PDF drawing key parameter information extraction method and system

Abstract

The invention provides a PDF drawing key information extraction method and system, and belongs to the technical field of information. The method comprises the steps of obtaining an execution process of a current task and detailed text description of a current construction process, and extracting keywords by utilizing a TextRank algorithm; searching in a metadata tag of a drawing library by utilizing the keyword, and quickly screening out a drawing file highly related to the current construction process; for the screened PDF drawing file, converting image characters in the PDF into editable text information by adopting an OCR (Optical Character Recognition) technology, and preprocessing the text information; key parameters are extracted from the preprocessed text according to predefined parameter extraction rules, and reasonability analysis is carried out. Through the automatic information extraction process, the key parameters matched with the current project execution process are rapidly and accurately extracted from a large number of PDF drawings, and the information processing efficiency is greatly improved. And strict rationality analysis is carried out on key parameter information, so that the accuracy and reliability of information extraction are further ensured.

Description

Technical Field

[0001] The invention belongs to the field of information technology, and in particular relates to a method and system for extracting key parameter information of a PDF drawing. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] In engineering construction, mechanical design and various engineering project management, PDF drawings, as a common file format, are widely used to store and transmit detailed design, construction or product manufacturing information. These drawings contain a wealth of key information such as technical parameters, material requirements, specifications and construction processes, which are of great significance for ensuring the smooth progress of the project, the effective procurement of materials and the control of construction quality.

[0004] However, in actual application, the traditional manual method of extracting these key information from PDF drawings is not only inefficient, but also prone to errors. Especially when the project is large-scale and the number of drawings is large, the manual review and extraction of key information is time-consuming and laborious, and it is difficult to meet the efficiency and accuracy requirements of project management. In addition, the existing PDF information extraction technology can only perform simple text recognition and extraction, and it is difficult to automatically identify specific layers, component types and their association with text information in the drawings, and it is also impossible to dynamically extract key parameters that match the current process according to the execution process of the project. This technical limitation leads to the limitation of extracting information from PDF drawings, and it is difficult to meet the complex parameter analysis and risk assessment requirements in the project management process. Summary of the invention

[0005] In order to overcome the deficiencies of the above-mentioned prior art, the present invention provides a method and system for extracting key parameter information of PDF drawings. Through an automated information extraction process, key parameters matching the current project execution progress can be quickly and accurately extracted from a large number of PDF drawings, thereby greatly improving the efficiency of information processing. A variety of technical means are used to ensure the accuracy and reliability of the extracted information. At the same time, the accuracy of the information is further verified by analyzing the rationality of the extracted key parameter information and comparing it with industry standards, and based on the extracted key parameter information, suggestions based on the current progress are given.

[0006] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions:

[0007] The first aspect of the present invention provides a method for extracting key parameter information of a PDF drawing;

[0008] A method for extracting key parameter information of a PDF drawing, comprising:

[0009] Obtain the overall cycle of project tasks based on the project plan; determine the execution progress of the current task based on the current time point and task cycle;

[0010] Based on the execution process, a detailed text description of the current construction process is obtained, and keywords are extracted from the text description using the TextRank algorithm;

[0011] The keyword is used to search in the metadata tags of the drawing library, and the drawing files that are highly relevant to the current construction process are quickly screened out by comparing the matching degree between the keyword and the metadata tags; the drawing library includes all PDF drawing files in the project, and each PDF drawing file is added with metadata tags;

[0012] For the selected PDF drawing files that are highly relevant to the current construction process, OCR technology is used to convert the image text in the PDF into editable text information, and the text information is pre-processed;

[0013] According to the predefined parameter extraction rules, key parameters are extracted from the preprocessed text, and the rationality of the obtained key parameters is analyzed.

[0014] As a further technical solution, the process of obtaining a detailed text description of the current construction process based on the execution process and extracting keywords from the text description using the TextRank algorithm is as follows;

[0015] Obtain a detailed text description of the current construction process; perform cleaning and word segmentation on the text description;

[0016] The segmented words are used as nodes of the graph, the co-occurrence relationship between words is defined by setting the window size, edges are constructed based on the co-occurrence relationship between words, and weights are assigned to edges based on the co-occurrence frequency;

[0017] Assign an initial weight to each node in the graph;

[0018] The TextRank algorithm is executed iteratively. In each iteration, the weight of the current node is updated according to the weights of the node's neighbor nodes and the edge weights between them until the node weights converge to a stable state. The nodes are sorted according to the final calculated node weights, and several nodes with the highest weights are selected as keywords.

