Intelligent file scanning and storing integrated cabinet

Abstract

The invention discloses an intelligent file scanning and storage integrated cabinet, which relates to the field of file management, and comprises a file cabinet main body which is used as a carrying main body of each function end and module and is used for editing and executing a control instruction; the touch control display end is used as a touch interaction interface and is used for displaying interaction information and archive data of an abnormal character area and corrective measures; the scanning end is used for receiving entity archive data, scanning and reading digital information and executing report prompt and prediction supplement of missing data; according to the method, an analysis model is constructed through deep learning, abnormal capture is performed on scanned archives, abnormal reason analysis is performed on abnormal character areas, correction measures are correspondingly provided for different abnormal reasons through a prediction recognition mechanism, so that entity archives are automatically checked and then classified and stored in the recording process, the quality of the recorded archives is improved, and the recording efficiency is improved. And the error rate in the file scanning process and the error rate of the file data are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of archive management, in particular to an intelligent archive scanning and storage integrated cabinet. Background Art

[0002] With the development of informatization, all walks of life are accumulating a large number of electronic and physical archives. These archives involve many fields such as law, business, medicine, education, etc., and the demand for management and storage is increasing. Traditional archive management methods mostly use paper storage, which leads to problems such as large space occupation, low information search efficiency and high management costs. Digital archive management has become an important means to improve work efficiency. The intelligent archive scanning and storage cabinet can quickly convert physical archives into digital format, which not only improves the accessibility of information, but also reduces the use of physical space; Currently, many archive management systems still rely on manual operations, including steps such as file classification, information entry, and exception checking. This approach is not only inefficient, but also prone to errors due to human negligence. There is a lack of effective mechanisms for data error checking, which can easily lead to inaccurate data. When errors occur, it is difficult to automatically identify the causes of the exceptions and correct the corresponding causes of the exceptions, resulting in the input data being stored in the wrong version. Even if data errors are discovered during the scanning phase, manual correction is still required, which delays processing efficiency. It lacks intelligent analysis and learning capabilities, and it is difficult to dynamically adjust processing methods based on historical data. Summary of the invention

[0003] 1. Technical issues to be resolved In view of the above-mentioned shortcomings of the prior art, the present invention provides an intelligent archive scanning and storage integrated cabinet, which can effectively solve the problems of the prior art.

[0004] (II) Technical solution To achieve the above objectives, the present invention is implemented through the following technical solutions: The present invention discloses an intelligent archive scanning and storage integrated cabinet, comprising: The main body of the filing cabinet, as the main body of each functional terminal and module, edits and executes the control instructions; The touch display terminal, as a touch interactive interface, displays interactive information and archival data of abnormal text areas and corrective measures; The scanning end is used to receive physical archive data, scan and read digital information, and perform reporting prompts and prediction supplements for missing data; The storage end is used to classify and store physical archives and electronic archive data; The scanning end includes: an image recognition module, a model building model, an abnormality capture module, a prediction and recognition module, a determination module and a feedback module, wherein: An image recognition module is used to convert received physical files into images, extract text from the images, aggregate the extracted data into a text data set, and issue classification labels based on the content of the text data set; A model building module, used for receiving the historical judgment-post-labeling data set and the historical prediction-correction data set of the feedback module as training parameters, and building an analysis model based on the training parameters through a deep learning algorithm; The exception capture module is used to obtain the analysis model constructed by the model construction module, perform exception capture based on the analysis model, perform real-time recognition of the text extraction area and semantic analysis of the text data set, obtain the abnormal text area accordingly, mark the abnormal text area, and output the marked data set; A determination module is used to receive the labeled data set output by the abnormal capture module, set an abnormal determination threshold, and perform determination and screening on the labeled data set submitted by the abnormal capture module based on the set abnormal determination threshold to obtain an abnormal data set determined to be abnormal; The prediction and identification module is used to receive the abnormal data set submitted by the determination module, identify the abnormal cause of the abnormal data set and its corresponding prediction and correction processing through the analysis model constructed by the model construction module, and output the prediction and correction data set; The feedback module is used to receive the determination and screening results of the determination module and submit them to the touch display terminal for interaction.

