Intelligent self-adaptive time sequence diagram analysis and structuring method

Abstract

The invention discloses an intelligent self-adaptive time sequence diagram analysis and structuring method. The method comprises the following steps: preprocessing an input time sequence diagram; identifying and extracting basic structure elements in the time sequence diagram; extracting text information in the time sequence diagram by using an optical character recognition technology; the recognized signal lines are analyzed, and signal state changes and key event points are extracted through a neural network; detecting arrows representing time sequence constraints, and performing arrow detection and classification by using a target detection network; establishing an association relationship between the graphic elements and the text; the association relationship and the identified elements are verified and corrected by using grammar rules and global layout information of the time sequence diagram, and unreasonable results are filtered; converting the time sequence relation model into a standardized structured data format; and the overall recognition performance is continuously optimized based on interactive correction and feedback of confidence. According to the method, the workload of manually converting the time sequence diagrams by engineers is greatly reduced, the accuracy of time sequence constraint extraction is improved, and various time sequence diagram formats and styles are supported.

Description

Technical Field

[0001] The present invention relates to the technical field of computer vision and image processing, and in particular to an intelligent adaptive timing graph parsing and structuring method. Background Art

[0002] Timing diagrams are an essential tool in digital system design and analysis. They clearly and visually depict the timing relationships between signals, including their triggering sequence, duration, and synchronous and asynchronous relationships. Consequently, they are widely used in many key areas, including chip design, protocol specification, and interface definition. During chip design, timing diagrams are used to ensure that signal interactions between modules meet design requirements, thereby guaranteeing proper chip operation.

[0003] However, traditional timing diagrams are typically presented in the form of images, PDF files, or paper documents, which have significant limitations. Due to the lack of machine-readable structured representation, these timing diagrams cannot be directly recognized and processed by automated tools, severely limiting the capabilities of automated verification and analysis.

[0004] In this situation, engineers are forced to manually interpret timing diagrams and convert them into formal descriptions for subsequent automated tool processing. This process is not only time-consuming and labor-intensive, but also prone to errors. Engineers need to carefully analyze every key element in the timing diagram, such as the signal waveform, timing arrow, and constraint parameter, and accurately convert them into a formal language or data model. Due to the complexity of timing diagrams and the uncertainty of manual operation, omissions, misunderstandings, or errors are easily made during the conversion process, resulting in inaccurate results for subsequent verification and analysis, and even the introduction of new design issues.

[0005] While existing image recognition methods can be used to identify basic elements in diagrams, most of these methods are designed for general-purpose diagrams and are not specifically optimized for the specific structure and semantics of timing diagrams. Timing diagrams have unique characteristics, such as the shape and variation of signal waveforms, the direction and meaning of timing arrows, and the annotation of constraint parameters. These elements carry important timing information. However, existing image recognition methods are generally unable to accurately identify these key elements or establish logical relationships between them.

[0006] Therefore, in order to improve the efficiency and accuracy of digital system design and verification, there is an urgent need for a technical solution that can automatically parse timing diagrams and generate structured data models to support the automated process of digital system design and verification. Summary of the Invention

[0007] Purpose of the invention: The present invention provides an intelligent adaptive timing diagram parsing and structuring method, which can automatically extract signal waveforms, timing constraints, parameter annotations and other information from timing diagram images, and construct a structured timing relationship model.

[0008] Technical solution: The intelligent adaptive timing diagram parsing and structuring method described in the present invention includes the following steps:

[0009] (1) Preprocessing the input timing diagram image;

[0010] (2) Identify and extract the basic structural elements in the timing diagram, including the horizontal signal lines and the vertical time axis, and determine the basic grid structure based on the identified signal lines and time axis;

[0011] (3) Use optical character recognition technology to extract text information from the timing diagram;

[0012] (4) Analyze the identified signal lines and use neural networks to extract signal state changes and key event points;

[0013] (5) Detect arrows representing temporal constraints and use the target detection network for arrow detection and classification;

