A door and window frame rapid splicing structure quality detection method and system

Abstract

The application discloses a door and window frame rapid splicing structure quality detection method and system, relates to the door and window detection technical field, and the method comprises the following steps: obtaining initial frame structure measurement results, mechanical arm end dynamic behavior information, laser echo signal space distribution information and local temperature field information; predicting measurement result deviation distribution; calculating total measurement deviation value and measurement fluctuation range according to the measurement result deviation distribution; correcting the initial frame structure measurement results according to the total measurement deviation value to obtain target frame structure measurement results; determining a qualified judgment limit according to the measurement fluctuation range; and generating a quality detection result according to the target frame structure measurement results and the qualified judgment limit. The application can combine the deviation distribution, the measurement deviation value and the measurement fluctuation range to generate the quality detection result, so that the structure quality detection is realized, and the detection accuracy and product quality are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of door and window detection, and particularly relates to a door and window frame rapid splicing structure quality detection method and system. BACKGROUND

[0002] In modern industrial production, accurate detection of product quality is a key link to ensure production efficiency and product reliability. Existing methods use a fixed process for detection. However, under long-time and high-intensity continuous operation of mechanical systems, cumulative wear and tear of moving parts is inevitable, resulting in a slight and slow drift in the accuracy of the mechanical arm when performing repeated positioning tasks. Existing methods are difficult to adapt to changes in performance caused by long-term operation of equipment and detection difficulties caused by the introduction of new materials, affecting the reliability and accuracy of quality detection, and possibly leading to confusion in the production process and waste of resources, low product quality.

[0003] To sum up, the technical problems in the related art need to be improved. SUMMARY

[0004] The main purpose of the embodiments of the present application is to provide a door and window frame rapid splicing structure quality detection method and system, which can combine deviation distribution, measurement deviation value and measurement fluctuation range to generate a quality detection result, so as to realize structure quality detection and improve detection accuracy and product quality.

[0005] In one aspect, the embodiments of the present application provide a door and window frame rapid splicing structure quality detection method, comprising the following steps:

[0006] Obtain initial frame structure measurement results, mechanical arm end dynamic behavior information, laser echo signal spatial distribution information and local temperature field information, wherein the mechanical arm end dynamic behavior information includes tremor and deformation, and the local temperature field information includes temperature distribution and temperature gradient;

[0007] According to the mechanical arm end dynamic behavior information, the laser echo signal spatial distribution information, the local temperature field information and the dynamic influence map, predict the measurement result deviation distribution;

[0008] According to the measurement result deviation distribution, calculate the total measurement deviation value and the measurement fluctuation range;

[0009] According to the total measurement deviation value, correct the initial frame structure measurement results to obtain target frame structure measurement results;

[0010] According to the measurement fluctuation range, determine the qualified judgment limit;

[0011] According to the target frame structure measurement results and the qualified judgment limit, generate a quality detection result.

[0012] In another aspect, the embodiment of the present application provides a door and window frame rapid splicing structure quality detection system, comprising:

[0013] a data acquisition module configured to acquire initial frame structure measurement results, mechanical arm end dynamic behavior information, laser echo signal spatial distribution information, and local temperature field information, wherein the mechanical arm end dynamic behavior information includes tremor and deformation, and the local temperature field information includes temperature distribution and temperature gradient;

[0014] a deviation distribution prediction module configured to predict measurement result deviation distribution according to the mechanical arm end dynamic behavior information, the laser echo signal spatial distribution information, the local temperature field information, and a dynamic influence atlas;

[0015] a deviation calculation module configured to calculate a total measurement deviation value and a measurement fluctuation range according to the measurement result deviation distribution;

[0016] a measurement result correction module configured to correct the initial frame structure measurement results according to the total measurement deviation value to obtain target frame structure measurement results;

[0017] a limit determination module configured to determine a qualified judgment limit according to the measurement fluctuation range;

[0018] a quality judgment module configured to generate a quality detection result according to the target frame structure measurement results and the qualified judgment limit.

[0019] The embodiment of the present application has at least the following beneficial effects: the embodiment of the present application first acquires initial frame structure measurement results, mechanical arm end dynamic behavior information, laser echo signal spatial distribution information, and local temperature field information, then predicts measurement result deviation distribution, calculates a total measurement deviation value and a measurement fluctuation range, corrects the initial frame structure measurement results according to the total measurement deviation value to obtain target frame structure measurement results, determines a qualified judgment limit according to the measurement fluctuation range, and finally generates a quality detection result according to the target frame structure measurement results and the qualified judgment limit, so that the quality detection result can be generated by combining the deviation distribution, the measurement deviation value, and the measurement fluctuation range to realize structure quality detection and improve detection accuracy and product quality.

[0020] Other features and advantages of the present application will be set forth in the following description, and in part will become apparent to those skilled in the art from the description, or can be learned by practice of the present application. The objects and other advantages of the present application can be achieved and obtained by means of the structures particularly pointed out in the description and the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiment description will be briefly introduced.

[0022] Figure 1 A flow chart of a quality detection method for a door and window frame rapid assembly structure according to an embodiment of the present application;

[0023] Figure 2 A structural schematic diagram of a quality detection system for a door and window frame rapid assembly structure according to an embodiment of the present application. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solutions and advantages of the present application more clear, the present application will be further described in detail below with reference to the drawings and embodiments.

[0025] In related art, in modern industrial production, accurate detection of product quality is a key link to ensure production efficiency and product reliability. Especially in the field of door and window frame manufacturing, with the popularization of automated production lines, it is necessary to maintain the accuracy of the detection system under high throughput and variable working conditions. The existing detection method often cannot adapt to the performance changes caused by long-term operation of the equipment and the detection problems caused by the introduction of new materials, which directly affects the reliability of quality judgment and may lead to confusion in production process and waste of resources.

[0026] For example, in a modern door and window manufacturing factory, a high-speed automated production line undertakes the task of batch production of door and window frame rapid assembly structures. One of the core links of this production line is the precision quality detection system equipped with it. The system uses a set of laser measurement-based devices and is accurately scanned and positioned by a multi-axis robot arm. The task of the robot arm is to guide the laser sensor to collect high-precision data on key geometric features and connection points on the rapidly assembled door and window frame, such as detecting the perpendicularity of the corner points, the width of the assembly gap, and the flatness of the profile surface, etc.

[0027] However, any mechanical system under long-time, high-intensity continuous operation, its moving parts, especially bearings and guide rails, will inevitably accumulate wear and tear. This wear and tear is not a sudden failure, but a gradual and slow process. This cumulative wear and tear will cause the robot arm to drift slightly and slowly in accuracy when performing repeated positioning tasks. For example, if the robot arm is programmed to accurately scan the center position of a certain corner point on the frame, as the wear and tear increases, it may slightly deviate from the center with each scan, for example, by a few microns. In the early stage of the production line, when the system mainly handles conventional aluminum alloy or polyvinyl chloride profiles, the measurement error caused by the slight positioning drift caused by mechanical wear is usually considered as acceptable random noise by the signal processing algorithm inside the system.

[0028] In response to market demands for improved product appearance and performance, the production line introduced a new type of aluminum profile with a special anodizing treatment. This new profile exhibits a unique matte texture and superior corrosion resistance. However, its specially treated surface has a significantly lower reflectivity to the specific wavelength laser used in the detection system compared to conventional profiles. This means that the signal strength returning to the receiver after laser emission is greatly reduced. For example, if a conventional profile can reflect 80% of the laser energy, the new profile may only reflect 30% to 40%. To ensure effective measurement data acquisition even with significantly reduced signal strength, the embedded software in the detection system automatically activates a compensation mechanism. This mechanism typically includes increasing the gain of the signal receiving circuit and extending the signal integration or processing time to extract useful information from the weak echoes.

[0029] This automatic adjustment, intended to compensate for weak signals, inadvertently amplifies all potential measurement uncertainties in the system. When the system increases signal gain, it amplifies not only the weak laser echo signal but also any minute noise and fluctuations in the signal path. At this point, the positioning drift, previously imperceptible due to its small size and caused by cumulative wear of the robotic arm, is significantly amplified. What was originally considered random background noise now manifests as a identifiable, systematic deviation in the amplified signal. This amplified robotic arm drift effect, combined with the increased measurement uncertainty due to signal attenuation, forms a significant and directional systematic measurement bias. This affects the reliability and accuracy of quality inspection and may lead to production process disruptions, resource waste, and low product quality.

