A method and system for intelligent photoelectric online detection of multi-layered pieces in industrial sewing

Abstract

The present application relates to a kind of intelligent photoelectric on-line detection method and system of multi-layered piece in industrial sewing.The method includes S1: fabric modeling;S2: calibration;S3: working measurement detection;S4: on-line self-learning.The system includes photoelectric sensor module, intelligent measurement and control module and man-machine interaction module.The method of the present application is based on photoelectric sensor and intelligent self-learning mode, by establishing fabric model, the detection of fabric thickness is realized, the working efficiency and detection accuracy of fabric thickness detection are improved, and it is suitable for the needs of modern industrial sewing intelligent flexible manufacturing.

Description

Technical Field

[0001] This invention belongs to the field of automated online detection technology for the thickness of multi-layer fabrics based on photoelectric sensing, specifically relating to an intelligent photoelectric online detection method and system for multi-layer cut pieces in industrial sewing. Background Technology

[0002] Industrial sewing typically involves layering multiple fabric pieces and reinforcing edges. Whether the fabric pieces are placed automatically or manually, the accuracy of their placement and the correct number of layers are crucial; any defect or error can lead to product quality issues or complete scrapping. Currently, quality control before the sewing process commonly uses traditional contact-type thickness gauges, which use cylinders or manual inspection at the sewing points to determine if the multiple layers of fabric pieces are correctly placed. While this method generally offers high accuracy, it is time-consuming, inefficient, and costly, failing to meet the demands of intelligent and flexible manufacturing. Especially with numerous sewing points, complex processes, and diverse fabric materials, it is difficult to balance speed and accuracy, becoming one of the main factors restricting the improvement of high-reliability industrial sewing efficiency.

[0003] The challenges of online inspection of multi-layer fabric pieces in industrial sewing lie in: limited space for fabric inspection after fixture setup; uncertain pre-pressure of multi-layer fabric pieces; single-sided fabric coating; fabric structural elasticity; and diverse process layering. Currently, among commonly used fabric thickness measurement methods, the accuracy of contact-based eddy current and capacitive methods is significantly affected by the area and pressure of the presser foot, especially for multi-layered textured fabrics. Non-contact light reflection methods are difficult to determine due to the gaps between multi-layer fabric pieces. Ultrasonic thickness measurement is difficult to obtain effective signals due to the fabric structure. X-rays have too low attenuation for ordinary fabrics and are unsuitable for densely populated production workshops. Laser reflection thickness measurement is suitable for single-layer thin materials but is costly. Optical interferometry is suitable for transparent and semi-transparent films. Furthermore, infrared transmission methods are generally difficult to perform stable quantitative measurements because the transmitted incident energy does not completely obey the Lambert-Beer law and is affected by factors such as fabric weight per square meter, wavelength, color, fabric tightness, and temperature. Patent CN201721065670.7 mentions "a through-beam photoelectric sensor to detect the thickness of the fabric piece and convert it into the corresponding number of fabric layers," but it does not explain how to detect and effectively convert the thickness. Summary of the Invention

[0004] This invention addresses the above-mentioned problems by proposing an intelligent photoelectric online detection method and system for multi-layer fabric pieces in industrial sewing, based on the study of fabric light transmittance. This invention utilizes the absorption, reflection, and transmission relationships of near-infrared light penetrating fabric, combined with intelligent control and self-learning methods, to achieve non-contact online detection of multi-layer fabric pieces. This invention effectively solves the quality inspection and confirmation problem of multi-layer fabric pieces after tooling and before sewing in automated industrial sewing systems, and has the advantages of real-time online operation and wide applicability, thus contributing to improving the level of automation in industrial sewing.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A method for intelligent photoelectric online detection of multi-layer cut pieces in industrial sewing, the specific steps of which are as follows:

[0007] S1: Fabric Modeling; Cut pieces with varying numbers of layers and varying coatings are randomly placed on a worktable and fixed using fixtures at the sewing points. Near-infrared light is focused through a lens and penetrates the cut pieces. A photoelectric sensor module directly beneath the cut piece receives the transmitted infrared light signal. The photoelectric sensor module converts the light signal into an electrical signal, which is then processed and converted into a corresponding digital quantity before being output. A model is established between the fabric and the output quantity based on different combinations of fabric pieces and their corresponding output values. After multiple modeling iterations, a reasonable threshold for detecting the number of fabric layers is set.

[0008] S2: Calibration; When the number of fabric layers and coating combination are determined, the corresponding parameters need to be calculated. Based on the determined fabric model, the system's measurement values ​​will produce detection deviations during operation, so periodic calibration is required to ensure the accuracy of the system's detection. In addition, when the number of fabric layers and coating combination, fabric type, fabric material, etc., change, the system needs to remodel the fabric.

