Method and system for rapidly extracting archive information

Abstract

The invention discloses an archive information rapid extraction method and system, and relates to the technical field of data extraction. Comprising the following steps: uniquely identifying a file by using an RFID tag or a bar code, and extracting data of the file after identification is completed; performing real-time processing and pre-analysis on the archive data by utilizing an edge computing technology; after the network connection is recovered, automatically carrying out deep analysis on the extracted content, and carrying out dynamic classification according to the types of the archives; changes in the archive system are automatically detected, when new archive information is scanned or modified, an updating process is automatically triggered, and the newest archive information is uploaded. According to the invention, through dynamic management and intelligent analysis, the overall efficiency of the system and the user experience are improved, the labor cost is reduced, and the working efficiency and the data security are improved.

Description

Technical Field

[0001] The invention relates to the technical field of data extraction, in particular to a method system for quickly extracting archive information. Background Art

[0002] In recent years, with the rapid development of information technology, the digital transformation of archive management has gradually become an important trend in various industries. Traditional archive management methods often rely on paper documents and manual operations, which are inefficient and prone to errors. Therefore, archive information extraction methods based on barcodes and RFID (radio frequency identification) technologies have emerged. These technologies make it possible to uniquely identify and quickly read archives, and improve the automation and information level of archive management. In particular, RFID technology, which uses radio waves for non-contact identification, can not only achieve fast scanning, but also work effectively in complex environments.

[0003] However, the existing technologies still have many shortcomings in practical applications. For example, although RFID and barcode technologies perform well in archival identification, they are often restricted by network stability and data transmission speed during the archival information extraction process. When the network connection is unstable or disconnected, traditional extraction methods are difficult to ensure timely processing and updating of data, resulting in delayed archival information and inability to meet real-time application needs. Summary of the invention

[0004] In view of the problems existing in the prior art, the present invention is proposed.

[0005] Therefore, the problem to be solved by the present invention is how to solve the delay and information inconsistency problems existing in traditional archive management.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: In the first aspect, an embodiment of the present invention provides a method for quickly extracting archival information, which includes using RFID tags or barcodes to uniquely identify archives, and extracting archival data after identification; using edge computing technology to perform real-time processing and pre-analysis of archival data; after the network connection is restored, automatically performing in-depth analysis of the extracted content and dynamically classifying it according to the type of archive; automatically detecting changes in the archival system, and automatically triggering an update process when new archival information is scanned or modified, and uploading the latest archival information.

[0007] As a preferred solution of the method for rapid extraction of archival information described in the present invention, the real-time processing and pre-analysis of archival data include: the extracted archival data will be divided into multiple small blocks for storage; when the network is disconnected, all archival data will be preferentially stored in the local memory of the handheld device, and a timestamp will be generated for each piece of data to record the storage time and the unique identifier of the source archive; if an interruption or failure occurs during the storage process, the system will automatically perform data verification; a dynamic cache area is set to preferentially store recently extracted and accessed archival data, and the cache area is dynamically managed using edge computing technology.

[0008] As a preferred solution of the method for rapid extraction of archive information of the present invention, the cache area is dynamically managed by edge computing technology, including: assigning a priority score to each archive based on the frequency of archive access, the last access time and the data size :

[0009] in, is the access frequency of the archive, is the time since the last visit, is the data size of the archive, , and Adjust the weight factors respectively; according to each file Value, first clean up low-priority files, where the threshold of the cleanup operation is set as:

[0010] in, is the average priority of the files in the cache, is the cleaning coefficient.

[0011] As a preferred solution of the method for rapid extraction of archival information described in the present invention, the automatic in-depth analysis of the extracted content and dynamic classification according to the type of archives include the following steps: the extracted archival data is cleaned and preprocessed through natural language processing; feature extraction is performed on the preprocessed text data; the archival data is trained and classified using a machine learning algorithm; after the newly extracted archival data has been preprocessed and feature extracted, the archives are dynamically classified according to the classification model; based on the classification results, the accuracy of the classification is verified through a verification mechanism.

