Laser coding and detecting integrated system based on spectral self-sensing technology

Abstract

The application belongs to the technical field of laser coding management and control, and specifically relates to a laser coding and detection integrated system based on spectral self-sensing technology, which comprises a spectral feature sensing and analyzing module, a material category judging module, a laser parameter self-adaptive matching module, a laser coding execution module, a multi-dimensional coding effect detection module, a defective product rejection control module and a background monitoring end; the application realizes high-precision identification of materials through the material category judging module and outputs material identification, generates an optimal laser coding parameter group based on the material identification and through a built-in mapping model, controls coding based on the optimal laser coding parameter group, and after coding is completed, multi-dimensional coding effect analysis is performed to identify defective coding products and automatically reject them, realizing intelligent management and control of the whole process of laser coding from material identification, parameter matching, coding execution to effect detection and defective rejection, and significantly improving the precision, efficiency and quality stability of coding operation.

Description

Technical Field

[0001] This invention relates to the field of laser marking control technology, specifically to an integrated laser marking and detection system based on spectral self-sensing technology. Background Technology

[0002] In the field of industrial laser marking and quality control, although the development of integrated equipment has achieved the integration of marking and inspection processes, most solutions are still at the level of mechanical splicing. For example, Chinese patent CN207223202U discloses an integrated laser marking and visual inspection machine. This invention achieves automated detection and alarm after marking by linking a laser generator and a visual sensor, which solves the problems of low efficiency and high misjudgment rate of traditional manual inspection. Only one person is needed to complete the marking and inspection operations, which has certain progress in terms of labor cost control and inspection efficiency improvement.

[0003] However, the above technical solutions still have significant limitations: First, no sensing and recognition module is set for the material of the product being coded. The laser coding parameters need to be manually preset based on a fixed product type. When the material of the product to be coded changes (such as switching from a plastic shell to metal parts), the preset parameters cannot be adjusted adaptively, which can easily lead to insufficient coding clarity or damage to the product surface.

[0004] Meanwhile, the visual inspection of the above-mentioned technical solutions only focuses on the basic verification of "whether the coding is correct", without involving multi-dimensional quality analysis such as clarity and character integrity. It does not establish a linkage feedback mechanism with coding parameters, cannot optimize coding parameters based on the detection results, and is difficult to reasonably analyze the performance of laser coding and remind staff to intervene and control it in a timely manner. This is not conducive to achieving intelligent management and control of the entire coding process and significantly improving the accuracy, efficiency and quality stability of coding operations.

[0005] To address the aforementioned technical shortcomings, a solution is proposed. Summary of the Invention

[0006] The purpose of this invention is to provide an integrated laser marking and detection system based on spectral self-sensing technology to address the technical deficiencies mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: an integrated laser marking and detection system based on spectral self-sensing technology, including a spectral feature sensing and analysis module, a material category determination module, a laser parameter adaptive matching module, a laser marking execution module, a multi-dimensional detection module for marking effect, a defective product rejection control module, and a back-end monitoring terminal;

[0008] The spectral feature perception and analysis module collects the reflectance spectral signal of the surface of the product to be coded in real time, and outputs a standardized spectral feature vector S after noise reduction and feature extraction. The material category determination module determines the material category of the product to be coded based on the standardized spectral feature vector S and outputs the material identifier C.

[0009] The laser parameter adaptive matching module generates the optimal laser marking parameter set for the product to be marked based on the material category identifier and product thickness parameters. The laser marking execution module receives the optimal laser marking parameter set and retrieves the marking pattern, controls the laser generator to adjust the output power to the set value, and simultaneously controls the marking terminal to complete the marking on the marking area on the product surface at the set speed and frequency, and sends the marking control information to the back-end monitoring terminal. After marking is completed, the marking effect multi-dimensional detection module analyzes the marking effect of the marked products and sends the marking effect analysis results to the back-end monitoring terminal. When marking defective products are identified, the defective product rejection control module controls the rejection mechanism to perform rejection operations and remove the marking defective products from the production line.