[0005] Furthermore, in the process of constructing the analysis model, the model construction module uses the historical judgment and labeled data set as the training parameter input, and the expression of the trained analysis model is: ; In the formula, represents the analysis model obtained during the training process, represents the initial analysis model, represents the loss function, represents the i-th input feature, represents the i-th label, represents the training parameter set, Represents model parameters.

[0006] Furthermore, the analysis model performs the work logic of capturing exceptions and determining and screening the submitted labeled data set through the exception capturing module, including: The text extracted from the input image is converted into a set of feature vectors including word frequency and word embedding features; Use the trained analysis model Identify abnormal samples for the newly extracted feature vector set and output anomaly scores; The abnormal judgment threshold set by the judgment module is D. The text is marked by the abnormal judgment threshold D. When the abnormal score corresponding to a text area in the text reaches or exceeds the abnormal judgment threshold, the abnormal mark is performed. The expression is: ; In the formula, represents the dataset marked as abnormal, represents the jth abnormal sample, Representatives and The corresponding anomaly score, represents the feature vector set of the current input, and D represents the abnormality judgment threshold.

[0007] Furthermore, the analysis model obtains several normal data sets, abnormal data sets and historical prediction and correction data sets of the analysis model historical output to perform recognition model training in the process of abnormal cause identification and corresponding prediction and correction processing of the determined labeled data set, and the process is as follows: ; Based on the abnormal cause output K of the recognition model, the new abnormal sample To identify the cause: = ; Due to several abnormal reasons , formulate corresponding corrective action set , apply the generated corrective action set to the abnormal samples acquired in real time for correction, and output the corrected electronic file data to the storage end; In the formula, Represents the abnormal cause output of the identification model, Represents the abnormal reason label, indicating the abnormal sample The corresponding exception reason, Representatives and The corresponding corrective measures, represents the loss function, Represents the training data set, which contains several normal data sets, abnormal data sets and historical prediction correction data sets of historical output.

[0008] Furthermore, the extraction unit is internally provided with sub-modules, including an extraction unit, and the extraction unit is interactively connected to the feedback module via a wireless network. The extraction unit is used to record the interactive results of the judgment and screening results submitted by the feedback module to the touch display terminal, record the addition and elimination of corrective measures, and submit them to the model construction module through the feedback module as a type of training reference for the analysis model.

[0009] Furthermore, the determination module supports the user to directly remove and add abnormal text areas through the touch display terminal, and the operation content is submitted to the feedback module as reference data, and the feedback module submits it to the model construction module as a second-category training reference for the analysis model.

[0010] Furthermore, the storage end includes: a deployment module, a local storage module and a cloud storage module, the deployment module and the local storage module are connected via an electrical medium, and the deployment module and the cloud storage module are interactively connected via a wireless network, wherein: The deployment module is used to receive the physical archive data and electronic archive data submitted by the storage end, read the classification label issued by the image recognition module, and generate classification instructions; The local storage module, as a loading area for entity archive data, performs entity area segmentation, receives classification instructions from the deployment module, and stores the entity archive data in corresponding segmented areas; The cloud storage module, as the loading area of ​​electronic archives, receives classification instructions from the deployment module and stores the physical archive data in a distributed manner.

[0011] Furthermore, a touch display terminal is installed at the top of the front side of the filing cabinet body, a scanning terminal is installed in the middle of the filing cabinet body, and a storage terminal is installed at the bottom of the filing cabinet body.

[0012] Furthermore, the image recognition module is interactively connected to the exception capture module through a wireless network, the determination module is interactively connected to the exception capture module, the prediction and recognition module and the feedback module through a wireless network, and the model building module is interactively connected to the exception capture module, the prediction and recognition module and the feedback module through a wireless network.

[0013] (III) Beneficial effects Compared with the known prior art, the technical solution provided by the present invention has the following beneficial effects: By building an analysis model through deep learning, anomalies in the scanned files are captured, and automatic analysis is used to determine whether there are abnormal text areas in the current scanned files. The abnormal text areas are marked, and the causes of the abnormalities are analyzed for the marked abnormal text areas. Through the predictive recognition mechanism, corrective measures are provided for different abnormal causes. During the entry process, the physical files are automatically proofread and then classified and stored, thereby improving the quality of the entered files, reducing the error rate in the file scanning process and the error rate of the file data itself, and improving the accuracy and consistency of the file data.

