Aluminum profile machining quality detection system and method based on multi-dimensional data analysis

Abstract

The invention relates to the technical field of aluminum profile machining detection, and discloses an aluminum profile machining quality detection system and method based on multi-dimensional data analysis, and the method comprises the steps: S1, collecting cutting end face image data through an industrial camera, collecting cutting contour size data through a laser sensor, and collecting machining vibration signal data through a vibration sensor; through cooperative acquisition of the industrial camera, the laser sensor and the vibration sensor, the limitation of a single sensing dimension on surface quality, size precision and dynamic stability monitoring is overcome, the processing quality of a single aluminum profile is effectively and comprehensively judged and analyzed, and the processing quality of the single aluminum profile is effectively and comprehensively judged and analyzed by utilizing score feature extraction and multi-window trend analysis in a sliding window. The stability of the aluminum profile machining process is further judged and early warned, the comprehensiveness and accuracy of aluminum profile quality evaluation are improved, and process abnormity can be recognized before the quality problem is dominated.

Description

Technical Field

[0001] This application relates to the field of aluminum profile processing and inspection technology, and in particular to an aluminum profile processing quality inspection system and method based on multidimensional data analysis. Background Technology

[0002] Aluminum profiles are widely used in construction, automotive, aerospace and other fields due to their advantages such as light weight, high strength and easy processing. The processing quality of aluminum profiles, especially the cutting quality, directly affects the structural strength, assembly accuracy and aesthetics of the final product. Therefore, efficient and accurate quality inspection and control of the aluminum profile processing process is a key link in the manufacturing industry.

[0003] In existing technologies, the quality inspection of aluminum profile processing mainly relies on single-dimensional technologies such as machine vision, dimensional measurement, or vibration analysis. These methods all have limitations: visual inspection is easily affected by environmental interference and cannot detect dimensional deformation; dimensional measurement is mostly based on sampling inspection and is difficult to reflect the dynamic stability of the processing; vibration analysis is difficult to directly correlate with the surface and dimensional quality of the workpiece. This single-dimensional quality inspection mode leads to a one-sided quality evaluation of aluminum profile processing and cannot provide effective early warning before quality problems occur. Summary of the Invention

[0004] To overcome the limitations of single detection technologies and achieve comprehensive evaluation and proactive early warning of aluminum profile processing quality, this application provides an aluminum profile processing quality detection system and method based on multidimensional data analysis.

[0005] Firstly, this application provides a method for inspecting the processing quality of aluminum profiles based on multidimensional data analysis, employing the following technical solution:

[0006] A system and method for inspecting the processing quality of aluminum profiles based on multidimensional data analysis, the method comprising:

[0007] S1. Acquire image data of the cutting end face through an industrial camera, acquire cutting contour dimension data through a laser sensor, and acquire processing vibration signal data through a vibration sensor;

[0008] S2. Perform comprehensive analysis on the collected multi-source data to obtain the comprehensive quality score of the current aluminum profile processed parts, and make quality judgment on the aluminum profile processed parts based on the comprehensive quality score.

[0009] S3. Preset a sliding window with the number of workpieces as the unit, extract features from the comprehensive quality score sequence of all processed parts within the sliding window, and obtain the score feature value of the preset sliding window.

[0010] S4. Perform trend analysis on the scoring feature value sequence of a preset number of continuous sliding windows. When the downward trend of the scoring feature value is detected to exceed the preset threshold, generate a processing anomaly warning.

[0011] By adopting the above technical solution, comprehensive quality inspection based on multi-source data is realized. Feature extraction and trend analysis of the comprehensive quality scoring sequence are performed through a preset sliding window. In this way, when the downward trend of the scoring feature value is detected to exceed the threshold, a processing anomaly warning is generated, which improves the comprehensiveness and timeliness of aluminum profile processing quality inspection.

[0012] Optionally, in step S2, the process of determining the quality of aluminum profile processed parts based on the comprehensive quality score includes:

[0013]

[0014] The overall quality score is obtained through formula analysis and calculation. ;

[0015] in, Scoring of surface defects in aluminum profiles. Scoring for dimensional deviations of aluminum profiles. Scoring for vibration signal spectrum analysis. , , The first weighting coefficient;

[0016] Comprehensive quality score Compared with the preset non-compliance scoring threshold Perform a comparison;

[0017] like If the aluminum profile is found to be defective, it will be marked and a defective signal will be generated.