[0030] On the high-speed production line for rapid assembly of door and window frames, facing the slow drift in positioning accuracy caused by the cumulative wear of multi-axis robotic arms, the weak signal amplification effect caused by the low reflectivity of new materials, and the inability of the initial calibration model to adapt to these dynamic and complex working conditions, it is necessary to identify, quantify, and compensate for the resulting systematic measurement deviations in real time to ensure the accuracy of quality judgment, avoid misjudgments and omissions, and ensure the reliability of the production and manufacturing execution system scheduling.

[0031] The embodiments of this application will be explained in detail below with reference to the accompanying drawings:

[0032] Figure 1 This is an optional flowchart of a method for quality inspection of a rapid splicing structure of door and window frames provided in an embodiment of this application. Figure 1 The method may include, but is not limited to, steps S101 to S106.

[0033] Step S101: Obtain the initial frame structure measurement results, dynamic behavior information of the robotic arm end effector, spatial distribution information of laser echo signal, and local temperature field information. The dynamic behavior information of the robotic arm end effector includes vibration and deformation, and the local temperature field information includes temperature distribution and temperature gradient.

[0034] Step S102: Based on the dynamic behavior information of the robotic arm end effector, the spatial distribution information of the laser echo signal, the local temperature field information, and the dynamic influence map, predict the distribution of the measurement result deviation;

[0035] Step S103: Calculate the total measurement deviation value and measurement fluctuation range based on the deviation distribution of the measurement results;

[0036] Step S104: Based on the total measurement deviation value, the initial frame structure measurement results are corrected to obtain the target frame structure measurement results;

[0037] Step S105: Determine the pass / fail threshold based on the measured fluctuation range;

[0038] Step S106: Generate quality inspection results based on the target frame structure measurement results and the acceptance criteria.

[0039] Steps S101 to S106 as shown in the embodiments of this application can combine deviation distribution, measured deviation value and measured fluctuation range to generate quality inspection results, so as to realize structural quality inspection and improve inspection accuracy and product quality.

[0040] In some embodiments, steps S101-S106 may first acquire the initial frame structure measurement results, the dynamic behavior information of the robotic arm end effector, the spatial distribution information of the laser echo signal, and the local temperature field information. The initial frame structure measurement results can be acquired in various ways. For example, traditional contact probe measurement can be used, where a mechanical probe contacts the surface of the door / window frame to acquire geometric dimensional data; alternatively, non-contact optical measurement can be used, such as using a line laser scanner to scan the door / window frame, generating point cloud data, and then extracting geometric features from the point cloud data. The dynamic behavior information of the robotic arm end effector includes jitter and deformation, which can be monitored in real time by installing high-precision accelerometers and displacement sensors at the end effector. Accelerometers can capture high-frequency jitter, and displacement sensors can monitor low-frequency deformation. The spatial distribution information of the laser echo signal can be acquired using lidar or laser triangulation sensors, which can provide data such as the intensity, flight time, and spot shape of the laser beam after reflection from the surface of the object being measured. Local temperature field information, including temperature distribution and temperature gradient, can be obtained by scanning the measured area of ​​the door and window frame with an infrared thermal imager. The infrared thermal imager can provide a two-dimensional distribution map of the surface temperature.

[0041] Then, based on the dynamic behavior information of the robotic arm's end effector, the spatial distribution information of the laser echo signal, the local temperature field information, and the dynamic influence map, the distribution of measurement result deviation is predicted. The dynamic influence map can be a pre-trained machine learning model, such as a neural network or support vector machine, which is trained with a large amount of experimental data to learn the influence of different combinations of dynamic behavior, laser signal, and temperature field on the measurement result deviation. In practical applications, the real-time acquired dynamic behavior information of the robotic arm's end effector, the spatial distribution information of the laser echo signal, and the local temperature field information can be used as input to the model, and the model can then output the predicted distribution of measurement result deviation.

[0042] Next, based on the deviation distribution of the measurement results, calculate the total measurement deviation value and the measurement fluctuation range. The measurement result deviation distribution refers to a set containing multiple deviation values, each corresponding to the predicted deviation of a specific measurement point on the door / window frame. The total measurement deviation value is a comprehensive index obtained by statistically averaging or weighted averaging these deviation values, reflecting the overall systematic deviation. The measurement fluctuation range is determined by calculating the standard deviation or the range of maximum and minimum values ​​of the deviation distribution, reflecting the degree of uncertainty in the measurement results.

[0043] Based on the total measurement deviation value, the initial frame structure measurement results are corrected to obtain the target frame structure measurement results. An addition / subtraction method can be used, that is, subtracting the corresponding total measurement deviation value from the initial frame structure measurement results to eliminate systematic deviations. For example, if the predicted measurement value for a certain dimension has a deviation of +0.1mm, then the actual measurement value is corrected by subtracting 0.1mm.

[0044] The acceptance threshold is determined based on the measurement fluctuation range. The acceptance threshold is dynamically adjusted based on the standard tolerance range and the measurement fluctuation range. For example, if the standard tolerance is ±0.2mm and the measurement fluctuation range is ±0.05mm, then the acceptance threshold can be tightened to ±0.15mm to ensure accurate judgment of product quality even under measurement uncertainty.

[0045] Finally, based on the measurement results of the target frame structure and the acceptance criteria, a quality inspection result is generated. The quality inspection result can be a binary judgment of "qualified" or "unqualified," or it can be a more detailed quality level assessment. For example, if the measurement results of the target frame structure fall within the acceptance criteria, it is judged as qualified; otherwise, it is judged as unqualified.

[0046] Through the above technical solution, this embodiment comprehensively captures dynamic factors affecting measurement accuracy by acquiring dynamic behavior information (vibration and deformation) of the robotic arm end effector, spatial distribution information of laser echo signals, and local temperature field information. This information is input into a dynamic influence map, which can predict how these dynamic factors collectively lead to the deviation distribution of the measurement results. This embodiment calculates the total measurement deviation value and measurement fluctuation range based on the measurement result deviation distribution, and uses the total measurement deviation value to correct the initial measurement results, thereby obtaining more accurate measurement results of the target frame structure. Simultaneously, the acceptance threshold is dynamically determined based on the measurement fluctuation range, making quality judgment more flexible and reliable. Therefore, this embodiment can effectively compensate for measurement deviations caused by long-term equipment operation and the introduction of new materials, significantly improving the accuracy and reliability of detection. This adaptive detection method not only avoids rework and resource waste caused by misjudgment in traditional methods, but also ensures that unqualified products do not flow into subsequent processes, thereby comprehensively improving the intelligence level and product quality control capabilities of the door and window frame production line.

[0047] In some embodiments, in step S102, predicting the distribution of measurement result deviation based on the dynamic behavior information of the robotic arm end effector, the spatial distribution information of the laser echo signal, the local temperature field information, and the dynamic influence map may include, but is not limited to, the following steps:

[0048] The geometric feature calibration reference block is scanned to obtain feature scanning information. The geometric feature calibration reference block contains various door and window frame geometric features.

[0049] Based on feature scanning information, identify the interaction patterns of geometric features;

[0050] Adjust the path parameters in the dynamic influence map based on the interaction pattern of geometric features;

[0051] Based on the adjusted dynamic influence map, the dynamic behavior information of the robotic arm end effector, the spatial distribution information of the laser echo signal, and the local temperature field information, the distribution of the measurement result deviation is predicted.

[0052] In some embodiments, since the geometric features of door and window frames may vary, if the dynamic influence map fails to fully consider the specificity of these geometric features and their impact on the measurement process, it may lead to insufficient prediction accuracy of the measurement result deviation distribution, thereby affecting the final quality inspection results.

[0053] To this end, feature scanning can be performed on the geometric feature calibration reference block to obtain feature scanning information. The geometric feature calibration reference block contains various geometric features of door and window frames. A geometric feature calibration reference block is a pre-made standard part or model with known and diverse geometric features of door and window frames. This reference block is used to simulate various complex structures of actual door and window frames, such as right-angle connections, beveled connections, T-shaped connections, transition areas of profiles with different thicknesses, and weld areas. Its purpose is to provide a controlled environment for obtaining real response data of different geometric features during the measurement process. Feature scanning refers to the process of comprehensively acquiring data from the geometric feature calibration reference block using high-precision sensors (such as laser scanners, structured light sensors, or high-resolution cameras). Through feature scanning, detailed information such as the surface geometry, size, and texture of the reference block can be obtained and converted into processable feature scanning information. This information can include point cloud data, mesh models, or image data.

