General intelligent oscilloscope system based on integration of LabVIEW and AI and data acquisition and analysis method

Abstract

The application provides a general intelligent oscilloscope system based on LabVIEW and AI integration, which comprises: a hardware layer for collecting original waveform data, serving as a general digital oscilloscope; a data acquisition layer for uploading the original waveform data to an upper computer through a VISA communication interface; a LabVIEW control layer deployed on the upper computer, used for receiving the original waveform data and completing preprocessing; an AI analysis layer integrating a lightweight machine learning model, used for receiving the preprocessed waveform data, automatically performing waveform classification, feature parameter extraction and waveform signal quality evaluation; a result output layer for superimposing and visualizing AI analysis results and original waveforms; and a Python script calling node between the LabVIEW control layer and the AI analysis layer, used for completing data interaction and model reasoning. The application realizes data interaction and model reasoning through the Python script calling node, so that the general digital oscilloscope can complete waveform classification, feature extraction and signal quality evaluation in real time without hardware modification, and the ordinary PC can complete waveform classification, feature extraction and signal quality evaluation in real time without relying on GPU, the equipment investment and deployment threshold are greatly reduced, millisecond-level real-time response is realized, and the automation degree and efficiency of waveform analysis are improved.

Description

Technical Field

[0001] This invention relates to the field of oscilloscope technology, and in particular to a general-purpose intelligent oscilloscope system and a data acquisition and analysis method. Background Technology

[0002] Oscilloscopes are fundamental instruments for observing electrical signal waveforms and analyzing signal characteristics in electronic measurements, physics experiments, and scientific research testing. Currently, while commercially available general-purpose digital oscilloscopes can achieve high-fidelity acquisition and basic display of raw waveforms, tasks such as signal filtering, feature parameter extraction, waveform classification, anomaly detection, and quality assessment heavily rely on manual experience. Testers must manually read waveform data, use external software for post-processing, or rely on visual judgment of signal quality, resulting in low data processing efficiency, large human errors, and an inability to achieve real-time online intelligent diagnostics. Especially in university batch experiments or continuous production line testing scenarios, manually analyzing waveforms one by one has become a bottleneck to efficiency.

[0003] On the other hand, the few existing oscilloscopes with intelligent analysis capabilities typically employ dedicated hardware architectures (such as built-in FPGAs or embedded AI chips) and customized algorithms. This results in high equipment purchase and maintenance costs, complex deployment processes, and incompatibility with the large number of general-purpose oscilloscopes already in the laboratory. Furthermore, while some solutions attempt to combine LabVIEW with AI models, these often require additional computing servers or high-performance GPUs, lacking sufficient lightweight design and making real-time operation on ordinary PCs difficult, thus limiting their versatility and practicality.

[0004] Therefore, there is an urgent need for a solution that is compatible with general-purpose oscilloscopes and does not require additional computing hardware. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention proposes a universal intelligent oscilloscope system and data acquisition and analysis method based on LabVIEW and AI integration, which solves the problem of limited universality and practicality of existing oscilloscopes with intelligent analysis functions.

[0006] To achieve the above objectives, the technical solution of the present invention is implemented as follows:

[0007] A general-purpose intelligent oscilloscope system based on LabVIEW and AI integration includes: a hardware layer, a general-purpose digital oscilloscope for acquiring raw waveform data; a data acquisition layer, which uploads the raw waveform data to a host computer via a VISA communication interface; a LabVIEW control layer, deployed on the host computer, for receiving the raw waveform data and performing preprocessing; an AI analysis layer, integrating a lightweight machine learning model, receiving the preprocessed waveform data, and automatically performing waveform classification, feature parameter extraction, and waveform signal quality assessment; and a result output layer, which overlays the AI ​​analysis results with the raw waveform for visualization. The LabVIEW control layer and the AI ​​analysis layer interact and perform model inference via a Python script call node. This invention achieves data interaction and model inference through a Python script call node between the LabVIEW control layer and the AI ​​analysis layer, enabling general-purpose digital oscilloscopes to perform waveform classification, feature extraction, and signal quality assessment in real time without hardware modification and ordinary PCs without relying on GPUs. This significantly reduces equipment investment and deployment barriers, while achieving millisecond-level real-time response and improving the automation and efficiency of waveform analysis.