[0018] Conversely, if the aluminum profile is not found to be qualified, it will be marked and a qualified signal will be generated.

[0019] By adopting the above technical solution, a comprehensive quality score is obtained through formulaic calculation and automatically compared with a preset non-conforming score threshold, thus realizing objective and efficient quality judgment of aluminum profile processed parts and reducing subjective errors.

[0020] Optionally, the process of obtaining the surface defect score of the aluminum profile includes:

[0021]

[0022] The surface defect score of aluminum profiles is obtained through analysis and calculation using the above formula. ;

[0023] in, This is the constant value for the maximum score in the scoring system. This represents the total number of defects currently detected on the surface of the aluminum profile. Let be the type weight coefficient for the i-th defect. Let be the projected area of ​​the i-th defect. This represents the total surface area of ​​the aluminum profile.

[0024] By adopting the above technical solution, and by considering the weight of defect type and the ratio of defect projection area to total surface area, the surface defect score of aluminum profile is calculated, which makes the score more accurately reflect the severity of surface defects and improves the accuracy of surface quality assessment.

[0025] Optionally, the process for obtaining the aluminum profile dimensional deviation score includes:

[0026]

[0027] The aluminum profile dimensional deviation score is obtained through analysis and calculation using the above formula. ;

[0028] in, The total number of key dimensional parameters being measured. For the j-th dimensional parameter, Let j be the theoretical design value of the j-th dimensional parameter. The weighting coefficient for the j-th dimension parameter. The maximum permissible composite relative deviation threshold. To take the maximum value within the parentheses.

[0029] By adopting the above technical solution, the relative deviations of multiple key dimensional parameters are calculated, and combined with weighting coefficients and the maximum permissible deviation threshold, a dimensional deviation score is obtained, thereby achieving a comprehensive quantitative assessment of the dimensional quality of aluminum profiles and improving the reliability of dimensional control.

[0030] Optionally, the process of obtaining the vibration signal spectrum analysis score includes:

[0031]

[0032] The vibration signal spectrum analysis score is obtained through the above formula analysis and calculation. ;

[0033] in, The number of characteristic frequency bands, Let be the amplitude of the current vibration signal in the kth characteristic frequency band. To reference the corresponding amplitude under normal conditions, The weighting coefficients for the k-th frequency band are... This is the maximum allowable amplitude deviation threshold.

[0034] By adopting the above technical solution, and by comparing the amplitude deviation of the current vibration signal with that of the reference normal state in the characteristic frequency band, the vibration signal spectrum analysis score is calculated, which can effectively identify abnormal vibrations in the processing process, thereby reflecting the equipment status and processing quality.

[0035] Optionally, in step S3, the process of obtaining the scoring feature value of the preset sliding window includes:

[0036]

[0037]

[0038]

[0039]

[0040] The scoring characteristic value of the preset sliding window is obtained by combining the above formulas and performing simultaneous analysis. ;

[0041] in, The defect rate of aluminum profile workpieces within the sliding window. This represents the average of all comprehensive quality scores within the sliding window. To account for the volatility of the overall quality score within the sliding window, , , This is the second weighting coefficient. This represents the number of defective aluminum profile workpieces within the sliding window. This represents the total number of aluminum profile workpieces within the sliding window. The overall quality score for the l-th aluminum profile workpiece within the sliding window.

[0042] By adopting the above technical solution, the non-conformance rate, average score, and volatility within the sliding window are calculated using simultaneous formulas to obtain the score feature value. This achieves comprehensive feature extraction of quality data within the window and provides key indicators for trend analysis.

[0043] Optionally, in step S4, the process of performing trend analysis on the scoring feature value sequence of a preset number of continuous sliding windows includes:

[0044]

[0045] The slope of the trend of the scoring feature value of a preset number of continuous sliding windows over time was obtained by analyzing and calculating using the above formula. ;

[0046] in, The preset number of consecutive sliding windows, Number the time sequence of the sliding window sequence. For the sliding window time indicator, The rating feature value corresponding to the h-th sliding window;

[0047] The slope of the trend of the scoring feature values ​​of a preset number of continuous sliding windows over time. Compared with the preset change threshold Perform a comparison;

[0048] when If an abnormal downward trend in the quality of aluminum profiles is detected during the processing, a processing anomaly warning will be generated.

