic equilibrium coupling mechanisms, the layout of the infrared temperature sensor array is automatically adjusted and the input-output ratio is optimized, resulting in an optimized infrared temperature sensor array comprising:

[0025] Acquire measurement data of the physical space of the high-speed wire coil site; based on statistical analysis, within a preset sliding time window, analyze the measurement data according to the environmental complexity index, calculate the two-dimensional environmental complexity distribution function, which is used to comprehensively reflect the temperature gradient change rate, flow field chaos, humidity change frequency and the degree of information obstruction caused by spatial obstacles to monitoring.

[0026] Using the two-dimensional environment complexity distribution function as input parameters, the node topology of the self-organizing map model is controlled, and the initial distribution density of each infrared temperature sensor node position is adjusted in the two-dimensional array space.

[0027] Establish a connection matrix between nodes, and output the node density distribution function from the self-organizing map model;

[0028] The two-dimensional environmental complexity distribution function is used as the monitoring demand function, and the node density distribution function is used as the monitoring supply function. The marginal benefit function and marginal cost function are calculated.

[0029] By comparing the marginal benefit function and the marginal cost function, the marginal difference function is calculated at a local scale, and the input-output ratio of the current infrared temperature sensor array layout is determined.

[0030] A dynamic equilibrium principle is established: when the input-output ratio of the infrared temperature sensor array layout is greater than zero, an instruction to increase node density is generated; otherwise, an instruction to decrease node density is generated.

[0031] Based on the principle of dynamic equilibrium, the marginal difference function and input-output ratio are transformed into parameter correction vectors for the self-organizing mapping model;

[0032] The parameter correction vector is input into the weight update mechanism of the self-organizing map model to change the node spacing and density distribution; the training process of the self-organizing map model is rerun to obtain the optimized infrared temperature sensor array.

[0033] Furthermore, the two-dimensional environmental complexity distribution function is used as the monitoring demand function, and the node density distribution function is used as the monitoring supply function. The marginal benefit function and marginal cost function are calculated as follows:

[0034] By using a weighted linear combination, the two-dimensional environmental complexity distribution function is transformed into a monitoring demand function; by combining the single-node monitoring capability constant, the node density distribution function is transformed into a monitoring supply function.

[0035] The supply-demand difference function is obtained by performing a point-by-point difference calculation on the monitored demand function and the monitored supply function.

[0036] The marginal benefit function is obtained by coupling the supply and demand difference corresponding to the supply and demand difference function with the improvement of monitoring accuracy.

[0037] The system acquires current sensor hardware cost data, unit node energy consumption data, data transmission energy consumption parameters, and maintenance cost parameters, and sums them up node by node to obtain the marginal cost function.

[0038] Furthermore, the monitoring demand function is as follows:

[0039] ;

[0040] In the formula, Q d To monitor the demand function, G T Let H be the distribution function of the rate of change of the temperature gradient. F Let F be the chaotic distribution function of the airflow field. H Let S be the frequency distribution function of humidity variation. O Let w be the distribution function of the degree of obstacle occlusion. T w F w H w O For different weighting coefficients;

[0041] The monitoring supply function is:

[0042] ;

[0043] In the formula, Q s To monitor the supply function, M cap D represents the monitoring capability per unit area of ​​a single infrared temperature sensor node. init This is the node density distribution function.

[0044] Furthermore, based on the principle of dynamic equilibrium, the marginal difference function and input-output ratio are transformed into parameter correction vectors for the self-organizing map model, including:

[0045] The dynamic equilibrium principle is applied. When the input-output ratio is greater than zero, the density of nodes with a positive difference between marginal benefit and cost is increased and converted into a correction amount according to a proportional coefficient. The correction amount for nodes with a non-positive difference is zero. When the input-output ratio is less than or equal to zero, the density of nodes with a negative difference between marginal benefit and cost is decreased and converted into a correction amount according to a proportional coefficient. The correction amount for nodes with a non-negative difference is zero. At the same time, the node correction amounts are collected to obtain a set of node density correction magnitudes.

[0046] The physical coordinates of each node are mapped to the topological index of the self-organizing map model, and the corresponding node density correction magnitude is converted into weight correction amount, and the mapping method adopts a linear proportional mapping method.

[0047] The weight correction amount is integrated with the corresponding node index number and position coordinate information to obtain the parameter correction vector.

[0048] Furthermore, the image contour of the coil is obtained, and combined with image processing algorithms, a 3D model of the coil is constructed, including:

[0049] Multi-angle image data of the coil were acquired, and edge detection was performed on each image using a two-way threshold segmentation method to obtain the outline of the coil.

[0050] A 3D reconstruction algorithm based on multi-angle images is used, combined with the depth information of the images, to generate a 3D model of the coil.

[0051] Furthermore, by utilizing the optimized infrared temperature sensor array, the temperature field distribution of the coil is obtained, including:

[0052] The optimized infrared temperature sensor array is used to monitor discrete temperature points of the coil and map them into a two-dimensional matrix according to the pre-calibrated spatial coordinates.

[0053] An interpolation algorithm is used to fill the blank areas in a two-dimensional matrix to generate the temperature field distribution of the coil.

[0054] Furthermore, by correlating the three-dimensional model of the coil with its temperature field distribution, a real-time temperature field model of the coil is obtained, including:

[0055] The temperature field distribution data of the coil is transformed into the world coordinate system through coordinate transformation and then aligned with the three-dimensional model of the coil.

[0056] Temperature values ​​are assigned as vertex attributes to each vertex of the coiled 3D model, and color mapping is used to convert the temperature values ​​into corresponding colors.

[0057] Furthermore, the deviation between the real-time temperature field model and historical data is calculated, and by comparing it with a preset tolerance threshold, coils exceeding the standard are selected, including:

[0058] The deviation between the real-time temperature field model and historical data is calculated. When the deviation exceeds the preset tolerance threshold, it is determined that the measurement is out of standard, and the corresponding reel is regarded as the out-of-standard reel.

[0059] Early warning information is generated based on the type and degree of exceeding the standard.

[0060] Furthermore, a real-time temperature field model is regenerated for the coils exceeding the standard, and the deviation between the real-time temperature field model and historical data is recalculated to obtain the retest results of the coils; if the retest results still exceed the standard, the coils are judged to be severely exceeding the standard, including:

[0061] The substandard coils are diverted to the re-inspection station; photoelectric switch sensing technology is used to realize the automatic diversion of coils;

[0062] If the coils that exceed the standard are retested, and the retest results are still found to be coils that exceed the standard, they are judged to be seriously exceeding the standard and reworked.

[0063] According to a second aspect of the present invention, a high-speed wire coil shape online measurement system is provided, comprising:

[0064] The sensor layout module is used to automatically adjust the layout of the infrared temperature sensor array and optimize the input-output ratio based on the self-organizing mapping and dynamic equalization coupling mechanism, so as to obtain the optimized infrared temperature sensor array.