[0009] S3: Working Measurement and Inspection; After the fabric to be inspected is placed on the worktable, a fabric model is selected on the human-machine interface, and the output signal is collected and compared with the threshold corresponding to the output signal in the fabric model. The distribution of the current threshold in the threshold sequence of the fabric model can be observed to evaluate the accuracy of the current inspection result. Based on the selected fabric model, the fabric layer count is detected, and the detected fabric layer count and coating information are output.

[0010] S4: Online self-learning; when the fabric model is determined, the output signal D changes with factors such as system working environment, equipment aging, and fabric batches. i The value of will fluctuate, causing the current output signal D to... i The mapped threshold deviates from the threshold sequence in the model. The self-learning module bases its learning on the output signal D. i The system automatically adjusts the preset threshold based on the deviation between the mapping threshold and the threshold sequence of the model.

[0011] Furthermore, the specific steps of step S1 are as follows:

[0012] S11: Different levels N i With or without coating C i The assembled cut pieces are laid flat on the workbench and secured with fixtures at the sewing points.

[0013] S12: Using a near-infrared light source Iin The near-infrared light emitted by the light source is focused by a lens onto the surface of the cut piece to be tested, and the infrared light source I is continuously adjusted. in The power and wavelength are determined to ensure that the photoelectric sensor placed directly below the cut piece receives the light signal after it penetrates the cut piece.

[0014] S13: Of the infrared light illuminating the fabric surface, part is reflected and absorbed by the fabric piece, while the other part penetrates the fabric piece and converges onto the photoelectric sensor. After multiple measurements and analyses, the light intensity I of the infrared light source is calculated using the following formula. in And the photoelectric sensor receives the light signal light intensity I out Mapping relationship between them:

[0015]

[0016] Among them, I offset α represents the luminous flux loss from the lens to the fabric surface, α is the attenuation coefficient of infrared light penetrating each layer of fabric, β is the light intensity attenuation coefficient of the coating on the top surface of the fabric, and γ is the light intensity attenuation coefficient of the coating on the bottom surface of the fabric. This indicates the number of coating layers on the top surface of the fabric. This indicates the number of coating layers on the underside of the fabric.

[0017] S14: The photoelectric sensor module placed directly below the cut piece receives infrared light I transmitted through the cut piece. out and the optical signal I out Converted to electrical signal X i The cut piece is made of non-transparent fabric and has a certain thickness. The intensity of infrared light passing through the cut piece is I. out The smaller value results in a converted electrical signal X. i It's also very small, so the electrical signal X needs to be... i Enlarge it.

[0018] S15: The system needs to filter out irrelevant signals from electrical signal X during operation. i The influence of this makes the system more resistant to interference. Since the infrared light source band in the system is set to a fixed frequency, a bandpass filter needs to be designed to retain the corresponding electrical signal X' in this band. i The filtered signal X' i Since it is an analog signal, the system processor cannot use it directly; it needs to be converted into a digital signal (D) via an A / D converter. i Output the results.

[0019] S16: Once the number of fabric layers N and the coating combination are determined, the emission power of the infrared light source is dynamically adjusted to change its light intensity I. in The system compares the emission intensity I0 of the infrared light source with the output signal D. iEstablish the current number of cut pieces N and the output signal D. i The model is defined, and a mapping threshold A is set. i .

[0020] S17: Based on different levels N i With or without coating C i The assembled fabric, using the number of cut pieces N and the output signal D in step S16. i The mapping relationship is used to create a threshold sequence A for detecting different fabric layers. ij .

[0021] Furthermore, the specific steps of step S2 are as follows:

[0022] S21: When the number of fabric layers and the coating combination are determined, calculate the light intensity attenuation coefficient in the formula of step S13 according to step S16.

[0023] S22: Through multiple experiments, data analysis and nonlinear fitting, the light intensity attenuation coefficient α of infrared light penetrating each layer of fabric, the light intensity attenuation coefficient β of the coating on the top surface of the fabric, and the light intensity attenuation coefficient γ of the coating on the bottom surface of the fabric are obtained in step S13.

[0024] S23: Based on a defined cloth model, the system will be affected by various factors during operation, leading to changes in the output signal D. i The corresponding threshold and the preset threshold A i The deviation increased, therefore the system needs to be calibrated periodically, and the threshold A needs to be changed. i The range.

[0025] S24: When the number of cloth layers, coating combination, cloth type, cloth material, etc. change, the cloth model created in step S1 is no longer applicable to the current detection, so the cloth model needs to be recreated.

[0026] S25: The newly created cloth model needs to recalculate the light intensity attenuation coefficient in the formula of step S13 according to step S21.

[0027] Furthermore, the specific steps of step S3 are as follows:

[0028] S31: When placing the fabric to be tested on the worktable, according to the number of fabric layers N i With or without coating C i Combine, and select the pre-created cloth model on the human-computer interaction terminal.