[0049] By adopting the above technical solution, and by calculating the trend slope of the continuous sliding window scoring feature value and comparing it with a preset threshold, the abnormal downward trend of quality can be detected in a timely manner, generating a processing abnormality warning, thus realizing predictive quality monitoring.

[0050] Optionally, the method further includes:

[0051] S5. Analyze the scoring feature values ​​of a preset number of continuous sliding windows and dynamically adjust the size of the sliding windows.

[0052] By adopting the above technical solution and dynamically adjusting the size of the sliding window, trend analysis can adapt to data changes, thereby improving the flexibility and adaptability of the quality inspection system.

[0053] Optionally, the dynamic adjustment process includes:

[0054]

[0055]

[0056] The size of the sliding window can be obtained by simultaneously calculating and analyzing the above formulas. ;

[0057] in, To predetermine the standard deviation of the rating feature values ​​of the most recent n consecutive sliding windows, This is the average of the rating feature values ​​of the most recent n consecutive sliding windows. The preset baseline window size, This is the floor function.

[0058] By adopting the above technical solution, the window size is dynamically calculated based on the standard deviation and average value of the most recent continuous sliding window score feature values, making the window adjustment more in line with the data fluctuation characteristics and optimizing the accuracy of trend analysis.

[0059] Secondly, this application provides an aluminum profile processing quality inspection system based on multidimensional data analysis, employing the following technical solution:

[0060] A multidimensional data analysis-based aluminum profile processing quality inspection system, wherein the system is applied to the multidimensional data analysis-based aluminum profile processing quality inspection method described in any one of the above-mentioned methods, and the system comprises:

[0061] The data acquisition module is used to collect multi-source data during the processing through industrial cameras, laser sensors, and vibration sensors;

[0062] The comprehensive scoring calculation module is used to perform comprehensive analysis on the collected multi-source data to obtain the comprehensive quality score of the aluminum profile processed parts and to make quality judgments.

[0063] The sliding window feature extraction module is used to extract features from the comprehensive quality score sequence based on a preset sliding window to obtain score feature values;

[0064] The trend analysis module is used to perform trend analysis on the scoring feature value sequence of a continuous sliding window and generate early warnings of processing anomalies.

[0065] The dynamic window adjustment module is used to dynamically adjust the size of the sliding window based on the scoring feature value;

[0066] The early warning module is used to execute early warning signals.

[0067] By adopting the above technical solution, a detection system for implementing the above method is provided. Through modular design, the specific implementation and application of the aluminum profile processing quality detection method are ensured, and the practicality and automation level of the system are improved.

[0068] In summary, this application includes at least one of the following beneficial technical effects:

[0069] (1) This invention overcomes the limitations of a single sensing dimension in monitoring surface quality, dimensional accuracy and dynamic stability by using industrial cameras, laser sensors and vibration sensors to collect data in a coordinated manner. It effectively conducts a comprehensive evaluation and analysis of the processing quality of individual aluminum profiles. Furthermore, by using the scoring feature extraction within the sliding window and multi-window trend analysis, it further evaluates and warns of the stability of the aluminum profile processing process. This not only improves the comprehensiveness and accuracy of aluminum profile quality evaluation, but also enables the identification of process anomalies before quality problems become apparent. It effectively addresses the defects of single-dimensional detection, sampling lag and inability to provide dynamic early warning, and provides an effective guarantee for continuous quality control of aluminum profile processing.

[0070] (2) By constructing quantitative scoring models for surface defects, dimensional deviations and vibration stability respectively and integrating them into a comprehensive quality index, this invention realizes a multi-dimensional and interpretable unified evaluation of the cutting quality of aluminum profiles. It can not only distinguish whether aluminum profiles are qualified in real time, but also provide data basis for process optimization through a structured scoring system. It effectively overcomes the defects of the single detection dimension in the background technology, which is one-sided and difficult to fully reflect the processing quality, and improves the systematicness and reliability of quality judgment.