[0065] The coil model construction module is used to acquire the image contour of the coil and, in combination with image processing algorithms, construct a three-dimensional model of the coil; it uses an optimized infrared temperature sensor array to acquire the temperature field distribution of the coil; and it performs correlation analysis between the three-dimensional model of the coil and the temperature field distribution of the coil to obtain a real-time temperature field model of the coil.

[0066] The out-of-standard coil identification module is used to calculate the deviation between the real-time temperature field model and historical data, and to filter out out-of-standard coils by comparing them with a preset tolerance threshold.

[0067] The coil retest module is used to regenerate the real-time temperature field model for coils that exceed the standard, and to recalculate the deviation between the real-time temperature field model and historical data to obtain the retest results of the coils; if the retest results are still exceeding the standard, the coils are judged to be seriously exceeding the standard.

[0068] According to a third aspect of the present invention, an electronic device is provided, the electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the online measurement method for the shape of high-speed wire coils according to any embodiment of the present invention.

[0069] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the online measurement method for the shape of high-speed wire coils according to any embodiment of the present invention.

[0070] The beneficial effects of this invention are as follows:

[0071] (1) By adopting a multi-view three-dimensional reconstruction module and machine vision recognition technology, the shortcomings of traditional contact measurement devices affected by the deformation of the coil surface are overcome, significantly improving the accuracy of size detection. The detection error is controlled within ±5mm, which is 4 times higher than the traditional method. It effectively avoids material jamming accidents caused by size deviation and improves the warehousing qualification rate from 92% to 99%.

[0072] (2) The present invention realizes real-time online measurement of the shape of the coil, and can monitor the length and diameter of the coil package in real time before the coil enters the automated warehouse, which greatly shortens the inspection time from 5 minutes to 30 seconds for manual sampling, significantly improving production efficiency.

[0073] (3) By setting up the intelligent early warning decision module, the detection of exceeding the standard can be automatically diverted to the re-inspection station without manual intervention, avoiding interference caused by human factors and improving the accuracy and stability of the detection results.

[0074] (4) The temperature field synchronous monitoring module is adopted to realize the monitoring of real-time temperature changes of the coil, which can promptly detect the risk of dimensional deviation after cold storage, effectively solving the problem of lack of predictive ability in the existing technology.

[0075] (5) Through multi-dimensional data fusion, dynamic benchmark self-correction and online closed-loop control, comprehensive and high-precision measurement of the coil shape is achieved, overcoming the limitation of existing technologies that can only achieve two-dimensional measurement.

[0076] (6) By introducing non-contact measurement technology, errors caused by manual operation in traditional methods are avoided, the influence of human factors on the detection results is reduced, and the accuracy and reliability of the detection are further improved. This method can not only adapt to different types of coil shapes, but also work stably in complex environments, overcoming the limitations of traditional equipment under harsh conditions.

[0077] (7) By combining the two-dimensional environmental complexity distribution function with the self-organizing mapping model and introducing the dynamic equilibrium principle, the automatic adjustment and economic optimization of the infrared temperature sensor array layout are realized. Dynamically matching the monitoring demand and supply ensures that the position density distribution of sensor nodes is always in the optimal state, reducing redundant deployment and energy consumption, improving the input-output ratio, and achieving a balance between performance and cost optimization. Attached Figure Description

[0078] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0079] Figure 1 This is a flowchart of an online measurement method for the shape of a high-speed wire coil according to an embodiment of the present invention;

[0080] Figure 2 This is a schematic block diagram of a high-speed wire coil shape online measurement system according to an embodiment of the present invention;

[0081] Figure 3 This is a schematic diagram of the structure of a computer device according to an embodiment of the present invention;

[0082] Figure 4 This is a flowchart of optimizing an infrared temperature sensor array in a high-speed wire coil shape online measurement method according to an embodiment of the present invention;

[0083] Figure 5 This is a schematic diagram of the outer contour of a steel coil according to an embodiment of the present invention. Detailed Implementation

[0084] To further illustrate the various embodiments, the present invention provides accompanying drawings, which are part of the disclosure of the present invention. These drawings are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these drawings, those skilled in the art should be able to understand other possible implementation methods and the advantages of the present invention. The components in the drawings are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0085] According to embodiments of the present invention, a method, system, device and medium for online measurement of the shape of high-speed wire coils are provided, namely, an intelligent early warning method for three-dimensional online measurement of the shape of high-speed wire coils and real-time temperature monitoring based on machine vision recognition.

[0086] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments, such as... Figure 1 As shown, an embodiment of the online measurement method for the shape of high-speed wire coils according to the present invention includes:

[0087] S1. Based on the self-organizing mapping and dynamic equilibrium coupling mechanism, the layout of the infrared temperature sensor array is automatically adjusted and the input-output ratio is optimized to obtain the optimized infrared temperature sensor array.

[0088] S2. Obtain the image outline of the coil and construct a three-dimensional model of the coil using image processing algorithms; use an optimized infrared temperature sensor array to obtain the temperature field distribution of the coil.

[0089] S3. Correlation analysis is performed between the three-dimensional model of the coil and the temperature field distribution of the coil to obtain the real-time temperature field model of the coil.

[0090] S4. Calculate the deviation between the real-time temperature field model and historical data, and filter out the coils that exceed the tolerance threshold by comparing them with the preset tolerance threshold.

[0091] S5. Regenerate the real-time temperature field model for the coils that exceed the standard, and recalculate the deviation between the real-time temperature field model and historical data to obtain the retest results of the coils; if the retest results are still exceeding the standard, the coils are judged to be seriously exceeding the standard.

[0092] In one embodiment, based on a self-organizing mapping and dynamic equalization coupling mechanism, the layout of the infrared temperature sensor array is automatically adjusted and the input-output ratio is optimized to obtain an optimized infrared temperature sensor array comprising:

[0093] like Figure 4 As shown, measurement data of the physical space at the high-speed wire coiling site are acquired, including temperature field distribution data, airflow data, humidity distribution data, obstacle space information, and time series change records. Based on statistical analysis, within a preset sliding time window, the measurement data is analyzed according to the environmental complexity index to calculate a two-dimensional environmental complexity distribution function, which comprehensively reflects the temperature gradient change rate, flow field chaos, humidity change frequency, and the degree of information obstruction caused by spatial obstacles. Using the two-dimensional environmental complexity distribution function as input parameters, the node topology of the self-organizing map model is controlled, and the initial distribution density of each infrared temperature sensor node position is adjusted in the two-dimensional array space. A node connection relationship matrix is ​​established, and the node density distribution function is output by the self-organizing map model. The two-dimensional environmental complexity is then analyzed. The distribution function is used as the monitoring demand function, and the node density distribution function is used as the monitoring supply function. Marginal benefit and marginal cost functions are calculated. By comparing the marginal benefit and marginal cost functions, the marginal difference function is calculated at a local scale, and the input-output ratio of the current infrared temperature sensor array layout is determined. A dynamic equilibrium principle is constructed: when the input-output ratio of the infrared temperature sensor array layout is greater than zero, an instruction to increase node density is generated; otherwise, an instruction to decrease node density is generated. Based on the dynamic equilibrium principle, the marginal difference function and input-output ratio are transformed into parameter correction vectors for the self-organizing map model. The parameter correction vectors are input into the weight update mechanism of the self-organizing map model to change the node spacing and density distribution. The training process of the self-organizing map model is rerun to obtain the optimized infrared temperature sensor array.