[0014] By recording the model output data of the judgment module and the feedback module and the user's active interaction data through the touch display as training references, the analysis model is in a continuous learning state during the working process. Based on the collected training references, the information can be updated and corrected in a timely manner through the training of the analysis model, ensuring the system's learning and adaptability, and continuously improving the accuracy and efficiency of the analysis model's recognition and prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0016] Figure 1 It is the overall three-dimensional structure diagram of the present invention; Figure 2 It is a schematic diagram of the framework of the present invention.

[0017] The numbers in the figure represent: 1. filing cabinet body; 2. touch display end; 3. scanning end; 4. storage end. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0019] The present invention will be further described below in conjunction with the embodiments. Example

[0020] The intelligent file scanning and storage integrated cabinet of this embodiment is as follows: Figure 1 and Figure 2 As shown, including: The filing cabinet body 1, as the main body of each functional terminal and module, edits and executes the control instructions; The touch display terminal 2, as a touch interactive interface, displays the interactive information and archive data of the abnormal text area and corrective measures; the abnormal text area and its corrective measures are displayed in real time, which enhances the convenience of operation and the visualization of information; Scanning terminal 3 is used to receive physical archive data, scan and read digital information, and perform reporting prompts and prediction supplements for missing data; Storage terminal 4, used for classified storage of physical archives and electronic archives data; A touch display terminal 2 is installed at the top of the front of the filing cabinet body 1, a scanning terminal 3 is installed in the middle of the filing cabinet body 1, and a storage terminal 4 is installed at the bottom of the filing cabinet body 1; The scanning end 3 includes: an image recognition module, a model building model, an abnormal capture module, a prediction and recognition module, a determination module and a feedback module. The image recognition module is interactively connected with the abnormal capture module through a wireless network, the determination module is interactively connected with the abnormal capture module, the prediction and recognition module and the feedback module through a wireless network, and the model building module is interactively connected with the abnormal capture module, the prediction and recognition module and the feedback module through a wireless network, wherein: An image recognition module is used to convert received physical files into images, extract text from the images, aggregate the extracted data into a text data set, and issue classification labels based on the content of the text data set; A model building module, used for receiving the historical judgment-post-labeling data set and the historical prediction-correction data set of the feedback module as training parameters, and building an analysis model based on the training parameters through a deep learning algorithm; In the process of building the analysis model in the model building module, the historical judgment and labeled data set is used as the training parameter input. The expression of the trained analysis model is: ; In the formula, represents the analysis model obtained during the training process, represents the initial analysis model, represents the loss function, which is used to measure the difference between the model output and the true label. Represents the i-th input feature, which represents the data or features of each sample in the data set. Represents the i-th label, which is consistent with the input feature Related real categories, Represents the training parameter set, which contains the labeled data and label information after historical judgment, and is used in the model training process. Represents model parameters, indicating the parameters that need to be optimized during model training; The exception capture module is used to obtain the analysis model constructed by the model construction module, perform exception capture based on the analysis model, perform real-time recognition of the text extraction area and semantic analysis of the text data set, obtain the abnormal text area accordingly, mark the abnormal text area, and output the marked data set; A determination module is used to receive the labeled data set output by the abnormal capture module, set an abnormal determination threshold, and perform determination and screening on the labeled data set submitted by the abnormal capture module based on the set abnormal determination threshold to obtain an abnormal data set determined to be abnormal; The analysis model uses the exception capture module to perform exception capture and determine the submitted labeled data set. The working logic includes: The text extracted from the input image is converted into a set of feature vectors including word frequency and word embedding features; Use the trained analysis model Identify abnormal samples for the newly extracted feature vector set and output anomaly scores; The abnormal judgment threshold set by the judgment module is D. The text is marked by the abnormal judgment threshold D. When the abnormal score corresponding to a text area in the text reaches or exceeds the abnormal judgment threshold, the abnormal mark is performed. The expression is: ; In the formula, represents the dataset marked as abnormal, represents the jth abnormal sample, indicating the data point identified as an abnormality in the new data set, Representatives and The corresponding anomaly score, Represents the feature vector set of the current input. D represents the abnormal judgment threshold, which is used to define the abnormal judgment standard. Samples with a judgment score exceeding this value are considered abnormal. They work together in the key steps of training, abnormality identification, and correction processing to achieve a complete closed-loop process from data input to abnormality processing. The prediction and identification module is used to receive the abnormal data set submitted by the determination module, identify the abnormal cause of the abnormal data set and its corresponding prediction and correction processing through the analysis model constructed by the model construction module, and output the prediction and correction data set; The analysis model performs abnormal cause identification and corresponding prediction and correction processing on the marked data set after determination, and obtains several normal data sets, abnormal data sets and historical prediction and correction data sets of the analysis model historical output for recognition model training. The process is as follows: ; Based on the abnormal cause output K of the recognition model, the new abnormal sample To identify the cause: = ; Due to several abnormal reasons , formulate corresponding corrective action set , applying the generated corrective action set to the abnormal samples acquired in real time for correction, and outputting the corrected electronic file data to the storage terminal 4; In the formula, Represents the abnormal cause output of the identification model, Represents the abnormal reason label, indicating the abnormal sample The corresponding exception reason, Representatives and The corresponding corrective measures, Represents the loss function, which is used to measure the difference between the recognition model output and the true label. Representative training data sets, including several normal data sets, abnormal data sets and historical prediction correction data sets of historical output, are used in the model training process to effectively identify the causes of anomalies and generate corresponding corrective measures for each abnormal sample, which not only improves the accuracy of anomaly detection, but also further optimizes the post-abnormal processing process, thereby ensuring the stability and reliability of the system; The feedback module is used to receive the determination and screening results of the determination module and submit them to the touch display terminal 2 for interaction, ensuring that the user gets instant feedback, thereby enhancing the interactivity and feedback reliability of the system.