[0014] (6) Establishing the relationship between graphic elements and text;

[0015] (7) Build a complete time sequence diagram model, use the grammatical rules and global layout information of the time sequence diagram to verify and correct the association relationships and identified elements generated in step (6), and filter out unreasonable results;

[0016] (8) Convert the temporal relationship model into a standardized structured data format;

[0017] (9) After outputting the structured data format in step (8), the confidence of the recognition result records of the time sequence diagram in steps (2) to (6) is evaluated; a confidence threshold is set, and if the confidence of all steps (2) to (6) is higher than the threshold, the output result of the time sequence diagram is considered correct; if the confidence of any step from (2) to (6) is lower than the threshold, the time sequence diagram and its preliminary recognition result are added to the "pool to be reviewed"; the data in the "pool to be reviewed" are manually corrected internally, and a user feedback interface is provided to allow external users to report and correct recognition errors; correction data from all sources are collected and fed back to the processing module corresponding to the initial recognition confidence lower than the threshold, so as to update or retrain its internal model to continuously optimize the overall recognition performance.

[0018] Furthermore, the preprocessing in step (1) includes grayscale, binarization, noise removal and tilt correction.

[0019] Furthermore, when performing the feature extraction tasks of steps (2) to (5), a multi-scale pyramid representation of the image is constructed, and the results are fused after feature extraction at different scales to improve the recognition robustness of elements of different sizes.

[0020] Furthermore, the implementation process of step (2) is as follows:

[0021] Signal line identification: Pixels in the pre-processed image are projected vertically, and the accumulated pixel values ​​of each row are counted. Candidate signal line regions are determined based on the peak positions on the projection curve. Connected domain analysis and screening are performed on the candidate regions to extract accurate horizontal signal lines.

[0022] Timeline recognition: First, using image segmentation techniques, the image is divided into multiple regions by analyzing the local or global features of pixels. The set of pixels that constitute the potential vertical timeline is classified as one or more independent candidate regions, separated from other image content. These candidate regions are then post-processed to screen and verify their geometric properties, ultimately extracting the main vertical straight line segments in the image that meet the definition of the timeline.

[0023] Determination of basic grid structure: The intersection points of signal lines and time axes on the image and the areas they constitute form the basic grid structure of the timing diagram.

[0024] Furthermore, the implementation process of step (3) is as follows:

[0025] A text detection algorithm is applied to locate areas containing text in the image. For each detected text area image block, an OCR engine is applied to perform text recognition. The OCR recognition results are processed, including: spelling correction based on a predefined dictionary, identification and standardization of special timing symbols, and parsing of numerical values ​​and units. Each recognized text block is classified into a signal name, timing parameter name, parameter value, and time unit based on the content characteristics of the text and its position in the image.

[0026] Furthermore, the implementation process of step (4) is as follows:

[0027] Traverse all identified signal lines and time axes and calculate their geometric intersections; extract a local image area of ​​a predetermined size as the analysis unit with each intersection as the center;

[0028] The extracted local image regions are used as input and sent to a pre-trained state classifier; the classifier outputs the corresponding signal state or key event type for each local image region, including high level, low level, rising edge, falling edge, and indeterminate state.

[0029] Furthermore, the implementation process of step (5) is as follows:

[0030] The YOLOv9 object detection model is used to process the input sequence image to directly identify and locate the arrow objects in the image, and output the bounding box of each detected arrow and its preliminary morphological classification, which includes left-pointing arrows, right-pointing arrows, and bidirectional arrows.

[0031] For each detected arrow, the image area within its corresponding bounding box is further analyzed, and the precise geometric features of the arrow are extracted through image processing or morphological analysis methods; including its precise starting point, end point, main axis direction, length, and specific shape details of the arrow part, for subsequent association and model building.

