An OCT laser welding data processing and visualization system and method based on C# and Python integration

Abstract

The application provides an OCT laser welding data processing and visualization system and method based on C# and Python integration, wherein the system comprises: a data acquisition module for acquiring original OCT data through C#; a data processing module for processing the original OCT data to obtain target data; a data analysis module for analyzing the target data to obtain depth statistical analysis; a visualization module for visualizing the target data to obtain multi-modal display; a scale processing module for performing scale processing according to the depth statistical analysis and the multi-modal display; an output control module for performing output control based on the scale processing result; and an interactive module for automatically generating a Python analysis script and executing based on algorithm parameters set by a user. The OCT laser welding data processing and visualization system and method based on C# and Python integration improve the continuity of data processing and the robustness of the whole system.

Description

Technical Field

[0001] This invention relates to the field of OCT measurement technology, and in particular to an OCT laser welding data processing and visualization system and method based on C# and Python integration. Background Technology

[0002] Optical coherence tomography (OCT) technology is increasingly used in laser welding processes. Its unique advantage lies in its ability to achieve high-resolution, non-contact, real-time detection of weld penetration, effectively ensuring welding quality. However, as industrial manufacturing demands ever higher levels of inspection efficiency and intelligence, existing OCT inspection systems still face numerous technical challenges in practical deployment.

[0003] First, the spectral data collected by OCT during welding is enormous. Data transmission and processing typically rely on traditional file interaction or static interfaces, resulting in low data exchange efficiency and impacting the response speed of real-time monitoring and analysis. Second, traditional systems primarily use fixed threshold methods for anomaly detection and filtering of depth data, which cannot flexibly adapt to dynamic changes under different welding process conditions, easily leading to misjudgments or missed detections. Furthermore, the data source is singular and lacks an intelligent switching mechanism; if cached data becomes abnormal or invalid, the system is prone to interruption or errors, thus reducing the robustness and reliability of the entire welding quality monitoring system.

[0004] In view of this, there is an urgent need for an OCT laser welding data processing and visualization system and method based on C# and Python integration, in order to at least solve the above-mentioned shortcomings. Summary of the Invention

[0005] One objective of this invention is to provide an OCT laser welding data processing and visualization system and method based on the integration of C# and Python. The system uses C# for real-time data acquisition and storage, and calls dynamically generated Python scripts to achieve efficient data analysis and result visualization. Simultaneously, it introduces an adaptive signal processing algorithm to perform LOF anomaly detection, Savitzky-Golay filtering, and Otsu adaptive threshold adjustment on depth data, effectively improving the stability and accuracy of the depth data. Furthermore, the system has an intelligent data source switching mechanism that automatically reverts to the original data when anomalies in cached data are detected, enabling priority judgment and intelligent switching of data sources, thereby ensuring the continuity of data processing and the overall robustness of the system.

[0006] This invention provides an OCT laser welding data processing and visualization system based on C# and Python integration, comprising:

[0007] The data acquisition module is used to acquire raw OCT data using C#.

[0008] The data processing module is used to process the raw OCT data based on an adaptive signal processing algorithm to obtain the target data;

[0009] The data analysis module is used to perform data analysis based on target data and obtain in-depth statistical analysis.

[0010] The visualization module is used to visualize target data and obtain multimodal displays;

[0011] The scale processing module is used to perform scale processing based on depth statistical analysis and multimodal display.

[0012] The output control module is used to control the output based on the scale processing results.

[0013] The interactive module is used to automatically generate and execute Python analysis scripts based on user-defined algorithm parameters.

[0014] Preferably, the data acquisition module performs the following operations:

[0015] The software and hardware are controlled using C# to acquire and store raw spectral data from the OCT system in real time.

[0016] The original spectral data is converted and preliminarily filtered to obtain the original OCT data;

[0017] The OCT system is configured with a data buffering mechanism.

[0018] Preferably, the data processing module performs the following operations:

[0019] A fixed threshold segmentation technique is used to identify the effective signal region in the raw OCT data;

[0020] The LOF algorithm is used to detect and filter out abnormal points in the effective signal area, and the DBSCAN density clustering algorithm is used to identify the location of the molten pool surface and the bottom of the weld.