[0094] In one embodiment, the two-dimensional environmental complexity distribution function is used as the monitoring demand function, the node density distribution function is used as the monitoring supply function, and the marginal benefit function and marginal cost function are calculated as follows:

[0095] The two-dimensional environmental complexity distribution function is transformed into a monitoring demand function through a weighted linear combination. The node density distribution function is transformed into a monitoring supply function by combining the single-node monitoring capability constant. The monitoring demand function and the monitoring supply function are then subjected to coordinate-point difference calculation to obtain the supply-demand difference function. The supply-demand difference function is coupled with the improvement of monitoring accuracy to obtain the marginal benefit function. The current sensor hardware cost data, unit node energy consumption data, data transmission energy consumption parameters, and maintenance cost parameters are obtained and summed node by node to obtain the marginal cost function.

[0096] In one embodiment, the monitoring requirement function is:

[0097] ;

[0098] In the formula, Q d To monitor the demand function, G T Let H be the distribution function of the rate of change of the temperature gradient. F Let F be the chaotic distribution function of the airflow field. H Let S be the frequency distribution function of humidity variation. O Let w be the distribution function of the degree of obstacle occlusion. T w F w H w O For different weighting coefficients;

[0099] The monitoring supply function is:

[0100] ;

[0101] In the formula, Q s To monitor the supply function, M cap D represents the monitoring capability per unit area of ​​a single infrared temperature sensor node. init This is the node density distribution function.

[0102] In one embodiment, transforming the marginal difference function and input-output ratio into a parameter correction vector for the self-organizing map model based on the dynamic equilibrium principle includes:

[0103] The dynamic equilibrium principle is applied. When the input-output ratio is greater than zero, the density of nodes with a positive difference between marginal benefit and cost is increased and converted into a correction amount according to a proportional coefficient, while the correction amount for nodes with a non-positive difference is zero. When the input-output ratio is less than or equal to zero, the density of nodes with a negative difference between marginal benefit and cost is decreased and converted into a correction amount according to a proportional coefficient, while the correction amount for nodes with a non-negative difference is zero. At the same time, the node correction amounts are collected to obtain a set of node density correction magnitudes. The physical coordinates of each node are mapped to the topological index of the self-organizing mapping model, and the corresponding node density correction magnitude is converted into a weight correction amount, with the mapping method adopting a linear proportional mapping method. The weight correction amount is integrated with the corresponding node index number and position coordinate information to obtain a parameter correction vector.

[0104] In one embodiment, acquiring the image outline of the reel and constructing a 3D model of the reel using image processing algorithms includes:

[0105] Multi-angle image data of the coil were acquired, and edge detection was performed on each image using a bidirectional threshold segmentation method to obtain the outline of the coil. A three-dimensional reconstruction algorithm based on multi-angle images was used, combined with the depth information of the images, to generate a three-dimensional model of the coil.

[0106] In one embodiment, obtaining the temperature field distribution of the coil using an optimized infrared temperature sensor array includes:

[0107] An optimized infrared temperature sensor array is used to monitor discrete temperature points of the coil and map them into a two-dimensional matrix based on pre-calibrated spatial coordinates. An interpolation algorithm is then used to fill the blank areas in the two-dimensional matrix to generate the temperature field distribution of the coil.

[0108] In one embodiment, the real-time temperature field model of the coil is obtained by correlating the three-dimensional model of the coil with its temperature field distribution through analysis, including:

[0109] The temperature field distribution data of the coil is transformed into the world coordinate system through coordinate transformation and aligned with the 3D model of the coil. The temperature value is assigned as a vertex attribute to each vertex of the 3D model of the coil, and the temperature value is converted into the corresponding color through color mapping.

[0110] In one embodiment, the deviation between the real-time temperature field model and historical data is calculated, and the non-compliant coils are screened out by comparing them with a preset tolerance threshold.

[0111] The deviation between the real-time temperature field model and historical data is calculated. When the deviation exceeds the preset tolerance threshold, it is determined that the measurement is out of standard, and the corresponding reel is designated as the out-of-standard reel. Early warning information is generated based on the type and degree of the out-of-standard.

[0112] In one embodiment, a real-time temperature field model is regenerated for the coils exceeding the standard, and the deviation between the real-time temperature field model and historical data is recalculated to obtain the retest results of the coils; if the retest results still exceed the standard, the coils are determined to be severely exceeding the standard, including:

[0113] The substandard coils are diverted to the re-inspection station; photoelectric switch sensing technology is used to realize the automatic diversion of the coils; the substandard coils are retested, and if the retest result is still a substandard coil, it is judged as seriously substandard and reworked.

[0114] like Figure 2 As shown, another embodiment of the high-speed wire coil shape online measurement system of the present invention is presented. The high-speed wire coil shape online measurement system includes:

[0115] Sensor layout module 1 is used to automatically adjust the layout of the infrared temperature sensor array and optimize the input-output ratio based on self-organizing mapping and dynamic equalization coupling mechanism, so as to obtain the optimized infrared temperature sensor array.

[0116] The coil model construction module 2 is used to acquire the image contour of the coil and, in combination with image processing algorithms, construct a three-dimensional model of the coil; use an optimized infrared temperature sensor array to acquire the temperature field distribution of the coil; and perform correlation analysis between the three-dimensional model of the coil and the temperature field distribution of the coil to obtain a real-time temperature field model of the coil.

[0117] The out-of-standard coil identification module 3 is used to calculate the deviation between the real-time temperature field model and historical data, and to filter out out-of-standard coils by comparing them with a preset tolerance threshold.

[0118] The coil re-inspection module 4 is used to regenerate the real-time temperature field model for coils that exceed the standard, and to recalculate the deviation between the real-time temperature field model and historical data to obtain the re-test results of the coils; if the re-test results are still exceeding the standard, the coils are judged to be seriously exceeding the standard.

[0119] To facilitate understanding of the above technical solutions of the present invention, the working principle of the present invention in actual process will be described in detail below.

[0120] This invention combines a smart early warning method with multi-dimensional data fusion, dynamic benchmark self-correction, and online closed-loop control. It not only needs to achieve high-speed and accurate measurement of coil shape, but also needs to have real-time temperature monitoring and intelligent prediction functions to improve the accuracy and efficiency of coil shape detection, reduce the impact of human factors on the detection results, and thus ensure further improvement in product quality and production efficiency.

[0121] Existing technologies suffer from low accuracy in coil shape detection, low process efficiency, lack of predictive capabilities, and inability to achieve multi-dimensional data fusion. Therefore, to address these issues, this invention employs a multi-view 3D reconstruction module and machine vision recognition technology to achieve real-time online measurement of coil shape; it also develops an intelligent early warning decision module and a temperature field synchronous monitoring module. Through multi-dimensional data fusion, dynamic benchmark self-correction, and online closed-loop control, it achieves comprehensive and high-precision measurement of coil shape. Specifically, it includes:

[0122] 1. Initialize system configuration, including starting the measurement system, establishing communication connections, initializing the 3D reconstruction and temperature monitoring models, and loading the historical dimensional tolerance database.