[0008] Furthermore, in order to ensure real-time inference while avoiding resource waste caused by repeatedly loading models, the AI ​​analysis layer uses a pre-trained convolutional neural network (CNN) model or a support vector machine (SVM) model. The lightweight model is stored locally as a file and is loaded once by the LabVIEW control layer when the program starts and remains resident in memory.

[0009] Furthermore, in order to reduce environmental noise interference and match the waveform data format with the input requirements of the AI ​​model, the data preprocessing submodule built into the LabVIEW control layer includes DC removal, digital filtering, and downsampling functions to convert the raw waveform data into a standard format that is compatible with the AI ​​model input.

[0010] Furthermore, to enable users to flexibly replace AI models for different experimental scenarios without modifying the LabVIEW main program, the LabVIEW control layer is configured with a custom model import interface, supporting users to import custom AI models trained for different experimental scenarios.

[0011] Furthermore, in order to quantitatively assess the availability of waveform data and automatically identify abnormal transitions and baseline shifts, waveform signal quality assessment includes abnormal transition identification, baseline shift discrimination, and signal validity determination, and outputs anomaly scores and confidence levels.

[0012] A data acquisition and analysis method for a general-purpose intelligent oscilloscope system based on LabVIEW and AI integration includes the following steps:

[0013] S1: The LabVIEW control layer establishes communication with the general-purpose digital oscilloscope via the VISA interface, configures the acquisition parameters, and starts acquisition. S2: The general-purpose digital oscilloscope acquires raw waveform data and uploads it to the LabVIEW control layer in real time. S3: The LabVIEW control layer performs DC removal, filtering, and downsampling preprocessing on the received raw waveform data. S4: The LabVIEW control layer calls the lightweight machine learning model of the AI ​​analysis layer via a Python script node, inputting the preprocessed data into the model to complete waveform classification, feature parameter extraction, and signal quality assessment. S5: The results output layer overlays the AI ​​analysis results with the raw waveform and automatically generates a structured analysis report. S6: When the AI ​​analysis layer determines that the signal quality does not meet preset requirements, the LabVIEW control layer automatically adjusts the oscilloscope's sampling rate or trigger threshold and reacquires the data. The automatic acquisition and analysis method of the present invention sets up automatic signal quality judgment and closed-loop adjustment steps. When the signal-to-noise ratio is too low or the trigger is unstable, it actively adjusts the sampling rate and trigger threshold and re-acquires data, forming a closed-loop control. This effectively avoids the problem of invalid data caused by improper acquisition parameters, greatly improves the acquisition success rate and data consistency of batch testing, and at the same time takes into account the flexibility of manual adjustment by users.

[0014] Furthermore, in order to clarify the triggering conditions for automatic closed-loop adjustment and avoid invalid resampling, situations in which the signal quality does not meet the preset requirements include a signal-to-noise ratio lower than the set threshold or an unstable triggering state.

[0015] Furthermore, to facilitate users in archiving and reviewing the analysis results, the structured analysis report in step S5 is in PDF format, automatically recording the acquisition parameters, waveform feature data, signal quality assessment conclusions, and trend prediction information.

[0016] Furthermore, in order to balance the flexibility of automation and user intervention and meet the needs of manual verification and secondary analysis in scientific research scenarios, users can manually adjust the acquisition parameters or preprocessing parameters through the LabVIEW control layer and export waveform data and analysis reports.

[0017] Furthermore, the method is applied to at least one of the following experimental scenarios: dynamic hysteresis loop measurement of ferromagnetic materials, sound velocity measurement, study of the resonance characteristics of RLC series circuits, and Fourier decomposition and synthesis experiments of periodic electrical signals.