[0021] Based on the location of the molten pool surface and the weld bottom, Savitzky-Golay filtering and moving average filtering are used to smooth the curves and obtain the target data, which includes: the molten pool depth profile curve and the reflectivity distribution curve.

[0022] Preferably, the data analysis module performs the following operations:

[0023] Calculate the real-time depth data between the surface of the molten pool and the bottom of the weld based on the target data;

[0024] Based on the depth statistics information from real-time depth data and the threshold monitoring mechanism, a warning is issued when the depth exceeds the preset threshold.

[0025] The curve smoothing coefficient is dynamically adjusted based on the welding condition.

[0026] Preferably, the visualization module performs the following operations:

[0027] Generates real-time dynamic display of the molten pool surface and weld bottom curves;

[0028] The depth distribution is displayed based on the depth heatmap mode, which supports the addition and customization of reference lines;

[0029] Generate depth data and display it synchronously with the timeline.

[0030] Preferably, the interaction module performs the following operations:

[0031] Based on historical algorithm parameter settings data and the interaction logic of algorithm parameter settings, a template is extracted to determine the closed-loop interaction logic sequence;

[0032] Based on the sequence distribution of adjacent and identical closed-loop interaction logic in the closed-loop interaction logic sequence, a standard local continuous sequence is obtained.

[0033] If successful, external factors are constructed based on the observed index values ​​after the closed-loop interaction logic interaction is completed at the last sequence position of the standard local continuous sequence.

[0034] Based on external factors, determine the target external factors;

[0035] Take appropriate action based on the external factors affecting the target.

[0036] Preferably, the preset conditions satisfied by the standard locally continuous sequence in the interaction module include:

[0037] The closed-loop interaction logic is the same in standard locally continuous sequences;

[0038] A standard locally continuous sequence is a continuous local subsequence of a closed-loop interactive logic sequence;

[0039] The closed-loop interaction logic for the last sequence position of a standard locally continuous sequence is the closed-loop interaction logic for the last sequence position of the closed-loop interaction logic sequence.

[0040] Preferably, the interaction module performs remote control according to the processing strategy, including:

[0041] Before remote control is implemented according to the processing strategy, the threshold monitoring parameters and the ideal future welding trajectory are determined at the current moment.

[0042] Simulation application processing strategies are used to determine the cost of modifying simulation early warning monitoring parameters and future ideal welding trajectories;

[0043] Based on the observed indicator values, the first impact cost is determined, and the first impact cost is the historical impact cost;

[0044] Based on the cost of modifying the simulated early warning monitoring parameters and the future ideal welding trajectory, the second impact cost is obtained. The second impact cost is the sum of the quality loss cost and the modification cost corresponding to the simulated early warning monitoring parameters.

[0045] If the cost of the first impact is less than the cost of the second impact, then remote control is abandoned; otherwise, the corresponding remote control is continued.

[0046] This invention provides a method for OCT laser welding data processing and visualization based on C# and Python integration, comprising:

[0047] Step 1: Collect raw OCT data using C#;

[0048] Step 2: Process the raw OCT data using an adaptive signal processing algorithm to obtain the target data;

[0049] Step 3: Perform data analysis based on the target data to obtain in-depth statistical analysis;

[0050] Step 4: Visualize the target data to obtain a multimodal display;

[0051] Step 5: Perform scale processing based on depth statistical analysis and multimodal display;

[0052] Step 6: Perform output control based on the ruler processing results;

[0053] Step 7: Automatically generate and execute a Python analysis script based on the algorithm parameters set by the user.

[0054] The beneficial effects of this invention are as follows:

[0055] This invention uses C# for real-time data acquisition and storage, and calls dynamically generated Python scripts to achieve efficient data analysis and result visualization. It also introduces an adaptive signal processing algorithm to perform LOF anomaly detection, Savitzky-Golay filtering, and Otsu adaptive threshold adjustment on depth data, effectively improving the stability and accuracy of the depth data. Furthermore, the system has an intelligent data source switching mechanism that automatically reverts to the original data when cached data anomalies are detected, enabling priority judgment and intelligent switching of data sources, thereby ensuring the continuity of data processing and the overall robustness of the system.