[0029] S32: After determining the fabric model, obtain the output signal D. i The output signal D i Compare the current threshold with the threshold corresponding to the output signal in the cloth model, and observe the distribution of the current threshold in the threshold sequence of the cloth model.

[0030] S33: By observing the distribution of the current threshold in the threshold sequence of the fabric model, analyze the deviation between the current threshold and the system's preset threshold sequence to evaluate the accuracy of the current detection result.

[0031] S34: Use the cloth model from step S16 to find a mapping threshold A i and with the threshold sequence A in step S17 ij By comparing the data, we can determine the number of layers and coating information of the current fabric.

[0032] S35: The detected fabric layer number and coating information are output through the communication interface, and users can view the current detection information in real time through the human-machine interface.

[0033] Furthermore, the specific steps of step S4 are as follows:

[0034] S41: When the number of fabric pieces and the coating combination are determined, the output signal D can be obtained based on the fabric model created in step S16. i and mapping threshold A i .

[0035] S42: In step S17, threshold sequence A ij Find the threshold A' of the current piece combination in the middle. i According to the mapping threshold A obtained in step S41 i The thresholds of the two are compared, and the calculation method is as follows:

[0036] ΔA i =|A' i -A i |

[0037] Where, ΔA i This is the absolute value of the threshold difference.

[0038] S43: Based on the threshold difference ΔA obtained in step S42 i The preset threshold A of the dynamic adjustment system i , so that ΔA i The value reaches its minimum.

[0039] S44: When the number of layers and coating combination of the fabric pieces change, the self-learning module adjusts the output signal D. i The degree of change determines whether the current combination of cut pieces has changed.

[0040] To achieve the above method, this invention provides an intelligent photoelectric online inspection system for multi-layer fabric pieces in industrial sewing, comprising a photoelectric sensor module, an intelligent measurement and control module, and a human-machine interaction module. The photoelectric sensor module receives the light intensity transmitted through a focusing lens from an infrared light source onto the fabric. The intelligent measurement and control module calculates the relationship between the infrared light source and the infrared light intensity received by the photoelectric sensor, processes the output signal of the photoelectric sensor, and establishes a fabric model; furthermore, it performs online self-learning and dynamically adjusts the system's measurement thresholds. The human-machine interaction module enables user interaction with the equipment, displays data, receives measurement and inspection results, sets fabric model parameters, calibration parameters, measurement thresholds, and provides alarm prompts for measurement anomalies.

[0041] In this invention, the photoelectric sensor module is used to receive infrared light sources of different power and wavelength from the system, and to focus them through a focusing lens and transmit them through the fabric. The photoelectric sensor placed directly below the fabric receives the transmitted light flux and converts the light signal into an electrical signal for output.

[0042] In this invention, the intelligent measurement and control module is used to collect electrical signals from the photoelectric sensor, calculate the relationship between the infrared light source and the intensity of the infrared light received by the photoelectric sensor, and process the collected electrical signals by amplification, filtering, and A / D conversion to make them into digital quantities that the system can directly use. This module can establish a fabric thickness detection model based on the infrared light source and the collected and processed photoelectric sensor electrical signals, and set the detection threshold. The online self-learning part of the module can automatically compensate for the measurement output caused by external factors during the system detection process based on the fabric model, and can dynamically adjust the measurement threshold according to the actual detection results to improve the accuracy of the system detection.

[0043] In this invention, the human-computer interaction module is used to realize the interaction between the user and the device. System data can be viewed in real time on the display interface, and the system's control parameters, calibration parameters, etc. can also be modified and set through the display interface. The module receives the fabric thickness detection results, and when the detection error is large, the system will alarm to prompt the on-site staff to check.

[0044] The working modes of this invention are divided into three categories: multi-point fixed detection, trajectory tracking detection, and fixed-point scanning detection, to adapt to different cut piece sizes, sewing point tooling fixtures, production line deployment conditions, etc. The specific methods and processes of each mode are as follows:

[0045] The multi-point fixed detection in the working mode is suitable for tooling with multiple fixed sewing points. At each sewing point of the cut piece, a set of photoelectric sensing modules is pre-fixed. Then, multiple photoelectric sensing modules are connected to the intelligent measurement and control module, and then to the human-machine interaction module. During normal production, after the cut piece tooling fixture is in place, the photoelectric sensing modules are triggered to start detection through position sensors, encoders, or manual control commands to detect the number of cut piece layers at the sewing point.

[0046] The trajectory tracking detection in the described working mode is suitable for tooling for complex edge sewing. Through an XY guide gantry (such as X guide rail, Y guide rail), photoelectric sensors and supports move on the plane of the cutting piece tooling, and perform detection along a predefined cutting piece sewing trajectory to obtain real-time measurement data and detect the number of cutting piece layers on the sewing trajectory.