[0010] Furthermore, when the product enters the preset detection area along the conveyor line, the spectral sensor built into the spectral feature perception and analysis module emits probe light to the product surface, receives the reflected spectral signal and converts it into a raw electrical signal. The original signal is then denoised using a wavelet threshold noise reduction algorithm based on the sym8 wavelet base to remove signal interference caused by ambient light and conveyor line vibration. The top ten principal components are extracted as key spectral features through principal component analysis, and then Z-score standardization is performed on the key features to finally output a standardized spectral feature vector S, which is then transmitted to the material category determination module.

[0011] Furthermore, the material category determination module has a built-in pre-built target material spectral database. The target material spectral database stores standard spectral feature vectors Sstdi for several common materials, where i={1,2,3,…,n}, n is the number of material types, and Sstdi is stored after the spectral signal of the corresponding material standard sample is processed by the spectral feature perception and analysis module.

[0012] After receiving the standardized spectral feature vector S of the product to be coded, the similarity γi between S and each Sstdi is calculated, and the similarity threshold γth is set to 0.92. All similarity γi are sorted from largest to smallest, and each similarity γi is compared with the similarity threshold γth. The corresponding Ci with the largest γi and γi≥γth is selected as the material category C, and the material identifier C of the product to be coded is transmitted to the laser parameter adaptive matching module. If all γi < γth, an "unknown material" warning signal is generated and sent to the background monitoring terminal.

[0013] Furthermore, the laser parameter adaptive matching module receives the material identifier C of the product to be marked, and simultaneously collects the thickness d of the marking area through the integrated laser displacement sensor. It determines the core marking parameters of the product to be marked through the built-in material-parameter mapping model, including laser power P, laser frequency f and marking speed v, forming a complete optimal laser marking parameter set (P,f,v) and transmitting it to the laser marking execution module.

[0014] Furthermore, the specific analysis and judgment process of the multi-dimensional detection module for the masking effect is as follows:

[0015] After coding is completed, a high-resolution industrial camera is activated to capture images of the coding area and perform contour matching integrity detection. The actual coding pattern is compared with the built-in standard coding pattern template. If the contour overlap between the actual coding pattern and the standard coding pattern template does not exceed the preset overlap threshold or there are missing strokes or broken codes in the actual coding pattern, the coded product is marked as a coding defective product.

[0016] If the overlap between the actual coding pattern and the standard coding pattern template exceeds a preset overlap threshold and there are no missing strokes or broken codes in the actual coding pattern, then positional accuracy is checked. The center coordinates (xact, yact) of the actual coding pattern are extracted using an edge detection algorithm, and the deviation is calculated between the actual coding pattern and the preset standard center coordinates (xstd, ystd). If Δx ≤ Δth and Δy ≤ Δth, then the positional accuracy is deemed acceptable, and coding clarity is evaluated and analyzed. If Δx > Δth or Δy > Δth, then the positional accuracy is deemed unacceptable, and the coded product is marked as a defective product.

[0017] Furthermore, the specific analysis process for evaluating the clarity of the masking is as follows:

[0018] After converting the actual coding pattern to grayscale, the difference between the grayscale value of pixel p in the actual coding pattern and the grayscale value of pixel p in the standard coding pattern template is calculated and the absolute value is taken. The absolute value result is then compared with the grayscale value of pixel p in the standard coding pattern template, and the ratio result is multiplied by 100% to obtain the grayscale deviation percentage.

[0019] The grayscale deviation percentage of all pixels is obtained and the average is calculated to obtain the sharpness anomaly value. The percentage of pixels whose grayscale deviation exceeds the preset grayscale deviation percentage threshold is marked as the pixel out-of-bounds value. The sharpness anomaly value and the pixel out-of-bounds value are compared with the preset sharpness anomaly threshold and the preset pixel out-of-bounds threshold respectively. If the sharpness anomaly value or the pixel out-of-bounds value exceeds the corresponding preset threshold, the coding sharpness is judged to be unqualified and the coded product is marked as a coding defective product.