[0054] Then, based on the feature scanning information, the interaction patterns of geometric features are identified. These interaction patterns refer to the ways and rules by which different geometric features interact with each other and with the measurement system (such as a laser beam or robotic arm) during the measurement process. For example, at acute edges, the laser may undergo multiple reflections; at narrow gaps, the laser may have difficulty penetrating completely; and at the interface between different materials, the laser echo signal may abruptly change. Identifying these interaction patterns aims to understand how specific geometric features affect the generation and propagation of measurement signals.

[0055] Then, based on the interaction pattern of geometric features, the path parameters in the dynamic influence map are adjusted. Parameters related to the measurement path, signal propagation path, or error accumulation path in the dynamic influence map can be corrected or optimized. For example, the incident angle of the laser beam, scanning speed, data sampling density, or the error weighting factor of a specific geometric region can be adjusted.

[0056] Finally, based on the adjusted dynamic influence map, the dynamic behavior information of the robotic arm end effector, the spatial distribution information of the laser echo signal, and the local temperature field information, the distribution of measurement deviation is predicted to more accurately reflect the measurement deviation in this area.

[0057] This embodiment, by introducing geometric feature calibration reference blocks and feature scanning, can systematically acquire response data of different door and window frame geometric features under actual measurement environments. This allows for the identification of the influence patterns of these geometric features on the dynamic behavior of the robotic arm's end effector, the spatial distribution of laser echo signals, and local temperature field information—that is, the interaction patterns of the geometric features. The explicit identification of these interaction patterns enables the dynamic influence map to be adjusted in a targeted manner. This adjustment ensures that the dynamic influence map can be adaptively optimized according to the specific geometric features of the door and window frame under test. In this way, the dynamic influence map can more accurately reflect the propagation and accumulation mechanisms of measurement errors in specific geometric feature regions, thereby significantly improving the prediction accuracy of the measurement result deviation distribution.

[0058] To illustrate this technical solution more clearly, a specific example is used below. Suppose we need to inspect the quality of a new type of door and window frame. This frame contains various complex geometric features, such as L-shaped connectors with chamfers, welding areas of profiles with different thicknesses, and embedded sealing grooves. To improve the prediction accuracy of the measurement result deviation distribution, a geometric feature calibration reference block is first prepared. This reference block accurately replicates the geometry of the L-shaped connectors, welding areas, and sealing grooves. Next, a high-precision laser scanner is used to scan the geometric feature calibration reference block, acquiring detailed three-dimensional point cloud data to form feature scanning information. By analyzing this point cloud data, the interaction patterns of the geometric features can be identified. For example, in the chamfered area of ​​the L-shaped connector, the reflection path of the laser beam will undergo a specific deflection; in the welding area, due to surface roughness and material inhomogeneity, the intensity and width of the laser echo signal will change; inside the sealing groove, the laser may experience multiple reflections or obstructions. These specific reflection, scattering, and attenuation patterns are the interaction patterns of the geometric features.

[0059] Based on these identified interaction patterns, the path parameters in the dynamic influence map are adjusted. Specifically, for chamfered areas, a correction factor considering laser reflection deflection is added to the dynamic influence map; for welded areas, the weighting parameters of the laser echo signal are adjusted to reflect its uncertainty; and for sealing grooves, the laser scanning path and data filtering algorithm are optimized. Finally, when actually inspecting the door and window frames under test, the adjusted dynamic influence map, combined with real-time acquired information on the dynamic behavior of the robotic arm end effector, the spatial distribution of the laser echo signal, and the local temperature field, will be used to more accurately predict the distribution of measurement result deviations. For example, for an L-shaped connector with slight deformation, the adjusted map can more accurately predict the deviations caused by the combined effects of geometric features and dynamic factors during measurement, thus providing a more reliable basis for subsequent correction processing.

[0060] Through the above technical solution, this embodiment achieves precise adjustment of path parameters in the dynamic influence map by scanning the features of the geometric feature calibration reference block and recognizing the geometric feature interaction mode, resulting in more refined and accurate prediction of the deviation distribution. This not only improves the reliability of quality inspection of rapid splicing structures for door and window frames, but also significantly enhances the correction accuracy of measurement results for door and window frames with diverse geometric features, thereby ensuring more accurate quality judgment.

[0061] In some embodiments, in step S103, calculating the total measurement deviation value and the measurement fluctuation range based on the measurement result deviation distribution may include, but is not limited to, the following steps:

[0062] Obtain the geometric feature type of the door and window frame to be tested;

[0063] Select target distribution morphology feature parameters that match the geometric feature type from a preset set of distribution morphology feature parameters;

[0064] The rate of change of parameters is monitored, including the rate of change of dynamic behavior information of the robotic arm end effector, the rate of change of spatial distribution information of laser echo signal, and the rate of change of local temperature field information.

[0065] The confidence level for fluctuation extraction is determined based on the rate of parameter change.

[0066] Based on the fluctuation, confidence level and target distribution morphology parameters are extracted, and the measurement fluctuation range is extracted from the measurement result deviation distribution.

[0067] Determine the defect sensitivity based on the type of defect in the door and window frame to be tested;

[0068] Determine the weighting of deviation extraction based on defect sensitivity;

[0069] Based on the deviation extraction weights and target distribution morphological characteristic parameters, the total measurement deviation value is extracted from the measurement result deviation distribution.

[0070] In some embodiments, due to the diverse geometric features of door and window frames and the complex and variable dynamic behavior of the measurement environment and robotic arm, calculations based solely on a single deviation distribution may not adequately account for the impact of these factors on the accuracy and volatility of the measurement results, resulting in insufficient precision or specificity in the calculated total measurement deviation value and measurement volatility range.

[0071] To this end, the geometric feature type of the door and window frame to be tested can be obtained first. This can be achieved through CAD model data, image recognition technology, or pre-set detection path information. This allows identification of the specific geometric morphological features of the door and window frame currently undergoing quality inspection. For example, features could include straight lines, corners, connector areas, or other complex curved surfaces; profile cross-sectional shape, connection method, and size specifications; or rectangular frames, curved frames, and complex structures with specific connectors. The purpose is to provide targeted feature information for subsequent deviation calculations.

[0072] Then, target distribution shape feature parameters matching the geometric feature type are selected from a preset set of distribution shape feature parameters. Based on the acquired geometric feature type, a set of deviation distribution shape parameters that best represent the geometric feature can be retrieved and selected from a pre-established database or parameter set. These parameters may include statistical characteristics such as mean, variance, skewness, and kurtosis, or specific mathematical model parameters, with the aim of making the deviation distribution analysis more consistent with the actual physical properties and manufacturing tolerance requirements of the specific geometric feature.

[0073] The rate of parameter change is then monitored, including the rate of change of dynamic behavior information of the robotic arm end effector, the rate of change of spatial distribution information of laser echo signal, and the rate of change of local temperature field information. This allows for real-time tracking of the rate of change of these three parameters over time, aiming to assess the dynamic stability of the measurement process and the degree of environmental disturbance.

[0074] The confidence level for fluctuation extraction is determined based on the rate of parameter change. This allows assessment of the reliability of extracting the measurement fluctuation range from the distribution of measurement result deviations. For example, a high rate of parameter change indicates instability in the measurement environment or equipment, potentially resulting in a lower confidence level for fluctuation extraction, and vice versa. This confidence level guides subsequent fluctuation range extraction to avoid misjudgments caused by environmental instability.

[0075] Based on the confidence level and target distribution morphological characteristics parameters extracted from the fluctuation, the measurement fluctuation range is extracted from the measurement result deviation distribution. The dynamic stability of the measurement process and the geometric characteristics of the door and window frames can be comprehensively considered to determine the measurement fluctuation range more accurately from the predicted measurement result deviation distribution. For example, when the confidence level is low, the extracted fluctuation range can be appropriately expanded to cover greater uncertainty; while when the confidence level is high, the fluctuation range can be defined more precisely.

[0076] The defect sensitivity is determined based on the type of defect in the door / window frame being tested. The impact of each defect on the overall quality or function of the door / window frame can be assessed according to its type (e.g., scratches, dents, cracks, misalignments, etc.) to obtain the defect sensitivity. For example, cracks may have a higher sensitivity, while minor scratches may have a lower sensitivity.

[0077] Based on defect sensitivity, the weights for deviation extraction are determined. Different weights are assigned to different parts or types of deviations in the measurement result deviation distribution according to the sensitivity corresponding to different defect types. Deviations corresponding to defects with higher sensitivity will receive higher weights to ensure they occupy a more significant position in the calculation of the total measurement deviation value.