[0021] Compared with the existing technology, the intelligent archive scanning and storage integrated cabinet has realized a complete closed loop from data input, anomaly detection to corrective processing, which greatly improves work efficiency and accuracy. Through the deep learning algorithm, the analysis model is continuously optimized. Compared with the traditional manual processing mode, it can significantly reduce human errors, improve the intelligence and accuracy of anomaly detection, capture abnormal data in real time and feedback to users, ensure the timeliness of information processing, avoid potential problems caused by delays, and generate corresponding corrective measures for each abnormal sample, which not only reduces the subsequent labor cost, but also improves the pertinence and effectiveness of exception processing, thereby enhancing the reliability and stability of the integrated cabinet.

[0022] Example 2 In other aspects, this embodiment also provides an extraction unit, such as Figure 2 As shown, the extraction unit is internally provided with submodules, including an extraction unit, the extraction unit and the feedback module are interactively connected via a wireless network, and the extraction unit is used to record the interactive results of the judgment and screening results submitted by the feedback module to the touch display terminal 2, record the addition and elimination of corrective measures, and submit them to the model building module through the feedback module as a type of training reference for the analysis model; The judgment module supports users to directly remove and add abnormal text areas through the touch display terminal 2. The operation content is submitted to the feedback module as reference data, and submitted to the model construction module by the feedback module as a second-category training reference for the analysis model. The judgment module supports users to directly remove and add abnormal text areas through the touch display terminal 2. The operation content is submitted to the feedback module as reference data, and submitted to the model construction module by the feedback module as a second-category training reference for the analysis model.

[0023] In this embodiment, through the judgment module, the user can directly remove and add abnormal text areas. This interactivity allows the user to directly participate in the operation of the system and make real-time adjustments to abnormal data, thereby improving the flexibility and response speed of the system. The extraction unit can record the judgment and screening results submitted by the feedback module and the increase and decrease of corrective measures, ensuring the traceability of each step of the operation, which is conducive to the optimization and improvement of the subsequent analysis model. Two types of training references are provided to promote the diversity and accuracy of the analysis model. The analysis model can learn different data processing situations more comprehensively, improve the intelligence of anomaly detection, and dynamically adjust the analysis model based on real-time feedback and data analysis results to quickly adapt to changing needs and input data. By introducing the user's direct operation as a training reference, the analysis model can better understand the user's needs and actual situation, further improve the accuracy of identifying abnormal data, and reduce the risk of misjudgment and missed judgment.