[0020] Furthermore, the back-end monitoring terminal communicates with the manual intervention emergency decision-making module. The manual intervention emergency decision-making module analyzes the laser marking operation performance during the detection period and determines whether a high-emergency signal for manual intervention is generated. When a high-emergency signal for manual intervention is generated, it is sent to the back-end monitoring terminal. When the back-end monitoring terminal receives the high-emergency signal for manual intervention, it issues an early warning.

[0021] Furthermore, the specific analysis process for the manual intervention emergency decision-making module includes:

[0022] The system obtains the total number of coded products and the number of coded defective products during the detection period. It calculates the coded defect count by comparing the number of coded defective products with the total number of coded products. The coded defect count is then compared with a preset coded defect count threshold. If the coded defect count exceeds the preset coded defect count threshold, a high-urgent signal for manual intervention is generated.

[0023] Furthermore, if the number of poor coding detection values ​​does not exceed the preset poor coding detection threshold, the number of times the execution is in a non-matching state during the detection period is obtained and marked as the execution non-matching detection value. The execution non-matching detection value is compared with the preset execution non-matching detection threshold. If the execution non-matching detection value exceeds the preset execution non-matching detection threshold, a high-urgent signal for manual intervention is generated.

[0024] If the non-match detection value does not exceed the preset non-match detection threshold, the average value of all coding status coefficients within the detection period is calculated to obtain the coding status feature value, and the coding status coefficient with the largest value within the detection period is marked as the coding status aberration value. The artificial intervention emergency coefficient is calculated by weighted summation of the non-match detection value, the coding status feature value, and the coding status aberration value. The artificial intervention emergency coefficient is compared with the preset artificial intervention emergency coefficient threshold. If the artificial intervention emergency coefficient exceeds the preset artificial intervention emergency coefficient threshold, a high artificial intervention emergency signal is generated.

[0025] Furthermore, the analysis and judgment process for non-matching states is as follows:

[0026] The actual laser power, actual laser frequency, and actual coding speed of the laser coding execution module are collected during coding. The difference between the actual laser power and the laser power P in the corresponding optimal laser coding parameter group is calculated and the absolute value is taken to obtain the laser power characteristic value. Similarly, the laser frequency characteristic value and coding speed characteristic value are obtained. The coding state coefficient is obtained by weighted summation of the laser power characteristic value, laser frequency characteristic value, and coding speed characteristic value. The coding state coefficient is compared with the preset coding state coefficient threshold. If the coding state coefficient exceeds the preset coding state coefficient threshold, it is determined that the current execution is in a non-matching state.

[0027] Compared with the prior art, the beneficial effects of the present invention are:

[0028] 1. In this invention, a high-precision standardized spectral feature vector is output through a spectral feature perception and analysis module. Based on the spectral feature vector, material identification is performed and a material identifier is output. Based on the material identifier, an optimal laser marking parameter set is generated through a built-in mapping model. Marking control is performed based on the optimal laser marking parameter set. After marking is completed, multi-dimensional marking effect analysis is performed to identify and automatically remove marking defects. This realizes intelligent management of the entire laser marking process from material identification, parameter matching, marking execution to effect detection and defect removal, significantly improving the accuracy, efficiency and quality stability of marking operations.

[0029] 2. In this invention, the laser marking operation performance during the detection period is analyzed by the emergency decision-making module with manual intervention. When an abnormality occurs, a high emergency signal for manual intervention is generated and fed back to the background monitoring terminal, which promptly reminds the staff to intervene and adjust. This effectively avoids the generation of batch defective products due to system abnormalities, further ensuring the continuous and efficient operation of the production line. It is conducive to realizing the automation, intelligence and high-quality control of marking operations, reducing management difficulty, and improving production efficiency and overall product quality stability. Attached Figure Description

[0030] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings;

[0031] Figure 1 This is a system block diagram of Embodiment 1 of the present invention;

[0032] Figure 2 This is a system block diagram of Embodiment 2 of the present invention. Detailed Implementation

[0033] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0034] Example 1: As Figure 1 As shown, the laser marking and detection integrated system based on spectral self-sensing technology proposed in this invention includes a spectral feature sensing and analysis module, a material category determination module, a laser parameter adaptive matching module, a laser marking execution module, a multi-dimensional detection module for marking effect, a defective product rejection control module, and a back-end monitoring terminal.