[0032] Furthermore, the implementation process of step (6) is as follows:

[0033] Proximity calculation and screening: Calculate the spatial distance between each text area and each graphic element (based on the distance threshold, screen out several candidate associated graphic elements that are spatially closest to each text area);

[0034] Association feature evaluation: Evaluate the association features between text and each candidate graphic element, including relative layout and semantic matching. Relative layout analyzes the orientation and alignment of the text; semantic matching determines whether the text type and graphic element type conform to the expected association rules.

[0035] Determine the best association: Use weighted scoring to calculate the total association score for each text-candidate graphic element pair; select the graphic element with the highest score that exceeds the threshold as its final association object for each text, and record the association.

[0036] Furthermore, the implementation process of step (7) is as follows:

[0037] Create a structured signal object for each identified signal line;

[0038] Create a structured timing constraint object for each identified arrow, which contains the constraint type, the signal object or specific event point on the signal, and the constraint value;

[0039] According to the signal objects or event points on them connected by the starting and ending points of the timing constraint arrows, clarify the timing dependency between them;

[0040] Integrate all signal objects and timing constraint objects and their dependencies into a unified, graph-structured or object-oriented timing relationship graph model.

[0041] Furthermore, the correction data from all sources in step (9) include correction data from internal correction and user feedback.

[0042] Beneficial effects: Compared with the existing technology, the beneficial effects of the present invention are as follows: the present invention realizes the automatic parsing of timing diagrams, greatly reducing the workload of engineers in manually converting timing diagrams; improves the accuracy of timing constraint extraction and reduces human errors; supports multiple timing diagram formats and styles and has good adaptability; generates standardized structured data models, which is convenient for integration with downstream verification tools; accelerates the digital system design and verification process and improves work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 is a flow chart of the present invention;

[0044] Figure 2 This is an example of a timing diagram in the chip manual. DETAILED DESCRIPTION

[0045] The present invention is described in further detail below with reference to the accompanying drawings:

[0046] like Figure 1 As shown, the present invention provides an intelligent adaptive timing diagram parsing and structuring method, which realizes the complete parsing of the timing diagram through a multi-stage processing flow, including image preprocessing, signal line recognition, waveform state extraction, arrow detection and classification, text recognition and association, and timing relationship diagram construction; it can process timing diagrams in various formats, extract key timing parameters and constraint relationships, and generate structured output, providing automated support for digital system design and verification.

[0047] Figure 2 This is an example of a timing diagram from a typical chip manual. This type of image is the input object that the present invention is intended to process. Specifically, the following steps are included:

[0048] Step 1: Preprocess the input timing diagram image to improve the accuracy of subsequent recognition.

[0049] Grayscale conversion: If the input is a color image, it is first converted into a single-channel 8-bit grayscale image using a standard weighted averaging method, such as the brightness formula Y = 0.299R + 0.587G + 0.114B, to reduce computational complexity and unify the processing basis.

[0050] Adaptive Binarization: For each pixel, a binarization threshold T(x, y) is dynamically calculated based on the mean and standard deviation of the pixels within its neighborhood window. The pixel value is compared with T(x, y) to generate a binary image that clearly separates foreground and background. The window size and algorithm-specific parameters should be determined based on experience or experimentation.

[0051] Noise removal: Eliminate isolated pixels or small non-structural breaks that may be produced after binarization by applying a morphological opening operation, which is defined as erosion followed by dilation of the image.

[0052] Tilt correction: Detect the tilt angle θ. Use the Hough transform to detect the tilt angle distribution of the dominant lines in the image and take the peak or average value. Once the angle θ is determined, apply the rotation matrix from the affine transformation to perform a rotation operation around the image center or a specified point.

[0053] Step 2: Identify the basic structural elements in the timing diagram.

[0054] Horizontal Signal Line Identification and Extraction: The preprocessed binary timing image is vertically projected to generate a one-dimensional vertical projection histogram to count the number of foreground pixels in each row. Peak regions in this histogram are analyzed to locate rows with potentially dense signal line populations, which are then designated as candidate signal line regions. Connected domain analysis is then applied to each candidate region, calculating the geometric properties of each connected domain. Filtering conditions are set based on the slender, horizontally extending characteristics of signal lines to filter out noise or non-signal line connected domains. This allows accurate horizontal signal line extraction and allows for the merging of collinear breakpoints.