[0056] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in this application.

[0057] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0058] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0059] Figure 1 This is a schematic diagram of an OCT laser welding data processing and visualization system based on C# and Python integration, as described in an embodiment of the present invention.

[0060] Figure 2 This is a schematic diagram of the interaction logic extraction template in an embodiment of the present invention;

[0061] Figure 3 This is a schematic diagram of an OCT laser welding data processing and visualization method based on C# and Python integration in an embodiment of the present invention. Detailed Implementation

[0062] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0063] This invention provides an OCT laser welding data processing and visualization system based on the integration of C# and Python, such as... Figure 1 As shown, it includes:

[0064] Data acquisition module 1 is used to acquire raw OCT data using C#.

[0065] The data acquisition module 1 performs the following operations:

[0066] The software and hardware are controlled using C# to acquire and store raw spectral data from the OCT system in real time.

[0067] The raw spectral data undergoes format conversion and preliminary filtering to obtain raw OCT data; the OCT system is configured with a data buffering mechanism.

[0068] Among them, hardware and software control via C# refers to using the C# programming language to build the hardware and software co-operation system of the OCT system and using C# programming language commands for control;

[0069] Among these processes, format conversion is performed, for example, from bytes to floating-point numbers; preliminary filtering involves preprocessing the original spectrum in the frequency or time domain to suppress noise and enhance the signal.

[0070] Data processing module 2 is used to process the raw OCT data based on an adaptive signal processing algorithm to obtain the target data;

[0071] The data processing module 2 performs the following operations:

[0072] A fixed threshold segmentation technique is used to identify the effective signal region in the raw OCT data;

[0073] The LOF algorithm is used to detect and filter out abnormal points in the effective signal area, and the DBSCAN density clustering algorithm is used to identify the location of the molten pool surface and the bottom of the weld.

[0074] Based on the location of the molten pool surface and the bottom of the weld, Savitzky-Golay filtering and moving average filtering are used to smooth the curve and obtain the target data, which includes: the molten pool depth profile curve and the reflectivity distribution curve.

[0075] Data analysis module 3 is used to perform data analysis based on target data to obtain in-depth statistical analysis;

[0076] The data analysis module 3 performs the following operations:

[0077] Calculate the real-time depth data between the surface of the molten pool and the bottom of the weld based on the target data;

[0078] Based on the depth statistics information from real-time depth data and the threshold monitoring mechanism, a warning is issued when the depth exceeds the preset threshold.

[0079] The curve smoothing coefficient is dynamically adjusted according to the welding condition.

[0080] The in-depth statistical information includes: maximum value, minimum value, average value, and median value;

[0081] Visualization module 4 is used to visualize the target data and obtain multimodal displays;

[0082] The visualization module 4 performs the following operations:

[0083] Generates real-time dynamic display of the molten pool surface and weld bottom curves;

[0084] The depth distribution is displayed based on the depth heatmap mode, which supports the addition and customization of reference lines;

[0085] Generate depth data and display it synchronously with the timeline;

[0086] Among them, reference lines are, for example: threshold lines, average depth lines;

[0087] The scale processing module 5 is used to perform scale processing based on depth statistical analysis and multimodal display.

[0088] Output control module 6 is used to control the output based on the ruler processing results;

[0089] Interactive module 7 is used to automatically generate and execute Python analysis scripts based on user-defined algorithm parameters;

[0090] The algorithm parameters include: filter window size, clustering parameters, and threshold.

[0091] The working principle and beneficial effects of the above technical solution are as follows:

[0092] The operation method of the OCT laser welding data processing and visualization system based on C# and Python integration includes:

[0093] OCT data acquisition and preprocessing steps: The OCT system acquires and stores raw spectral data in real time; the spectral data undergoes format conversion (byte to floating point) and preliminary filtering; a data buffering mechanism is implemented to support historical data backtracking and anomaly recovery.