[0078] Finally, based on the deviation extraction weights and target distribution morphological characteristic parameters, the total measurement deviation value is extracted from the measurement result deviation distribution. After considering the geometric characteristics of the door and window frames and the sensitivity to potential defects, the total measurement deviation value is calculated from the measurement result deviation distribution. By introducing deviation extraction weights, greater attention can be paid to deviations in critical defects or important geometric areas, resulting in a more representative and instructive total measurement deviation value.

[0079] This embodiment achieves refined analysis of measurement result deviation distribution by incorporating considerations of the geometric feature types of door and window frames, the rate of change of dynamic parameters, and defect types. First, by acquiring the geometric feature types and matching them with corresponding target distribution morphological feature parameters, the deviation analysis becomes more targeted. Second, by monitoring the parameter change rate and determining the confidence level for fluctuation extraction, the stability of the measurement process can be dynamically assessed. This allows for adaptive adjustments based on actual measurement conditions when extracting the measurement fluctuation range, improving the accuracy of the fluctuation range. Finally, by identifying defect types and determining defect sensitivity, and then assigning weights to deviation extraction, the calculation of the total measurement deviation value highlights the impact of key defects, ensuring that the focus of quality assessment is emphasized. These multi-dimensional considerations and adaptive adjustment mechanisms make the calculated results of the total measurement deviation value and measurement fluctuation range more realistic and instructive.

[0080] To illustrate this technical solution more clearly, a specific example is used below. Suppose we need to perform quality inspection on a door / window frame containing right-angle corners and straight line segments. First, when a right-angle corner area is detected, the system identifies its geometric feature type as "corner" and selects target distribution morphology feature parameters from a preset set of distribution morphology feature parameters that match "corner." These parameters may focus more on angle deviation and local stress concentration. Simultaneously, if the rate of change of the dynamic behavior information at the robotic arm's end effector is high, indicating significant disturbance in the measurement process, the system will determine a lower confidence level for fluctuation extraction. This allows for a wider range to be extracted from the measurement result deviation distribution to cover greater uncertainty. When a straight line segment area is detected, the geometric feature type is identified as "straight line segment," and corresponding target distribution morphology feature parameters are selected. These parameters may focus more on linearity and flatness. Furthermore, if a minor scratch and a potential internal crack are detected on the frame surface, the system will set the sensitivity of the "internal crack" to be much higher than that of the "minor scratch," based on a preset defect sensitivity, thus assigning a higher extraction weight to the deviation related to the "internal crack." Ultimately, when calculating the total measurement deviation value, the deviation contribution of the internal crack will be amplified, thereby ensuring the accurate assessment of critical defects.

[0081] Through the above technical solution, this embodiment can significantly improve the calculation accuracy and reliability of the total measurement deviation and measurement fluctuation range in the quality inspection of rapid splicing structures of door and window frames. This embodiment can adaptively and precisely extract deviations based on the geometric characteristics of the door and window frames, the dynamic changes during the measurement process, and the severity of potential defects, thereby overcoming the problems of potential generalization and insufficient accuracy in calculation results. Therefore, it can provide a more accurate and robust data foundation for subsequent quality correction and acceptance judgment, effectively improving the overall efficiency and reliability of door and window frame quality inspection.

[0082] In some embodiments, in step S104, the initial frame structure measurement results are corrected based on the total measurement deviation value to obtain the target frame structure measurement results. This may include, but is not limited to, the following steps:

[0083] Obtain the geometric feature type of the door and window frame to be tested;

[0084] Based on the geometric feature type, the door and window frame to be tested is divided into regions to obtain multiple local regions;

[0085] Based on multiple local regions, the initial frame structure measurement results are decomposed to obtain the corresponding local measurement results;

[0086] Based on the dynamic behavior information of the robotic arm end effector, the spatial distribution information of the laser echo signal, the local temperature field information, and the dynamic influence map, the local deviation distribution corresponding to each local area is predicted.

[0087] Calculate the local measurement deviation value based on the local deviation distribution;

[0088] The total measurement deviation and multiple local measurement deviations are verified to obtain the verification results.

[0089] If the verification result is that the verification passes, then the deviation is corrected based on the local measurement results and the corresponding local measurement deviation values.

[0090] The measurement results of the target frame structure are obtained by integrating the local measurement results after multiple deviation corrections.

[0091] In some embodiments, due to the structural complexity and material diversity of door and window frames, as well as the dynamic environmental influences that may occur during the measurement process, measurement deviations may exhibit non-uniform distribution characteristics in different areas. Simply employing global correction may not be able to accurately eliminate errors in local areas, and may even introduce new deviations in some areas, thereby affecting the accuracy and reliability of the final measurement results.

[0092] To achieve this, the geometric feature type of the door / window frame to be tested can be obtained first. Based on this geometric feature type, the frame can be divided into multiple local regions. The entire door / window frame structure can be logically or physically divided into several smaller, independently analyzable sub-parts according to its geometric characteristics, functional zoning, or potential error-sensitive areas. For example, straight segments, corners, and connection points of the frame can be divided into different local regions.

[0093] Then, based on multiple local regions, the initial frame structure measurement results are decomposed to obtain corresponding local measurement results. The initial measurement data acquired for the entire door and window frame can be refined and assigned according to the divided local regions, ensuring that each local region has its own independent measurement data reflecting the actual conditions of that region. Furthermore, based on the dynamic behavior information of the robotic arm's end effector, the spatial distribution information of the laser echo signal, the local temperature field information, and the dynamic influence map, the local deviation distribution corresponding to each local region is predicted. A mechanism similar to predicting the overall measurement deviation distribution can be used, but for the specific conditions and data of each local region, the spatial distribution of potential measurement errors within that local region can be predicted. This takes into account the influence of the local environment and structure on measurement accuracy.

[0094] Next, based on the local deviation distribution, the local measurement deviation value is calculated. The magnitude of the overall or key point measurement deviation in each local area can be quantified based on the predicted deviation distribution. The total measurement deviation value and multiple local measurement deviation values ​​are then verified to obtain the verification results. The purpose is to ensure the consistency and reasonableness between the global and local deviations. For example, by comparing the total measurement deviation value with the sum or weighted average of all local measurement deviation values, it can be determined whether there are significant differences or anomalies, thus verifying the accuracy of the deviation prediction.

[0095] If the verification result is successful, deviation correction is performed based on the local measurement results and the corresponding local measurement deviation values. After the verification confirms the effectiveness of the deviation prediction, precise local correction is performed for each local area using its own local measurement results and the corresponding local measurement deviation values ​​to eliminate the measurement error in that area.

[0096] Finally, the local measurement results after multiple deviation corrections are integrated to obtain the target frame structure measurement results. All locally corrected sub-region measurement data can be recombined to form a complete and highly accurate measurement result for the door and window frame structure.

[0097] This embodiment obtains geometric feature types and divides regions, decomposing the complex overall structure into multiple independently analyzable local regions, thus enabling targeted handling of the characteristic differences in different regions. Secondly, it independently predicts the local deviation distribution and calculates the local measurement deviation value for each local region, ensuring the precision and accuracy of deviation correction and preventing global deviations from masking local details. This embodiment verifies the accuracy and consistency of the deviation prediction model by checking the total measurement deviation value and multiple local measurement deviation values, thereby improving the reliability of the entire correction process. If the verification passes, precise local correction is performed based on the local measurement results and corresponding local measurement deviation values, ensuring that the measurement accuracy of each local region is maximized. Finally, by integrating these locally corrected results, a more accurate and reliable overall measurement result of the target frame structure is obtained.

[0098] To illustrate this technical solution more clearly, a specific example is used below. Assume the window / door frame under test is a complex structure with curved corners and multiple connectors. First, the system acquires the geometric feature type of the window / door frame, identifying it as containing straight segments, curved segments, and connector areas. Next, based on these geometric feature types, the entire frame is divided into multiple local regions; for example, each straight segment, each curved segment, and each connector area is treated as an independent local region. Then, for each local region, the system decomposes the initial frame structure measurement results to obtain the corresponding local measurement results. Simultaneously, using the dynamic behavior information of the robotic arm's end effector, the spatial distribution information of the laser echo signal, the local temperature field information, and the dynamic influence map, the system predicts the local deviation distribution of each local region and calculates the corresponding local measurement deviation value. Based on this, the system verifies the total measurement deviation value and all local measurement deviation values. For example, if the total measurement deviation value and the sum of all local measurement deviation values ​​are within a preset error range, the verification passes.