[0055] Vertical timeline identification and extraction: Optional image preprocessing is performed, such as applying vertical morphological operations or edge detection to enhance vertical structure. Next, pixel-level image segmentation techniques are applied to classify image pixels into "timeline" and "background / other." This can be achieved using a pre-trained deep learning semantic segmentation model. Connected domain analysis is performed on the segmentation results to extract independent candidate timeline regions. By calculating their geometric properties, such as aspect ratio, principal direction, and length, vertical straight line segments that meet the timeline characteristics are selected as the identified timeline.

[0056] Determination of the basic grid structure: After extracting all horizontal signal lines and vertical time axes, by calculating the coordinates of the geometric intersection of each pair of signal lines and time axes, all identified signal line sets, time axis sets, and the intersection sets between them are collectively defined as the basic grid structure of the timing diagram. This structure provides a spatial framework for subsequent analysis.

[0057] Step 3: Use optical character recognition technology to extract text information from the timing diagram.

[0058] Text Region Detection: First, apply a text detection algorithm to locate the regions containing text in the image. Use deep learning-based scene text detection models (such as EAST, DBNet, CRAFT, etc.), which can directly output bounding boxes or polygonal regions of text lines.

[0059] Optical Character Recognition (OCR): For each detected text region image block, an OCR engine is used to perform text recognition. You can choose from established open-source OCR libraries or commercial OCR services. When configuring the OCR engine, you can specify the language and page segmentation mode. It may also require specific training or fine-tuning for common fonts and symbols found in time series diagrams.

[0060] Post-recognition processing and classification: Post-processing the raw OCR results to improve accuracy, including spelling correction based on a predefined dictionary, identification and standardization of special timing symbols, and parsing of numerical values ​​and units. Finally, each recognized text block is classified into semantic types such as signal name, timing parameter name, parameter value, and time unit based on its content characteristics and position in the image.

[0061] Step 4: Use the pre-trained convolutional neural network to identify signal waveform patterns and extract signal state changes and key event points.

[0062] Locating the intersection of signal lines and time axes and extracting local image regions: Based on the basic grid structure determined in step 2, traverse each identified horizontal signal line. For each signal line, determine its geometric intersection with all identified vertical time axes in turn. Centered on each calculated intersection coordinate, extract a local image region of fixed size (e.g., N×M pixels) from the preprocessed timing image. This local image region captures the visual features surrounding the signal at that specific time point.

[0063] Local image region feature extraction and state / event classification: Each extracted local image region is fed into a pre-trained state / event classifier. This classifier can be a deep learning-based model that learns features that distinguish different states directly from raw pixels. The classifier's task is to classify the input local image features into predefined signal states (e.g., high level, low level, rising edge, falling edge, indeterminate state / high impedance state) or key event types (e.g., data valid point, specific clock edge trigger point, etc.).

[0064] Waveform state sequence construction and key event point recording: For each signal line, collect the state or event labels obtained by classifying the corresponding local image area in the order of its intersection with the time axis (usually from left to right). Arrange these labels in chronological order to form a discrete waveform state sequence for the signal line. At the same time, record the intersection positions classified as "key event points" and their corresponding event types. By analyzing the changes in adjacent states in the state sequence, state transition points (such as the rising edge from low level to high level) can be further identified. Ultimately, each signal line will be associated with a waveform state sequence that describes its dynamic behavior and a set of time points marked with key events.

[0065] Step 5: Detect the arrows representing timing constraints and classify them.

[0066] Arrow object detection and preliminary morphological classification: The preprocessed time series image is fed into a pre-trained YOLOv9 object detection model, which can directly identify and locate arrow objects in the image. The output of the object detection model is typically the bounding box coordinates of each detected arrow and a preliminary morphological classification label, which indicates the basic geometric shape of the arrow, such as "left-pointing single arrow", "right-pointing single arrow", "bidirectional arrow", "vertical upward arrow", "vertical downward arrow", and other predefined categories.