[0094] Signal processing and depth extraction steps: Use fixed threshold segmentation to identify valid signal regions; apply the LOF algorithm for outlier detection and filtering; use the DBSCAN density clustering algorithm to identify surface and bottom locations; use Savitzky-Golay filtering and moving average filtering to smooth curves.

[0095] Depth calculation and statistical analysis steps: Calculate the distance between the surface and the bottom to obtain real-time depth data; generate depth statistics (maximum, minimum, average, median); implement a threshold monitoring mechanism to issue a warning when the depth exceeds the preset threshold; support an adaptive depth processing algorithm to dynamically adjust the smoothing coefficient according to the welding status.

[0096] Multimodal visualization steps: generate real-time dynamic display of surface and bottom curves; provide a depth heatmap mode to intuitively display depth distribution; support the addition and customization of various reference lines (threshold lines, average depth lines); realize synchronous display of depth data on the time axis for easy process analysis.

[0097] Interactive script generation steps: Automatically generate Python analysis scripts based on user-defined parameters; support customization of key parameters (filter window size, clustering parameters, threshold); provide a script template replacement mechanism to ensure the consistency and reliability of generated scripts; and provide real-time feedback on script execution results for easy parameter tuning.

[0098] Data storage and export steps: Supports saving raw data, processed data, and analysis results in multiple formats; provides in-depth data export functions in formats such as CSV and Excel; achieves high-quality saving of image results, supporting multiple formats such as PNG and JPEG; includes automatic naming and batch saving functions for easy management of large amounts of data.

[0099] System optimization and performance improvement steps: Employ multi-threaded parallel processing technology to separate data acquisition, processing, and display processes; optimize memory management to reduce data copying and transmission overhead; provide GPU acceleration options to speed up the processing of large-scale data; support real-time adjustment of processing parameters to ensure stable system operation under different operating conditions.

[0100] Integrated control and feedback steps: It establishes a communication interface with the laser welding system to support real-time feedback of depth data; it provides an automated control interface to adjust welding parameters based on depth monitoring results; it implements a multi-level alarm mechanism to provide graded alerts for abnormal depths and system faults; and it supports remote monitoring and data sharing for collaborative analysis across multiple terminals.

[0101] This invention utilizes C# for real-time data acquisition and storage, and calls dynamically generated Python scripts to achieve efficient data analysis and result visualization. It also introduces an adaptive signal processing algorithm to perform LOF anomaly detection, Savitzky-Golay filtering, and Otsu adaptive threshold adjustment on depth data, effectively improving the stability and accuracy of the depth data. Furthermore, the system features an intelligent data source switching mechanism that automatically reverts to the original data when cached data anomalies are detected, enabling priority judgment and intelligent switching of data sources, thereby ensuring the continuity of data processing and the overall robustness of the system. This invention not only achieves efficient cross-platform data processing collaboration but also achieves significant improvements in anomaly detection, filtering, and adaptive adjustment of welding depth data, greatly enhancing the real-time performance, accuracy, and reliability of welding quality monitoring systems and improving the quality control precision in industrial laser welding processes.

[0102] In one embodiment, the interaction module performs the following operations:

[0103] Based on historical algorithm parameter settings data and the interaction logic of algorithm parameter settings, a template is extracted to determine the closed-loop interaction logic sequence;

[0104] The historical algorithm parameter setting data includes records of various algorithm-related parameter adjustments made historically, such as adjustments to DBSCAN parameters, LOF thresholds, and filtering parameters based on observed indicators. The interaction logic extraction template serves as a template for generating closed-loop interaction logic for algorithm settings by comparing historical algorithm parameter setting records. The interaction logic extraction template is as follows: Figure 2As shown; the closed-loop interaction logic is, for example: showing the staff the visualization interface → the staff view the observation indicators from the visualization interface → the staff sees the abnormal observation indicators (e.g., blurred melt pool boundary) → adjusting the DBSCAN parameters → reprocessing the data → updating the visualization → showing the visualization interface to the staff, which constitutes a logical closed loop; the closed-loop interaction logic sequence is obtained by arranging the closed-loop interaction logic according to the chronological order of the setting time of the corresponding extracted historical algorithm parameter settings data;

[0105] Based on the sequence distribution of adjacent and identical closed-loop interaction logic in the closed-loop interaction logic sequence, a standard local continuous sequence is obtained.