[0018] The beneficial effects of this invention are:

[0019] 1. This invention uses a general-purpose digital oscilloscope as the hardware foundation, requiring no hardware modification to the oscilloscope, nor the purchase of a dedicated smart oscilloscope or external computing server. It can achieve intelligent waveform analysis simply by integrating a lightweight AI model through the host computer LabVIEW environment, which significantly reduces the equipment investment cost in university laboratories and general scientific research testing scenarios.

[0020] 2. This invention directly calls a locally stored lightweight machine learning model through the built-in Python script node in LabVIEW. The model is loaded once when the program starts and remains resident in memory. The same model is reused for each subsequent frame of data. Combined with the direct memory connection data interaction method, the delay caused by hard disk read / write and repeated loading is avoided, and millisecond-level real-time inference response is achieved.

[0021] 3. This invention deeply integrates the LabVIEW control layer with the AI ​​analysis layer. After waveform data acquisition, it automatically completes preprocessing such as DC removal, digital filtering, downsampling, and outlier removal, and directly sends the data to the AI ​​model for waveform classification, feature parameter extraction, and signal quality assessment. No manual intervention or offline post-processing is required, which greatly improves the efficiency of signal analysis.

[0022] 4. This invention can automatically output a structured analysis report containing waveform feature parameters, signal quality assessment conclusions, and trend predictions. At the same time, it overlays the AI ​​analysis results with the original waveform for visualization, enabling testers to intuitively obtain analysis conclusions and reducing errors and workload of manual interpretation.

[0023] 5. When the AI ​​analysis layer determines that the signal quality does not meet the preset requirements, the LabVIEW control layer can automatically adjust the oscilloscope's sampling rate, trigger threshold, or trigger coupling mode and re-acquire the signal, forming a closed-loop control until the signal quality meets the requirements, which significantly improves the acquisition success rate and data consistency in batch testing.

[0024] 6. This invention supports the import of user-defined AI models. Users can load targeted special models according to different experimental scenarios (such as hysteresis loop measurement of ferromagnetic materials, RLC resonant circuit testing, optical signal detection, etc.) without modifying the LabVIEW main program, and has good scene adaptability and scalability.

[0025] 7. The lightweight AI model and data format conversion used in this invention are all completed by the local CPU of the host computer, without the need for GPU or dedicated acceleration hardware, which lowers the deployment threshold and can run stably on ordinary PCs, making it easy to promote and use in low-cost scientific research and teaching environments.

[0026] 8. This invention automates the entire process from oscilloscope communication configuration, waveform acquisition, data preprocessing, AI intelligent analysis, result visualization to report output and closed-loop parameter adjustment, simplifying the operation steps for testers and reducing the probability of human error. Detailed Implementation

[0027] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0028] Example 1

[0029] The aforementioned general-purpose intelligent oscilloscope system based on LabVIEW and AI integration enables automatic acquisition of waveform data, real-time intelligent analysis, and result visualization without modifying existing general-purpose digital oscilloscope hardware or relying on additional computing servers or dedicated GPUs.

[0030] The system comprises a hardware layer, a data acquisition layer, a LabVIEW control layer, an AI analysis layer, and a results output layer. The hardware layer uses a general-purpose digital oscilloscope, such as the commercially available GDS-1102B model, which connects to the host computer via its standard VISA interface. The oscilloscope is only responsible for acquiring the raw voltage or current waveforms of the target under test and does not perform any onboard intelligent analysis. The data acquisition layer establishes a data transmission link between the oscilloscope and the host computer based on the VISA communication protocol. The VISA interface supports bus methods such as USB, LAN, or GPIB. The host computer sets the oscilloscope's sampling rate, acquisition channels, trigger mode, and trigger threshold via VISA commands and issues acquisition commands. The oscilloscope returns the raw waveform data, typically a binary byte stream.