[0067] Arrowhead Fine Geometric Feature Extraction: For each arrow identified and located by the object detection model, a more detailed analysis is performed on the image region within its corresponding bounding box to extract precise geometric features. This includes applying image processing techniques to accurately determine the position of the arrow's shaft axis and arrowhead; calculating the arrow's precise start and end coordinates; determining the direction of the arrow's main axis; and measuring the arrow's length. These fine features are crucial for subsequently understanding the arrow's specific meaning and connectivity.

[0068] In order to process time series graphs of different resolutions and complexities, during the feature extraction tasks in steps 2 to 5, a multi-scale pyramid representation of the image is constructed, and the results are fused after feature extraction at different scales to improve the recognition robustness of elements of different sizes.

[0069] Step 6: Establish an association between graphic elements and text.

[0070] Spatial relationship evaluation: For each text region and each graphic element identified by step 1, the shortest distance between the edge of the text region bounding box and the pixel set of the graphic element is calculated, and the direction of the vector pointing from the center point of the graphic element to the center point of the text region is calculated. The orientation of the center point of the text region relative to the center point of the graphic element and the degree of alignment of the two in the horizontal or vertical direction are analyzed to determine the most spatially relevant candidate graphic element.

[0071] Inferring association types using semantic features: Perform semantic analysis on the extracted text content to infer its functional type and the type of graphic elements it may be associated with. By using a keyword list or pattern matching, identify whether the text content represents a signal name, timing parameter name, or signal status. Based on the identified text type, preliminarily determine its expected associated graphic element type.

[0072] Comprehensive decision-making to determine associations: Based on the previously calculated distance, analyzed relative position and alignment, and inferred semantic match, a score is assigned to the potential association between each text region and its candidate graphic elements. This scoring process uses a weighted summation approach, with smaller distances, better alignment, and a strong semantic type match receiving higher scores. Ultimately, the graphic element with the highest score, exceeding a predetermined threshold, is selected as the associated object for each text region.

[0073] Step 7: Build a complete timing relationship graph model, use the grammatical rules and global layout information of the timing graph to verify and correct the association relationships and identified elements generated in step 6, and filter out unreasonable results.

[0074] Create signal objects: Create a signal object for each identified signal line, storing the associated signal name and the analyzed waveform state sequence.

[0075] Create constraint objects: Create a constraint object for each recognized arrow, storing its constraint type, associated parameter values, and geometric information.

[0076] Establish timing dependencies: By mapping the start / end coordinates of the constraint object to the waveform event points of the corresponding signal object, record its source / target events in the constraint object to clarify the dependencies between signals.

[0077] Model integration: Organize all signal objects and constraint objects containing dependency links into a unified data structure to form the final timing relationship graph model.

[0078] Step 8: Convert the constructed time series relationship model into a standardized structured data format.

[0079] Design a standardized data model for representing timing relationships, including elements such as signals, events, and constraints; convert the timing relationship graph model into a defined data model format; generate output in different formats such as JSON, XML, and dedicated timing description languages ​​based on application requirements; provide interfaces with downstream verification tools to support automated timing verification.

[0080] Step 9: Support interactive correction and adaptive learning.

[0081] Comprehensive confidence assessment and pending review marking: After the structured data format is output in step 8, for each processed timing diagram, the recognition confidence of each element recorded in the timing diagram in key steps such as structure extraction (step 2), text recognition (step 3), waveform analysis (step 4), arrow detection and classification (step 5), and association analysis (step 6) is recorded. Set an independent confidence threshold for each step. If the confidence of the timing diagram in all steps is higher than the respective thresholds, the parsing result of the diagram is preliminarily considered to be highly reliable. Conversely, if the confidence of any key step is lower than the corresponding threshold, the timing diagram and its current parsing result are automatically marked and sent to the "pending review pool".