[0019] As a further technical solution, the TextRank algorithm is:

[0020]

[0021] In the formula, V represents the node, W(V i ) represents the weight value of node i, d represents the learning rate, Vi represents the i-th node, V j Indicates the jth node, In(V i ) represents the incoming link set of the i-th node, w ji represents the weight value between the jth node and the ith node, V k Indicates the window size k of the node, that is, at most k nodes co-occur, Out(V j ) represents the outbound link set of the jth node, w jk represents the weight value between the jth node and the window size k, W(V j ) represents the weight value of the jth node.

[0022] As a further technical solution, the process of converting the image text in the PDF into editable text information by using OCR technology and preprocessing the text information is as follows:

[0023] Use the PyMuPDF library in the drawing library to parse the file and extract the vector data of the layers in the drawing;

[0024] Render the vector data into images through the graphics processing library and preprocess the images, including scaling, cropping and denoising;

[0025] Input the preprocessed image into the pre-trained image classification model to identify the component type represented by the layer;

[0026] Using the coordinate and size information in the vector data and the results of image classification, the specific position and size of each component are determined, and the text information is extracted using OCR technology, and the text information is preprocessed to associate the preprocessed text information with the component data; wherein the text information preprocessing process includes noise removal, format standardization, and word segmentation processing;

[0027] As a further technical solution, the process of inputting the processed image into a pre-trained image classification model to identify the type of component represented by the layer is as follows:

[0028] The image classification model uses the PyTorch deep learning framework to implement the Inception module by defining the Inception class, uses the defined Inception module to build the image classification model, and uses a large-scale data set to train the image classification model;

[0029] After the training is completed, the processed image is input into the image classification model, and the model's prediction results on the image content are obtained;

[0030] According to the task requirements, create a mapping table from the model output category number to the specific component type name;

[0031] Use the category number predicted by the model as the index to query the mapping table to obtain the corresponding component type name.

[0032] As a further technical solution, the process of extracting text information using OCR technology, preprocessing the text information, and associating the preprocessed text information with component data is as follows:

[0033] Based on the acquired component vector data and the component type of the image;

[0034] Use Tesseract OCR to perform text recognition on the preprocessed image and extract the name and size of the component;

[0035] According to the coordinates and size information of the components in the vector data, the corresponding component position is found in the image, and the component name and size recognized by OCR are matched with the components in the vector data to ensure that the corresponding text information can be found for each component;

[0036] The vector data, OCR recognition results, and image classification results are stored in the database, and the association relationship between the data tables is established.

[0037] As a further technical solution, the process of analyzing the rationality of the acquired key parameters is as follows:

[0038] Obtain standard parameters from industry standards, design specifications and technical manuals;

[0039] Compare the extracted key parameter information with the standard parameters to check whether they meet the requirements.

[0040] A second aspect of the present invention provides a system for extracting key parameter information of a PDF drawing.

[0041] A PDF drawing key parameter information extraction system, comprising:

[0042] The execution process acquisition module is configured to: acquire the overall cycle of the project task based on the project plan; determine the execution process of the current task according to the current time point and the task cycle;

[0043] The keyword extraction module is configured to: obtain a detailed text description of the current construction process based on the execution process, and extract keywords from the text description using a TextRank algorithm;

[0044] The PDF drawing screening module is configured to: use the keyword to search in the metadata tags of the drawing library, and quickly screen out the drawing files that are highly relevant to the current construction process by comparing the matching degree between the keyword and the metadata tags; the drawing library includes all PDF drawing files in the project, and each PDF drawing file is added with metadata tags;

[0045] The image and text conversion module is configured to: for the selected PDF drawing files that are highly relevant to the current construction process, use OCR technology to convert the images and texts in the PDF into editable text information, and pre-process the text information;

[0046] The key parameter extraction module is configured to extract key parameters from the preprocessed text according to predefined parameter extraction rules, and perform rationality analysis on the obtained key parameters.