[0123] 2. Perform three-dimensional online measurement of the coil shape, including acquiring multi-view image data, reconstructing the three-dimensional model after preprocessing, optimizing the infrared temperature sensor array, monitoring the temperature in real time, and obtaining the temperature field model through correlation analysis.

[0124] 3. Determine if the measurement result exceeds the standard. If it does, proceed to step 4; otherwise, proceed to step 5. The determination of whether the result exceeds the standard is based on the standard size data in the historical database. Calculate the relative error and record the type and degree of exceeding the standard.

[0125] 4. Conduct a re-inspection of the coil shape, divert the coils that exceed the standard to the re-inspection station, remeasure and compare with the original results to determine whether rework is required.

[0126] 5. Complete the warehousing operation, generate warehousing instructions based on the measurement results, move the instructions to the warehousing location, complete the warehousing, and update the database.

[0127] This invention achieves real-time measurement and intelligent early warning of the coil shape through machine vision recognition and temperature monitoring, ensuring the accuracy of measurement results and temperature safety.

[0128] Example 1:

[0129] 1. Initialize system configuration:

[0130] (1) Start the measurement system and establish a communication connection. The measurement system includes a multi-view 3D reconstruction module, a temperature field synchronous monitoring module, and an intelligent early warning decision module. The multi-view 3D reconstruction module consists of two infrared thermal imaging cameras, installed directly above the center of the wire coil transport roller conveyor, at a vertical distance of 2.0 meters from the conveyor roller conveyor. The imaging equipment has a sampling frequency of 3 frames / second, and the field of view of the infrared thermal imaging camera must be greater than the width of the wire coil transport roller conveyor. The temperature field synchronous monitoring module uses a non-contact infrared temperature sensor array, evenly arranged around the wire coil, with a temperature measurement range of 0-200℃. The intelligent early warning decision module includes an industrial control computer and data processing software.

[0131] (2) Initialize the three-dimensional reconstruction model and temperature monitoring model. The original images were denoised using the standard median filtering method, and then the edge contour of the coil was obtained using the bidirectional threshold segmentation method. The maximum allowable curvature change threshold was set to 0.01, and the equivalent ellipse was obtained by fitting the curve using the least squares fitting algorithm. The center point position and major axis length of the calculated equivalent ellipse were used as the center position and diameter of the wire coil layer corresponding to the curve segment combination.

[0132] (3) Load the historical dimensional tolerance database. Organize and classify the collected historical measurement data to establish a dynamic dimensional tolerance database, which includes historical data on key dimensions such as coil length, diameter, and thickness, with more than 10,000 sets of data.

[0133] 2. Perform three-dimensional online measurement of the coil shape:

[0134] (1) Multi-angle image data of the coil were acquired using a multi-view 3D reconstruction module. Edge detection was performed on each image using a bidirectional threshold segmentation method to obtain the coil's contour line. A line-by-line scan was performed from the left to the right of the field of view to obtain all left boundary points; conversely, a scan was performed from the right to the left of the field of view to obtain all right boundary points. All left boundary points were combined to form the left contour edge curve, and all right boundary points were combined to form the right contour edge curve.

[0135] (2) Preprocess the acquired image data. The standard median filtering method is used to reduce noise in each image, and then the image enhancement algorithm is used to adjust the brightness and contrast of the image.

[0136] (3) Reconstruct the three-dimensional model of the coil using image processing algorithms. A three-dimensional reconstruction algorithm based on multi-view images is used, combined with the depth information of the images, to generate a three-dimensional geometric model of the coil.

[0137] (4) Real-time monitoring of the temperature distribution on the surface of the coil, combined with a temperature field synchronous monitoring module. A non-contact infrared temperature sensor array is used to monitor the temperature field distribution on the surface of the coil in real time, with a temperature measurement range of 0-200℃ and a temperature measurement frequency of 10Hz.

[0138] The monitoring system integrates optics, electronics, and software algorithms to achieve non-contact, fully automated, high-frequency scanning of the temperature field on the coil surface. Details are as follows:

[0139] System architecture and sensor array layout, sensor selection and array design:

[0140] Based on the dimensions of the coil, such as diameter and height, and the required temperature resolution, select a specific model of infrared thermometer, typically a thermopile or microbolometer. Employ a one-dimensional linear array or a two-dimensional area array layout. For long coils, a one-dimensional linear array sensor is commonly used, mechanically scanning to cover the entire surface. Determine the number of sensors in the array to ensure their combined field of view completely covers the radial scanning range of the coil. Install and position the sensor to ensure it is protected from vibration. Precisely adjust the module's mounting angle, height, and distance to ensure seamless alignment of all sensor measurement fields of view, collectively covering the entire surface scanning area of ​​the coil and avoiding measurement blind spots.

[0141] An automatic layout and cost-benefit optimization of an infrared temperature sensor array is performed based on a self-organizing mapping and dynamic equilibrium coupling mechanism. At a high-speed wire coiling site, multi-source measurement data of the physical space is acquired using technologies such as temperature field sampling, flow field measurement, humidity monitoring, and laser ranging obstacle identification. This data includes temperature values ​​and their temporal variations at different spatial locations, airflow velocity fields and vorticity characteristics, time-series information on humidity distribution and changes, and the three-dimensional coordinates and shape information of obstacles in the space. After data acquisition, an environmental complexity index is calculated within a set time span using a sliding time window statistical analysis method, and a two-dimensional environmental complexity distribution function is generated. This function comprehensively reflects the rate of change of the temperature gradient, the degree of chaos in the airflow field, the frequency of humidity changes, and the degree of information obstruction caused by spatial obstacles. By setting weighting factors corresponding to the temperature gradient, airflow, humidity changes, and obstacle obstruction degree for the above environmental characteristic data, and performing a weighted linear combination, a monitoring demand function is formed. As initial input, this function determines the node topology of the self-organizing map model and provides the initial distribution density of infrared temperature sensor nodes in the two-dimensional array space, thereby establishing the connection matrix between nodes. Through training iterations of the self-organizing map model, the density distribution function of the current node is output based on the initial density and the connection matrix, and combined with the monitoring capability constant per unit area of ​​a single node, it is converted into a monitoring supply function.