[0031] The LabVIEW control layer is deployed on a host computer, which is a regular PC running a Windows operating system and the LabVIEW development environment. The LabVIEW program includes a data receiving module, a preprocessing module, a visualization module, and a format conversion module. The data receiving module reads the raw waveform data returned by the oscilloscope using the VISA Read function and parses it into a LabVIEW numerical array. The preprocessing module sequentially performs DC removal (i.e., subtracting the array mean), low-pass digital filtering (e.g., using a Butterworth filter, with the cutoff frequency set according to the bandwidth of the measured signal), downsampling (reducing the number of data points as needed to match the input length of the AI ​​model), and outlier removal (e.g., linear interpolation correction for data points exceeding three standard deviations). The preprocessed data is stored in an internal buffer. The visualization module displays the raw waveform and the superimposed curve of the subsequent AI analysis results on the front panel in real time.

[0032] The format conversion module solves the data type compatibility problem between LabVIEW's native arrays and Python AI models. LabVIEW uses its built-in "Python script node" to call Python code, converting the preprocessed LabVIEW arrays into NumPy's ndarray format and reshaping their shape according to the input dimension requirements of different AI models. For example, for convolutional neural network models, it may be necessary to reconstruct one-dimensional time-series data into a three-dimensional tensor of [batch, channels, length].

[0033] The AI ​​analysis layer contains a lightweight machine learning model. This embodiment employs a pre-trained support vector machine (SVM) or convolutional neural network (CNN) model, with the model file stored in .pkl or .pt format in a designated path on the host computer's local hard drive. When the LabVIEW program starts, the model is loaded into memory once via a Python script node. Each subsequent frame of data reuses this model, avoiding latency and resource waste caused by repeated loading. The model input is preprocessed waveform time-series data, and the output includes waveform classification results (e.g., sine wave, square wave, hysteresis loop, etc.), feature parameters, and signal quality assessment results. For hysteresis loops, feature parameters include output saturation magnetic induction, coercivity, and remanence. For general periodic signals, feature parameters include output amplitude, frequency, and duty cycle. Signal quality assessment specifically includes anomaly jump identification, baseline offset discrimination, signal-to-noise ratio estimation, and comprehensive confidence scoring. These assessment results are output as anomaly scores and confidence scores. It should be noted that the specific types of the feature parameters depend on the experiment being tested. This embodiment uses a hysteresis loop as an example, but the present invention is also applicable to other physical experiments, such as sound speed measurement, RLC series circuit resonance characteristics research, Fourier decomposition and synthesis of periodic electrical signals, etc., simply by replacing the corresponding AI model or adjusting the characteristic parameter output.

[0034] The output layer overlays the feature parameters obtained from AI analysis, signal quality assessment conclusions, and the original waveform curves onto the LabVIEW front panel. Simultaneously, the system automatically generates a structured analysis report, which includes at least the acquisition time, device parameters (sampling rate, channels), waveform feature parameters, signal quality assessment conclusions, and trend predictions based on historical data.

[0035] Data interaction between the LabVIEW control layer and the AI ​​analysis layer uses a direct memory connection: LabVIEW arrays are passed to the Python process via Python script nodes. Python internally calls NumPy to perform format conversion before directly feeding the data into the model for inference. The inference results are then returned to LabVIEW via nodes. There is no intermediate disk file read / write during this process, and data processing latency is controlled within milliseconds, meeting the needs of real-time analysis. Both the data format conversion and model inference calculations are performed using the host computer's local CPU, requiring no GPU or dedicated computing hardware. Actual measurements show that the processing time for a single frame is controlled within milliseconds, satisfying real-time analysis requirements.

[0036] Example 2

[0037] Building upon Embodiment 1, this embodiment further provides a custom model import function. The LabVIEW control layer is configured with a model import interface, implemented as a file selection control and a model loading sub-VI. Users can select locally stored custom AI model files (supporting common formats such as .pkl, .pt, or .onnx) on the LabVIEW front panel, and the system automatically reads the model file and replaces the currently running model. This interface allows users to load dedicated models for different experimental scenarios. For example, in RLC resonant circuit testing, users can import a model trained for resonant frequency and quality factor extraction; in optical signal detection, they can import a model trained for photocurrent waveform peak recognition. Through this custom import interface, the system of this invention can flexibly adapt to various physical experiments and scientific research testing scenarios without rewriting the LabVIEW main program.