[0047] A third aspect of the present invention provides a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps in a method for extracting key parameter information of a PDF drawing as described in the first aspect of the present invention.

[0048] The fourth aspect of the present invention provides an electronic device, including a memory, a processor, and a program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps in the method for extracting key parameter information of a PDF drawing as described in the first aspect of the present invention are implemented.

[0049] One or more of the above technical solutions have the following beneficial effects:

[0050] The present invention uses an automated information extraction process, advanced OCR technology and TextRank algorithm, and can quickly and accurately extract key parameters that match the current project execution process from a large number of PDF drawings, thereby greatly improving the efficiency of information processing; and performs strict rationality analysis on the acquired key parameters to further ensure the accuracy and reliability of the extracted information.

[0051] At the same time, the accuracy of the information is further verified by analyzing the rationality of the extracted parameters and comparing them with industry standards. Based on the extracted key parameters, the abnormal risks in the task execution process are evaluated in real time, and suggestions based on the current progress are given. This helps project managers to promptly identify potential problems and take corresponding measures, thereby reducing risks during project execution.

[0052] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0054] Figure 1 This is a flow chart of the method of the first embodiment.

[0055] Figure 2 It is a system structure diagram of the second embodiment. DETAILED DESCRIPTION

[0056] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0057] It should be noted that the terms used herein are for describing specific embodiments only and are not intended to be limiting of exemplary embodiments according to the present invention.

[0058] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0059] In engineering construction, mechanical design and various engineering project management, a large number of PDF drawings are often required to store relevant design, construction or product manufacturing information. In actual application, manual methods are usually used to extract key information from PDF drawings, which is inefficient and prone to errors, and it is difficult to meet the efficiency and accuracy requirements of project management. At the same time, existing PDF extraction technologies can often only perform simple text information recognition and advance, and cannot automatically identify specific layers, component types and the relationship between text information in PDF drawings, nor can they dynamically extract key parameters that match the current process according to the execution process of the project. Based on the above problems, the present invention proposes a method and system for extracting key parameter information from PDF drawings, which will be described in detail in conjunction with specific embodiments below.

[0060] Embodiment 1

[0061] This embodiment discloses a method for extracting key parameter information of a PDF drawing; Figure 1 As shown, a method for extracting key information from a PDF drawing includes:

[0062] Step S1, obtaining the overall cycle of the project task based on the project plan; determining the execution progress of the current task according to the current time point and the task cycle.

[0063] In step S1, first, access or import the project plan document. The project plan document often details various stages, tasks, and their estimated completion times of the project. Based on the above plan document, the overall cycle of the project can be obtained, and the current time point can be tracked in real time, so as to accurately determine the execution process of the current task. The information sources of the execution process are diverse, including but not limited to the project plan document, progress report, real-time data in project management software, and work logs of project team members. By integrating these data sources, the execution status of the current task, the completed work content, and the upcoming tasks can be obtained in real time. These information are important basis for determining the description of the current construction process, providing a clear goal orientation for subsequent information extraction.

[0064] Step S2, based on the execution process, obtain the detailed text description of the current construction process, and extract keywords from the text description using the TextRank algorithm.

[0065] In step S2, obtain the detailed description of the current construction process from the project management system, progress report, or work log source. This description should contain sufficient information to accurately extract keywords related to the drawing content in subsequent steps;

[0066] Clean the obtained detailed description text to remove irrelevant characters and stop words. Irrelevant characters include punctuation marks and special symbols, which are not substantially helpful for keyword extraction. Stop words refer to those words with extremely high frequencies but little contribution to the text theme, including "de", "shi", "zai". The cleaned text will be purer, which is beneficial for subsequent word segmentation and keyword extraction;

[0067] Perform word segmentation on the cleaned text, splitting it into individual words or phrases. The quality of word segmentation directly affects the accuracy of subsequent keyword extraction. Therefore, when selecting a word segmentation tool, its word segmentation accuracy, speed, and whether it supports custom dictionaries need to be considered; in this embodiment, the Jieba word segmentation tool is selected. Jieba word segmentation has high word segmentation accuracy and can accurately split Chinese text into individual words or phrases. It is fast and can meet the needs of large-scale text processing. At the same time, Jieba word segmentation supports custom dictionaries, which is very helpful for text processing in specific fields. In the text related to the construction process, if there are some professional terms or specific vocabulary, the accuracy of word segmentation can be improved through custom dictionaries.