[0047] The fixed-point scanning detection in the working mode is applicable when several sewing points or sewing trajectories are on a line, and the cutting piece fixture moves on a conveyor belt or track. The photoelectric sensing module continuously measures and obtains continuous layer thickness values ​​to detect the number of cutting piece layers on the sewing trajectory.

[0048] Compared with the prior art, the present invention has the following advantages:

[0049] This invention employs a photoelectric sensor and an intelligent self-learning method for detecting fabric thickness, overcoming the drawbacks of traditional contact-based thickness measurement methods, such as long processing time, low efficiency, and high cost. Furthermore, it adapts to the needs of modern industrial sewing intelligent and flexible manufacturing. In fabric thickness detection, the system improves detection efficiency by establishing a fabric model; and through online self-learning, it adjusts the detection threshold promptly based on detection deviations, thereby improving the system's detection accuracy. Attached Figure Description

[0050] Figure 1 This is a flowchart of an intelligent photoelectric online detection method for multi-layer cut pieces in industrial sewing, according to the present invention.

[0051] Figure 2 This is a schematic diagram of the structure of an intelligent photoelectric online detection system for multi-layer cut pieces in industrial sewing according to the present invention.

[0052] Figure 3 This is a schematic diagram of the fabric model creation process of the present invention.

[0053] Figure 4 This is a schematic diagram of the system workflow of the present invention. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of this invention application clearer, the specific embodiments of this invention will be further described and explained below in conjunction with the accompanying drawings.

[0055] like Figure 1 As shown, an intelligent photoelectric online detection method for multi-layer cut pieces in industrial sewing is implemented through the following steps:

[0056] S1: Fabric modeling, near-infrared light selection and signal processing. Fabric pieces with different layers and varying coatings are randomly placed on the worktable and fixed using fixtures at the sewing points. A near-infrared light source, focused through a lens, penetrates the fabric piece. A photoelectric sensor module directly beneath the fabric piece receives the transmitted infrared light signal. The photoelectric sensor module converts the light signal into an electrical signal, which is then processed and converted into a corresponding digital quantity before being output. A model is established between the fabric and the output quantity based on different combinations of fabric pieces and their corresponding output values. After multiple modeling iterations, a reasonable threshold for detecting the number of fabric layers is set, such as... Figure 3 As shown, the specific steps are as follows:

[0057] S11: Different levels N i With or without coating C i The assembled cut pieces are laid flat on the workbench and secured with fixtures at the sewing points.

[0058] S12: Using a near-infrared light source I o The near-infrared light emitted by the light source is focused by a lens onto the surface of the cut piece to be tested, and the infrared light source I is continuously adjusted. o The power and wavelength are determined to ensure that the photoelectric sensor placed directly below the cut piece receives the light signal after it penetrates the cut piece.

[0059] S13: Of the infrared light illuminating the fabric surface, part is reflected and absorbed by the fabric piece, while the other part penetrates the fabric piece and converges onto the photoelectric sensor. Through multiple measurements and analyses, this invention derives the infrared light source intensity I. in And the photoelectric sensor receives the light signal light intensity I out The calculation formula between them is:

[0060]

[0061] Among them, I offset α represents the luminous flux loss from the lens to the fabric surface, α is the attenuation coefficient of infrared light penetrating each layer of fabric, β is the light intensity attenuation coefficient of the coating on the top surface of the fabric, and γ is the light intensity attenuation coefficient of the coating on the bottom surface of the fabric. This indicates the number of coating layers on the top surface of the fabric. This indicates the number of coating layers on the underside of the fabric.

[0062] S14: The photoelectric sensor module placed directly below the cut piece receives infrared light I transmitted through the cut piece. out and the optical signal I out Converted to electrical signal X iThe cut piece is made of non-transparent fabric and has a certain thickness. The intensity of infrared light passing through the cut piece is I. out The smaller value results in a converted electrical signal X. i It's also very small, so the electrical signal X needs to be... i Enlarge it.

[0063] S15: The system needs to filter out irrelevant signals from electrical signal X during operation. i The influence of this makes the system more resistant to interference. Since the infrared light source band in the system is set to a fixed frequency, a bandpass filter needs to be designed to retain the corresponding electrical signal X' in this band. i The filtered signal X' i Since it is an analog signal, the system processor cannot use it directly; it needs to be converted into a digital signal (D) via an A / D converter. i Output the results.

[0064] S16: Once the number of fabric layers N and the coating combination are determined, the emission power of the infrared light source is dynamically adjusted to change its light intensity I. in The system compares the emission intensity I0 of the infrared light source with the output signal D. i Establish the current number of cut pieces N and the output signal D. i The model is defined, and a mapping threshold A is set. i .