[0082] Correction data from all sources (including internal corrections and user feedback) are collected. For internal corrections, the system maintains a "pending review pool" containing all time series diagrams that do not meet the confidence requirements after evaluation; internal operators or designated review teams can access this pool to conduct detailed reviews and manual corrections of the parsing results. For user feedback corrections, the system provides a feedback interface for end users. When using the parsing results, if users find any errors or inaccuracies, even if the results are not automatically marked as low confidence, they can submit error reports and suggested correction information through this interface.

[0083] Unified Recording: Whether it is internally corrected data or correction information collected through the user feedback interface, the system will record it in detail together with the original recognition results, forming a high-quality annotated dataset in a unified format and marking its source.

[0084] Targeted Feedback and Model Update Triggering: The system integrates correction data from all sources. For each correction record, the system compares it with the original low-confidence recognition result and accurately feeds this correction data back to the corresponding processing module that originally produced the low-confidence result or was pointed out by the user. This feedback data is used to expand the training set of the corresponding module for subsequent model updates or retraining.

[0085] Based on a feedback mechanism, high-quality annotated samples are collected through both internal corrections and user feedback. The system not only collects corrected data but also records the original low-confidence recognition result corresponding to the sample and the specific deep learning model(s) that performed poorly on the sample. This builds a dynamic training sample library tailored to specific weaknesses.

[0086] Based on the sample library, the system can trigger fine-tuning or incremental learning of a specific model periodically or after collecting enough correction samples for that specific model. For example, if a large number of correction samples involve recognition errors for a specific type of arrow or complex waveform transition, these samples will be prioritized for optimizing the corresponding deep learning model. In this way, the model can learn and improve its performance in scenarios that were previously poorly recognized, thereby continuously improving the recognition accuracy and robustness of the overall system.

[0087] The above content merely illustrates the technical idea of ​​the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of this technical solution belong to the technical idea proposed by the present invention and fall within the protection scope of the claims of the present invention.

Claims

1. An intelligent adaptive timing diagram parsing and structuring method, characterized in that: The following steps are involved: (1) Preprocessing the input timing diagram image; (2) Identify and extract the basic structural elements in the timing diagram, including the horizontal signal lines and the vertical time axis, and determine the basic grid structure based on the identified signal lines and time axis; (3) Use optical character recognition technology to extract text information from the timing diagram; (4) Analyze the identified signal lines and use neural networks to extract signal state changes and key event points; (5) Detect arrows representing temporal constraints and use the target detection network for arrow detection and classification; (6) Establishing the relationship between graphic elements and text; (7) Build a complete time sequence diagram model, use the grammatical rules and global layout information of the time sequence diagram to verify and correct the association relationships and identified elements generated in step (6), and filter out unreasonable results; (8) Convert the temporal relationship model into a standardized structured data format; (9) After outputting the structured data format in step (8), the confidence of the recognition result records of the time sequence diagram in steps (2) to (6) is evaluated; a confidence threshold is set. If the confidence of all steps (2) to (6) is higher than the threshold, the output result of the time sequence diagram is considered correct; if the confidence of any step (2) to (6) is lower than the threshold, the time sequence diagram and its preliminary recognition result are added to the "to be reviewed pool"; The data in the "pending review pool" is manually corrected internally, and a user feedback interface is provided to allow external users to report and correct recognition errors. Corrected data from all sources is collected and fed back to the processing modules whose initial recognition confidence is below the threshold, so as to update or retrain their internal models and continuously optimize the overall recognition performance.

2. The intelligent adaptive timing diagram parsing and structuring method according to claim 1, characterized in that: The preprocessing in step (1) includes grayscale, binarization, noise removal and tilt correction.

3. The intelligent adaptive timing diagram parsing and structuring method according to claim 1, characterized in that: When performing the feature extraction tasks in steps (2) to (5), a multi-scale pyramid representation of the image is constructed, and the results are fused after feature extraction at different scales to improve the recognition robustness of elements of different sizes.