[0071] (3) This invention achieves overall quantitative assessment of batch processing quality by defining window feature values ​​that integrate non-conforming rate, mean and standard deviation; and by using slope detection of continuous window feature values, it achieves early identification of quality degradation trend, extending the quality monitoring perspective from a single workpiece to the process sequence, and can detect potential systemic degradation problems before they have a significant impact on the production line. It effectively overcomes the shortcomings of traditional detection methods proposed in the background technology, which can only make post-judgment judgments and cannot achieve pre-warning, and provides effective support for predictive quality control of aluminum profile processing. Attached Figure Description

[0072] Figure 1 This is a flowchart of the steps of an aluminum profile processing quality inspection method based on multidimensional data analysis disclosed in this invention.

[0073] Figure 2 This is a schematic diagram of an aluminum profile processing quality inspection system based on multidimensional data analysis disclosed in this invention. Detailed Implementation

[0074] The embodiments of this application are described in detail below, and examples of the embodiments are shown in the accompanying drawings.

[0075] In the description of this specification, the references to "certain embodiments," "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples" refer to specific features, structures, materials, or characteristics described in connection with the described embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0076] This application discloses a method for inspecting the processing quality of aluminum profiles based on multidimensional data analysis, referring to... Figure 1 The method includes:

[0077] S1. Acquire image data of the cut end face through an industrial camera to analyze surface roughness, burrs and texture uniformity; acquire cutting contour dimension data through a laser sensor, including but not limited to length deviation, flatness and perpendicularity; acquire vibration signals of the spindle or worktable of the processing equipment through a vibration sensor to reflect the dynamic stability of the processing.

[0078] S2. Perform comprehensive analysis on the collected multi-source data to obtain the comprehensive quality score of the current aluminum profile processed parts, and make quality judgment on the aluminum profile processed parts based on the comprehensive quality score.

[0079] S3. A sliding window with the number of continuously processed workpieces as the unit is preset. The comprehensive quality score sequence of all processed workpieces in the sliding window is subjected to feature extraction to obtain the score feature value of the preset sliding window. The extracted features include, but are not limited to, the score mean, standard deviation, and descent slope, which are used to characterize the quality stability of the current window.

[0080] S4. Perform trend analysis on the scoring feature value sequence of a preset number of continuous sliding windows. Generally, a linear regression algorithm is used to analyze the trend of the scoring feature value changes. When the downward trend of the scoring feature value is detected to exceed the preset threshold, a processing anomaly warning is generated.

[0081] Through the above technical solution, this embodiment provides a method for inspecting the processing quality of aluminum profiles based on multidimensional data analysis. The method overcomes the limitations of single sensing dimensions in monitoring surface quality, dimensional accuracy, and dynamic stability by using the collaborative acquisition of industrial cameras, laser sensors, and vibration sensors. It effectively performs comprehensive evaluation and analysis of the processing quality of individual aluminum profiles, and further evaluates and warns of the stability of the aluminum profile processing process by using scoring feature extraction within a sliding window and multi-window trend analysis. This not only improves the comprehensiveness and accuracy of aluminum profile quality evaluation, but also identifies process anomalies before quality problems become apparent. It effectively addresses the shortcomings of single-dimensional detection, sampling lag, and lack of dynamic early warning, providing an effective guarantee for continuous quality control of aluminum profile processing.

[0082] In one embodiment, step S2, the process of determining the quality of aluminum profile processed parts based on a comprehensive quality score, includes:

[0083]

[0084] The overall quality score is obtained through formula analysis and calculation. ;

[0085] in, For image analysis-based scoring of aluminum profile surface defects, For scoring the dimensional deviation of aluminum profiles based on laser measurement, Scoring for vibration signal spectrum analysis. , , The first weighting coefficient is usually determined by production quality experts based on historical quality data or customer requirements. Alternatively, machine learning methods can be used to fit the importance of each dimension of quality.

[0086] Comprehensive quality score Compared with the preset non-compliance scoring threshold Compare and preset the non-compliance scoring threshold. It can be set based on experience, generally set at 70-80 points, and can be dynamically adjusted according to product grade requirements and historical pass rate data;

[0087] like If the aluminum profile is deemed unqualified, it will be marked and a non-qualified signal will be generated, triggering a rejection or rework process.