[0038] Example 3

[0039] This embodiment details the specific working method of signal quality assessment. The signal quality assessment output by the AI ​​analysis layer includes the following dimensions: First, abnormal jump identification: The model analyzes whether there are abrupt changes in continuous waveform data, such as an abnormal spike with an amplitude far exceeding the average amplitude suddenly appearing in a normal periodic waveform. The model outputs the location and amplitude deviation rate of this abnormal jump. Second, baseline offset judgment: The model calculates the trend of the DC component of the waveform data over time. When the DC component deviates from the initial calibration value by more than a preset threshold (e.g., 5% of full scale), it is judged as a baseline offset, and the offset direction and offset amount are output. Third, signal validity determination: The model integrates the estimated signal-to-noise ratio, waveform periodic stability, and trigger consistency to output a valid signal confidence level between 0 and 1. When the confidence level is lower than 0.6, the system determines that the signal quality does not meet the requirements. All the above assessment results are recorded in the structured analysis report.

[0040] Example 4

[0041] This embodiment describes the system's automatic closed-loop acquisition control method. When the AI ​​analysis layer determines that the signal quality does not meet preset requirements, such as a signal-to-noise ratio lower than a set threshold (e.g., 20dB) or unstable triggering (jitter exceeding 10% of the sampling period during multiple consecutive triggering frames), the LabVIEW control layer automatically performs adjustment operations. These adjustments include: reducing the sampling rate to decrease noise bandwidth, or increasing the sampling rate to capture faster transient features; adjusting the trigger threshold level; and changing the trigger coupling method (DC coupling or AC coupling). After each adjustment, the system automatically re-acquires a frame of data and re-evaluates the signal quality. If the requirements are still not met, adjustments continue until the signal quality meets the preset requirements or the maximum number of adjustments is reached. This closed-loop control process requires no user intervention, significantly improving the acquisition success rate in batch testing.

[0042] Example 5

[0043] This embodiment illustrates how the present invention can be applied to other typical physical experiments besides hysteresis loop measurement, to demonstrate its versatility.

[0044] Sound velocity measurement experiment: In sound velocity measurement, an oscilloscope typically receives a sinusoidal signal output from a transducer and measures the sound time by observing the Lissajous figure or phase difference. Using the system of this invention, the LabVIEW control layer acquires the signals from oscilloscope channels 1 and 2 (transmitted and received waves, respectively). A pre-trained CNN model is loaded into the AI ​​analysis layer. This model takes dual-channel waveform data as input and outputs the phase difference or time delay to calculate the sound velocity. Signal quality assessment can detect waveform distortion and amplitude stability, and automatically adjusts the oscilloscope's vertical sensitivity and sampling rate when the signal-to-noise ratio is too low.

[0045] Research on the Resonant Characteristics of RLC Series Circuits: In RLC resonance experiments, an oscilloscope measures the voltage across the resistor (in phase with the current) and the signal source voltage, obtaining the amplitude-frequency characteristics through frequency sweeping. This invention's system can control the oscilloscope to automatically sweep frequencies (via an external signal source or the oscilloscope's built-in function generator), acquiring voltage amplitudes at different frequencies. An AI model (such as SVM regression or a lightweight neural network) directly outputs the resonant frequency and quality factor. Compared to manually reading data from multiple frequency points, this invention significantly improves experimental efficiency.