[0012] As a preferred solution of the method for rapidly extracting archive information of the present invention, the feature extraction of the pre-processed text data comprises the following steps: for each archive, a feature vector is generated. , represents the content characteristics of the text. Assume that the extracted archive data contains m archives, and the system generates a feature vector for each archive:

[0013] The eigenvalues ​​are calculated by the TF-IDF algorithm:

[0014]

[0015] Where N is the total number of documents, For specific words in the archive, For the current file, represents the weight of feature j in file i.

[0016] As a preferred solution of the method for rapid extraction of archival information of the present invention, the verification of the accuracy of classification by the inspection mechanism includes: adopting a hierarchical classification strategy, including: the first layer of coarse-grained classification: classifying archival data according to major categories, using loose thresholds; the second layer of fine-grained classification: on the basis of coarse classification, more detailed classification is performed through contextual semantic analysis or embedded similarity, specifically including: extracting keywords from archival data, assuming that the keyword set is K, and the keyword vector of each archive is , then the similarity of the comparison is calculated as:

[0017] Through natural language processing technology, the contextual semantics of the archive content is analyzed. If the semantics of the newly classified archive is inconsistent with the typical semantics of the category, it is automatically corrected and the corrected label is fed back to the classification model: Let the classification error of the current model be E. When an incorrect classification is found, the model parameter θ is adjusted through the feedback mechanism:

[0018] in, is the learning rate, is the gradient of the classification error.

[0019] As a preferred solution of the method for rapid extraction of archival information described in the present invention, the automatic triggering of the update process when new archival information is scanned or modified includes: when new archival information is scanned, recording its unique identification, creation time and content summary; when existing archival information is changed, recording the changed archival unique identification and modification time, and performing a difference comparison between the new and old archival contents to generate a change log; when new or modified archival information is detected, the update process is automatically triggered, and the update interface is called in an event-driven manner to prepare the changed archival information for uploading to a central server or an archival management system; the triggering of the update process is based on the following conditions: after the network connection is restored, the upload operation is automatically triggered, and if the network is unstable or interrupted, the update operation is put into a queue to be processed and retried after the network is restored; the changed archival information is data verified and packaged into an upload file to generate a data packet.

[0020] In the second aspect, an embodiment of the present invention provides a system for rapid extraction of archival information, which includes: an archive identification module, which is used to uniquely identify archives and extract relevant data; an edge computing module, which is used to use edge computing technology to perform real-time processing and pre-analysis of archival data; a deep analysis module, which is used to automatically perform deep analysis on the extracted content after the network connection is restored, and dynamically classify the content according to the type of archive; an automatic update module, which is used to automatically detect changes in the archive system, and when new archival information is scanned or modified, automatically trigger the update process to upload the latest archival information.

[0021] In a third aspect, an embodiment of the present invention provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of the method for quickly extracting archival information as described in the first aspect of the present invention are implemented.

[0022] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of the method for rapidly extracting archival information as described in the first aspect of the present invention are implemented.

[0023] The beneficial effects of the present invention are as follows: the present invention ensures that archives can still be extracted and accessed in the case of unstable network, avoiding the problem of information being unable to be obtained due to network failure, and setting a dynamic cache area to manage the storage of archive data according to the access frequency, thereby improving the access efficiency of data, and optimizing cache management to ensure that commonly used data can be quickly called; adopting coarse-grained and fine-grained hierarchical classification strategies to effectively reduce the possibility of misclassification, ensure that complex content archives can be reasonably classified and processed, and improve management efficiency; after detecting a classification error, it will automatically correct it and feed back to the model for optimization, forming a closed-loop feedback mechanism, and further improving the intelligence level of the system; detecting changes in the archive system through an intelligent perception mechanism ensures timely updating of archive information. The present invention improves the overall performance and user experience of the system through dynamic management and intelligent analysis, reduces labor costs, and improves work efficiency and data security. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0025] Figure 1 Flowchart of the method for rapid extraction of archival information.

[0026] Figure 2 This is a structural diagram of the system for rapid extraction of archival information. DETAILED DESCRIPTION

[0027] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0028] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0029] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0030] Example 1 Reference Figure 1~Figure 2, which is the first embodiment of the present invention, and provides a method for quickly extracting archive information, such as Figure 1 As shown, including, S1: Use RFID tags or barcodes to uniquely identify archives, achieve contactless rapid identification and information acquisition, and after identification, extract the full text information of the archives, including text, pictures and charts.