[0068] The segmented words are used as nodes of the graph, and the edges are constructed according to the co-occurrence relationship between the words, and the TextRank algorithm is used to extract key parameters. The graph refers to an undirected weighted graph, and the nodes are composed of the words after cleaning and segmentation. Each word represents a node, and the edges are constructed according to the co-occurrence relationship between the words. If two words co-occur within the set window size, there is an edge between them. The weight of the edge is assigned according to the frequency of co-occurrence. The higher the co-occurrence frequency, the greater the weight of the edge, which means that the correlation between the two words is stronger. The graph constructed in this way can effectively use the TextRank algorithm to calculate the weight of the node, thereby extracting keywords in the text.

[0069] In the process of using segmented words as nodes of the graph, the co-occurrence relationship between words is defined by setting a reasonable window size. The window size determines which words in the text are considered to be related to each other. In the process of building edges, weights are assigned to edges according to the frequency of co-occurrence. The higher the weight, the stronger the association between the two words. An initial weight is assigned to each node in the graph. This initial weight is set according to the frequency of the word appearing in the text. The way the initial weight is allocated will affect the convergence speed of the TextRank algorithm and the final keyword extraction results.

[0070] The TextRank algorithm is a graph-based sorting algorithm that uses the connection relationship between nodes to calculate the weight of the node. In each iteration, the algorithm updates the weight of the current node based on the weight of the node's neighbor nodes and the edge weight between them. This process is repeated until the node weight converges to a stable state. The nodes are sorted according to the node weights calculated by the TextRank algorithm. The higher the weight of the node, the higher its importance in the text, and it is the required keyword. Finally, several nodes with the highest weight are selected as keyword output. These keywords will be used as clues for subsequent searches for related drawings from the drawing library.

[0071] The formula of TextRank keyword extraction algorithm is:

[0072]

[0073] In the formula, V represents the node, W(V i ) represents the weight value of node i, d represents the learning rate, V i represents the i-th node, V j Indicates the jth node, In(V i ) represents the incoming link set of the i-th node, w ji represents the weight value between the jth node and the ith node, V k Indicates the window size k of the node, that is, at most k nodes co-occur, Out(Vj ) represents the outbound link set of the jth node, w jk represents the weight value between the jth node and the window size k, W(V j ) represents the weight value of the jth node.

[0074] Step S3, using the keywords to search in the metadata tags of the drawing library, by comparing the matching degree between the keywords and the metadata tags, quickly screen out the drawing files that are highly relevant to the current construction process. This process greatly reduces the time cost of manual screening of drawings and improves the efficiency of information extraction.

[0075] In addition, in step S3, the drawing library includes all PDF drawing files in the project. This drawing library is the core of project information management and contains all design, construction and material description drawings required during the project execution. In order to facilitate subsequent information retrieval and extraction, each drawing file is given a detailed metadata tag. These tags include but are not limited to the drawing name, the project phase, the associated tasks, the bill of materials, and the design version number. The use of metadata tags provides a basis for quickly locating PDF drawings. Therefore, in the screening process, it is only necessary to compare the matching degree of keywords and metadata tags to quickly screen out PDF drawings that are highly relevant to the current construction progress.

[0076] Step S4, for the selected PDF drawing files that are highly relevant to the current construction process, OCR technology is used to convert the image text in the PDF into editable text information, and the text information is pre-processed.

[0077] Specifically, in this embodiment, step S4 further includes:

[0078] Step S41, use the PyMuPDF library in the drawing library to parse the file and extract the vector data of the layers in the drawing. PyMuPDF is a powerful PDF processing library that can read and parse the page content in the PDF file, including text, images and vector graphics. For the drawing file, the main focus is on the vector layer data, which is usually used to describe the precise graphics in the fields of architecture, machinery or other engineering.