[0065] S17: Based on different levels N i With or without coating C i The assembled fabric, using the number of cut pieces N and the output signal D in step S16. i The mapping relationship is used to create a threshold sequence A for detecting different fabric layers. ij .

[0066] S2: Calibration. When the number of fabric layers and coating combination are determined, the parameters in the formula of step S13 need to be calculated. Based on the determined fabric model, the system's measured values ​​will exhibit detection deviations during operation; therefore, periodic calibration is required to ensure the accuracy of the system's detection. Furthermore, when the number of fabric layers, coating combination, fabric type, or fabric material changes, the system needs to remodel the fabric. The specific steps are as follows:

[0067] S21: When the number of fabric layers and the coating combination are determined, calculate the light intensity attenuation coefficient in the formula of step S13 according to step S16.

[0068] S22: Through multiple experiments, data analysis and nonlinear fitting, the light intensity attenuation coefficient α of infrared light penetrating each layer of fabric, the light intensity attenuation coefficient β of the coating on the top surface of the fabric, and the light intensity attenuation coefficient γ of the coating on the bottom surface of the fabric are obtained in step S13.

[0069] S23: Based on a defined cloth model, the system will be affected by various factors during operation, leading to changes in the output signal D. i The corresponding threshold and the preset threshold A i The deviation increased, therefore the system needs to be calibrated periodically, and the threshold A needs to be changed. i The range.

[0070] S24: When the number of cloth layers, coating combination, cloth type, cloth material, etc. change, the cloth model created in step S1 is no longer applicable to the current detection, so the cloth model needs to be recreated.

[0071] S25: The newly created cloth model needs to recalculate the light intensity attenuation coefficient in the formula of step S13 according to step S21.

[0072] S3: Working Measurement and Inspection. After the fabric to be inspected is placed on the worktable, a fabric model is selected on the human-machine interface, and the output signal is collected and compared with the threshold corresponding to the output signal in the fabric model. Observing the distribution of the current threshold in the threshold sequence of the fabric model can evaluate the accuracy of the current inspection result. Based on the selected fabric model, the fabric layer count is detected, and the detected fabric layer count and coating information are output. The specific steps are as follows:

[0073] S31: When placing the fabric to be tested on the worktable, according to the number of fabric layers N i With or without coating C i Combine, and select the pre-created cloth model on the human-computer interaction terminal.

[0074] S32: After determining the fabric model, obtain the output signal D. i The output signal D i Compare the current threshold with the threshold corresponding to the output signal in the cloth model, and observe the distribution of the current threshold in the threshold sequence of the cloth model.

[0075] S33: By observing the distribution of the current threshold in the threshold sequence of the fabric model, analyze the deviation between the current threshold and the system's preset threshold sequence to evaluate the accuracy of the current detection result.

[0076] S34: Use the cloth model from step S16 to find a mapping threshold A i and with the threshold sequence A in step S17 ij By comparing the data, we can determine the number of layers and coating information of the current fabric.

[0077] S35: The detected fabric layer number and coating information are output through the communication interface, and users can view the current detection information in real time through the human-machine interface.

[0078] S4: Online self-learning. When the fabric model is determined, the output signal D changes with factors such as system operating environment, equipment aging, and fabric batches. i The value of will fluctuate, causing the current output signal D to... i The mapped threshold deviates from the threshold sequence in the model. The self-learning module bases its learning on the output signal D. i The system automatically adjusts the preset threshold based on the deviation between the mapping threshold and the model's threshold sequence. The specific steps are as follows:

[0079] S41: When the number of fabric pieces and the coating combination are determined, the output signal D can be obtained based on the fabric model created in step S16. i and mapping threshold A i .

[0080] S42: In step S17, threshold sequence A ij Find the threshold A' of the current piece combination in the middle. i According to the mapping threshold A obtained in step S41 i The thresholds of the two are compared, and the calculation method is as follows:

[0081] ΔA i =|A' i -A i |

[0082] Where, ΔA i This is the absolute value of the threshold difference.

[0083] S43: Based on the threshold difference ΔA obtained in step S42 i The preset threshold A of the dynamic adjustment system i , so that ΔA i The value reaches its minimum.

[0084] S44: When the number of layers and coating combination of the fabric pieces change, the self-learning module adjusts the output signal D. i Determine the degree of change to determine whether the current pattern combination has changed.