[0031] Specifically, open the archive management application on the mobile terminal, bring the RFID extractor or barcode scanner close to the archive, trigger the scanning operation (which may be a button or automatic trigger), extract the RFID tag or scan the barcode, obtain the unique identifier, and search for the corresponding archive information in the system database based on the unique identifier.

[0032] Integrate the extracted text, image and graphic information into a unified digital document, maintaining the structure and layout of the original archive.

[0033] S2: Use edge computing technology to perform real-time processing and pre-analysis of archival data on handheld devices. At the same time, the extracted data is temporarily stored in the local storage space of the device to ensure that archive extraction and access can still be performed when offline or the network is unstable.

[0034] Specifically, after the archival data is extracted, the handheld device starts the local storage management system and allocates space for the stored archival information. At this time, the system will create a corresponding local storage path based on the unique identifier of the archive (obtained by RFID or barcode scanning in step S1) to ensure that each piece of archival data can be managed separately to avoid conflicts.

[0035] In order to optimize storage space and speed up the extraction, the extracted archive data (text, pictures, charts) will be divided into multiple small blocks for storage. Text data can be stored in compressed text format, and pictures and charts are compressed and stored according to their formats. This process is performed in real time through edge computing devices to ensure that the storage operation does not affect the extraction speed of the archive.

[0036] When the network is disconnected, all archive data will be stored in the local storage of the handheld device first, and a timestamp will be generated for each piece of data to record the storage time and the unique identifier of the source archive.

[0037] If an interruption or failure occurs during the storage process, the system will automatically perform data verification (such as checksum or hash algorithm) to ensure the integrity of the stored file, and will automatically retry the storage operation during the next recovery.

[0038] Furthermore, a dynamic cache is set up locally on the device to prioritize the storage of recently extracted and accessed archive data. The cache is dynamically managed using edge computing technology. When the storage space approaches the upper limit, the system will automatically clean up or compress less used data based on the frequency of archive access. The specific operations are as follows: To optimize cache management, each archive is assigned a priority score based on its access frequency, last access time, and data size. , used to dynamically manage the retention of cached data:

[0039] in, is the access frequency of the archive, is the time since the last visit, is the data size of the archive, , and are used to adjust the weight factors respectively.

[0040] When the cache space is close to the upper limit, the value of each file is calculated according to the above formula. Value, first clean up low-priority files, where the threshold of the cleanup operation is set as:

[0041] in, is the average priority of the files in the cache, is the cleaning coefficient, which is usually in the range of [0.5, 0.8] to control the strictness of the cleaning.

[0042] In the case of unstable or completely disconnected network, users can still access and process the extracted archive information through local storage space. At this time, edge computing technology will continue to support data pre-analysis and processing operations, such as local text retrieval, simple image recognition, etc., and build a temporary index based on the locally stored archive information to support users to quickly search, classify and preview when offline.

[0043] S3: After the network connection is restored, the extracted content is automatically analyzed in depth through natural language processing (NLP) and machine learning algorithms, and dynamically classified according to the type of archive (such as contracts, financial reports, technical documents, etc.).

[0044] S3.1: The extracted archival data is first cleaned and preprocessed through NLP, mainly including: Process the archive text data through the word segmentation algorithm to generate a vocabulary set , where each Represents a word or phrase in text.

[0045] Filter stop word set , retain only the key content and generate the processed vocabulary set ,in:

[0046] The processed text is tagged with parts of speech to extract important information such as key nouns, verbs, time, etc. for subsequent classification.

[0047] S3.2: Perform feature extraction on the preprocessed text data through the bag-of-words model or TF-IDF.

[0048] For each file, generate a feature vector , represents the content features of the text. Assuming that the extracted archive data contains m archives, the system generates a feature vector for each archive:

[0049] The eigenvalues ​​are calculated by the TF-IDF algorithm:

[0050]

[0051] Where N is the total number of documents, For specific words in the archive, For the current file, represents the weight of feature j in file i.

[0052] S3.3: After extracting features, use machine learning algorithms to train and classify archival data.