[0088] Conversely, if the aluminum profile is not found to be qualified, it is marked and a qualified signal is generated, and it proceeds to the next process.

[0089] The process of obtaining the surface defect score for the aluminum profile includes:

[0090]

[0091] The surface defect score of aluminum profiles is obtained through analysis and calculation using the above formula. ;

[0092] in, This is the maximum score constant in the scoring system, usually set to 100. This represents the total number of defects detected on the surface of the aluminum profile, which can be automatically identified and counted using image segmentation algorithms or deep learning object detection models. This is the type weighting coefficient for the i-th defect. It is assigned a value by process experts based on the degree of impact of the defect type (e.g., crack, scratch, dent) on product performance. Cracks, being severe defects, are typically assigned a value of 1, while minor scratches can be assigned a value of 0.3. Let be the projected area of ​​the i-th defect. In image processing, the number of pixels enclosed by the defect outline is calculated through pixel statistics, and then converted into the actual physical area by combining the camera calibration parameters. This refers to the total surface area of ​​the aluminum profile, which is known from the product drawings or calculated through a 3D scanning model.

[0093] The process for obtaining the aluminum profile dimensional deviation score includes:

[0094]

[0095] The aluminum profile dimensional deviation score is obtained through analysis and calculation using the above formula. ;

[0096] in, This refers to the total number of key dimensional parameters being measured, typically including 3–8 key dimensions such as length, angle, and flatness. The actual measured value of the j-th dimensional parameter is obtained in real time through a laser sensor. The theoretical design value for the j-th dimensional parameter is obtained from CAD drawings or process specifications. The weighting coefficient for the j-th dimensional parameter is assigned based on the degree of influence of this dimension on assembly accuracy and structural function. The maximum permissible comprehensive relative deviation threshold is set empirically, typically to 0.05, but can be tightened to 0.02 for high-precision scenarios. To take the maximum value within the parentheses.

[0097] The process of obtaining the vibration signal spectrum analysis score includes:

[0098]

[0099] The vibration signal spectrum analysis score is obtained through the above formula analysis and calculation. ;

[0100] in, The number of characteristic frequency bands typically covers the spindle rotation fundamental frequency, tool engagement frequency, and their harmonics, approximately 3–5 characteristic frequency bands. To determine the amplitude of the current vibration signal in the k-th characteristic frequency band, the amplitude value is obtained by performing an FFT transform on the time-domain signal acquired by the vibration sensor and integrating within the corresponding frequency band. To reference the corresponding amplitude under normal conditions, multiple measurements were taken and the average value was calculated when the equipment was initially in good working order. The weighting coefficient for the k-th frequency band can be set empirically. For example, high-frequency bands (such as tool meshing bands) usually have a higher weight (0.6–0.8) because they are more sensitive to tool wear and impact. The maximum allowable amplitude deviation threshold is calculated based on historical normal operation data and is usually taken as 3 times the standard deviation of normal amplitude fluctuation.

[0101] Through the above technical solution, this embodiment provides a comprehensive scoring and judgment method for aluminum profile processing quality based on multi-source sensor data fusion. The method constructs quantitative scoring models for surface defects, dimensional deviations, and vibration stability respectively, and integrates them into a comprehensive quality index. This achieves a multi-dimensional, interpretable, and unified evaluation of aluminum profile cutting quality. It can not only distinguish whether aluminum profiles are qualified in real time, but also provide data basis for process optimization through a structured scoring system. This effectively overcomes the shortcomings of the single detection dimension in the background technology, which is one-sided and cannot fully reflect the processing quality, and improves the systematicness and reliability of quality judgment.