[0079] Step S42, the vector data in the drawing is rendered into an image through a graphics processing library, especially when image preprocessing technology is required, this step helps to convert the vector data into a format that is easier to process. Subsequently, the image is preprocessed, including scaling, cropping and denoising. The image is scaled as needed for subsequent processing, unnecessary parts of the image are removed, and only the areas related to the current extraction task are retained. The filter technology is used to remove noise in the image, improve the image quality, and lay the foundation for subsequent text recognition and other analysis tasks.

[0080] Step S43, input the preprocessed image into the pretrained image classification model to identify the component type represented by the layer; then, use the coordinate and size information in the vector data, combined with the result of image classification, to further determine the specific position and size of each component in the drawing. The vector data provides accurate geometric information, while the image classification provides identification of the component type. The two are combined to build a comprehensive database containing component types, positions, and sizes; wherein, the image classification model adopts the PyTorch deep learning framework, implements the Inception module by defining the Inception class, uses the defined Inception module to build the image classification model, and uses a large-scale data set to train the image classification model; after the training, input the processed image into the image classification model, and obtain the model's prediction result for the image content; according to the task requirements, create a mapping table from the model output category number to the specific component type name, use the model predicted category number as an index, and query the mapping table to obtain the corresponding component type name.

[0081] Subsequently, a series of preprocessing operations are performed on the text, including noise removal, format standardization, and word segmentation. These preprocessing steps lay the foundation for the subsequent key parameter extraction.

[0082] Step S44, after determining the position and size of the component, use OCR technology to extract the text information in the drawing. OCR technology can recognize the text in the image and convert it into an editable text format. For PDF drawings, the text information usually contains important design parameters, material descriptions, and construction requirements. Through OCR technology, these text information are associated with the previously extracted component data to form a complete drawing information database.

[0083] In step S44, the process of associating these text information with the previously extracted component data is as follows: the type information of each component in the drawing is obtained through the previous processing flow, and the vector data of the component is extracted from the drawing file. The vector data contains the precise geometric information of the component, which is crucial for the subsequent text information association.

[0084] Next, use an open source OCR engine such as Tesseract OCR to perform text recognition on the preprocessed image. The preprocessing steps include grayscale, binarization, denoising, and contrast enhancement to improve the OCR recognition accuracy. Tesseract can recognize the text in the image and convert it into an editable text format. During the recognition process, special attention is paid to the text information related to the component, including the name, size, and material description of the component, which are usually located near the component or in the annotations directly related to it.

[0085] According to the coordinates and size information of the components in the vector data, the corresponding component position is found in the image, and the component name and size recognized by OCR are matched with the components in the vector data to ensure that the corresponding text information can be found for each component; the vector data, OCR recognition results and image classification results are stored in the database, and the association relationship between the data tables is established.

[0086] Finally, the text information extracted by OCR is associated with the component data, which usually involves matching the keywords in the text information with the identifiers in the component data. Through this association, all relevant information of a component can be quickly queried, including its design parameters and construction requirements, providing strong support for project management and construction.

[0087] Step S5, extracting key parameters from the preprocessed text according to predefined parameter extraction rules. These key parameters include but are not limited to material name, type, specification and specific numerical requirements, which are crucial for guiding project execution and ensuring material procurement and construction standards; then the rationality of the acquired key parameters is analyzed to further improve the accuracy of subsequent analysis.

[0088] In step S5, according to the specific content of the drawings and project requirements, the system or project team predefines a set of parameter extraction rules. These rules specify which information is the key parameters to be extracted, how to accurately identify and extract these parameters from the text, and define regular expressions to match the text in the format of material names, specifications and numerical requirements. Finally, the system extracts the key parameters from the preprocessed text according to these rules and organizes them into a format that is easy to understand and use for reference and decision-making by project managers. In the process of rationality analysis, the text information recognized by OCR or manually entered is first converted from unstructured free text form to a structured data format, which usually involves the use of data parsing, regular expression matching or natural language processing technology to identify and extract key information, and organize the original unstructured and relatively messy text information into a format with a clear structure, easy to understand and use, such as converting the free text describing material information into a structured data format containing clear fields such as material name, type, specification, and numerical requirements. At the same time, it is necessary to correct spelling errors in the text to ensure the accuracy of the data, unify the units, avoid unit conflicts between data from different sources, and remove redundant information to improve data processing efficiency;

[0089] In order to evaluate the rationality of the extracted parameters, a set of reliable standard parameters is needed as a reference. These standard parameters can be obtained from relevant industry standards, design specifications, technical manuals or official databases. During the acquisition process, the applicability and update of the selected standards should be ensured to reflect the current best practices and technical requirements in the industry.