[0142] The monitoring demand function and monitoring supply function are processed at each coordinate point of the two-dimensional array. First, the difference between the corresponding points is calculated to obtain the supply-demand difference distribution. Then, the supply-demand difference is coupled with the monitoring accuracy improvement coefficient to obtain the marginal benefit function. Hardware costs, energy consumption per unit node, data transmission energy consumption parameters, and maintenance cost parameters of the current sensor node at different locations are collected and summed at each node to obtain the marginal cost function. Based on this, the difference function between marginal benefit and marginal cost is calculated at a local scale, and the input-output ratio of the overall layout of the infrared temperature sensor array is statistically derived. According to the dynamic equilibrium principle, when the input-output ratio is greater than zero, the density is increased only for nodes with a positive difference between marginal benefit and marginal cost, and the difference is converted into a correction amount according to a proportional coefficient. When the input-output ratio is less than or equal to zero, the density is decreased only for nodes with a negative difference, and the correction amount is converted into a correction amount according to a proportional coefficient. The correction amount is zero in all other cases. The density correction values ​​of all nodes are collected into a correction set, and the physical coordinates of the nodes are mapped to the topological index of the self-organizing map model. The density correction values ​​are converted into weight correction values ​​through a linear scaling method, and then integrated with the node index number and coordinate information into a parameter correction vector. This correction vector is input into the weight update mechanism of the self-organizing map model to adjust the spacing and density distribution of the nodes. The model training process is then rerun to output an optimized infrared temperature sensor array layout, thereby ensuring the economy of deployment and operation while meeting monitoring performance requirements.

[0143] Real-time data acquisition and synchronization signal triggering and synchronization acquisition:

[0144] The system is linked with the production line's PLC (Programmable Logic Controller). When the coil is transported to the temperature measurement station, a trigger signal is emitted by a photoelectric sensor or encoder. Signal conditioning and analog-to-digital conversion: The weak analog signal acquired is preprocessed by the signal conditioning circuit to improve the signal-to-noise ratio.

[0145] Temperature calculation and data preprocessing:

[0146] Temperature Value Calculation: The software within the main control unit converts the voltage digital signal into a preliminary temperature value based on the unique calibration coefficients of each sensor. These calibration coefficients are obtained in the laboratory using a blackbody furnace, ensuring the traceability of the measurement. The software also performs emissivity correction to compensate for measurement errors caused by varying degrees of surface oxidation. Environmental Compensation: An integrated ambient temperature sensor monitors the internal and surrounding ambient temperatures of the sensor module in real time, compensating for ambient temperature drift in the infrared readings and further improving long-term stability.

[0147] Temperature field reconstruction and visualization:

[0148] Data Mapping and Image Generation: The system maps discrete temperature points measured by each sensor into a two-dimensional matrix based on their pre-calibrated spatial coordinates. An interpolation algorithm fills in the blank areas of the matrix, generating a complete, continuous, and visualized pseudo-color temperature field image, i.e., a thermal image. Real-time Display and Alarm: The system automatically plots temperature distribution curves and marks the highest, lowest, and average temperatures. Temperature alarm thresholds can be set, and alarm information can be pushed to relevant personnel.

[0149] Data storage and transmission:

[0150] Recording and Traceability: The system synchronously stores the raw data of each frame, the calculated temperature field data, alarm records, and the corresponding reel ID number into the real-time database and historical database at a high frequency (10Hz). All data can be stored long-term and supports querying, playback, and export by time, reel number, etc., providing complete data support for process optimization, quality traceability, and fault analysis.

[0151] (5) Correlation analysis is performed between temperature data and three-dimensional model to obtain real-time temperature field model of coil. Using the correspondence between the spatial coordinates of temperature sensor and three-dimensional model, temperature data is mapped into three-dimensional geometric model of coil to generate real-time temperature field model.

[0152] Among them, the three-dimensional model and coordinate system are established:

[0153] High-precision 3D geometric modeling creates a digital 3D model in a virtual environment that perfectly matches the geometry and structure of the real coil, including the mandrel and interlayer components. This model serves as the carrier of temperature data. Unified spatial coordinate system calibration establishes a unified world coordinate system, ensuring that every point on the 3D model corresponds one-to-one with the spatial coordinates of the sensor array's measurement points. This forms the mathematical foundation for accurate mapping.

[0154] Data association and mapping:

[0155] Data acquisition and coordinate transformation involve converting the data measured by the sensor in its own coordinate system to a unified world coordinate system, which is then aligned with the coordinate system of the 3D model. Temperature interpolation calculation is crucial because the sensor measures discrete points, while the coiled surface is continuous. To generate a continuous and smooth temperature field, the temperature of all points on the model surface needs to be calculated using an interpolation algorithm based on the known temperatures of the measured points.

[0156] Model generation and visualization:

[0157] Real-time temperature field model generation assigns calculated temperature values ​​as vertex attributes to each vertex of the 3D model, thus constructing a temperature-geometry fusion model. Color rendering and visualization convert the temperature data into intuitive colors through color mapping, achieving visualization. Based on the generated 3D temperature field model, spatial analysis that is impossible to perform on a 2D plane can be conducted.

[0158] 3. Determine if the measurement results exceed the standard:

[0159] (1) Calculate the relative error of the measurement results based on the standard size data in the historical database. Use the size tolerance evaluation method based on the median and standard deviation to calculate the deviation of each measurement result from the historical data.

[0160] (2) If the error exceeds the preset threshold (i.e., the tolerance threshold), it is determined that the measurement is out of control. The dimensional tolerance threshold is set to ±5%. If any measurement result exceeds this threshold, it is determined that the measurement is out of control.

[0161] (3) Record the type and degree of exceeding the standard and generate early warning information. According to the degree of exceeding the standard, it is divided into three levels: minor, serious, and severe, and corresponding early warning information is generated.

[0162] 4. Conduct a re-inspection of the coil's appearance:

[0163] (1) The oversized coils are diverted to the re-inspection station. The automatic diversion of the coils is achieved by using photoelectric switch sensing technology. The photoelectric switch is used as the sensing organ of the system to detect the arrival and position status of the coils, and provides this signal to the control system (such as PLC) as the basis for triggering the subsequent diversion.

[0164] (2) Perform the measurement again and compare it with the original measurement results. Use the same measurement process to perform the remeasurement and generate a new three-dimensional model and temperature field model.

[0165] (3) Based on the comparison results, determine whether rework is required. If the retest results still exceed the standard, it is determined to be seriously exceeding the standard and manual intervention is required for rework.

[0166] 5. Complete the warehousing operation:

[0167] (1) Automatically generate inbound instructions based on measurement results. Automatically generate inbound instructions based on measurement results and inventory requirements.

[0168] (2) Move the coil to the storage location. Use an automatic guide device to move the coil to the storage location.

[0169] (3) Complete the inbound operation and update the database records. Record the inbound data such as inbound time and location information, and update the database.

[0170] Example 2:

[0171] 1. Initialize system configuration:

[0172] (1) Start the measurement system and establish a communication connection. The measurement system includes a multi-view 3D reconstruction module, a temperature field synchronous monitoring module, and an intelligent early warning decision module. The multi-view 3D reconstruction module consists of two infrared thermal imaging cameras, installed directly above the center of the wire coil conveyor rollers, at a vertical distance of 2.5 meters from the conveyor rollers. The imaging equipment has a sampling frequency of 5 frames / second, and the field of view of the infrared thermal imaging camera must be greater than the width of the wire coil conveyor rollers. The temperature field synchronous monitoring module uses a non-contact infrared temperature sensor array, evenly arranged around the wire coil, with a temperature measurement range of 0-300℃. The intelligent early warning decision module includes an industrial control computer and data processing software.