[0024] Example 3 In this embodiment, Figure 2 As shown, a storage terminal 4 is provided, including: a deployment module, a local storage module and a cloud storage module, the deployment module is connected to the local storage module through an electrical medium, and the deployment module is interactively connected to the cloud storage module through a wireless network, wherein: The deployment module is used to receive the physical archive data and electronic archive data submitted by the storage terminal 4, read the classification label issued by the image recognition module, and generate a classification instruction; The local storage module, as a loading area for entity archive data, performs entity area segmentation, receives classification instructions from the deployment module, and stores the entity archive data in corresponding segmented areas; The cloud storage module, as the loading area of ​​electronic archives, receives classification instructions from the deployment module and stores the physical archive data in a distributed manner.

[0025] In this embodiment, the user scans and inputs the physical archive data into the system, which is converted into electronic archive data. The allocation module receives the archive data and the corresponding classification labels and generates classification instructions. The local storage module divides the physical archive data into regions and stores them in preset locations according to the instructions of the allocation module. The cloud storage module distributes the electronic archive data in the cloud according to the same classification instructions. Through such a data storage mechanism, users can efficiently retrieve and manage physical and electronic archives, realize local and cloud data integration, and improve the flexibility and reliability of data storage and access.

[0026] Working principle: When the present invention is used, the filing cabinet body 1 is used as a carrier of each functional module and component. The user scans the physical file through the scanning terminal 3, and the scanning terminal 3 converts the physical file into a digital file. In the process, the abnormal errors in the physical file are identified and the causes are analyzed. Corrective actions are provided based on the analyzed causes. The touch display terminal 2 provides interaction between scanning display and corrective actions, and the corrected digital files and physical files are classified and stored in the storage terminal 4; It realizes the rapid digital scanning and reading of physical archives, and can promptly remind users of any erroneous or missing data, thus improving the accuracy and completeness of archive management, effectively integrating the data of physical and electronic archives, optimizing the storage and retrieval process, improving the efficiency of archive utilization, providing a friendly user interaction experience, and being able to display abnormal text areas and their corrective measures in real time, thus enhancing the convenience of operation and the visualization of information.

[0027] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. Intelligent file scanning and storage integrated cabinet, characterized by: include: The filing cabinet body (1) is used as the main body for carrying each functional terminal and module, and is used for editing and executing control instructions; A touch display terminal (2), serving as a touch interactive interface, displays interactive information and archive data of abnormal text areas and corrective measures; The scanning end (3) is used to receive physical archive data, scan and read digital information, and perform reporting prompts and prediction supplements for missing data; The storage end (4) is used for classified storage of physical archives and electronic archives data; The scanning end (3) comprises: an image recognition module, a model building model, an abnormality capture module, a prediction recognition module, a determination module and a feedback module, wherein: An image recognition module is used to convert received physical files into images, extract text from the images, aggregate the extracted data into a text data set, and issue classification labels based on the content of the text data set; A model building module, used for receiving the historical judgment-post-labeled data set and the historical prediction-corrected data set of the feedback module as training parameters, and building an analysis model based on the training parameters through a deep learning algorithm; The exception capture module is used to obtain the analysis model constructed by the model construction module, perform exception capture based on the analysis model, perform real-time recognition of the text extraction area and semantic analysis of the text data set, obtain the abnormal text area accordingly, mark the abnormal text area, and output the marked data set; A determination module is used to receive the labeled data set output by the abnormal capture module, set an abnormal determination threshold, and perform determination and screening on the labeled data set submitted by the abnormal capture module based on the set abnormal determination threshold to obtain an abnormal data set determined to be abnormal; The prediction and identification module is used to receive the abnormal data set submitted by the determination module, identify the abnormal cause of the abnormal data set and its corresponding prediction and correction processing through the analysis model constructed by the model construction module, and output the prediction and correction data set; The feedback module is used to receive the determination and screening results of the determination module and submit them to the touch display terminal (2) for interaction.