[0046] Fourier Decomposition and Synthesis Experiment of Periodic Electrical Signals: In the Fourier decomposition and synthesis experiment, an oscilloscope displays the waveforms of periodic signals such as square waves and triangular waves. The system of this invention acquires waveform data within one cycle. An AI model (e.g., a pre-trained CNN or a small Transformer) directly outputs the amplitude and phase of the fundamental wave and each harmonic, without requiring Fast Fourier Transform (FFT) calculations or serving as auxiliary verification. Users can customize the desired harmonic order to retain via the LabVIEW front panel and synthesize and reconstruct the waveform, then superimpose and compare it with the original waveform.

[0047] The system architecture and processes described in Examples 1 to 5 can be directly reused in all three experimental scenarios, requiring only the replacement of the AI ​​training model or adjustment of preprocessing parameters (e.g., the sampling rate needs to be set according to the signal frequency range). The custom model import interface (Example 2) allows teachers or researchers to train dedicated models for different experiments and easily deploy them. This embodiment fully demonstrates that the system and method of the present invention have good versatility and scalability.

[0048] Example 6

[0049] Based on the above embodiments, this embodiment provides the complete workflow of the general intelligent oscilloscope data automatic acquisition and analysis method of the present invention.

[0050] First, the host computer establishes a communication connection with the general-purpose digital oscilloscope via the LabVIEW control layer and a VISA interface. The operator sets the sampling rate, acquisition channels, trigger mode, and trigger threshold on the front end, or directly uses the system default parameters. LabVIEW then issues acquisition commands to initiate the acquisition of raw waveform data at the hardware layer.

[0051] Secondly, a general-purpose digital oscilloscope acquires the raw waveform data of the target signal under test, which is then uploaded to the LabVIEW control layer in real time via the data acquisition layer. After receiving the raw waveform data, the LabVIEW control layer performs DC removal, low-pass filtering, downsampling, and outlier removal operations through its built-in data preprocessing submodule. The preprocessed data is then stored in the data buffer.

[0052] Then, the LabVIEW control layer calls the lightweight machine learning model of the AI ​​analysis layer. Specifically, LabVIEW loads the locally stored pre-trained model (CNN or SVM) through a Python script node, converts the preprocessed waveform data into NumPy array format, and sends it into the model. The model performs automatic waveform classification, noise suppression, feature parameter extraction, and signal quality assessment in real time, outputting feature parameters (such as the saturation point, coercivity, and remanence of the hysteresis loop, or the sound time and speed in the sound velocity measurement experiment, or the resonant frequency and quality factor in the RLC resonance experiment, or the fundamental to harmonic amplitude ratio in the Fourier decomposition and synthesis experiment) and signal quality assessment results (including anomalous jumps, baseline offset, and confidence score).

[0053] Next, the output layer receives the raw waveform data from the LabVIEW control layer and the analysis results from the AI ​​analysis layer, and displays them overlaid on the LabVIEW front panel. Simultaneously, the system automatically generates a structured analysis report in PDF format, which details the acquisition parameters (sampling rate, channels, trigger thresholds), waveform characteristic data, signal quality assessment conclusions, and trend prediction information (such as characteristic parameter change curves from multiple consecutive acquisitions).

[0054] Finally, when the AI ​​analysis layer determines that the signal quality does not meet the preset requirements, the LabVIEW control layer automatically adjusts the oscilloscope's sampling rate or trigger threshold and reacquires the signal until the signal quality meets the requirements. Furthermore, users can manually adjust the acquisition parameters or preprocessing parameters through the LabVIEW control layer based on the information in the analysis report. They can also export the raw waveform data and analysis report for subsequent scientific research or data archiving.

[0055] The above complete workflow achieves full-chain automation from signal acquisition to intelligent analysis, result output, and closed-loop adjustment, enabling intelligent processing and diagnosis of waveform data without human intervention.