[0102] In one embodiment, step S3, the process of obtaining the scoring feature value of the preset sliding window includes:

[0103]

[0104]

[0105]

[0106]

[0107] The scoring characteristic value of the preset sliding window is obtained by combining the above formulas and performing simultaneous analysis. ;

[0108] in, The defect rate of aluminum profile workpieces within the sliding window reflects the overall compliance level of processing quality during that window period. The mean of all comprehensive quality scores within the sliding window represents the central tendency of the overall quality level. This represents the volatility of the overall quality score within the sliding window, reflecting the fluctuation and stability of the quality output. , , This is the second weighting coefficient, used to adjust the contribution of each feature to the overall evaluation. It can be determined through historical data regression analysis or expert experience. The number of defective aluminum profile workpieces within the sliding window is directly calculated from the results of step S2. This refers to the total number of aluminum profile workpieces within the sliding window, i.e., the preset window size. The overall quality score of the l-th aluminum profile workpiece within the sliding window can be calculated and analyzed according to step S2.

[0109] In step S4, the process of performing trend analysis on the scoring feature value sequence of a preset number of continuous sliding windows includes:

[0110]

[0111] The slope of the trend of the scoring feature value of a preset number of continuous sliding windows over time was obtained by analyzing and calculating using the above formula. ;

[0112] in, The preset number of consecutive sliding windows is typically set to 5-10. Too few windows can easily lead to false alarms, while too many will result in delayed warnings. Number the time sequence of the sliding window sequence. , For the sliding window time identifier, you can take the absolute timestamp at the end of the window, or directly use the window number sequence as the equally spaced time variable. The scoring feature value corresponding to the h-th sliding window can be obtained by calculation and analysis according to step S3;

[0113] The slope of the trend of the scoring feature values ​​of a preset number of continuous sliding windows over time. Compared with the preset change threshold The preset change threshold is compared. It can be determined by statistical analysis of the normal fluctuation range in historical data, and its value depends on the quality stability requirements;

[0114] when If an abnormal downward trend in the quality of aluminum profiles is detected during the processing, a processing anomaly warning will be generated.

[0115] Through the above technical solution, this embodiment provides a quality early warning method for aluminum profile processing based on sliding window and trend analysis. The method achieves overall quantitative evaluation of batch processing quality by defining window feature values ​​that integrate non-conforming rate, mean, and standard deviation. Furthermore, by utilizing the slope detection of continuous window feature values, it achieves early identification of quality degradation trends. This method extends the quality monitoring perspective from a single workpiece to a process sequence, enabling the detection of potential systemic degradation problems before they significantly impact the production line. It effectively overcomes the shortcomings of traditional detection methods proposed in the background art, which can only make post-event judgments and cannot achieve pre-event early warning, thus providing effective support for predictive quality control in aluminum profile processing.

[0116] In one embodiment, the method further includes:

[0117] S5. Analyze the scoring feature values ​​of a preset number of continuous sliding windows and dynamically adjust the size of the sliding windows.

[0118] The dynamic adjustment process includes:

[0119]

[0120]

[0121] The size of the sliding window can be obtained by simultaneously calculating and analyzing the above formulas. ;

[0122] in, The standard deviation of the scoring feature values ​​of the most recent n consecutive sliding windows is calculated based on the feature values ​​of the n consecutive sliding windows preceding the current time. It is a key indicator for measuring the stability of process quality. The higher the value, the more drastic the recent fluctuations in the processing process and the worse the stability. The average value of the rating feature values ​​of the most recent n consecutive sliding windows is calculated based on the feature values ​​of the n consecutive sliding windows before the current time. It represents the central trend of the recent quality level and serves as a benchmark reference value for stability in the formula. This is a preset baseline window size, an initial value set based on the specific production line cycle time and experience. For example, it can be set to an hourly output or a fixed batch size of 200 pieces. It serves as the baseline for adjusting the window size. This is the floor function, meaning it takes the largest integer not greater than the calculated result.

[0123] When the processing is stable, the ratio Approaching 1, at this point close to The system employs a large sliding window, which helps smooth out random fluctuations, avoids false alarms due to minor noise, and improves monitoring efficiency; when fluctuations occur in the process, the ratio Reduced, leading to The system automatically reduces the window size, increasing the frequency of quality assessment and trend analysis. This allows for faster and more sensitive detection of abnormal changes in the process, enabling early warning.