[0035] The spectral feature perception and analysis module collects the reflection spectrum signal of the surface of the product to be coded in real time. After noise reduction and feature extraction, it outputs a standardized spectral feature vector S. Through wavelet noise reduction and PCA feature extraction, environmental interference is effectively removed, material-specific spectral information is preserved, and standardization processing eliminates the influence of surface roughness differences of different products on the features, providing high-precision basic data for subsequent material identification and improving the accuracy of material identification from the source.

[0036] Specifically, when the product enters the preset detection area along the conveyor line, the spectral sensor built into the spectral feature perception and analysis module emits probe light to the product surface, receives the reflected spectral signal and converts it into a raw electrical signal. The original signal is denoised using a wavelet threshold noise reduction algorithm based on the sym8 wavelet base to remove signal interference caused by ambient light and conveyor line vibration. The first ten principal components are extracted as key spectral features through principal component analysis (PCA), and then the key features are standardized by Z-score. Finally, the standardized spectral feature vector S is output and transmitted to the material category determination module.

[0037] The material category determination module is based on the standardized spectral feature vector S to determine the material category of the product to be coded and output the material identifier C. It does not require manual intervention and achieves fast and high-precision material identification. In addition, the "unknown material" warning mechanism prevents products of non-target materials from entering the coding process, improves system security, and helps reduce invalid coding operations.

[0038] Specifically, the material category determination module has a built-in pre-built target material spectral database (such as storing common materials like pine, 304 stainless steel, and 6061 aluminum alloy). The target material spectral database stores standard spectral feature vectors Sstdi for several common materials, where i = {1, 2, 3, ..., n}, and n is the number of material types. Sstdi is stored after the spectral signal of the corresponding material standard sample is processed by the spectral feature perception and analysis module. Each Sstdi corresponds to a unique material category identifier Ci (such as C1 = pine, C2 = 304 stainless steel, etc.).

[0039] After receiving the standardized spectral feature vector S of the product to be coded, the similarity γi between S and each Sstdi is calculated (cosine similarity algorithm can be used for calculation), and the similarity threshold γth=0.92 is set (through experiments with 1000 sets of samples of different materials, the recognition accuracy is ≥99% under this threshold).

[0040] All similarity scores γi are sorted from largest to smallest, and each similarity score γi is compared with the similarity threshold γth. The corresponding Ci with the largest γi and γi≥γth is selected as the material category C, and the material identifier C of the product to be coded is transmitted to the laser parameter adaptive matching module. If all γi < γth, an "unknown material" warning signal is generated and sent to the backend monitoring terminal.

[0041] The laser parameter adaptive matching module generates the optimal laser marking parameter set (power, frequency, speed) for the product to be marked based on the material category identifier and product thickness parameters. It achieves fully automatic parameter matching, and the parameter matching is based on material characteristics and product thickness to avoid blurry marking (insufficient power) or product burn-out (excessive power). No manual adjustment is required, which significantly improves the efficiency of marking preparation. The parameter matching takes into account both material characteristics and product thickness, effectively avoiding blurry marking due to insufficient power or product burn-out due to excessive power, ensuring the consistency of marking effect and reducing the management difficulty of the marking process.

[0042] The laser marking execution module receives the optimal laser marking parameter set and retrieves the marking pattern. It controls the laser generator to adjust the output power to the set value and controls the marking terminal to complete the marking on the marking area of ​​the product surface at the set speed and frequency. The marking control information is sent to the back-end monitoring terminal, which allows back-end personnel to have a detailed understanding of the marking control process and facilitates manual control in special circumstances. This ensures the monitorability and controllability of the marking operation and ensures that the marking process is strictly executed according to the set parameters.