[0099] If the verification passes, the system will perform precise deviation correction for each local area based on the local measurement results and corresponding local measurement deviation values. For example, for a straight segment area, if the local measurement results show a slight bend, and the local deviation value indicates that the bend is caused by a temperature gradient, the system will correct the local measurement results according to the temperature gradient model. Finally, the measurement results of all locally corrected straight segments, curved segments, and connecting parts areas are integrated to obtain a highly accurate and fully corrected measurement result of the target frame structure.

[0100] Through the above technical solution, this embodiment, by introducing a local area division, local deviation prediction, and verification mechanism, can more accurately identify and eliminate measurement errors in different areas of the door and window frame. The correction effect is particularly significant for frames with complex geometry, uneven material properties, or local defects. Therefore, the final measurement results of the target frame structure have higher accuracy, providing a more solid data foundation for subsequent quality judgment and effectively avoiding misjudgments or omissions caused by improper correction, thereby improving the overall performance and practical value of the inspection system.

[0101] In some embodiments, step S105, determining the pass / fail threshold based on the measured fluctuation range, may include, but is not limited to, the following steps:

[0102] Step S201: Obtain the geometric feature type of the door and window frame to be tested;

[0103] Step S202: Search the process database for the downstream process compatibility requirements corresponding to the geometric feature type;

[0104] Step S203: Adjust the measurement fluctuation range according to the downstream process compatibility requirements;

[0105] Step S204: Determine the pass / fail threshold based on the adjusted measurement fluctuation range.

[0106] In some embodiments, since the acceptance criteria are set solely based on the measurement fluctuation range, the actual compatibility requirements of different door and window frame geometric features in subsequent process steps may not be fully considered, resulting in the acceptance criteria being too strict or too lenient, affecting production efficiency or product quality.

[0107] To achieve this, the geometric feature type of the door / window frame to be tested can be obtained first. Then, the downstream process compatibility requirements corresponding to this geometric feature type can be found in the process database. A pre-established database can be accessed, storing the specific requirements for dimensional accuracy, surface roughness, and geometric tolerances between various door / window frame geometric features and different downstream processes (such as welding, painting, assembly, and installation). For example, the dimensional tolerance requirements for connections requiring precision welding may be much higher than those for non-critical parts requiring simple assembly. The aim is to closely integrate quality inspection standards with actual production and usage needs.

[0108] Then, the measurement fluctuation range is adjusted according to downstream process compatibility requirements. For example, if the downstream process compatibility requirements for a certain geometric feature allow for a larger tolerance, the measurement fluctuation range can be appropriately widened; conversely, if extremely high accuracy is required, the measurement fluctuation range may need to be tightened. This adjustment can be linear, non-linear, or based on preset rules. The purpose is to make the measurement fluctuation range more consistent with actual process requirements, avoiding unnecessary rework or potential quality risks. Based on the adjusted measurement fluctuation range, a pass / fail threshold is determined. This threshold value can be used to compare with the measurement results of the target frame structure to generate the final quality inspection result.

[0109] This embodiment addresses the issue of acceptance criteria being detached from actual production needs by incorporating considerations of the geometric feature types of door and window frames and their corresponding downstream process compatibility requirements. First, by acquiring the geometric feature type of the door and window frame under test, different parts or designs can be distinguished. Second, downstream process compatibility requirements matching the geometric feature type are searched from the process database, allowing quality standards to be aligned with tolerance requirements in actual manufacturing and assembly processes. This introduction of downstream process compatibility requirements enables reasonable adjustment of measurement fluctuation ranges, ensuring that the final acceptance criteria meet both technical accuracy requirements and production efficiency and cost control.

[0110] To illustrate this technical solution more clearly, a specific example is used below. Suppose the geometric feature type of a connection part in the door / window frame under test is a "T-type welded joint," and this joint requires high-precision robotic welding in the downstream process. After obtaining this geometric feature type, the system will search the process database for the dimensional and geometrical tolerance requirements of the "T-type welded joint" under the high-precision robotic welding process, for example, requiring a flatness deviation of less than 0.1mm. If the initially calculated measurement fluctuation range is 0.15mm, then according to the downstream process compatibility requirements, this fluctuation range will be adjusted to 0.08mm to ensure welding quality. Ultimately, the acceptance threshold will be determined based on this adjusted 0.08mm fluctuation range, thus ensuring that only T-type joints that meet the high-precision welding requirements can pass quality inspection. Conversely, if another geometric feature type is a "decorative cover plate," whose downstream process compatibility requirements allow for larger tolerances, such as a flatness deviation of less than 0.5mm, then even if the initial measurement fluctuation range is 0.3mm, it may be adjusted to 0.4mm to avoid unnecessary rework and improve production efficiency.

[0111] By considering downstream process compatibility requirements, this embodiment avoids resource waste caused by overly strict judgment standards or subsequent quality problems caused by overly lenient standards. This method ensures that quality inspection results not only reflect measurement accuracy but also the product's functionality and reliability throughout its entire lifecycle, significantly improving the intelligence level and production adaptability of quality inspection.

[0112] In some embodiments, step S204, determining the pass / fail threshold based on the adjusted measurement fluctuation range, may include, but is not limited to, the following steps:

[0113] Step S301: Obtain the multi-material stacked structure information of the door and window frame to be tested. The multi-material stacked structure information includes the material interface location and material type.

[0114] Step S302: Based on the multi-material stacked structure information, identify the reflection and absorption characteristics of different material interfaces to the laser signal;

[0115] Step S303: Based on the reflection and absorption characteristics, construct the material interface-specific influencing factor;

[0116] Step S304: Correct the adjusted measurement fluctuation range according to the material interface specific influencing factor;

[0117] Step S305: Determine the pass / fail threshold based on the corrected measurement fluctuation range.

[0118] In some embodiments, due to the significant differences in the reflection and absorption characteristics of laser signals at different material interfaces during the actual manufacturing process of door and window frames, especially for door and window frames employing multi-material laminated structures, failure to fully consider these material interface specificities may result in inaccurate determination of the acceptance criteria, thereby affecting the reliability of the final quality inspection results.

[0119] To achieve this, we can first obtain information about the multi-material layered structure of the door / window frame to be tested. Multi-material layered structure information refers to the structural characteristics of a door / window frame composed of two or more different materials layered together, including the location of material interfaces between layers and the material type of each material. For example, a door / window frame may be composed of multiple materials such as metal profiles, glass, and sealing strips, with clearly defined interfaces between these materials.

[0120] Then, based on the information of the multi-material stacked structure, the reflection and absorption characteristics of different material interfaces to the laser signal can be identified. This allows analysis of the degree to which the energy of the laser beam is reflected and absorbed when it penetrates or irradiates different material interfaces. For example, metallic materials typically have high reflectivity, while some polymer materials may have high absorptivity.

[0121] Then, based on the reflection and absorption characteristics, a material interface-specific influence factor is constructed. The material interface-specific influence factor refers to a parameter that quantifies the degree of influence of different material interfaces on laser measurement results. It is constructed by comprehensively considering the reflection and absorption characteristics of the materials.

[0122] Finally, the adjusted measurement fluctuation range is corrected based on the material interface specificity influencing factor. Building upon the initial adjustment, the impact of material interface specificity on measurement fluctuation can be further considered to refine the measurement fluctuation range. Furthermore, the acceptable threshold is determined based on the corrected measurement fluctuation range.

[0123] This embodiment acquires information about the multi-material layered structure of the door and window frame under test and identifies the reflection and absorption characteristics of different material interfaces to the laser signal, enabling a more comprehensive understanding of the complexity of signal-material interaction during laser measurement. Therefore, the constructed material interface-specific influence factor accurately reflects the impact of these interfaces on measurement result fluctuations. By applying this influence factor to the adjusted measurement fluctuation range, measurement uncertainties caused by material interface differences can be effectively corrected, making the final determined pass / fail threshold closer to the actual physical properties and measurement accuracy requirements.

[0124] To illustrate this technical solution more clearly, a specific example is used below. Assume the window / door frame under test is composed of three layers: aluminum alloy profile, insulated glass, and silicone sealant. First, obtain information about the multi-material laminated structure of the window / door frame, identifying the interface locations between the aluminum alloy and glass, the glass and sealant, and the types of each material. Next, through experiments or database searches, identify the reflection and absorption characteristics of the aluminum alloy, glass, and silicone sealant for specific laser wavelengths. For example, aluminum alloy has high reflectivity to laser light, glass has certain transmission and reflection characteristics, while silicone sealant may have strong absorption of laser light. Based on these characteristics, a material interface-specific influence factor is constructed. This factor quantifies the impact of different interfaces on the spatial distribution of the laser echo signal. For example, at the aluminum alloy / glass interface, a sudden change in reflectivity may lead to a local increase in the measurement fluctuation range. Finally, this material interface-specific influence factor is applied to the measurement fluctuation range, which has been initially adjusted for downstream process compatibility requirements, to correct it. For example, if the influence factor is high at a certain interface, the measurement fluctuation range in that area can be appropriately expanded to more accurately reflect the uncertainties in actual measurements. In this way, the final determined acceptance criteria will more precisely adapt to the characteristics of multi-material laminated structures, improving the accuracy of quality inspection.