[0085] like Figure 2As shown, this invention provides an intelligent photoelectric online inspection system for multi-layer fabric pieces in industrial sewing, comprising a photoelectric sensor module, an intelligent measurement and control module, and a human-machine interaction module. The photoelectric sensor module receives the light intensity transmitted through a focusing lens from an infrared light source onto the fabric. The intelligent measurement and control module calculates the relationship between the infrared light source and the infrared light intensity received by the photoelectric sensor, processes the output signal of the photoelectric sensor, and establishes a fabric model; furthermore, it performs online self-learning and dynamically adjusts the system's measurement thresholds. The human-machine interaction module enables user interaction with the equipment, displays data, receives measurement and inspection results, sets fabric model parameters, calibration parameters, measurement thresholds, and provides alarm prompts for measurement anomalies.

[0086] The photoelectric sensing module mainly consists of a near-infrared light source 1-1, a focusing lens 1-2, and a photoelectric sensor 1-7. Cut pieces 1-4 with different layers and varying degrees of coating are placed on a workbench 1-5 and fixed using fixtures 1-3 and 1-6 at the sewing points. The near-infrared light source 1-1 passes through the focusing lens 1-2 and penetrates the cut piece 1-4. The photoelectric sensor 1-7, located directly below the cut piece 1-4, receives the transmitted infrared light signal.

[0087] The intelligent measurement and control modules 1-8 mainly consist of: a modulation drive module 1-9, used to control the emission power and pulse width of the near-infrared light source 1-1; a material model processing model 1-10, used to describe the light transmission characteristics and layering characteristics of sewn fabrics; a calibration module 1-11, used for measurement and calibration of new fabrics or new products and periodic production; a self-learning module 1-12, used for automatic online compensation of measurement drift caused by temperature changes, aging, fabric batch changes, etc.; a threshold management module 1-13, used to set the threshold sequence for detecting the number of fabric pieces corresponding to this workstation; a measurement and detection module 1-15, used to preprocess the acquired signals and compare them with the current threshold sequence to obtain the measurement results; a signal acquisition module 1-16, used to receive the output signals of the photoelectric sensor 1-7; and a signal output module 1-17, used to output the current measured value and the determination result of the number of fabric pieces through the communication interface and the IO interface.

[0088] The human-machine interaction modules 1-14 mainly consist of a computer or PLC with interface display, parameter and control input, control output, etc., used for real-time monitoring of online detection, parameter input during calibration, output control of detection results, and alarms.

[0089] like Figure 4 As shown, the working modes of this invention mainly employ three types: multi-point fixed detection, trajectory tracking detection, and fixed-point scanning detection, to adapt to different cut piece sizes, sewing point tooling fixtures, production line deployment conditions, etc., as detailed below:

[0090] The multi-point fixed detection in the described working mode is suitable for fixtures with multiple fixed sewing points. At each sewing point of the cut piece, a set of photoelectric sensor modules is pre-set and fixed. Then, multiple photoelectric sensor modules (e.g., 2-7, 2-6) are connected to an intelligent measurement and control module (e.g., 2-1), and then to a human-machine interaction module (e.g., 2-2). During normal production, after the cut piece fixture is in place, the photoelectric sensor modules are triggered to start detection via a position sensor, encoder, or manual control command, according to... Figure 1 The method shown is used to detect the number of fabric layers at the sewing point.

[0091] The trajectory tracking detection in the described working mode is suitable for fixtures used in complex edge sewing. Through an XY guide rail gantry (such as X guide rail 2-3, Y guide rail 2-4), the photoelectric sensor module 2-6 and the bracket 2-5 move on the plane of the cut piece fixture, detecting along a predefined cut piece sewing trajectory to obtain real-time measurement data. Figure 1 The method shown is used to detect the number of fabric layers on the sewing track.

[0092] The fixed-point scanning detection in the described working mode is suitable for situations where several sewing points or sewing trajectories are on a line, and the cut piece fixture moves along a conveyor belt or track. During this process, the photoelectric sensing module continuously measures and obtains continuous layer thickness values. Figure 1 The method shown is used to detect the number of fabric layers on the sewing track.

[0093] In summary, this invention proposes an intelligent photoelectric online detection method and device for multi-layer fabric pieces in industrial sewing. It employs photoelectric sensors and an intelligent self-learning method for detecting fabric thickness, overcoming the drawbacks of traditional contact-based thickness measurement methods, such as long processing time, low efficiency, and high cost. Furthermore, it adapts to the needs of modern intelligent and flexible manufacturing in industrial sewing. In fabric thickness detection, the system improves detection efficiency by establishing a fabric model; and through online self-learning, it adjusts the detection threshold promptly based on detection deviations, thereby improving the system's detection accuracy.

[0094] The specific implementation steps described in this invention are merely illustrative of the spirit of the invention. Those skilled in the art to which this invention pertains may make various modifications or additions to the described specific implementation steps or use similar methods to replace them, but without departing from the spirit of the invention or exceeding the scope defined by the appended claims.