[0043] Specifically, the laser parameter adaptive matching module receives the material identifier C of the product to be marked, and at the same time collects the thickness d of the marking area through the integrated laser displacement sensor. It determines the core marking parameters of the product to be marked through the built-in material-parameter mapping model, including laser power P, laser frequency f and marking speed v, forming a complete optimal laser marking parameter set (P,f,v) and transmitting it to the laser marking execution module.

[0044] After coding is completed, the coding effect of the coded products is analyzed, and the analysis results are sent to the back-end monitoring terminal. This covers core coding defects (positional misalignment, blurriness, missing strokes, and broken codes), resulting in a low false negative rate and high detection efficiency, meeting the production line cycle time requirements. Furthermore, when defective coded products are identified, the defective product rejection control module controls the rejection mechanism to perform a rejection operation, removing the defective products from the production line. This eliminates the need for manual screening, significantly reducing the labor intensity of workers, preventing defective products from flowing into subsequent processes, and ensuring the overall quality stability of the products. The specific analysis and judgment process is as follows:

[0045] The laser marking execution module sends the marking completion signal to the marking effect multi-dimensional detection module. After receiving the marking completion signal, the marking effect multi-dimensional detection module starts a high-resolution industrial camera to acquire images of the marking area, performs contour matching integrity detection, and compares the actual marking pattern with the built-in standard marking pattern template. If the contour overlap between the actual marking pattern and the standard marking pattern template does not exceed the preset overlap threshold or there are missing strokes or broken codes in the actual marking pattern, it indicates that the marking effect of the marked product is not good, and the marked product is marked as a marking defective product.

[0046] If the overlap between the actual coding pattern and the standard coding pattern template exceeds the preset overlap threshold and there are no missing strokes or broken codes in the actual coding pattern, then position accuracy detection is performed. The center coordinates (xact, yact) of the actual coding pattern are extracted by the edge detection algorithm, and the deviation is calculated with the preset standard center coordinates (xstd, ystd).

[0047] Where Δx = |xact-xstd|, Δy = |yact-ystd|; set a deviation threshold Δth, compare Δx and Δy with the deviation threshold Δth respectively, if Δx > Δth or Δy > Δth, then the position accuracy is judged to be unqualified and the coded product is marked as a coded defective product;

[0048] If Δx≤Δth and Δy≤Δth, then the positional accuracy is deemed acceptable and a coding clarity evaluation analysis is performed: After grayscale conversion of the actual coding pattern, the difference between the grayscale value of pixel p in the actual coding pattern and the grayscale value of pixel p in the standard coding pattern template is calculated and the absolute value is taken. The absolute value result is then compared with the grayscale value of pixel p in the standard coding pattern template, and the ratio result is multiplied by 100% to obtain the grayscale deviation percentage. The larger the grayscale deviation percentage, the worse the display performance of pixel p.

[0049] The grayscale deviation percentage of all pixels is obtained and the average is calculated to obtain the sharpness anomaly value. The percentage of pixels whose grayscale deviation exceeds the preset grayscale deviation percentage threshold is marked as the pixel out-of-bounds value. The sharpness anomaly value and the pixel out-of-bounds value are compared with the preset sharpness anomaly threshold and the preset pixel out-of-bounds threshold respectively. If the sharpness anomaly value or the pixel out-of-bounds value exceeds the corresponding preset threshold, it indicates that the display effect of the marked product is not good. Therefore, the marking clarity is judged to be unqualified and the marked product is marked as a marking defective product.

[0050] Example 2: Figure 2As shown, the difference between this embodiment and Embodiment 1 is that the background monitoring terminal is connected to the manual intervention emergency decision module. The manual intervention emergency decision module analyzes the laser marking operation performance during the detection period and determines whether to generate a high emergency signal for manual intervention. When a high emergency signal for manual intervention is generated, it is sent to the background monitoring terminal.