[0125] Through the above technical solution, this embodiment can fully consider the inherent characteristics of multi-material laminated structures in door and window frames when determining the acceptance criteria, avoiding judgment deviations caused by differences in material interfaces. This significantly improves the accuracy and reliability of the acceptance criteria, thereby making the quality inspection results of rapid splicing structures of door and window frames more accurate, especially suitable for door and window frames with complex multi-material structures, and improving the applicability and robustness of the overall inspection method.

[0126] In some embodiments, obtaining the multi-material laminated structure information of the door and window frame to be tested in step S301 may include, but is not limited to, the following steps:

[0127] High-frequency vibration scanning was performed on the surface of the door and window frame to obtain the vibration response signal inside the multi-material laminated structure;

[0128] The characteristic changes of the vibration response signal are analyzed to identify the abrupt change points and attenuation characteristics of the vibration response signal at different material interfaces. The characteristic changes include frequency changes, amplitude changes and phase changes.

[0129] Information on multi-material stacked structures can be identified based on abrupt change points and attenuation characteristics.

[0130] In some embodiments, a high-frequency vibration scan can be performed on the surface of the door / window frame to obtain the vibration response signal inside the multi-material laminated structure. Mechanical vibrations, typically in the kilohertz to megahertz range, can be applied to the surface of the door / window frame using non-contact or contact excitation devices. This high-frequency vibration can propagate within the material in the form of waves and interact with the internal structure. Thus, the vibration response signal inside the multi-material laminated structure can be obtained, carrying information about the material layers, interfaces, and internal defects.

[0131] Then, the characteristic changes of the vibration response signal are analyzed to identify abrupt changes and attenuation characteristics at different material interfaces. These characteristic changes include frequency, amplitude, and phase variations. Frequency variations can reflect the inherent vibration modes or resonance phenomena of a material; amplitude variations indicate the attenuation or enhancement of vibration energy during propagation, and usually occur significantly at material interfaces; phase variations provide information about the wave propagation path and the properties of the material medium. Abrupt changes refer to regions where the frequency, amplitude, or phase of the vibration signal changes drastically in space or time. These abrupt changes typically correspond to the physical interfaces between different material layers because differences in acoustic impedance lead to wave reflection and refraction. Attenuation characteristics describe the degree of energy loss when the vibration signal propagates in different materials. Different materials have different absorption and scattering capabilities for vibration waves, thus their attenuation characteristics also vary. Based on the abrupt changes and attenuation characteristics, information about multi-material layered structures is identified, such as the thickness, location, and material type of each layer.

[0132] This embodiment applies high-frequency vibration to the surface of the door and window frame, enabling the vibration energy to penetrate the interior of the multi-material laminated structure. When the vibration wave propagates at the interfaces of different materials, its frequency, amplitude, and phase characteristics change significantly due to differences in the acoustic impedance of the materials. Specifically, at the material interfaces, the vibration response signal exhibits abrupt changes, such as a sudden drop in amplitude or a phase jump. Simultaneously, the attenuation characteristics of the vibration signal vary depending on the sound absorption and scattering properties of the materials themselves as it propagates within different materials. By accurately analyzing these characteristic changes, the location of the material interfaces can be effectively pinpointed, and the material type and layer thickness can be inferred based on the attenuation characteristics, thereby accurately identifying the information of the multi-material laminated structure.

[0133] Through the above technical solution, this embodiment can provide more refined and accurate internal structural information, such as the precise location of material interfaces and the distribution of different materials. Therefore, when subsequently correcting and adjusting the measurement fluctuation range, a more accurate material interface-specific influencing factor can be constructed based on more reliable multi-material layered structural information. This makes the determination of the pass / fail judgment threshold more scientific and reasonable, significantly improving the accuracy and reliability of door and window frame quality inspection.

[0134] In some embodiments, in step S303, constructing a material interface-specific influencing factor based on reflection and absorption characteristics may include, but is not limited to, the following steps:

[0135] Obtain information on the material aging degree of the door and window frame to be tested, as well as the environmental humidity and temperature fluctuation information of the testing area;

[0136] The reflection and absorption characteristics are corrected based on environmental humidity information, temperature fluctuation information, and material aging information.

[0137] Based on the modified reflection and absorption properties, a material interface-specific influencing factor is constructed.

[0138] In some embodiments, the reflective and absorptive properties of materials are not static; they are significantly affected by the aging of the material itself and environmental factors in the testing area, such as humidity and temperature fluctuations. If these dynamic factors are not adequately considered, the constructed material interface-specific influencing factor may not accurately reflect the actual situation, resulting in insufficiently precise correction of the measurement fluctuation range and consequently affecting the reliability of the final pass / fail threshold.

[0139] To this end, information on the aging degree of the material in the window / door frame under test, the ambient humidity of the testing area, and temperature fluctuations can be obtained first. Material aging degree information refers to the degree of performance degradation of the material during its service life due to environmental exposure, stress, and other factors. This includes, for example, the degree of oxidation, embrittlement, and changes in surface roughness. This information can be obtained through non-destructive testing techniques (such as ultrasonic testing, eddy current testing, and infrared thermography) or by tracing historical material data. Its purpose is to quantify the impact of changes in material properties over time on optical characteristics. Ambient humidity information refers to the water vapor content in the air of the environment where the window / door frame is being tested, usually expressed as relative or absolute humidity. Temperature fluctuation information refers to the temperature changes in the testing area over a certain time range, including the average, maximum, and minimum temperatures, as well as the rate of change. These environmental parameters can be monitored in real time by environmental sensors integrated into the testing equipment, aiming to capture the instantaneous impact of environmental factors on the material's optical response.

[0140] Then, based on environmental humidity, temperature fluctuation, and material aging information, the reflection and absorption characteristics are corrected. This can be done by establishing a correction model or consulting a correction coefficient table. For example, experiments or simulations can be conducted beforehand to study the changes in reflectivity and absorptivity of a specific material to laser signals under different humidity levels, temperatures, and aging degrees, and this data can be stored in a database. Based on real-time acquired environmental humidity, temperature fluctuation, and material aging information, corresponding correction factors can be retrieved from the database or calculated using a model. These correction factors are then applied to the initially identified reflection and absorption characteristics to obtain corrected reflection and absorption characteristics. The aim is to make the optical characteristic parameters used more closely resemble the material state under actual testing conditions. Finally, based on the corrected reflection and absorption characteristics, a material interface-specific influencing factor is constructed.

[0141] To illustrate this technical solution more clearly, a specific example is used below. Assume the window / door frame under test is composed of layers of aluminum alloy and silicone sealant. When constructing the material interface-specific influencing factor, information on the aging degree of the aluminum alloy and silicone sealant is first obtained through infrared spectroscopy analysis or batch data provided by the material supplier. This includes, for example, changes in the oxide layer thickness of the aluminum alloy or the hardness of the silicone sealant. Simultaneously, temperature and humidity sensors are deployed in the detection area to monitor ambient humidity and temperature fluctuations in real time. For example, the relative humidity is recorded as 60%, and the temperature fluctuates between 20℃ ± 2℃. Subsequently, a pre-established correction model is used, describing how the laser reflectivity and absorptivity of the aluminum alloy and silicone sealant change with oxide layer thickness, silicone sealant hardness, ambient humidity, and temperature. For example, the model might indicate that for every 10% increase in humidity, the reflectivity of the aluminum alloy decreases by 0.5%; and for every 1℃ increase in temperature, the absorptivity of the silicone sealant increases by 0.2%. Using this real-time information on material aging degree, ambient humidity, and temperature fluctuations, the initially identified reflectivity and absorptivity characteristics of the aluminum alloy-silicone interface are corrected. For example, if the initial reflectivity is 80%, it may become 79.5% under the current environment after correction according to the modified model. Finally, based on these corrected reflection and absorption characteristics, a material interface-specific influence factor that better reflects the actual working conditions is constructed for subsequent correction of the measurement fluctuation range.