[0106] The sequence distribution is as follows: which closed-loop interaction logic is adjacent to and the same as which closed-loop interaction logic in the closed-loop interaction logic sequence; the standard local continuous sequence meets the preset conditions, which are: the closed-loop interaction logic in the standard local continuous sequence is the same and belongs to a continuous local subsequence of the closed-loop interaction logic sequence, and the closed-loop interaction logic at the last sequence position of the standard local continuous sequence is the closed-loop interaction logic at the last sequence position of the closed-loop interaction logic sequence.

[0107] If successful, external factors are constructed based on the observed index values ​​after the closed-loop interaction logic interaction is completed at the last sequence position of the standard local continuous sequence.

[0108] Among them, the external factor is the descriptive vector of the abnormal observation index, which is constructed according to the preset descriptive vector construction rules;

[0109] Based on external factors, determine the target external factors;

[0110] In determining the target external factor, the historical external factor basis factors in the historical invalid parameter adjustment cases are matched, and the historical external factors in the corresponding historical invalid parameter adjustment cases are determined as the target external factor.

[0111] Take appropriate action based on the external factors affecting the target.

[0112] When processing external factors, it determines whether the processing strategy is remotely controllable. If so, it performs remote control based on the processing strategy (e.g., remotely increasing laser power or welding speed). Otherwise, it sends the external factors to the staff's terminal device for reminder.

[0113] The working principle and beneficial effects of the above technical solution are as follows:

[0114] When the visualization interface is not performing well, staff will consider resetting the algorithm parameters to obtain a more accurate and effective visualization. However, the observation indicators in the visualization interface (such as the molten pool boundary and outliers) are affected not only by the algorithm parameter settings themselves, but also by other external factors. For example, besides unreasonable DBSCAN parameter settings, a blurred molten pool boundary may also be caused by insufficient laser power or excessive welding speed. For instance, if the laser power is too low, the temperature of the molten pool is insufficient to completely melt the material, resulting in an unclear molten pool boundary. Conversely, if the welding speed is too fast, the molten pool will cool down too quickly, and the molten pool boundary may not be fully formed, thus appearing as a blurred boundary in the image.

[0115] When existing technologies encounter such situations, they cannot effectively identify invalid parameter settings. It is still necessary to manually determine the external factors influencing abnormal indicators. The accuracy of this determination is highly dependent on the experience of the person making the determination and is not intelligent enough.

[0116] Therefore, this invention introduces an interactive logic extraction template for algorithm parameter settings to determine the closed-loop interactive logic sequence. When a standard local continuous sequence that meets preset conditions exists in the closed-loop interactive logic sequence, it indicates that the operator is facing this situation (a situation where external factors other than the algorithm parameter settings themselves affect the observed indicators). For example, if the operator adjusts the DBSCAN parameters three times consecutively, it means that the problem of blurred melt pool boundaries has not been improved despite multiple adjustments to the algorithm parameters. In this case, considering external influencing factors, a descriptive vector (external factor basis factor) for the abnormal observed indicator is constructed based on the observed indicator values ​​after continuous improvement. This external factor basis factor is then matched with historical external factor basis factors in historical invalid parameter adjustment situations. The historical external factor corresponding to the vector matching historical external factor basis factor is determined as the target external factor. Based on the target external factor, corresponding processing is performed. This avoids invalid parameter adjustments, reduces trial and error costs, transforms user interaction data into executable process knowledge, dynamically compensates for the impact of material and environmental changes, and is more intelligent.

[0117] In one embodiment, the interaction module performs remote control according to a processing strategy, including:

[0118] Before remote control is implemented according to the processing strategy, the threshold monitoring parameters and the ideal future welding trajectory are determined at the current moment.