Claims

1. A method for intelligent photoelectric online detection of multi-layer cut pieces in industrial sewing, characterized in that, The specific steps are as follows: S1: Fabric Modeling; Cut pieces with different layers and varying coatings are randomly placed on the workbench and fixed using fixtures at the sewing points. Near-infrared light is focused through a lens and penetrates the cut pieces. A photoelectric sensor module directly below the cut pieces receives the transmitted infrared light signal. The photoelectric sensor module converts the light signal into an electrical signal, which is then processed and converted into a corresponding digital quantity before being output. A model is established between the fabric and the output quantity based on different combinations of fabric pieces and their corresponding output values. After multiple modeling iterations, a reasonable threshold for detecting the number of fabric pieces is set. S2: Calibration; When the number of fabric layers and coating combination are determined, the corresponding parameters need to be calculated; Based on the determined fabric model, the system measurement values ​​will produce detection deviations during operation, so periodic calibration is required to ensure the accuracy of the system detection; In addition, when the number of fabric layers and coating combination, fabric type, or fabric material changes, the system needs to remodel the fabric. S3: Working measurement and detection; After the fabric to be tested is placed on the workbench, the fabric model is selected on the human-machine interface terminal, and the output signal is collected and compared with the threshold corresponding to the output signal in the fabric model. The distribution of the current threshold in the threshold sequence of the fabric model is observed, and the accuracy of the current detection result is evaluated. Based on the selected fabric model, the fabric layer count is detected, and the detected fabric layer count and coating information are output. S4: Online self-learning; when the fabric model is determined, the output signal changes with the system's working environment, equipment aging, and fabric batch factors. The value will fluctuate, affecting the current output signal. The mapping threshold deviates from the threshold sequence in the model; the self-learning module adjusts the output signal accordingly. The system automatically adjusts the preset threshold based on the deviation between the mapping threshold and the threshold sequence of the model. The specific steps of step S1 are as follows: S11: Different number of layers With or without coating The assembled cut pieces are laid flat on the workbench and secured with fixtures at the sewing points; S12: Uses a near-infrared light source The near-infrared light emitted by the light source is focused by a lens onto the surface of the cut piece to be tested, and the infrared light source is continuously adjusted. The power and wavelength are determined to ensure that the photoelectric sensor placed directly below the cut piece receives the light signal after it penetrates the cut piece; S13: Of the infrared light illuminating the fabric surface, part is reflected and absorbed by the fabric piece, while the other part penetrates the fabric piece and converges onto the photoelectric sensor. After multiple measurements and analyses, the light intensity of the infrared light source is calculated using the following formula. And photoelectric sensors receive light signals and light intensity Mapping relationship between them: ； in, This represents the light flux lost from the lens to the fabric surface. α The attenuation coefficient of infrared light penetrating each layer of fabric. β The light intensity attenuation coefficient of the coating on the top surface of the fabric. γ The light intensity attenuation coefficient of the coating on the bottom surface of the fabric. This indicates the number of coating layers on the top surface of the fabric. Indicates the number of coating layers on the underside of the fabric; S14: The photoelectric sensor module placed directly below the cut piece receives infrared light penetrating the cut piece. and the light signal Converted into electrical signals The cut piece is made of non-transparent fabric and has a certain thickness. The intensity of infrared light passing through the cut piece is... The smaller the value, the smaller the converted electrical signal. It's also very small, so the electrical signal needs to be... Enlarge; S15: The system needs to filter out irrelevant signals from electrical signals during operation. The influence of this makes the system more resistant to interference; since the infrared light source band in the system is set to a fixed frequency, a bandpass filter needs to be designed to retain the electrical signal corresponding to this band. Filtered signal Since it is an analog signal, the system processor cannot use it directly; it needs to be converted into a digital signal via an A / D converter. Output; S16: When the number of fabric layers N Once the coating combination is determined, the emission power of the infrared light source is dynamically adjusted to change its light intensity. The system compares the emission intensity of infrared light sources. and output signal Establish the current number of cut pieces layers. N and output signal The model is defined, and a mapping threshold is set. ; S17: According to different levels With or without coating The assembled fabric uses the number of cut pieces layers from step S16. N and output signal The mapping relationship is used to create threshold sequences for detecting different fabric layers. .

2. The intelligent photoelectric online detection method for multi-layer cut pieces in industrial sewing according to claim 1, characterized in that, The specific steps of step S2 are as follows: S21: When the number of fabric layers and the coating combination are determined, calculate the light intensity attenuation coefficient in the formula of step S13 according to step S16. S22: Through multiple experiments, data analysis, and nonlinear fitting, the light intensity attenuation coefficient of infrared light penetrating each layer of fabric in step S13 is obtained. α The light intensity attenuation coefficient of the coating on the top surface of the fabric β The light intensity attenuation coefficient of the coating on the underside of the fabric γ ; S23: Based on a defined cloth model, the system will be affected by various factors during operation, leading to changes in the output signal. The corresponding threshold and the preset threshold The deviation is increasing, therefore the system needs to be calibrated periodically and the threshold needs to be changed. Scope; S24: When the number of fabric layers, coating combination, fabric type, or fabric material changes, the fabric model created in step S1 is no longer applicable to the current detection, so the fabric model needs to be recreated. S25: The newly created cloth model needs to recalculate the light intensity attenuation coefficient in the formula of step S13 according to step S21.