[0090] Compare the sorted parameters with the parameters obtained from the standard one by one. This step is to check whether the extracted parameters are within a reasonable range or meet specific technical requirements. During the comparison process, the tolerance range of the parameters needs to be considered to cope with minor deviations in actual production. Parameters that do not meet the requirements should be marked and recorded for further analysis or correction.

[0091] Verification is performed based on the intrinsic relationship between material performance parameters. This usually involves using calculation formulas in the fields of physics, chemistry or engineering to check the rationality and consistency of the parameters, using the strength, density and size of the material to calculate its load-bearing capacity and comparing it with the design requirements. In this way, the rationality of the parameters can be evaluated more deeply and potential errors or inconsistencies can be discovered.

[0092] In this embodiment, it also includes evaluating the abnormal risks in the task execution process based on the rationality analysis results and providing suggestions based on the current progress according to the abnormal risk evaluation results. Specifically, in the process of evaluating the abnormal risks in the task execution process, firstly, in-depth research is conducted on the project plan to clarify the goals, milestones, dependencies and critical paths of each stage, and by comparing historical project data, common risk types in previous projects are identified. At the same time, close attention is paid to the current environment, including changes in the market environment, adjustments to policies and regulations, natural disaster warnings, as well as internal team dynamics and the speed of technological innovation, so as to comprehensively identify the risk points that affect project execution;

[0093] Secondly, for each risk point, a combination of qualitative and quantitative methods is used to assess the possibility of its occurrence. Qualitative assessment is a subjective judgment of the probability of risk occurrence based on experience and professional knowledge through expert judgment and group discussion, and the probability of risk occurrence is quantified by collecting historical data and conducting statistical analysis;

[0094] Thirdly, use the risk assessment matrix to classify risks according to probability and impact. The risk assessment matrix is ​​a tool used to classify risks according to probability and impact. It assesses risks from two key dimensions: the probability of risk occurrence and the impact of risk once it occurs. This two-dimensional classification method can more clearly present the overall situation of risk. As the project progresses, the execution phase of the task will change. Therefore, it is necessary to dynamically adjust the focus of risk assessment, pay close attention to the specific activities of the current execution phase, identify and pay special attention to those risk points that are closely related to the current phase, which will help the project team take timely measures to deal with risks, avoid the expansion of problems, and ensure that the project proceeds as planned.

[0095] Finally, the key parameter information extracted from the PDF drawings is checked and analyzed in detail, and compared with the project design requirements, industry standards and specifications to evaluate whether they meet the requirements. At the same time, the correlation between these key parameter information and whether changes in material specifications will affect the cost budget are also considered to identify potential risk points. Any non-conformities or potential risks found should be recorded and reported to the relevant departments or personnel in a timely manner so that necessary corrective or preventive measures can be taken.

[0096] In the process of giving suggestions based on the current progress according to the abnormal risk assessment results, the identified risk points are ranked according to their probability of occurrence and the degree of impact on the project according to the results of the risk assessment, and clearly divided into three levels: high, medium and low. High risk usually refers to those risks with a high probability of occurrence and a significant negative impact on the project, medium risk refers to those risks with a moderate probability of occurrence or a certain impact on the project, and low risk refers to those risks with a low probability of occurrence and a small impact. Such classification helps the project team clearly identify the risk points that need to be paid attention to and addressed first;

[0097] Formulating corresponding emergency plans for different levels of risks is the key to reducing the impact of risks and ensuring the smooth progress of projects. For high risks, detailed response measures and plans should be formulated, including emergency response processes, backup plans, and resource allocation plans to ensure that risks can be responded to quickly and effectively when they occur. For medium risks, more flexible response strategies should be formulated, including enhanced monitoring, regular evaluation, and timely adjustment of plans to control the development of risks. For low risks, general monitoring and management measures are taken to keep an eye on risks; the schedule is adjusted based on the risk assessment results, and resources are reallocated based on the resource bottlenecks found in the risk assessment.