[0119] Among them, the threshold monitoring associated parameters are: parameter values ​​of the relevant parameter types for which the warning threshold is set at the current time, such as: the current melt pool depth is 50μm;

[0120] Simulation application processing strategies are used to determine the cost of modifying simulation early warning monitoring parameters and future ideal welding trajectories;

[0121] Among them, the simulated early warning monitoring parameters are: the values ​​of the threshold monitoring related parameters after the simulation application of the processing strategy, which exceed their set early warning thresholds. For example, after simulating the increase of laser power, the predicted molten pool depth exceeds the molten pool depth early warning threshold (e.g., 60μm) and reaches 62μm; the modification cost of the future ideal welding trajectory is: the preset cost value of the difference between the length of the modified trajectory and the length of the future ideal welding trajectory. The larger the difference in length, the higher the modification cost.

[0122] Based on the observed indicator values, the first impact cost is determined, and the first impact cost is the historical impact cost;

[0123] The historical impact cost is the sum of the losses caused by historically neglecting the observed indicator values; the extent of the losses from the consequential events is assessed by staff.

[0124] Based on the cost of modifying the simulated early warning monitoring parameters and the future ideal welding trajectory, the second impact cost is obtained. The second impact cost is the sum of the quality loss cost and the modification cost corresponding to the simulated early warning monitoring parameters.

[0125] Among them, the more the simulated early warning monitoring parameters exceed their early warning thresholds, the greater the corresponding quality loss cost.

[0126] If the cost of the first impact is less than the cost of the second impact, then remote control is abandoned; otherwise, the corresponding remote control is continued.

[0127] The working principle and beneficial effects of the above technical solution are as follows:

[0128] When user interaction data is transformed into executable process knowledge, the execution result may conflict with the basic welding condition constraints. For example, the operator may intend to view a clearer boundary of the molten pool and monitor it together with the machine. However, the interaction data may identify that the user has adjusted the DBSCAN parameters three times, triggering an increase in laser power. At the same time, the machine has a higher recognition accuracy than the human eye and can accurately determine the boundary of the molten pool. If the control is blindly based on the feedback of the operator's interaction data, it is very easy to cause a deviation from the basic welding conditions, resulting in more serious consequences than the operator not being able to view the boundary of the molten pool, such as causing the welded part to break down.

[0129] Therefore, before formally implementing the processing strategy, this invention determines the first impact cost of historical consequences caused by ignoring observed index values ​​based on the observed index values, and simulates the application of the processing strategy to determine the second impact cost of the applied processing strategy. If the first impact cost is less than the second impact cost, remote control is abandoned; otherwise, the corresponding remote control continues to be executed. This achieves a control balance between feedback control of user interaction data on the control system and feedback control of basic welding condition constraints on the control system, making laser control more suitable.

[0130] This invention provides a method for OCT laser welding data processing and visualization based on the integration of C# and Python, such as... Figure 3 As shown, it includes:

[0131] Step 1: Collect raw OCT data using C#;

[0132] Step 2: Process the raw OCT data using an adaptive signal processing algorithm to obtain the target data;

[0133] Step 3: Perform data analysis based on the target data to obtain in-depth statistical analysis;

[0134] Step 4: Visualize the target data to obtain a multimodal display;

[0135] Step 5: Perform scale processing based on depth statistical analysis and multimodal display;

[0136] Step 6: Perform output control based on the ruler processing results;

[0137] Step 7: Automatically generate and execute a Python analysis script based on the algorithm parameters set by the user;

[0138] Step 7: Automatically generate and execute a Python analysis script based on the user-defined algorithm parameters, including:

[0139] Based on historical algorithm parameter settings data and the interaction logic of algorithm parameter settings, a template is extracted to determine the closed-loop interaction logic sequence;

[0140] Based on the sequence distribution of adjacent and identical closed-loop interaction logic in the closed-loop interaction logic sequence, a standard local continuous sequence is obtained.

[0141] If successful, external factors are constructed based on the observed index values ​​after the closed-loop interaction logic interaction is completed at the last sequence position of the standard local continuous sequence.