[0142] Through the above technical solution, this embodiment achieves higher accuracy and adaptability in constructing a material interface-specific influencing factor by dynamically correcting the reflection and absorption characteristics. This allows for more refined and intelligent correction of measurement fluctuation ranges. Consequently, the final determined acceptance criteria more accurately reflect the actual quality status of door and window frames, significantly improving the accuracy and reliability of rapid assembly structure quality inspection for door and window frames. This reduces the risk of misjudgment or omission, which is of great significance for ensuring product quality and production efficiency.

[0143] In some embodiments, in step S304, the adjusted measurement fluctuation range is corrected according to the material interface-specific influencing factor, which may include, but is not limited to, the following steps:

[0144] Ultrasonic scanning was performed on the surface of the door and window frame to obtain ultrasonic echo signals inside the multi-material laminated structure;

[0145] The signal characteristics of ultrasonic echo signals are analyzed to identify micro-stress concentration areas and fatigue damage areas inside the material. The signal characteristics include propagation time, attenuation characteristics and scattering characteristics.

[0146] Based on the micro-stress concentration region and fatigue damage region, an influence factor for internal defects in the material is constructed.

[0147] The adjusted measurement fluctuation range is corrected based on the material interface specificity influence factor and the material internal defect influence factor.

[0148] In some embodiments, since the correction mainly focuses on the influence of the material interface on the measurement fluctuation, it may not fully consider the micro-defects existing inside the material, such as micro-stress concentration areas or fatigue damage areas. These internal defects can also have a significant impact on the fluctuation range of the measurement results, which may result in the corrected measurement fluctuation range being inaccurate, thereby affecting the accuracy of the pass / fail judgment limit and thus failing to fully reflect the actual quality status of the door and window frame.

[0149] To this end, ultrasonic scanning can be performed on the surface of the door and window frame to obtain ultrasonic echo signals from within the multi-material laminated structure. The aim is to transmit ultrasonic waves into the door and window frame using non-contact or contact methods, and to receive the echo signals formed after propagation, reflection, and scattering within the multi-material laminated structure. This allows for the acquisition of structural information about the materials, providing a data foundation for subsequent defect identification.

[0150] Then, the signal characteristics of the ultrasonic echo signal are analyzed to identify microscopic stress concentration areas and fatigue damage areas within the material. These signal characteristics include propagation time, attenuation characteristics, and scattering characteristics. Key parameters such as propagation time, attenuation characteristics, and scattering characteristics of the ultrasonic echo signal can be analyzed in depth. Changes in these signal characteristics directly reflect the physical state and structural integrity within the material. For example, abnormal changes in propagation time may indicate changes in material density or sound velocity; enhanced attenuation characteristics may be related to increased absorption or scattering within the material; and scattering characteristics may reveal the presence of heterogeneity or defects within the material. Through comprehensive analysis of these characteristics, microscopic stress concentration areas and fatigue damage areas within the material can be identified, with the aim of accurately locating and assessing potential defects within the material.

[0151] Then, based on the micro-stress concentration regions and fatigue damage regions, an internal defect influence factor for the material is constructed. A quantitative model or evaluation system can be established based on information such as the type, size, distribution, and severity of the identified micro-stress concentration regions and fatigue damage regions to generate a numerical factor that characterizes the degree of influence of these internal defects on the measurement fluctuation range—that is, the material internal defect influence factor. The construction of this factor aims to transform the objective existence of internal material defects into a quantitative parameter that can be used to correct the measurement fluctuation range.

[0152] Finally, the adjusted measurement fluctuation range is corrected based on the material interface-specific influence factor and the material internal defect influence factor. The material interface-specific influence factor and the newly constructed material internal defect influence factor are considered comprehensively. These two influence factors can be fused using weighted averaging, coupled models, or other appropriate algorithms to achieve a more comprehensive and accurate correction to the adjusted measurement fluctuation range.

[0153] To illustrate this technical solution more clearly, a specific example is used below. Assume the window / door frame under test contains a multi-material laminated structure of aluminum alloy and fiberglass composite. When correcting for measurement fluctuations, the surface of the window / door frame is first scanned using an ultrasonic probe. The ultrasonic signal penetrates the material's interior, and when it encounters the interface between the aluminum alloy and the composite material, or any micro-cracks or voids within the material, it is reflected and scattered. The propagation time, amplitude attenuation, and scattering pattern of the received ultrasonic echo signals are recorded and analyzed. For example, in a certain area, the ultrasonic propagation time is abnormally prolonged and attenuated significantly, which may indicate the presence of micro-stress concentration or fatigue damage in that area. Based on these signal characteristics, specific micro-stress concentration areas and fatigue damage areas are identified, and an internal material defect influence factor is constructed according to their severity and distribution. For example, a quantitative model can be set up to map the size, density, and location of defects to influence factor values. Subsequently, this internal material defect influence factor is combined with a material interface-specific influence factor previously constructed based on the material interface's reflection and absorption characteristics, for example, through a weighted average or a more complex coupling model, to comprehensively correct the adjusted measurement fluctuation range. The revised measurement fluctuation range will more accurately reflect the actual measurement uncertainty of the door and window frame when internal defects are taken into account, thus providing a more solid foundation for the subsequent determination of the acceptance criteria.

[0154] Through the above technical solution, this embodiment not only considers the specific influencing factors of material interfaces but also further introduces the influencing factors of internal material defects. This allows the correction process to more comprehensively reflect the actual quality status of the door and window frames, especially quantifying and compensating for the measurement fluctuations caused by micro-stress concentration areas and fatigue damage areas within the material. This embodiment can significantly improve the accuracy of the measurement fluctuation range correction, thereby making the final determined acceptance criteria more precise and reliable, effectively avoiding potential quality risks caused by internal defects, and improving the overall reliability of the quality inspection of rapid assembly structures for door and window frames.

[0155] The beneficial effects of implementing the embodiments of the present invention include: the embodiments of this application first obtain the initial frame structure measurement results, the dynamic behavior information of the robotic arm end effector, the spatial distribution information of the laser echo signal, and the local temperature field information. Then, the measurement result deviation distribution is predicted, and the total measurement deviation value and the measurement fluctuation range are calculated. Based on the total measurement deviation value, the initial frame structure measurement results are corrected to obtain the target frame structure measurement results. Based on the measurement fluctuation range, the pass / fail judgment limit is determined. Finally, based on the target frame structure measurement results and the pass / fail judgment limit, the quality inspection results are generated. Thus, the quality inspection results can be generated by combining the deviation distribution, the measurement deviation value, and the measurement fluctuation range to achieve structural quality inspection, thereby improving the inspection accuracy and product quality.

[0156] like Figure 2 As shown, this embodiment of the invention also provides a quality inspection system for rapid assembly structures of door and window frames, comprising:

[0157] The data acquisition module 401 is used to acquire the initial frame structure measurement results, the dynamic behavior information of the robotic arm end effector, the spatial distribution information of the laser echo signal, and the local temperature field information. The dynamic behavior information of the robotic arm end effector includes vibration and deformation, and the local temperature field information includes temperature distribution and temperature gradient.

[0158] The deviation distribution prediction module 402 is used to predict the deviation distribution of the measurement results based on the dynamic behavior information of the robotic arm end effector, the spatial distribution information of the laser echo signal, the local temperature field information, and the dynamic influence map.

[0159] The deviation calculation module 403 is used to calculate the total measurement deviation value and the measurement fluctuation range based on the deviation distribution of the measurement results.

[0160] The measurement result correction module 404 is used to correct the initial frame structure measurement results based on the total measurement deviation value to obtain the target frame structure measurement results.

[0161] The boundary determination module 405 is used to determine the pass / fail judgment boundary based on the measurement fluctuation range.

[0162] The quality judgment module 406 is used to generate quality inspection results based on the measurement results of the target frame structure and the acceptance judgment threshold.