[0098] Embodiment 2

[0099] A PDF drawing key parameter information extraction system, comprising:

[0100] The execution process acquisition module is configured to: acquire the overall cycle of the project task based on the project plan; determine the execution process of the current task according to the current time point and the task cycle;

[0101] The keyword extraction module is configured to: obtain a detailed text description of the current construction process based on the execution process, and extract keywords from the text description using a TextRank algorithm;

[0102] The PDF drawing screening module is configured to: use the keyword to search in the metadata tags of the drawing library, and quickly screen out the drawing files that are highly relevant to the current construction process by comparing the matching degree between the keyword and the metadata tags; the drawing library includes all PDF drawing files in the project, and each PDF drawing file is added with metadata tags;

[0103] The image and text conversion module is configured to: for the selected PDF drawing files that are highly relevant to the current construction process, use OCR technology to convert the images and texts in the PDF into editable text information, and pre-process the text information;

[0104] The key parameter extraction module is configured to extract key parameters from the preprocessed text according to predefined parameter extraction rules, and perform rationality analysis on the obtained key parameters.

[0105] Embodiment 3

[0106] The purpose of this embodiment is to provide a computer-readable storage medium.

[0107] A computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the steps in a method for extracting key parameter information of a PDF drawing as described in Example 1 of the present disclosure.

[0108] Embodiment 4

[0109] The purpose of this embodiment is to provide an electronic device.

[0110] An electronic device includes a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, the steps in a PDF drawing key parameter information method as described in Embodiment 1 of the present disclosure are implemented.

[0111] The steps involved in the apparatuses of the above embodiments 2, 3 and 4 correspond to the method embodiment 1, and the specific implementation methods can refer to the relevant description part of embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood to include any medium that can store, encode or carry an instruction set for execution by a processor and enable the processor to execute any method in the present invention.

[0112] Those skilled in the art should understand that the modules or steps of the present invention described above can be implemented by a general-purpose computer device, or alternatively, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. The present invention is not limited to any specific combination of hardware and software.

[0113] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A method for extracting key parameter information of PDF drawings, characterized in that: include: Obtain the overall cycle of project tasks based on the project plan; determine the execution progress of the current task based on the current time point and task cycle; Based on the execution process, a detailed text description of the current construction process is obtained, and keywords are extracted from the text description using the TextRank algorithm; The keyword is used to search in the metadata tags of the drawing library, and the drawing files that are highly relevant to the current construction process are quickly screened out by comparing the matching degree between the keyword and the metadata tags; the drawing library includes all PDF drawing files in the project, and each PDF drawing file is added with metadata tags; For the selected PDF drawing files that are highly relevant to the current construction process, OCR technology is used to convert the image text in the PDF into editable text information, and the text information is pre-processed; According to the predefined parameter extraction rules, key parameters are extracted from the preprocessed text, and the rationality of the obtained key parameters is analyzed.

2. A method for extracting key parameter information of a PDF drawing as claimed in claim 1, characterized in that: The process of obtaining a detailed text description of the current construction process based on the execution process and extracting keywords from the text description using the TextRank algorithm is as follows; Obtain a detailed text description of the current construction process; perform cleaning and word segmentation on the text description; The segmented words are used as nodes of the graph, the co-occurrence relationship between words is defined by setting the window size, edges are constructed based on the co-occurrence relationship between words, and weights are assigned to edges based on the co-occurrence frequency; Assign an initial weight to each node in the graph; The TextRank algorithm is executed iteratively. In each iteration, the weight of the current node is updated according to the weights of the node's neighbor nodes and the edge weights between them until the node weights converge to a stable state. The nodes are sorted according to the final calculated node weights, and several nodes with the highest weights are selected as keywords.