[0142] Based on external factors, determine the target external factors;

[0143] Take appropriate action based on the external factors affecting the target;

[0144] The preset conditions that a standard locally continuous sequence must satisfy include:

[0145] The closed-loop interaction logic is the same in standard locally continuous sequences;

[0146] A standard locally continuous sequence is a continuous local subsequence of a closed-loop interactive logic sequence;

[0147] The closed-loop interaction logic of the last sequence position of the standard local continuous sequence is the closed-loop interaction logic of the last sequence position of the closed-loop interaction logic sequence.

[0148] The process of processing according to the target external factors includes:

[0149] Determine the processing strategy based on the target external factors, and determine whether the processing strategy can be handled by remote control.

[0150] If so, then remote control will be implemented according to the processing strategy;

[0151] Otherwise, the terminal device of the staff will send a reminder about the target external factor;

[0152] The remote control based on the processing strategy includes:

[0153] Before remote control is implemented according to the processing strategy, the threshold monitoring parameters and the ideal future welding trajectory are determined at the current moment.

[0154] Simulation application processing strategies are used to determine the cost of modifying simulation early warning monitoring parameters and future ideal welding trajectories;

[0155] Based on the observed indicator values, the first impact cost is determined, and the first impact cost is the historical impact cost;

[0156] Based on the cost of modifying the simulated early warning monitoring parameters and the future ideal welding trajectory, the second impact cost is obtained. The second impact cost is the sum of the quality loss cost and the modification cost corresponding to the simulated early warning monitoring parameters.

[0157] If the cost of the first impact is less than the cost of the second impact, then remote control is abandoned; otherwise, the corresponding remote control is continued.

[0158] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A C# and Python integrated OCT laser welding data processing and visualization system, characterized in that, include: The data acquisition module is used to acquire raw OCT data using C#. The data processing module is used to process the raw OCT data based on an adaptive signal processing algorithm to obtain the target data; The data analysis module is used to perform data analysis based on target data and obtain in-depth statistical analysis. The visualization module is used to visualize target data and obtain multimodal displays; The scale processing module is used to perform scale processing based on depth statistical analysis and multimodal display. The output control module is used to control the output based on the scale processing results. The interactive module is used to automatically generate and execute Python analysis scripts based on user-defined algorithm parameters. The data processing module performs the following operations: A fixed threshold segmentation technique is used to identify the effective signal region in the raw OCT data; The LOF algorithm is used to detect and filter out abnormal points in the effective signal area, and the DBSCAN density clustering algorithm is used to identify the location of the molten pool surface and the bottom of the weld. Based on the location of the molten pool surface and the bottom of the weld, Savitzky-Golay filtering and moving average filtering are used to smooth the curve and obtain the target data, which includes: the molten pool depth profile curve and the reflectivity distribution curve. The interaction module performs the following operations: Based on historical algorithm parameter settings data and the interaction logic of algorithm parameter settings, a template is extracted to determine the closed-loop interaction logic sequence; Based on the sequence distribution of adjacent and identical closed-loop interaction logic in the closed-loop interaction logic sequence, a standard local continuous sequence is obtained. If successful, construct the external factor based on the observation index value after the closed-loop interaction logic interaction is completed at the last sequence position of the standard local continuous sequence; where the external factor is the description vector of the abnormal observation index. Based on the external factors, the target external factors are determined; when determining the target external factors, the historical external factors in the historical invalid parameter adjustment cases are matched, and the historical external factors in the corresponding historical invalid parameter adjustment cases are determined as the target external factors. Take appropriate action based on the external factors affecting the target.

2. The OCT laser welding data processing and visualization system based on C# and Python integration as described in claim 1, characterized in that, The data acquisition module performs the following operations: The software and hardware are controlled using C# to acquire and store raw spectral data from the OCT system in real time. The original spectral data is converted and preliminarily filtered to obtain the original OCT data; The OCT system is configured with a data buffering mechanism.

3. The OCT laser welding data processing and visualization system based on C# and Python integration as described in claim 1, characterized in that, The data analysis module performs the following operations: Calculate the real-time depth data between the surface of the molten pool and the bottom of the weld based on the target data; Based on the depth statistics information from real-time depth data and the threshold monitoring mechanism, a warning is issued when the depth exceeds the preset threshold. The curve smoothing coefficient is dynamically adjusted based on the welding condition.