4. The intelligent adaptive timing diagram parsing and structuring method according to claim 1, characterized in that: The implementation process of step (2) is as follows: Signal line identification: Pixels in the pre-processed image are projected vertically, and the accumulated pixel values ​​of each row are counted. Candidate signal line regions are determined based on the peak positions on the projection curve. Connected domain analysis and screening are performed on the candidate regions to extract accurate horizontal signal lines. Timeline recognition: First, using image segmentation techniques, the image is divided into multiple regions by analyzing the local or global features of pixels. The set of pixels that constitute the potential vertical timeline is classified as one or more independent candidate regions, separated from other image content. These candidate regions are then post-processed to screen and verify their geometric properties, ultimately extracting the main vertical straight line segments in the image that meet the definition of the timeline. Determination of basic grid structure: The intersection points of signal lines and time axes on the image and the areas they constitute form the basic grid structure of the timing diagram.

5. The intelligent adaptive timing diagram parsing and structuring method according to claim 1, characterized in that: The implementation process of step (3) is as follows: A text detection algorithm is applied to locate the area containing text in the image. For each detected text area image block, an OCR engine is applied to perform text recognition. The OCR recognition results are processed, including: spelling correction based on a predefined dictionary, identification and standardization of special timing symbols, and parsing of numerical values ​​and units. Each recognized text block is classified into a signal name, timing parameter name, parameter value, and time unit according to the content characteristics of the text and its position in the image.

6. The intelligent adaptive timing diagram parsing and structuring method according to claim 1, characterized in that: The implementation process of step (4) is as follows: Traverse all identified signal lines and time axes and calculate their geometric intersections; extract a local image area of ​​a predetermined size as the analysis unit with each intersection as the center; The extracted local image regions are used as input and sent to a pre-trained state classifier; the classifier outputs the corresponding signal state or key event type for each local image region, including high level, low level, rising edge, falling edge, and indeterminate state.

7. The intelligent adaptive timing diagram parsing and structuring method according to claim 1, characterized in that: The implementation process of step (5) is as follows: The YOLOv9 object detection model is used to process the input sequence image to directly identify and locate the arrow objects in the image, and output the bounding box of each detected arrow and its preliminary morphological classification, which includes left-pointing arrows, right-pointing arrows, and bidirectional arrows. For each detected arrow, the image area within its corresponding bounding box is further analyzed to extract the precise geometric features of the arrow through image processing or morphological analysis methods; Including its precise starting point, end point, main axis direction, length and specific shape details of the arrow part for subsequent association and model construction.

8. The intelligent adaptive timing diagram parsing and structuring method according to claim 1, characterized in that: The implementation process of step (6) is as follows: Proximity calculation and screening: Calculate the spatial distance between each text area and each graphic element (based on the distance threshold, screen out several candidate associated graphic elements that are spatially closest to each text area); Association feature evaluation: Evaluate the association features between text and each candidate graphic element, including relative layout and semantic matching. Relative layout analyzes the orientation and alignment of the text; semantic matching determines whether the text type and graphic element type conform to the expected association rules. Determine the best association: Use weighted scoring to calculate the total association score for each text-candidate graphic element pair; select the graphic element with the highest score that exceeds the threshold as its final association object for each text, and record the association.

9. The intelligent adaptive timing diagram parsing and structuring method according to claim 1, characterized in that: The implementation process of step (7) is as follows: Create a structured signal object for each identified signal line; Create a structured timing constraint object for each identified arrow, which contains the constraint type, the signal object or specific event point on the signal, and the constraint value; According to the signal objects or event points on them connected by the starting and ending points of the timing constraint arrows, clarify the timing dependency between them; Integrate all signal objects and timing constraint objects and their dependencies into a unified, graph-structured or object-oriented timing relationship graph model.

10. The intelligent adaptive timing diagram parsing and structuring method according to claim 1, characterized in that: The correction data from all sources in step (9) include internal correction data and correction data from user feedback.

Images