3. A method for extracting key parameter information of a PDF drawing as claimed in claim 2, characterized in that: The TextRank algorithm is: In the formula, V represents the node, W(V i ) represents the weight value of node i, d represents the learning rate, V i represents the i-th node, V j Indicates the jth node, In(V i ) represents the incoming link set of the i-th node, w ji represents the weight value between the jth node and the ith node, V k Indicates the window size k of the node, that is, at most k nodes co-occur, Out(V j ) represents the outbound link set of the jth node, w jk represents the weight value between the jth node and the window size k, W(V j ) represents the weight value of the jth node.

4. A method for extracting key parameter information of a PDF drawing as claimed in claim 1, characterized in that: The process of using OCR technology to convert image text in PDF into editable text information and preprocessing the text information is as follows: Use the PyMuPDF library in the drawing library to parse the file and extract the vector data of the layers in the drawing; Render the vector data into images through the graphics processing library and preprocess the images, including scaling, cropping and denoising; Input the preprocessed image into the pre-trained image classification model to identify the component type represented by the layer; Using the coordinate and size information in the vector data, combined with the results of image classification, the specific position and size of each component are determined, the text information is extracted using OCR technology, the text information is preprocessed, and the preprocessed text information is associated with the component data; The text information preprocessing process includes noise removal, format standardization and word segmentation.

5. A method for extracting key parameter information of a PDF drawing as claimed in claim 4, characterized in that: The process of inputting the processed image into the pre-trained image classification model to identify the component type represented by the layer is as follows: The image classification model uses the PyTorch deep learning framework to implement the Inception module by defining the Inception class, uses the defined Inception module to build the image classification model, and uses a large-scale data set to train the image classification model; After the training is completed, the processed image is input into the image classification model, and the model's prediction results on the image content are obtained; According to the task requirements, create a mapping table from the model output category number to the specific component type name; Use the category number predicted by the model as the index to query the mapping table to obtain the corresponding component type name.

6. A method for extracting key parameter information of a PDF drawing as claimed in claim 4, characterized in that: The process of extracting text information using OCR technology, preprocessing the text information, and associating the preprocessed text information with component data is as follows: Based on the acquired component vector data and the component type of the image; Use Tesseract OCR to perform text recognition on the preprocessed image and extract the name and size of the component; According to the coordinates and size information of the components in the vector data, the corresponding component position is found in the image, and the component name and size recognized by OCR are matched with the components in the vector data to ensure that the corresponding text information can be found for each component; The vector data, OCR recognition results, and image classification results are stored in the database, and the association relationship between the data tables is established.

7. A method for extracting key parameter information of a PDF drawing as claimed in claim 1, characterized in that: The process of analyzing the rationality of the acquired key parameters is as follows: Obtain standard parameters from industry standards, design specifications and technical manuals; Compare the extracted key parameter information with the standard parameters to check whether they meet the requirements.

8. A PDF drawing key parameter information extraction system, characterized in that: include: The execution process acquisition module is configured to: acquire the overall cycle of the project task based on the project plan; determine the execution process of the current task according to the current time point and the task cycle; The keyword extraction module is configured to: obtain a detailed text description of the current construction process based on the execution process, and extract keywords from the text description using a TextRank algorithm; The PDF drawing screening module is configured to: use the keyword to search in the metadata tags of the drawing library, and quickly screen out the drawing files that are highly relevant to the current construction process by comparing the matching degree between the keyword and the metadata tags; the drawing library includes all PDF drawing files in the project, and each PDF drawing file is added with metadata tags; The image and text conversion module is configured to: for the selected PDF drawing files that are highly relevant to the current construction process, use OCR technology to convert the images and texts in the PDF into editable text information, and pre-process the text information; The key parameter information extraction module is configured to extract key parameters from the preprocessed text according to predefined parameter extraction rules, and perform rationality analysis on the obtained key parameters.

9. A computer-readable storage medium having a program stored thereon, characterized in that: When the program is executed by a processor, the steps in the method for extracting key parameter information of a PDF drawing as described in any one of claims 1 to 7 are implemented.

10. An electronic device comprising a memory, a processor, and a program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps in the method for extracting key parameter information of a PDF drawing as described in any one of claims 1-7 are implemented.

Images