[0051] When the back-end monitoring terminal receives a high-urgency signal requiring manual intervention, it issues an early warning, reminding staff to intervene promptly for inspection and control. This effectively prevents system anomalies from causing a batch of defective products, ensuring coding effectiveness and execution stability, further improving the system's intelligent management level, and guaranteeing the continuous and efficient operation of the production line. The specific analysis process of the manual intervention urgency decision-making module is as follows:

[0052] The system obtains the total number of coded products and the number of coded defective products during the detection period. It calculates the coded defect count by comparing the number of coded defective products with the total number of coded products. The coded defect count is then compared with a preset coded defect count threshold. If the coded defect count exceeds the preset threshold, it indicates that the coded effect during the detection period is abnormal and requires timely manual intervention and adjustment, thus generating a high-urgent-time signal for manual intervention.

[0053] Furthermore, if the number of defective coding tests does not exceed the preset defective coding threshold, the actual laser power, actual laser frequency, and actual coding speed of the laser coding execution module during coding are collected. The difference between the actual laser power and the laser power P in the corresponding optimal laser coding parameter group is calculated and the absolute value is taken to obtain the laser power characteristic value. Similarly, the laser frequency characteristic value and coding speed characteristic value are obtained.

[0054] The coding state coefficient is obtained by weighted summation of laser power characteristic value, laser frequency characteristic value and coding speed characteristic value; that is, the laser power characteristic value, laser frequency characteristic value and coding speed characteristic value are respectively assigned corresponding preset weight coefficients, and the laser power characteristic value, laser frequency characteristic value and coding speed characteristic value are respectively multiplied by the corresponding preset weight coefficients, and the sum of the three sets of product results is marked as the coding state coefficient.

[0055] It should be noted that the larger the value of the captcha status coefficient, the worse the overall real-time captcha status. The captcha status coefficient is compared with the preset captcha status coefficient threshold. If the captcha status coefficient exceeds the preset captcha status coefficient threshold, it indicates that the overall real-time captcha status is poor, and it is determined that the current state is in a non-matching state.

[0056] The number of times the execution is in a non-matching state during the detection period is obtained and marked as the execution non-matching detection value. The execution non-matching detection value is compared with the preset execution non-matching detection threshold. If the execution non-matching detection value exceeds the preset execution non-matching detection threshold, it indicates that the coding parameter execution performance during the detection period is poor and timely manual intervention and adjustment are required. In this case, a high emergency signal for manual intervention is generated.

[0057] If the non-match detection value does not exceed the preset non-match detection threshold, the average value of all coding status coefficients within the detection period is calculated to obtain the coding status feature value, and the coding status coefficient with the largest value within the detection period is marked as the coding status abnormal value.

[0058] The emergency coefficient for manual intervention is calculated by weighting and summing the non-matching detection value, the coding status feature value, and the coding status aberration value. Specifically, the non-matching detection value, the coding status feature value, and the coding status aberration value are each assigned a corresponding preset weight coefficient. The non-matching detection value, the coding status feature value, and the coding status aberration value are then multiplied by their respective preset weight coefficients, and the sum of the three products is marked as the emergency coefficient for manual intervention.

[0059] It should be noted that the higher the value of the emergency coefficient for manual intervention, the worse the overall performance of the coding parameters in the coding process, and the more timely the manual intervention check and adjustment is required. The emergency coefficient for manual intervention is compared with the preset emergency coefficient threshold. If the emergency coefficient for manual intervention exceeds the preset emergency coefficient threshold, it indicates that the overall performance of the coding parameters during the detection period is poor, and timely manual intervention check and adjustment is required, thus generating a high emergency signal for manual intervention.

[0060] The working principle of this invention is as follows: During use, the spectral feature perception and analysis module provides a high-precision standardized spectral feature vector to the material category determination module. The material category determination module, relying on a preset target material spectral database, automatically completes high-precision material identification and outputs a material identifier. The laser parameter adaptive matching module combines the material identifier with the product thickness parameters collected by the laser displacement sensor, and generates an optimal coding parameter set through a built-in mapping model. The laser coding execution module controls the coding based on the optimal laser coding parameter set to complete the product surface coding operation, avoiding coding blurring due to insufficient power or product burnout due to excessive power, ensuring coding effect and stability. After coding is completed, a multi-dimensional coding effect detection module performs multi-dimensional coding effect analysis, and immediately triggers a defective product rejection control module upon identifying defective products, automatically removing them from the production line. This achieves intelligent control of the entire laser coding process, from material identification, parameter matching, coding execution to effect detection and defective product rejection, significantly improving the accuracy, efficiency, and quality stability of the coding operation.