[0173] (2) Initialize the three-dimensional reconstruction model and temperature monitoring model. The median filtering algorithm is used to denoise the acquired original image, and then the threshold segmentation method is used to obtain the edge contour of the coil. The maximum allowable curvature change threshold is set to 0.02, and the least squares fitting algorithm is used to fit the equivalent ellipse. The center point position and major axis length of the calculated equivalent ellipse are used as the center position and diameter of the wire coil layer corresponding to the curve segment combination.

[0174] (3) Load the historical dimensional tolerance database. Organize and classify the collected historical measurement data to establish a dynamic dimensional tolerance database, which includes historical data on key dimensions such as coil length, diameter, and thickness, with more than 15,000 sets of data.

[0175] 2. Perform three-dimensional online measurement of the coil shape:

[0176] (1) Multi-angle image data of the coil were acquired using a multi-view 3D reconstruction module. Edge detection was performed on each image using a threshold segmentation method to obtain the outline of the coil. The coil was scanned line by line from the left side of the field of view to the right side to obtain all left boundary points; in the reverse direction, the coil was scanned from the right side of the field of view to the left side to obtain all right boundary points. All left boundary points were combined to form the left contour edge curve, and all right boundary points were combined to form the right contour edge curve.

[0177] (2) Preprocess the acquired image data. Median filtering algorithm is used to reduce noise in each image, and then image enhancement algorithm is used to adjust the brightness and contrast of the image.

[0178] (3) Reconstruct the three-dimensional model of the coil using image processing algorithms. A three-dimensional reconstruction algorithm based on multi-view images is used, combined with the depth information of the images, to generate a three-dimensional geometric model of the coil.

[0179] (4) Real-time monitoring of the temperature distribution on the surface of the coil, combined with a temperature field synchronous monitoring module. A non-contact infrared temperature sensor array is used to monitor the temperature field distribution on the surface of the coil in real time, with a temperature measurement range of 0-300℃ and a temperature measurement frequency of 15Hz.

[0180] (5) Correlation analysis is performed between temperature data and three-dimensional model to obtain real-time temperature field model of coil. Using the correspondence between the spatial coordinates of temperature sensor and three-dimensional model, temperature data is mapped into three-dimensional geometric model of coil to generate real-time temperature field model.

[0181] 3. Determine if the measurement results exceed the standard:

[0182] (1) Calculate the relative error of the measurement results based on the standard size data in the historical database. Use the size tolerance evaluation method based on the median and standard deviation to calculate the deviation of each measurement result from the historical data.

[0183] (2) If the error exceeds the preset threshold, it is determined that the measurement is out of control. The dimensional tolerance threshold is set to ±3%. If any measurement result exceeds this threshold, it is determined that the measurement is out of control.

[0184] (3) Record the type and degree of exceeding the standard and generate early warning information. According to the degree of exceeding the standard, it is divided into three levels: minor, serious, and severe, and corresponding early warning information is generated.

[0185] 4. Conduct a re-inspection of the coil's appearance:

[0186] (1) The oversized coils are diverted to the re-inspection station. Photoelectric switch sensing technology is used to realize the automatic diversion of the coils.

[0187] (2) Perform the measurement again and compare it with the original measurement results. Use the same measurement process to perform the remeasurement and generate a new three-dimensional model and temperature field model.

[0188] (3) Based on the comparison results, determine whether rework is required. If the retest results still exceed the standard, it is determined to be seriously exceeding the standard and manual intervention is required for rework.

[0189] 5. Complete the warehousing operation:

[0190] (1) Automatically generate inbound instructions based on measurement results. Automatically generate inbound instructions based on measurement results and inventory requirements.

[0191] (2) Move the coil to the storage location. Use an automatic guide device to move the coil to the storage location.

[0192] (3) Complete the inbound operation and update the database records. Record the inbound data such as inbound time and location information, and update the database.

[0193] This invention includes:

[0194] (1) Multi-dimensional data fusion: A spatiotemporal alignment algorithm for temperature field and three-dimensional topography is proposed to eliminate measurement errors caused by thermal expansion. The coil is no longer regarded as a rigid body, but as a flexible body that dynamically changes with the temperature field. By establishing a thermal expansion deformation model, the theoretical thermal expansion displacement vector of each point is calculated based on the real-time temperature field, thereby dynamically correcting the room temperature three-dimensional model or instantaneous three-dimensional scanning data to a thermal geometric model that perfectly matches the current physical state of the temperature field, achieving accurate spatiotemporal alignment of temperature and geometry.

[0195] By precisely aligning the temperature field and 3D topography data, deformation errors caused by uneven thermal expansion can be effectively reduced, thereby improving the accuracy and reliability of the measurement results. Furthermore, by considering the dynamic changes in the temperature field, the algorithm further optimizes the 3D topography reconstruction process, ensuring stable measurement results under different temperature conditions.

[0196] (2) Dynamic benchmark self-correction: A dynamic database of dimensional tolerances is established based on historical data, and the detection threshold is optimized in real time. The historical data consists of more than 10,000 samples. Using a large amount of historical data, the dimensional characteristics of wire coils of different batches and specifications are analyzed through machine learning algorithms, and the detection threshold is dynamically adjusted to adapt to the slight changes in the production process, ensuring the flexibility and adaptability of the detection. In addition, through continuous data updates and learning capabilities, the system can automatically adjust the threshold setting according to newly emerging sample data, further improving the accuracy and stability of the detection. That is, the traditional, fixed tolerance threshold, such as diameter ±10mm, is abandoned. Instead, the qualified size is considered to be a dynamic statistical range, which is distributed probabilistically around a target value, and the shape of this distribution, such as the mean and standard deviation, will change dynamically with the slow changes in the production process, i.e., process drift. The goal is to enable the detection system to be like an experienced master craftsman, able to perceive the subtle changes in the production process, such as mold wear and material batch differences, and intelligently adjust the judgment criteria, i.e., the threshold, thereby reducing false alarms and false negatives and achieving accurate quality inspection.

[0197] (3) Online closed-loop control: The detection results are directly linked to the automated warehouse scheduling system, and the oversized coils are automatically diverted to the re-inspection station. Through the real-time closed-loop control system, the detection results are combined with the automated scheduling of the production line, realizing the high efficiency and intelligence of the production process, reducing the need for manual intervention, and improving production efficiency and product quality stability. In addition, through intelligent scheduling algorithms, the operating speed of the production line and the configuration of workstations can be automatically adjusted according to the degree of abnormality of the detection results, ensuring the continuity and efficiency of the production process.

[0198] The algorithm dynamically balances production line speed and quality risk, with quality risk quantified by inspection results. When an anomaly is detected, it signifies increased volatility in the production process and a higher quality risk. The algorithm reduces production speed to allow more processing time for subsequent workstations, such as manual re-inspection. Simultaneously, it optimizes workstation configuration to isolate and handle defective products, preventing the problem from spreading. Ultimately, it aims for overall efficiency and optimization while ensuring product quality.