3. The intelligent photoelectric online detection method for multi-layer cut pieces in industrial sewing according to claim 1, characterized in that, The specific steps of step S3 are as follows: S31: When placing the fabric to be tested on the worktable, according to the number of fabric layers... With or without coating Combine the existing cloth models by selecting them on the human-computer interaction interface. S32: After determining the fabric model, obtain the output signal. , output signal Compare the current threshold with the threshold corresponding to the output signal in the cloth model, and observe the distribution of the current threshold in the threshold sequence of the cloth model; S33: By observing the distribution of the current threshold in the threshold sequence of the fabric model, analyze the deviation between the current threshold and the system's preset threshold sequence to evaluate the accuracy of the current detection result; S34: Use the cloth model from step S16 to find a mapping threshold. and the threshold sequence in step S17 By comparing the data, the number of layers and coating information of the current fabric can be determined; S35: The detected fabric layer number and coating information are output through the communication interface, and the user can view the current detection information in real time on the human-machine interface.

4. The intelligent photoelectric online detection method for multi-layer cut pieces in industrial sewing according to claim 1, characterized in that, The specific steps of step S4 are as follows: S41: When the number of fabric pieces and the coating combination are determined, the output signal is obtained based on the fabric model created in step S16. and mapping threshold ; S42: Threshold sequence in step S17 Find the threshold of the current piece combination in the middle. Based on the mapping threshold obtained in step S41 The thresholds of the two are compared, and the calculation method is as follows: ； in, The absolute value of the threshold difference; S43: Based on the threshold difference obtained in step S42 The preset threshold of the dynamic adjustment system ,make The value reaches its minimum; S44: When the number of layers and coating combinations of the fabric pieces change, the self-learning module adjusts the output signal accordingly. The degree of change determines whether the current combination of cut pieces has changed.

5. An intelligent photoelectric online detection system for multi-layer cut pieces in industrial sewing, implementing the intelligent photoelectric online detection method for multi-layer cut pieces in industrial sewing as described in claim 1, characterized in that, The intelligent photoelectric online detection system for multi-layer fabric pieces in industrial sewing includes a photoelectric sensor module, an intelligent measurement and control module, and a human-machine interaction module. The photoelectric sensor module receives the light intensity transmitted through a focusing lens from an infrared light source onto the fabric. The intelligent measurement and control module calculates the relationship between the infrared light source and the infrared light intensity received by the photoelectric sensor, processes the output signal of the photoelectric sensor, and establishes a fabric model. Furthermore, it performs online self-learning and dynamically adjusts the system's measurement thresholds. The human-machine interaction module enables user-device interaction, displays data, receives measurement and detection results, sets fabric model parameters, calibration parameters, and measurement thresholds, and provides alarm prompts for measurement anomalies.

6. The intelligent photoelectric online detection system for multi-layer cut pieces in industrial sewing according to claim 5, characterized in that, The photoelectric sensor module is used to receive infrared light sources of different power and wavelength from the system, and to focus the light through the focusing lens and transmit it through the fabric. The photoelectric sensor placed directly below the fabric receives the transmitted light flux and converts the light signal into an electrical signal for output.

7. The intelligent photoelectric online detection system for multi-layer cut pieces in industrial sewing according to claim 5, characterized in that, The intelligent measurement and control module is used to collect electrical signals from the photoelectric sensor, calculate the relationship between the infrared light source and the intensity of the infrared light received by the photoelectric sensor, and amplify, filter, and perform A / D conversion on the collected electrical signals to make them into digital quantities that the system can directly use. This module can establish a fabric thickness detection model based on the infrared light source and the collected and processed photoelectric sensor electrical signals, and set the detection threshold. The online self-learning part of the module can automatically compensate for the measurement output caused by external factors during the system detection process based on the fabric model, and dynamically adjust the measurement threshold according to the actual detection results to improve the accuracy of the system detection.

8. The intelligent photoelectric online detection system for multi-layer cut pieces in industrial sewing according to claim 5, characterized in that, The human-machine interaction module is used to enable interaction between users and equipment. System data can be viewed in real time on the display interface, and the system's control parameters and calibration parameters can be modified and set through the display interface. The module receives the fabric thickness detection results, and when the detection error is large, the system will alarm to prompt on-site personnel to check.

Images