[0061] In this invention, the threshold, preset value, or preset range settings are for result comparison and analysis to determine whether the result is good or bad. The magnitude of these values ​​is determined by a combination of large-scale model analysis of sample data and human experience, and can also be appropriately adjusted based on seasonal or common-sense influence conditions. Similarly, the preset weight coefficients and influence factors are assigned specific values ​​based on the magnitude of each parameter's influence on the result, ultimately reflecting the impact on the result. These settings are also determined by a combination of large-scale model analysis of sample data and human experience, and can also be appropriately adjusted based on seasonal or common-sense influence conditions.

[0062] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, enabling those skilled in the art to better understand and utilize it. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A laser marking and detection integrated system based on spectral self-sensing technology, characterized in that, It includes a spectral feature perception and analysis module, a material category determination module, a laser parameter adaptive matching module, a laser marking execution module, a multi-dimensional marking effect detection module, a defective product rejection control module, and a back-end monitoring terminal; The spectral feature perception and analysis module collects the reflectance spectral signal of the surface of the product to be coded in real time, and outputs a standardized spectral feature vector S after noise reduction and feature extraction. The material category determination module determines the material category of the product to be coded based on the standardized spectral feature vector S and outputs the material identifier C. The laser parameter adaptive matching module generates the optimal laser marking parameter set for the product to be marked based on the material category identifier and product thickness parameters. The laser marking execution module receives the optimal laser marking parameter set, retrieves the marking pattern, completes the marking on the marking area on the product surface, and sends the marking control information to the back-end monitoring terminal. After marking is completed, the marking effect multi-dimensional detection module analyzes the marking effect of the marked products and sends the marking effect analysis results to the back-end monitoring terminal. When marking defective products are identified, the defective product rejection control module removes the marking defective products from the production line. The actual laser power, actual laser frequency, and actual coding speed of the laser coding execution module are collected during coding. The difference between the actual laser power and the laser power P in the corresponding optimal laser coding parameter group is calculated and the absolute value is taken to obtain the laser power characteristic value. Similarly, the laser frequency characteristic value and coding speed characteristic value are obtained. The coding state coefficient is obtained by weighted summation of the laser power characteristic value, laser frequency characteristic value, and coding speed characteristic value. The coding state coefficient is compared with the preset coding state coefficient threshold. If the coding state coefficient exceeds the preset coding state coefficient threshold, it is determined that the current execution is in a non-matching state. The spectral feature perception and analysis module receives the reflection spectral signal and converts it into the original electrical signal. It uses a wavelet threshold denoising algorithm to denoise the original signal and extracts the top ten principal components as key spectral features through principal component analysis. Then, it performs Z-score standardization on the key features and finally outputs the standardized spectral feature vector S, which is then transmitted to the material category determination module. The material category determination module has a built-in pre-built target material spectral database. The target material spectral database stores standard spectral feature vectors Sstdi for several common materials, where i={1,2,3,…,n}, and n is the number of material types. The similarity γi between S and each Sstdi is calculated, and the corresponding Ci with the largest γi and γi≥γth is selected as the material category C. The material identifier C of the product to be coded is then transmitted to the laser parameter adaptive matching module.

2. The integrated laser marking and detection system based on spectral self-sensing technology according to claim 1, characterized in that, The laser parameter adaptive matching module receives the material identifier C of the product to be marked, and at the same time collects the thickness d of the marking area through the integrated laser displacement sensor. It determines the core marking parameters of the product to be marked through the built-in material-parameter mapping model, including laser power P, laser frequency f and marking speed v, forming a complete optimal laser marking parameter set (P,f,v) and transmitting it to the laser marking execution module.

Images