[0053] First, a training set is established through historical archive data. Each archive data will be used as a sample. The feature vector representing the sample is used to label the archive data according to the archive type and generate the label set corresponding to the training set.

[0054] Use supervised learning algorithms to train data and build a classification model f(X). Common machine learning algorithms include support vector machine (SVM), random forest, K-nearest neighbor (KNN) or neural network. The goal of classification model training is to minimize classification error:

[0055] Where m is the number of training samples, For the model to sample The predicted output is is the true label of the sample, and L is the loss function, such as cross entropy loss or mean square error.

[0056] By iteratively training the model, the model parameters are continuously adjusted to minimize the classification error, thereby improving the classification accuracy.

[0057] S3.4: After the newly extracted archive data has been preprocessed and feature extracted, the system dynamically classifies the archives according to the classification model. The feature vector of each archive data The predicted labels are obtained through classification model calculations. The classification results are stored in the classification result set. The classification results can be updated dynamically. The system archives and classifies the archival data according to its predicted labels.

[0058] S3.5: Based on the classification results, the accuracy of the classification is verified through a verification mechanism.

[0059] Preferably, the correctness of the classification is verified by keyword comparison of the archives or consistency check based on contextual semantics. If an erroneous classification is found, the system can automatically correct it and feed it back to the classification model to optimize and update the classification model.

[0060] Preferably, in order to improve the accuracy of classification, a hierarchical classification strategy can be adopted, that is, firstly classify the archives in a coarse-grained manner, and then further classify them in a fine-grained manner, including: The first level of coarse-grained classification: archival data is classified into broad categories (such as contracts, financial reports, technical documents, etc.), using relatively loose thresholds to ensure that all possible classifications are considered; The second level of fine-grained classification: Based on the coarse classification, more detailed classification is performed through methods such as contextual semantic analysis or embedding similarity, which can gradually narrow the scope and more accurately determine the specific subcategory to which the archive belongs, including: First, keywords are extracted from the archive data and compared with the typical keyword sets of each category. Assuming that the keyword set is K, the keyword vector of each archive is , then the similarity measure of the comparison can be calculated using cosine similarity:

[0061] Hierarchical classification can significantly reduce the possibility of misclassification, especially for archives with complex or ambiguous content, hierarchical classification can improve overall accuracy.

[0062] Furthermore, the contextual semantics of the archive content is analyzed through natural language processing (NLP) technology. If the semantics of the newly classified archive is obviously inconsistent with the typical semantics of the category, it will be suspected that the classification may be wrong.

[0063] If the system finds that a certain file is misclassified, it will automatically correct it and feed the corrected label back to the classification model to optimize the model parameters and update the classification rules. The optimized model will be reapplied to subsequent classification tasks. Assuming that the classification error of the current model is E, when an incorrect classification is found, the system adjusts the model parameter θ through the feedback mechanism to further reduce the classification error:

[0064] in, is the learning rate, is the gradient of the classification error.

[0065] It can be seen that in this step, the archival data is dynamically classified through machine learning algorithms, and combined with verification mechanisms such as keyword comparison and semantic analysis, the classification model is continuously optimized to ensure the accuracy and efficiency of classification. The process of model training, classification and verification can be adaptively adjusted to provide intelligent automatic classification functions for the archive management system.

[0066] S4: Automatically detect changes in the archive system (such as adding or modifying archives) through intelligent perception mechanisms. When new archive information is scanned or modified, the update process is automatically triggered to upload the latest archive information to the central server or archive management system.

[0067] It adopts an intelligent perception mechanism based on file change monitoring. By monitoring the status changes of archive files in handheld devices or local storage space, including addition, modification, deletion and other operations, this mechanism is combined with event-driven file system monitoring tools (such as inotify or fswatch) to monitor local file directories in real time to ensure that any changes in archive files are perceived at the first time.

[0068] Among them, the intelligent perception mechanism can identify different types of file changes, mainly including: When new archive information is scanned or manually added to the system, the system detects the new archive and records its unique ID, creation time, and content summary.

[0069] When existing archive information changes (such as content updates or metadata changes), the system records the changed archive unique identifier and modification time, compares the new and old archive contents, and generates a change log.