4. The OCT laser welding data processing and visualization system based on C# and Python integration as described in claim 1, characterized in that, The visualization module performs the following operations: Generates real-time dynamic display of the molten pool surface and weld bottom curves; The depth distribution is displayed based on the depth heatmap mode, which supports the addition and customization of reference lines; Generate depth data and display it synchronously with the timeline.

5. The OCT laser welding data processing and visualization system based on C# and Python integration as described in claim 1, characterized in that, The preset conditions that the standard locally continuous sequences in the interactive module must satisfy include: The closed-loop interaction logic is the same in standard locally continuous sequences; A standard locally continuous sequence is a continuous local subsequence of a closed-loop interactive logic sequence; The closed-loop interaction logic for the last sequence position of a standard locally continuous sequence is the closed-loop interaction logic for the last sequence position of the closed-loop interaction logic sequence.

6. The OCT laser welding data processing and visualization system based on C# and Python integration as described in claim 1, characterized in that, The interaction module performs corresponding processing based on the target external factors, including: Determine the processing strategy based on the target external factors, and determine whether the processing strategy can be handled by remote control. If so, then remote control will be implemented according to the processing strategy; Otherwise, the terminal device of the staff will send a reminder about the target external factor.

7. The OCT laser welding data processing and visualization system based on C# and Python integration as described in claim 6, characterized in that, The interaction module performs remote control based on the processing strategy, including: Before remote control is implemented according to the processing strategy, the threshold monitoring parameters and the ideal future welding trajectory are determined at the current moment. Simulation application processing strategies are used to determine the cost of modifying simulation early warning monitoring parameters and future ideal welding trajectories; Based on the observed indicator values, the first impact cost is determined, and the first impact cost is the historical impact cost; Based on the cost of modifying the simulated early warning monitoring parameters and the future ideal welding trajectory, the second impact cost is obtained. The second impact cost is the sum of the quality loss cost and the modification cost corresponding to the simulated early warning monitoring parameters. If the cost of the first impact is less than the cost of the second impact, then remote control is abandoned; otherwise, the corresponding remote control is continued.

8. A method for OCT laser welding data processing and visualization based on the integration of C# and Python, characterized in that, include: Step 1: Collect raw OCT data using C#; Step 2: Process the raw OCT data using an adaptive signal processing algorithm to obtain the target data; Step 3: Perform data analysis based on the target data to obtain in-depth statistical analysis; Step 4: Visualize the target data to obtain a multimodal display; Step 5: Perform scale processing based on depth statistical analysis and multimodal display; Step 6: Perform output control based on the ruler processing results; Step 7: Automatically generate and execute a Python analysis script based on the algorithm parameters set by the user; Step 2 involves processing the original OCT data using an adaptive signal processing algorithm to obtain the target data, including: A fixed threshold segmentation technique is used to identify the effective signal region in the raw OCT data; The LOF algorithm is used to detect and filter out abnormal points in the effective signal area, and the DBSCAN density clustering algorithm is used to identify the location of the molten pool surface and the bottom of the weld. Based on the location of the molten pool surface and the bottom of the weld, Savitzky-Golay filtering and moving average filtering are used to smooth the curve and obtain the target data, which includes: the molten pool depth profile curve and the reflectivity distribution curve. Step 7: Automatically generate and execute a Python analysis script based on the user-defined algorithm parameters, including: Based on historical algorithm parameter settings data and the interaction logic of algorithm parameter settings, a template is extracted to determine the closed-loop interaction logic sequence; Based on the sequence distribution of adjacent and identical closed-loop interaction logic in the closed-loop interaction logic sequence, a standard local continuous sequence is obtained. If successful, construct the external factor based on the observation index value after the closed-loop interaction logic interaction is completed at the last sequence position of the standard local continuous sequence; where the external factor is the description vector of the abnormal observation index. Based on the external factors, the target external factors are determined; when determining the target external factors, the historical external factors in the historical invalid parameter adjustment cases are matched, and the historical external factors in the corresponding historical invalid parameter adjustment cases are determined as the target external factors. Take appropriate action based on the external factors affecting the target.

Images