[0124] Through the above technical solution, this embodiment provides an adaptive aluminum profile processing quality monitoring method. The method introduces a dynamic window adjustment mechanism based on recent quality stability, achieving an intelligent balance between monitoring sensitivity and robustness. It overcomes the inherent defects of fixed window size, which may result in slow response when the process is stable and excessive sensitivity when the process fluctuates. This allows the entire early warning system to automatically optimize its monitoring strategy according to the actual state of the production process, improving not only the timeliness and accuracy of early warnings but also enhancing the system's practicality and intelligence in dealing with complex working conditions. It further solves the defects of the rigid single-point detection mode in the background technology, realizing a change from static detection to dynamic intelligent monitoring.

[0125] This application also discloses an aluminum profile processing quality inspection system based on multidimensional data analysis, referring to... Figure 2The system is applied to the aluminum profile processing quality inspection method based on multidimensional data analysis described in any one of the above-mentioned methods, and the system includes:

[0126] The data acquisition module is used to collect multi-source data during the processing through industrial cameras, laser sensors, and vibration sensors;

[0127] The comprehensive scoring calculation module is used to perform comprehensive analysis on the collected multi-source data to obtain the comprehensive quality score of the aluminum profile processed parts and to make quality judgments.

[0128] The sliding window feature extraction module is used to extract features from the comprehensive quality score sequence based on a preset sliding window to obtain score feature values;

[0129] The trend analysis module is used to perform trend analysis on the scoring feature value sequence of a continuous sliding window and generate early warnings of processing anomalies.

[0130] The dynamic window adjustment module is used to dynamically adjust the size of the sliding window based on the scoring feature value;

[0131] The early warning module is used to execute early warning signals.

[0132] Through the above technical solution, this embodiment provides an aluminum profile processing quality inspection system based on multidimensional data analysis. The system constructs a multi-level quality monitoring system from real-time judgment to process early warning through the collaborative work of data acquisition, comprehensive scoring calculation, sliding window feature extraction, trend analysis, dynamic window adjustment and early warning modules. It deeply integrates machine vision, precision measurement and vibration analysis, and introduces a sliding window-based dynamic trend monitoring and adaptive adjustment mechanism to achieve comprehensive perception, accurate evaluation and early warning of aluminum profile processing quality. It effectively solves the core defects of the background technology, such as one-dimensional detection bias, sampling inspection lag and inability to provide dynamic early warning, and significantly improves the intelligence level and foresight of quality control.

[0133] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A method for inspecting the processing quality of aluminum profiles based on multidimensional data analysis, characterized in that, The method includes: S1. Acquire image data of the cutting end face through an industrial camera, acquire cutting contour dimension data through a laser sensor, and acquire processing vibration signal data through a vibration sensor; S2. Perform comprehensive analysis on the collected multi-source data to obtain the comprehensive quality score of the current aluminum profile processed parts, and make quality judgment on the aluminum profile processed parts based on the comprehensive quality score. S3. Preset a sliding window with the number of workpieces as the unit, extract features from the comprehensive quality score sequence of all processed parts within the sliding window, and obtain the score feature value of the preset sliding window. S4. Perform trend analysis on the scoring feature value sequence of a preset number of continuous sliding windows. When the downward trend of the scoring feature value is detected to exceed the preset threshold, generate a processing anomaly warning.

2. The method for inspecting the processing quality of aluminum profiles based on multidimensional data analysis according to claim 1, characterized in that, In step S2, the process of determining the quality of aluminum profile processed parts based on the comprehensive quality score includes: The overall quality score is obtained through formula analysis and calculation. ; in, Scoring of surface defects in aluminum profiles. Scoring for dimensional deviations of aluminum profiles. Scoring for vibration signal spectrum analysis. , , The first weighting coefficient; Comprehensive quality score Compared with the preset non-compliance scoring threshold Perform a comparison; like If the aluminum profile is found to be defective, it will be marked and a defective signal will be generated. Conversely, if the aluminum profile is not found to be qualified, it will be marked and a qualified signal will be generated.

3. The method for inspecting the processing quality of aluminum profiles based on multidimensional data analysis according to claim 2, characterized in that, The process of obtaining the surface defect score for the aluminum profile includes: The surface defect score of aluminum profiles is obtained through analysis and calculation using the above formula. ; in, This is the constant value for the maximum score in the scoring system. This represents the total number of defects currently detected on the surface of the aluminum profile. Let be the type weight coefficient for the i-th defect. Let be the projected area of ​​the i-th defect. This represents the total surface area of ​​the aluminum profile.