[0070] When the system detects new or modified archival information, the intelligent perception mechanism will automatically trigger the update process. Through event-driven means, it will call the update interface and prepare to upload the changed archival information to the central server or archival management system. This process does not require manual intervention by the user and is automatically executed by the system background.

[0071] The update process can be triggered based on the following conditions: The system automatically triggers the upload operation after the network connection is restored. If the network is unstable or interrupted, the system will put the update operation into the pending queue and retry after the network is restored; Perform data verification (such as hash verification) on the changed archive information and package it into an upload file to generate a data package.

[0072] Furthermore, once the upload process is triggered, the system uploads the latest archive information to the central server or archive management system. The upload process is as follows: Before uploading the data, the system will perform an integrity check on the archive data and generate a checksum to ensure that the uploaded data is not lost or damaged. After the upload is completed, the server will recalculate the checksum for the uploaded data and compare it with the local checksum to verify the accuracy of the data transmission. The uploaded file information will be synchronized to the server's file database. If a new file is added, the system will add a new record to the server database; if a file is modified, the system will overwrite and update the existing file data and save the modification history for traceability.

[0073] After the file upload and synchronization are completed, the system will feedback the upload results to the user end and update the file status mark of the local database (such as uploaded, synchronization completed). At the same time, the server will generate a synchronization log to record the upload time, data volume and verification results of the file to ensure the integrity and consistency of the file.

[0074] Furthermore, this embodiment also provides a system for quickly extracting archive information, comprising: The file identification module is used to uniquely identify the files and extract relevant data.

[0075] The edge computing module is used to utilize edge computing technology to perform real-time processing and pre-analysis of archival data; at the same time, the extracted data is temporarily stored in the local storage space of the device.

[0076] The deep analysis module is used to automatically perform deep analysis on the extracted content after the network connection is restored, and dynamically classify it according to the type of archive.

[0077] The automatic update module is used to automatically detect changes in the archive system. When new archive information is scanned or modified, the update process is automatically triggered to upload the latest archive information.

[0078] This embodiment also provides a computer device suitable for the method of quickly extracting archival information, including a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the method of quickly extracting archival information proposed in the above embodiment.

[0079] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0080] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for quickly extracting archive information as proposed in the above embodiment is implemented.

[0081] In summary, the present invention ensures that archives can still be extracted and accessed in the case of unstable network, avoiding the problem of information being unable to be obtained due to network failure, and setting a dynamic cache area to manage the storage of archive data according to the access frequency, improve the access efficiency of data, and optimize the cache management to ensure that commonly used data can be quickly called; adopting coarse-grained and fine-grained hierarchical classification strategies to effectively reduce the possibility of misclassification, ensure that complex content archives can be reasonably classified and processed, and improve management efficiency; after detecting classification errors, they will automatically correct them and feed back to the model for optimization, forming a closed-loop feedback mechanism, and further improving the intelligence level of the system; through the intelligent perception mechanism, changes in the archive system are detected to ensure the timely update of archive information. The present invention improves the overall performance and user experience of the system through dynamic management and intelligent analysis, reduces labor costs, and improves work efficiency and data security.

[0082] Example 2 Referring to Table 1, which is a second embodiment of the present invention, this embodiment provides a method for quickly extracting archive information. In order to verify the beneficial effects of the present invention, a scientific demonstration is carried out through simulation experiments.

[0083] The experiment selected the archives management department of a large enterprise as the test site and selected 1,000 archives of different types as test samples, including contracts, financial reports, technical documents, etc.

[0084] First, the experimenters followed step S1 in the invention content and affixed RFID tags or barcodes to 1,000 files, using professional RFID tag printers and barcode printers to ensure that each tag had a unique identification code. At the same time, the electronic versions of these files were entered into the database of the file management system, including text, pictures and chart information.

[0085] Next, the experimenters used a mobile terminal device equipped with an RFID reader and a barcode scanner to conduct a test of archival information extraction. A specially developed archive management application was installed on the mobile terminal, which implemented the functions described in the invention content. When performing archive extraction, the operator only needs to bring the device close to the archive to trigger the scanning operation, and the system can automatically identify the unique identifier of the archive and extract the corresponding archive information from the database.