[0056] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any modifications to the technical solutions described in the foregoing embodiments, or equivalent substitutions of some or all of the technical features thereof, within the spirit and principles of the present invention, do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A general-purpose intelligent oscilloscope system based on LabVIEW and AI integration, characterized in that: include: The hardware layer consists of a general-purpose digital oscilloscope used to acquire raw waveform data. The data acquisition layer uploads the raw waveform data to the host computer via the VISA communication interface; The LabVIEW control layer, deployed on the host computer, receives the raw waveform data and performs preprocessing. The AI ​​analysis layer integrates a lightweight machine learning model, receives the preprocessed waveform data, and automatically performs waveform classification, feature parameter extraction, and waveform signal quality assessment. The result output layer overlays the AI ​​analysis results with the raw waveform for visualization. The LabVIEW control layer and the AI ​​analysis layer interact with each other and perform model inference through Python script calling nodes.

2. The general-purpose intelligent oscilloscope system based on LabVIEW and AI integration according to claim 1, characterized in that, The AI ​​analysis layer uses a pre-trained convolutional neural network (CNN) model or a support vector machine (SVM) model. The lightweight model is stored locally as a file and is loaded once by the LabVIEW control layer when the program starts and remains resident in memory.

3. The general-purpose intelligent oscilloscope system based on LabVIEW and AI integration according to claim 1 or 2, characterized in that, The built-in data preprocessing submodule of the LabVIEW control layer includes DC removal, digital filtering, and downsampling functions, which are used to convert raw waveform data into a standard format that is compatible with AI model input.

4. The general-purpose intelligent oscilloscope system based on LabVIEW and AI integration according to claim 3, characterized in that, The LabVIEW control layer is configured with a custom model import interface, which allows users to import custom AI models trained for different experimental scenarios.

5. The general-purpose intelligent oscilloscope system based on LabVIEW and AI integration according to claim 1, characterized in that, The waveform signal quality assessment includes abnormal transition identification, baseline offset discrimination, and signal validity determination, and outputs anomaly score and confidence level.

6. An automatic data acquisition and analysis method for a general-purpose intelligent oscilloscope data system integrating LabVIEW and AI as described in any one of claims 1 to 5, characterized in that, Includes the following steps: S1: The LabVIEW control layer establishes communication with the general-purpose digital oscilloscope via the VISA interface, configures the acquisition parameters, and starts acquisition; S2: The general-purpose digital oscilloscope acquires raw waveform data and uploads it to the LabVIEW control layer in real time; S3: The LabVIEW control layer performs DC removal, filtering, and downsampling preprocessing on the received raw waveform data; S4: The LabVIEW control layer calls the lightweight machine learning model of the AI ​​analysis layer through a Python script calling node, inputs the preprocessed data into the model, and completes waveform classification, feature parameter extraction, and signal quality assessment; S5: The results output layer overlays the AI ​​analysis results with the original waveform and automatically generates a structured analysis report; S6: When the AI ​​analysis layer determines that the signal quality does not meet the preset requirements, the LabVIEW control layer automatically adjusts the oscilloscope's sampling rate or trigger threshold and re-acquires the signal.

7. The data acquisition and analysis method for a general-purpose intelligent oscilloscope system based on LabVIEW and AI integration as described in claim 6, characterized in that, The situations in which the signal quality does not meet the preset requirements include a signal-to-noise ratio lower than a set threshold or an unstable trigger state.

8. The data acquisition and analysis method for a general-purpose intelligent oscilloscope system based on LabVIEW and AI integration as described in claim 6, characterized in that, The structured analysis report mentioned in step S5 is in PDF format and automatically records the acquisition parameters, waveform feature data, signal quality assessment conclusions, and trend prediction information.

9. The data acquisition and analysis method for a general-purpose intelligent oscilloscope system based on LabVIEW and AI integration as described in claim 6, characterized in that, Users can manually adjust acquisition or preprocessing parameters through the LabVIEW control layer and export waveform data and analysis reports.

10. The data acquisition and analysis method for a general-purpose intelligent oscilloscope system based on LabVIEW and AI integration as described in claim 6, characterized in that, The method is applicable to at least one of the following experimental scenarios: dynamic hysteresis loop measurement of ferromagnetic materials, sound velocity measurement, study of the resonance characteristics of RLC series circuits, and Fourier decomposition and synthesis experiments of periodic electrical signals.