[0163] The content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0164] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

Claims

1. A method for quality inspection of rapid splicing structures for door and window frames, characterized in that, Includes the following steps: The system acquires initial frame structure measurement results, dynamic behavior information of the robotic arm end effector, spatial distribution information of laser echo signals, and local temperature field information. The dynamic behavior information of the robotic arm end effector includes vibration and deformation, and the local temperature field information includes temperature distribution and temperature gradient. Based on the dynamic behavior information of the robotic arm end effector, the spatial distribution information of the laser echo signal, the local temperature field information, and the dynamic influence map, the measurement result deviation distribution is predicted. The dynamic influence map is a pre-trained machine learning model. By taking the dynamic behavior information of the robotic arm end effector, the spatial distribution information of the laser echo signal, and the local temperature field information as input, the machine learning model outputs the measurement result deviation distribution. Based on the deviation distribution of the measurement results, calculate the total measurement deviation value and the measurement fluctuation range; Based on the total measurement deviation value, the initial frame structure measurement results are corrected to obtain the target frame structure measurement results; Based on the measured fluctuation range, the pass / fail threshold is determined; Based on the measurement results of the target frame structure and the acceptance criteria, a quality inspection result is generated; The step of predicting the distribution of measurement result deviation based on the dynamic behavior information of the robotic arm end effector, the spatial distribution information of the laser echo signal, the local temperature field information, and the dynamic influence map includes: Feature scanning is performed on the geometric feature calibration reference block to obtain feature scanning information. The geometric feature calibration reference block contains various door and window frame geometric features. Based on the feature scanning information, the interaction pattern of the geometric features is identified; Adjust the path parameters in the dynamic influence map according to the interaction mode of the geometric features; Based on the adjusted dynamic influence map, the dynamic behavior information of the robotic arm end effector, the spatial distribution information of the laser echo signal, and the local temperature field information, the deviation distribution of the measurement results is predicted. The step of calculating the total measurement deviation value and the measurement fluctuation range based on the deviation distribution of the measurement results includes: Obtain the geometric feature type of the door and window frame to be tested; Select target distribution morphology feature parameters that match the geometric feature type from a preset set of distribution morphology feature parameters; The rate of change of monitored parameters includes the rate of change of dynamic behavior information of the robotic arm end effector, the rate of change of spatial distribution information of laser echo signal, and the rate of change of local temperature field information. The confidence level for fluctuation extraction is determined based on the rate of change of the parameters. Based on the confidence level of the fluctuation and the characteristic parameters of the target distribution, the measurement fluctuation range is extracted from the deviation distribution of the measurement results; Determine the defect sensitivity based on the type of defect in the door and window frame to be tested; Based on the aforementioned defect sensitivity, the deviation extraction weight is determined; Based on the deviation extraction weights and the target distribution morphological characteristic parameters, the total measurement deviation value is extracted from the measurement result deviation distribution.

2. The method according to claim 1, characterized in that, The step of correcting the initial frame structure measurement results based on the total measurement deviation value to obtain the target frame structure measurement results includes: Obtain the geometric feature type of the door and window frame to be tested; Based on the geometric feature type, the door and window frame to be tested is divided into regions to obtain multiple local regions; Based on the multiple local regions, the initial frame structure measurement results are decomposed to obtain the corresponding local measurement results; Based on the dynamic behavior information of the robotic arm end effector, the spatial distribution information of the laser echo signal, the local temperature field information, and the dynamic influence map, the local deviation distribution corresponding to each local region is predicted; Calculate the local measurement deviation value based on the local deviation distribution; The total measurement deviation value and multiple local measurement deviation values ​​are verified to obtain the verification result; If the verification result is that the verification passes, then deviation correction is performed based on the local measurement results and the corresponding local measurement deviation values. The measurement results of the target frame structure are obtained by integrating the local measurement results after multiple deviation corrections.

3. The method according to claim 1, characterized in that, The step of determining the pass / fail threshold based on the measured fluctuation range includes: Obtain the geometric feature type of the door and window frame to be tested; Search the process database for the downstream process compatibility requirements corresponding to the geometric feature type; The measurement fluctuation range is adjusted according to the downstream process compatibility requirements. The qualified judgment threshold is determined based on the adjusted measurement fluctuation range.

4. The method according to claim 3, characterized in that, The step of determining the pass / fail threshold based on the adjusted measurement fluctuation range includes: Obtain multi-material laminated structure information of the door and window frame to be tested, wherein the multi-material laminated structure information includes material interface location and material type; Based on the multi-material stacked structure information, the reflection and absorption characteristics of different material interfaces on laser signals are identified; Based on the aforementioned reflection and absorption characteristics, a material interface-specific influencing factor is constructed. The adjusted measurement fluctuation range was corrected based on the material interface-specific influencing factor. The qualified judgment threshold is determined based on the corrected measurement fluctuation range.

5. The method according to claim 4, characterized in that, The acquisition of multi-material layered structure information of the door and window frame under test includes: High-frequency vibration scanning was performed on the surface of the door and window frame to obtain the vibration response signal inside the multi-material laminated structure; The characteristic changes of the vibration response signal are analyzed to identify the abrupt change points and attenuation characteristics of the vibration response signal at different material interfaces. The characteristic changes include frequency changes, amplitude changes, and phase changes. The information of the multi-material stacked structure is identified based on the abrupt change point and the attenuation characteristics.

6. The method according to claim 4, characterized in that, The construction of a material interface-specific influencing factor based on the reflection and absorption characteristics includes: Acquire information on the material aging degree of the door and window frame to be tested, as well as the environmental humidity and temperature fluctuation information of the testing area; The reflection and absorption characteristics are corrected based on the environmental humidity information, the temperature fluctuation information, and the material aging information. Based on the corrected reflection and absorption properties, the material interface-specific influencing factor is constructed.

7. The method according to claim 4, characterized in that, The step of correcting the adjusted measurement fluctuation range based on the material interface-specific influencing factor includes: Ultrasonic scanning was performed on the surface of the door and window frame to obtain ultrasonic echo signals inside the multi-material laminated structure; The signal characteristics of the ultrasonic echo signal are analyzed to identify micro-stress concentration areas and fatigue damage areas inside the material. The signal characteristics include propagation time, attenuation characteristics, and scattering characteristics. Based on the micro-stress concentration region and the fatigue damage region, an internal defect influence factor for the material is constructed. The adjusted measurement fluctuation range is corrected based on the material interface specificity influence factor and the material internal defect influence factor.

8. A rapid assembly structure quality inspection system for door and window frames, characterized in that, include: The data acquisition module is used to acquire the initial frame structure measurement results, the dynamic behavior information of the robotic arm end effector, the spatial distribution information of the laser echo signal, and the local temperature field information. The dynamic behavior information of the robotic arm end effector includes flutter and deformation, and the local temperature field information includes temperature distribution and temperature gradient. The deviation distribution prediction module is used to predict the measurement result deviation distribution based on the dynamic behavior information of the robotic arm end effector, the spatial distribution information of the laser echo signal, the local temperature field information, and the dynamic influence map. The dynamic influence map is a pre-trained machine learning model. By taking the dynamic behavior information of the robotic arm end effector, the spatial distribution information of the laser echo signal, and the local temperature field information as input, the machine learning model outputs the measurement result deviation distribution. The deviation calculation module is used to calculate the total measurement deviation value and the measurement fluctuation range based on the deviation distribution of the measurement results; The measurement result correction module is used to correct the initial frame structure measurement results based on the total measurement deviation value to obtain the target frame structure measurement results; The boundary determination module is used to determine the pass / fail judgment boundary based on the measured fluctuation range; The quality judgment module is used to generate quality inspection results based on the measurement results of the target frame structure and the pass / fail judgment threshold; The step of predicting the distribution of measurement result deviation based on the dynamic behavior information of the robotic arm end effector, the spatial distribution information of the laser echo signal, the local temperature field information, and the dynamic influence map includes: Feature scanning is performed on the geometric feature calibration reference block to obtain feature scanning information. The geometric feature calibration reference block contains various door and window frame geometric features. Based on the feature scanning information, the interaction pattern of the geometric features is identified; Adjust the path parameters in the dynamic influence map according to the interaction mode of the geometric features; Based on the adjusted dynamic influence map, the dynamic behavior information of the robotic arm end effector, the spatial distribution information of the laser echo signal, and the local temperature field information, the deviation distribution of the measurement results is predicted. The step of calculating the total measurement deviation value and the measurement fluctuation range based on the deviation distribution of the measurement results includes: Obtain the geometric feature type of the door and window frame to be tested; Select target distribution morphology feature parameters that match the geometric feature type from a preset set of distribution morphology feature parameters; The rate of change of monitored parameters includes the rate of change of dynamic behavior information of the robotic arm end effector, the rate of change of spatial distribution information of laser echo signal, and the rate of change of local temperature field information. The confidence level for fluctuation extraction is determined based on the rate of change of the parameters. Based on the confidence level of the fluctuation and the characteristic parameters of the target distribution, the measurement fluctuation range is extracted from the deviation distribution of the measurement results; Determine the defect sensitivity based on the type of defect in the door and window frame to be tested; Based on the aforementioned defect sensitivity, the deviation extraction weight is determined; Based on the deviation extraction weights and the target distribution morphological characteristic parameters, the total measurement deviation value is extracted from the measurement result deviation distribution.

Images