4. The method for inspecting the processing quality of aluminum profiles based on multidimensional data analysis according to claim 3, characterized in that, The process for obtaining the aluminum profile dimensional deviation score includes: The aluminum profile dimensional deviation score is obtained through analysis and calculation using the above formula. ; in, The total number of key dimensional parameters being measured. For the j-th dimensional parameter, Let j be the theoretical design value of the j-th dimensional parameter. The weighting coefficient for the j-th dimension parameter. The maximum permissible composite relative deviation threshold. To take the maximum value within the parentheses.

5. The method for inspecting the processing quality of aluminum profiles based on multidimensional data analysis according to claim 4, characterized in that, The process of obtaining the vibration signal spectrum analysis score includes: The vibration signal spectrum analysis score is obtained through the above formula analysis and calculation. ; in, The number of characteristic frequency bands, Let be the amplitude of the current vibration signal in the kth characteristic frequency band. To reference the corresponding amplitude under normal conditions, The weighting coefficients for the k-th frequency band are... This is the maximum allowable amplitude deviation threshold.

6. The method for inspecting the processing quality of aluminum profiles based on multidimensional data analysis according to claim 5, characterized in that, In step S3, the process of obtaining the scoring feature values ​​of the preset sliding window includes: The scoring characteristic value of the preset sliding window is obtained by combining the above formulas and performing simultaneous analysis. ; in, The defect rate of aluminum profile workpieces within the sliding window. This represents the average of all comprehensive quality scores within the sliding window. To account for the volatility of the overall quality score within the sliding window, , , This is the second weighting coefficient. This represents the number of defective aluminum profile workpieces within the sliding window. This represents the total number of aluminum profile workpieces within the sliding window. The overall quality score for the l-th aluminum profile workpiece within the sliding window.

7. The method for inspecting the processing quality of aluminum profiles based on multidimensional data analysis according to claim 6, characterized in that, In step S4, the process of performing trend analysis on the scoring feature value sequence of a preset number of continuous sliding windows includes: The slope of the trend of the scoring feature value of a preset number of continuous sliding windows over time was obtained by analyzing and calculating using the above formula. ; in, The preset number of consecutive sliding windows, Number the time sequence of the sliding window sequence. For the sliding window time indicator, The rating feature value corresponding to the h-th sliding window; The slope of the trend of the scoring feature values ​​of a preset number of continuous sliding windows over time. Compared with the preset change threshold Perform a comparison; when If an abnormal downward trend in the quality of aluminum profiles is detected during the processing, a processing anomaly warning will be generated.

8. The method for inspecting the processing quality of aluminum profiles based on multidimensional data analysis according to claim 7, characterized in that, The method further includes: S5. Analyze the scoring feature values ​​of a preset number of continuous sliding windows and dynamically adjust the size of the sliding windows.

9. The method for inspecting the processing quality of aluminum profiles based on multidimensional data analysis according to claim 8, characterized in that, The dynamic adjustment process includes: The size of the sliding window can be obtained by simultaneously calculating and analyzing the above formulas. ; in, To predetermine the standard deviation of the rating feature values ​​of the most recent n consecutive sliding windows, This is the average of the rating feature values ​​of the most recent n consecutive sliding windows. The preset baseline window size, This is the floor function.

10. A quality inspection system for aluminum profile processing based on multidimensional data analysis, characterized in that, The system is applied to the aluminum profile processing quality inspection method based on multidimensional data analysis as described in any one of claims 1-9, and the system comprises: The data acquisition module is used to collect multi-source data during the processing through industrial cameras, laser sensors, and vibration sensors; The comprehensive scoring calculation module is used to perform comprehensive analysis on the collected multi-source data to obtain the comprehensive quality score of the aluminum profile processed parts and to make quality judgments. The sliding window feature extraction module is used to extract features from the comprehensive quality score sequence based on a preset sliding window to obtain score feature values; The trend analysis module is used to perform trend analysis on the scoring feature value sequence of a continuous sliding window and generate early warnings of processing anomalies. The dynamic window adjustment module is used to dynamically adjust the size of the sliding window based on the scoring feature value; The early warning module is used to execute early warning signals.

Images