[0199] The dimensional inspection accuracy has been improved to ±5mm, a fourfold increase compared to traditional methods; inspection time has been reduced from 5 minutes of manual sampling to within 30 seconds; the warehousing pass rate has increased from 92% to 99%, and production line efficiency has increased by 60%. Through the application of machine vision recognition technology, real-time monitoring and data feedback of the shape of coiled wire can be achieved during the production process, ensuring the stability of the production process and the consistency of product quality. Furthermore, this method can also alert operators to potential quality problems through an intelligent early warning mechanism, further improving the controllability of the production process and the overall quality of the product. The innovation of this method lies in its combination of advanced machine vision recognition technology and an intelligent early warning mechanism, which can provide high-precision and high-efficiency measurement and monitoring services in high-speed production environments, ensuring product quality and the safety of the production process. In this way, enterprises can significantly reduce the impact of human error, improve product consistency and reliability, and meet the high standards of industrial production.

[0200] The existing inspection device has low accuracy and causes issues with material leaving the warehouse after it has been received. Furthermore, if the existing inspection device detects non-compliant (NG) material, it requires a complete re-processing, which is cumbersome and inefficient. Requirements: 1. Implement a new inspection device at the front end to pre-detect NG materials. If an NG material is detected, the original inspection and warehousing process should be stopped. 2. Add a temperature detection function capable of detecting the temperature of the upper, middle, and lower sections of the steel coil. Accuracy: ±5mm; Time: <1min; Coil length: 1100-2100mm; Coil diameter: 1250-1450mm.

[0201] Process and Functions: A new visual inspection and temperature detection function is added to the original detection device. Upon receiving the customer's PF line arrival signal, a 5-10 second delay is made until the workpiece comes to a stop. The module then starts scanning and imaging the workpiece. Once the module returns to its origin, a completion signal is output to the PF line, and the process continues. Simultaneously, the detection results are uploaded to the central control room.

[0202] Testing Procedure: The standby position of the device is near the electrical control cabinet, offset from the steel coil production line to avoid interference and collisions. Upon receiving the start signal, the traverse module moves at a constant speed from the standby position towards the steel coil. Upon reaching the preset testing position, the module continues its constant speed movement, and four 3D cameras begin detecting the outer contour of the steel coil. Simultaneously, two temperature measuring instruments (one above and one below) also begin measuring the temperature of the steel coil. After testing, the 3D cameras and temperature measuring instruments output the test data to the central control unit for comparison to ensure compliance with requirements and for data archiving based on the steel coil number. Then, the traverse module returns to the standby position, waiting for the next product to arrive before repeating the above actions. The electrical control cabinet has an alarm function; if the device malfunctions, the alarm light activates, and personnel can press the stop button to troubleshoot the problem. By using four 3D cameras to photograph the product, comprehensive detection of the outer contour of the product under test is achieved. The system determines the compliance of the product's hump, protrusions, outer diameter, and length, and outputs the test data. This invention employs two thermometers, installed at the top and bottom of the product respectively. The upper thermometer measures the temperature at points 1 and 2, while the lower thermometer measures the temperature at point 3. This invention can detect the product's external dimensions and temperature, outputting the data to a central control unit for archiving. This solution can identify non-compliant products and eliminate them before loading into the grid, improving production efficiency and avoiding the risks to personnel and the work area caused by defective products falling during packaging due to insufficient suction force in the clamping system caused by factors such as humps. The detection data is input into the product nameplate, allowing for real-time access to information such as the product's external dimensions from warehousing and shipping to sales and delivery to the customer, facilitating a detailed understanding of each product and enhancing product reliability.

[0203] Equipment Function: The equipment of this invention is developed based on the user's logistics line automated warehouse inbound requirements and the user's product appearance and size requirements. It can solve the customer's problem of controlling the status of steel coils and providing timely alarm prompts.

[0204] The host computer implements the following functions: 1. To ensure accurate entry of customer steel coils into the automated storage and retrieval system (AS / RS), it measures and outputs the three-dimensional dimensions of the steel coil's outer contour. 2. To detect and determine the hump and coil shape of the steel coil, it uses camera scanning to output 2D images from both side and top views, and calculates the difference between the maximum height and minimum outer diameter of the steel coil's outer contour. This value is used to determine if the steel coil has a hump. 3. For abnormal protrusions in the wire rod, the outer contour can be obtained by stitching together images from the camera. Foreign objects connected to the outer contour are detected in designated areas outside the outer contour, indicating a protruding steel rod. 4. The temperature of the upper, middle, and lower regions of the steel coil is detected by two thermal imagers, and the detected temperature images can be integrated and displayed on the host computer. 5. A database management system is implemented on the local computer to store the original image information, detection parameters, and judgment results of each steel coil. An Excel spreadsheet can be generated daily or weekly according to customer needs. Code: The steel coil number transmitted via the PF line. Length: The maximum length of the three-dimensional outer contour. Outer Diameter: The maximum inner diameter of the outer contour solid dimension. Front View Hump: The value of D2-D1 in the front view. Top View Hump: The value of D2-D1 in the top view. Top Coil Temperature: Detected by thermal imager. Interval Temperature of Coil: Detected by thermal imager. Bottom Coil Temperature: Detected by thermal imager. Warehousing Judgment: Determines whether the steel coil meets warehousing requirements by comparing the detected length and outer diameter data with the set maximum warehousing standard size. Hump Judgment: Determines whether there is a hump phenomenon by comparing the calculated hump value with the set value. Protrusion Judgment: Determines whether there is a steel bar protrusion by checking for foreign objects on the outer contour. Temperature Judgment: Determines whether the standard is met by comparing with the highest warehousing temperature. In addition to the detection data, 3D files of each detected steel coil and 2D image files of each view can be saved on the local computer. 6. Detection data is uploaded to the customer's nearest secondary system via Ethernet for customer use. Steel coil outer contour plane, such as... Figure 5 As shown.

[0205] The beat details are shown in Table 1:

[0206] Table 1. Beat Details

[0207] The configuration list is shown in Table 2:

[0208] Table 2. Beat Details

[0209] This invention improves dimensional inspection accuracy to ±5mm, a fourfold increase over traditional methods; reduces inspection time from 5 minutes for manual sampling to within 30 seconds; increases the warehousing pass rate from 92% to 99%; and improves production line efficiency by 60%. Combining advanced machine vision recognition technology and an intelligent early warning mechanism, it provides high-precision, high-efficiency measurement and monitoring services in high-speed production environments, ensuring product quality and production process safety, and meeting the demands of high-standard industrial production. The mechanical configuration list is shown in Table 3, and the electrical configuration list is shown in Table 4.

[0210] Table 3 Mechanical Configuration List

[0211] Table 4 List of Electrical Configuration Brands

[0212] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 3 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores static and dynamic information data. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements the steps in the above method embodiments.