[0086] In order to test the effect of edge computing technology, the experimenters installed a software module that supports edge computing on the mobile terminal, which can perform real-time processing and pre-analysis of the extracted archival data locally. At the same time, local storage space was set up to store the extracted archival information when offline or when the network is unstable.

[0087] The experiment also tested the system's performance when the network was disconnected. By manually disconnecting the network connection, an unstable network environment was simulated to test whether the system could normally perform file extraction and local storage. At the same time, the process of automatically synchronizing local storage data to the central server after the system restored the network connection was recorded.

[0088] In order to verify the intelligent classification function of the system, the experimenters prepared a pre-labeled archive data set for training the classification model. This dataset contains 5,000 archives of different types, each of which is manually labeled with the correct category. A classification model based on a support vector machine (SVM) was trained using this dataset. Then, the 1,000 test archives in the experiment were used to evaluate the classification performance of the model.

[0089] Finally, the experimenters tested the system's intelligent perception and automatic update functions by simulating operations such as adding new files and modifying existing files to observe whether the system can automatically detect these changes and trigger the update process in a timely manner.

[0090] During the experiment, key indicators such as the time and accuracy of each operation were recorded. At the same time, in order to compare the differences between the method of the present invention and the traditional method, the experimenters also used multiple existing manual file management methods to process the same files and recorded the corresponding data.

[0091] The following is a table of some experimental data: Table 1 Data comparison table Treatment Recognition time (seconds) Recognition accuracy (%) Data extraction time (seconds) Data integrity (%) Offline processing speed (documents / minute) Classification accuracy (%) Update response time (seconds) Traditional manual method 45.2 92.5 180.3 95.8 2.3 85.7 300.5 Barcode + manual classification 15.8 98.3 90.5 97.2 4.5 88.9 180.2 RFID+Machine Learning (this invention) 3.2 99.8 30.1 99.5 12.8 96.5 5.3 RFID+Simple Automation 5.6 99.5 45.7 98.7 8.2 92.3 30.6 OCR+human assistance 20.3 97.1 60.4 98.1 6.7 90.1 120.8 QR code + cloud processing 8.7 99.1 55.2 98.9 5.9 93.7 15.4 From the table above, we can see that in terms of recognition time, the method of the present invention only takes 3.2 seconds, which is much lower than the 45.2 seconds of the traditional manual method and the 15.8 seconds of the barcode method. This significant difference reflects the superiority of RFID technology in non-contact rapid recognition. At the same time, the recognition accuracy of the present invention reaches 99.8%, the highest among all methods, which not only ensures the accuracy of subsequent processing, but also greatly reduces the need for manual intervention.

[0092] In terms of data extraction time, the method of the present invention only takes 30.1 seconds, which is 6 times faster than the traditional manual method and 3 times faster than the barcode method. This efficient data extraction capability is due to the high-performance image acquisition equipment and optimized data processing algorithm integrated in the present invention. More importantly, the method of the present invention has achieved 99.5% data integrity, which means that almost all text, pictures and chart information are completely captured and processed, greatly improving the availability and reliability of archival information.

[0093] Comprehensive analysis of the table data shows that the rapid archival information extraction method of the present invention is significantly superior to the prior art in multiple key indicators such as recognition speed, accuracy, data extraction efficiency, offline processing capability, classification accuracy, and update response speed. These advantages can not only greatly improve the efficiency of archival management, but also significantly improve the integrity and availability of archival information. In particular, the outstanding performance in offline processing and rapid updating enables the method to adapt to various complex working environments and provide a comprehensive solution for modern archival management.

[0094] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for quickly extracting archive information, characterized in that: include, Use RFID tags or barcodes to uniquely identify the files, and after identification, extract the data of the files; Use edge computing technology to perform real-time processing and pre-analysis of archival data; After the network connection is restored, the extracted content is automatically analyzed in depth and dynamically classified according to the type of archive; Automatically detect changes in the archive system. When new archive information is scanned or modified, the update process is automatically triggered to upload the latest archive information.

2. The method for rapidly extracting archive information according to claim 1, characterized in that: The real-time processing and pre-analysis of archival data includes: The extracted archive data will be divided into multiple small blocks for storage; In the case of network disconnection, all archive data will be stored in the local storage of the handheld device first, and a timestamp will be generated for each piece of data to record the storage time and the unique identification of the source archive; If there is an interruption or failure during the storage process, data verification will be automatically performed; Set up a dynamic cache area to prioritize storage of recently extracted and accessed archival data, and dynamically manage the cache area using edge computing technology.

3. The method for rapidly extracting archive information according to claim 2, characterized in that: The cache area is dynamically managed using edge computing technology including: Assign a priority score to each archive based on how often it is accessed, when it was last accessed, and the size of the data : in, is the access frequency of the archive, is the time since the last visit, is the data size of the archive, , and are the adjustment weight factors respectively; According to each file Value, first clean up low-priority files, where the threshold of the cleanup operation is set as: in, is the average priority of the files in the cache, is the cleaning coefficient.

4. The method for rapidly extracting archive information according to claim 3, characterized in that: The automatic in-depth analysis of the extracted content and dynamic classification according to the type of archives includes the following steps: The extracted archival data were cleaned and preprocessed using natural language processing; Perform feature extraction on the preprocessed text data; Use machine learning algorithms to train and classify archival data; After the newly extracted archive data has been preprocessed and feature extracted, the archives are dynamically classified according to the classification model; Based on the dynamic classification results, the accuracy of the classification is verified through a verification mechanism.

5. The method for rapidly extracting archive information according to claim 4, characterized in that: The feature extraction of the pre-processed text data comprises the following steps: For each file, generate a feature vector , represents the content characteristics of the text. Assume that the extracted archive data contains m archives, and the system generates a feature vector for each archive: The eigenvalues ​​are calculated by the TF-IDF algorithm: Where N is the total number of documents, For specific words in the archive, For the current file, represents the weight of feature j in file i.

6. The method for rapidly extracting archive information according to claim 5, characterized in that: The verification of the accuracy of the classification by the verification mechanism includes: Adopt a hierarchical classification strategy, including: The first level of coarse-grained classification: classify the archival data into broad categories; The second layer of fine-grained classification: Based on the coarse classification, more detailed classification is performed through contextual semantic analysis or embedding similarity, including: Extract keywords from the archive data. Let the keyword set be K and the keyword vector of each archive be , the similarity is calculated as: Through natural language processing technology, the contextual semantics of the archive content is analyzed. If the semantics of the newly classified archive is inconsistent with the typical semantics of the category, it is automatically corrected and the corrected label is fed back to the classification model: Assume that the classification error of the current model is E. When an incorrect classification is found, the model parameter θ is adjusted through the feedback mechanism: in, is the learning rate, is the gradient of the classification error.

7. The method for rapidly extracting archive information according to claim 6, characterized in that: When new file information is scanned or modified, the update process is automatically triggered including: When new archival information is scanned, its unique identification, creation time and content summary are recorded; When existing archive information is changed, the unique identifier and modification time of the changed archive are recorded, and the difference between the new and old archive contents is compared to generate a change log; When new or modified archive information is detected, the update process is automatically triggered. In an event-driven manner, the update interface is called to prepare and upload the changed archive information to the central server or archive management system; The update process is triggered based on the following conditions: After the network connection is restored, the upload operation is automatically triggered. If the network is unstable or interrupted, the update operation will be put into the pending queue and retried after the network is restored; The changed archive information is verified and packaged into an upload file to generate a data package.

8. A system for rapid extraction of archival information, based on the method for rapid extraction of archival information according to any one of claims 1 to 7, characterized in that: Also includes, The file identification module is used to uniquely identify the file and extract relevant data; Edge computing module, used to perform real-time processing and pre-analysis of archival data using edge computing technology; A deep analysis module, which is used to automatically perform deep analysis on the extracted content and dynamically classify it according to the type of archive after the network connection is restored; The automatic update module is used to automatically detect changes in the archive system. When new archive information is scanned or modified, the update process is automatically triggered to upload the latest archive information.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for quickly extracting archival information described in any one of claims 1 to 7 are implemented.

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

Images