[0213] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the computer device to which the present invention is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0214] In addition, the present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0215] In addition, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0216] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0217] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for online measurement of the shape of high-speed wire coils, characterized in that, include: Based on self-organizing mapping and dynamic equilibrium coupling mechanisms, the layout of an infrared temperature sensor array is automatically adjusted and the input-output ratio is optimized to obtain an optimized infrared temperature sensor array. Specifically, this includes: acquiring measurement data of the physical space at the high-speed wire coiling site; based on statistical analysis, within a preset sliding time window, parsing the measurement data according to the environmental complexity index to calculate a two-dimensional environmental complexity distribution function, which comprehensively reflects the temperature gradient change rate, flow field chaos, humidity change frequency, and the degree of information obstruction caused by spatial obstacles; using the two-dimensional environmental complexity distribution function as input parameters to control the node topology of the self-organizing mapping model and adjust the initial distribution density of each infrared temperature sensor node in the two-dimensional array space; establishing a node connection matrix, and outputting the node density based on the initial distribution density from the self-organizing mapping model. The distribution function is used as the monitoring demand function, and the node density distribution function is used as the monitoring supply function. Marginal benefit and marginal cost functions are calculated. By comparing the marginal benefit and marginal cost functions, the marginal difference function is calculated at a local scale, and the input-output ratio of the current infrared temperature sensor array layout is determined. A dynamic equilibrium principle is constructed: when the input-output ratio of the infrared temperature sensor array layout is greater than zero, an instruction to increase node density is generated; otherwise, an instruction to decrease node density is generated. Based on the dynamic equilibrium principle, the marginal difference function and input-output ratio are transformed into parameter correction vectors for the self-organizing map model. The parameter correction vectors are input into the weight update mechanism of the self-organizing map model to change the node spacing and density distribution. The training process of the self-organizing map model is rerun to obtain the optimized infrared temperature sensor array. The image outline of the coil is obtained, and a three-dimensional model of the coil is constructed by combining image processing algorithms; the temperature field distribution of the coil is obtained by using an optimized infrared temperature sensor array. By correlating the three-dimensional model of the coil with the temperature field distribution of the coil, a real-time temperature field model of the coil is obtained. The deviation between the real-time temperature field model and historical data is calculated, and the non-compliant coils are screened out by comparing them with the preset tolerance threshold. The real-time temperature field model of the coils exceeding the standard is regenerated, and the deviation between the real-time temperature field model and historical data is recalculated to obtain the retest results of the coils; if the retest results are still exceeding the standard, the coils are judged to be seriously exceeding the standard. The calculation of the marginal benefit function includes: The supply-demand difference function is obtained by performing point-by-point difference calculations on the monitoring demand function and the monitoring supply function; the supply-demand difference corresponding to the supply-demand difference function is coupled with the improvement of monitoring accuracy to obtain the marginal benefit function.

2. The method for online measurement of the shape of high-speed wire coils according to claim 1, characterized in that, Using the two-dimensional environmental complexity distribution function as the monitoring demand function and the node density distribution function as the monitoring supply function, the marginal cost function is calculated as follows: By using a weighted linear combination, the two-dimensional environmental complexity distribution function is transformed into a monitoring demand function; by combining the single-node monitoring capability constant, the node density distribution function is transformed into a monitoring supply function. The system acquires current sensor hardware cost data, unit node energy consumption data, data transmission energy consumption parameters, and maintenance cost parameters, and sums them up node by node to obtain the marginal cost function.

3. The method for online measurement of the shape of high-speed wire coils according to claim 2, characterized in that, The monitoring requirement function is: ； In the formula, Q d To monitor the demand function, G T Let H be the distribution function of the rate of change of the temperature gradient. F Let F be the chaotic distribution function of the airflow field. H Let S be the frequency distribution function of humidity variation. O Let w be the distribution function of the degree of obstacle occlusion. T w F w H w O For different weighting coefficients; The monitoring supply function is: ； In the formula, Q s To monitor the supply function, M cap D represents the monitoring capability per unit area of ​​a single infrared temperature sensor node. init This is the node density distribution function.

4. The method for online measurement of the shape of high-speed wire coils according to claim 1, characterized in that, The process of transforming the marginal difference function and input-output ratio into a parameter correction vector for the self-organizing map model based on the dynamic equilibrium principle includes: The dynamic equilibrium principle is applied. When the input-output ratio is greater than zero, the density of nodes with a positive difference between marginal benefit and cost is increased and converted into a correction amount according to a proportional coefficient. The correction amount for nodes with a non-positive difference is zero. When the input-output ratio is less than or equal to zero, the density of nodes with a negative difference between marginal benefit and cost is decreased and converted into a correction amount according to a proportional coefficient. The correction amount for nodes with a non-negative difference is zero. At the same time, the node correction amounts are collected to obtain a set of node density correction magnitudes. The physical coordinates of each node are mapped to the topological index of the self-organizing map model, and the corresponding node density correction magnitude is converted into weight correction amount, and the mapping method adopts a linear proportional mapping method. The weight correction amount is integrated with the corresponding node index number and position coordinate information to obtain the parameter correction vector.

5. The method for online measurement of the shape of high-speed wire coils according to claim 1, characterized in that, The process of acquiring the image contour of the reel and constructing a 3D model of the reel using image processing algorithms includes: Multi-angle image data of the coil were acquired, and edge detection was performed on each image using a two-way threshold segmentation method to obtain the outline of the coil. A 3D reconstruction algorithm based on multi-angle images is used, combined with the depth information of the images, to generate a 3D model of the coil.

6. The method for online measurement of the shape of high-speed wire coils according to claim 1, characterized in that, The process of acquiring the temperature field distribution of the coil using the optimized infrared temperature sensor array includes: The optimized infrared temperature sensor array is used to monitor discrete temperature points of the coil and map them into a two-dimensional matrix according to the pre-calibrated spatial coordinates. An interpolation algorithm is used to fill the blank areas in a two-dimensional matrix to generate the temperature field distribution of the coil.

7. The method for online measurement of the shape of high-speed wire coils according to claim 1, characterized in that, The step of correlating the three-dimensional model of the coil with its temperature field distribution to obtain a real-time temperature field model of the coil includes: The temperature field distribution data of the coil is transformed into the world coordinate system through coordinate transformation and then aligned with the three-dimensional model of the coil. Temperature values ​​are assigned as vertex attributes to each vertex of the coiled 3D model, and color mapping is used to convert the temperature values ​​into corresponding colors.

8. The method for online measurement of the shape of high-speed wire coils according to claim 1, characterized in that, The deviation between the real-time temperature field model and historical data is calculated, and the discrepancies are compared with a preset tolerance threshold to identify coils that exceed the tolerance. The deviation between the real-time temperature field model and historical data is calculated. When the deviation exceeds the preset tolerance threshold, it is determined that the measurement is out of standard, and the corresponding reel is regarded as the out-of-standard reel. Early warning information is generated based on the type and degree of exceeding the standard.

9. The method for online measurement of the shape of high-speed wire coils according to claim 1, characterized in that, The process involves regenerating the real-time temperature field model for the substandard coils and recalculating the deviation between the real-time temperature field model and historical data to obtain the retest results of the coils. If the retest result still exceeds the standard, the coil is judged to be seriously out of standard, including: The substandard coils are diverted to the re-inspection station; photoelectric switch sensing technology is used to realize the automatic diversion of coils; If the coils that exceed the standard are retested, and the retest results are still found to be coils that exceed the standard, they are judged to be seriously exceeding the standard and reworked.

Images