2. The intelligent file scanning and storage integrated cabinet according to claim 1, characterized in that: In the process of constructing the analysis model by the model construction module, the historical judgment and labeled data set is used as the training parameter input, and the expression of the trained analysis model is: ； In the formula, represents the analysis model obtained during the training process, represents the initial analysis model, represents the loss function, represents the i-th input feature, represents the i-th label, represents the training parameter set, Represents model parameters.

3. The intelligent file scanning and storage integrated cabinet according to claim 1, characterized in that: The analysis model performs the work logic of capturing exceptions and judging and screening the submitted labeled data set through the exception capture module, including: The text extracted from the input image is converted into a set of feature vectors including word frequency and word embedding features; Use the trained analysis model Identify abnormal samples for the newly extracted feature vector set and output anomaly scores; The abnormal judgment threshold set by the judgment module is D. The text is marked by the abnormal judgment threshold D. When the abnormal score corresponding to a text area in the text reaches or exceeds the abnormal judgment threshold, the abnormal mark is performed. The expression is: ； In the formula, represents the dataset marked as abnormal, represents the jth abnormal sample, Representatives and The corresponding anomaly score, represents the feature vector set of the current input, and D represents the abnormality judgment threshold.

4. The intelligent file scanning and storage integrated cabinet according to claim 1, characterized in that: The analysis model performs abnormal cause identification and corresponding prediction and correction processing on the determined labeled data set, and obtains several normal data sets, abnormal data sets and historical prediction and correction data sets output by the analysis model to perform recognition model training. The process is as follows: ； Based on the abnormal cause output K of the recognition model, the new abnormal sample To identify the cause: = ； Due to several abnormal reasons , formulate corresponding corrective action set , applying the generated corrective action set to the abnormal samples acquired in real time for correction, and outputting the corrected electronic file data to the storage end (4); In the formula, Represents the abnormal cause output of the identification model, Represents the abnormal reason label, indicating the abnormal sample The corresponding exception reason, Representatives and The corresponding corrective measures, represents the loss function, Represents the training data set, which contains several normal data sets, abnormal data sets and historical prediction correction data sets of historical output.

5. The intelligent file scanning and storage integrated cabinet according to claim 1, characterized in that: The extraction unit is internally provided with submodules, including an extraction unit, wherein the extraction unit is interactively connected with the feedback module via a wireless network, and the extraction unit is used to record the interactive results of the judgment and screening results submitted by the feedback module to the touch display terminal (2), record the addition and elimination of corrective measures, and submit them to the model construction module via the feedback module as a type of training reference for the analysis model.

6. The intelligent file scanning and storage integrated cabinet according to claim 1, characterized in that: The determination module supports the user to directly remove or add abnormal text areas through the touch display terminal (2), and the operation content is submitted to the feedback module as reference data, which is then submitted to the model construction module as a second-category training reference for the analysis model.

7. The intelligent file scanning and storage integrated cabinet according to claim 1, characterized in that: The storage end (4) comprises: a deployment module, a local storage module and a cloud storage module, wherein the deployment module is connected to the local storage module via an electrical medium, and the deployment module is interactively connected to the cloud storage module via a wireless network, wherein: A dispatching module, used to receive the physical archive data and electronic archive data submitted by the storage end (4), read the classification label issued by the image recognition module, and generate a classification instruction; The local storage module, as a loading area for entity archive data, performs entity area segmentation, receives classification instructions from the deployment module, and stores the entity archive data in corresponding segmented areas; The cloud storage module, as the loading area of ​​electronic archives, receives classification instructions from the deployment module and stores the physical archive data in a distributed manner.

8. The intelligent file scanning and storage integrated cabinet according to claim 1, characterized in that: A touch display terminal (2) is installed at the top of the front of the filing cabinet body (1), a scanning terminal (3) is installed in the middle of the filing cabinet body (1), and a storage terminal (4) is installed at the bottom of the filing cabinet body (1).

9. The intelligent file scanning and storage integrated cabinet according to claim 1, characterized in that: The image recognition module is interactively connected to the exception capture module through a wireless network, the determination module is interactively connected to the exception capture module, the prediction recognition module and the feedback module through a wireless network, and the model construction module is interactively connected to the exception capture module, the prediction recognition module and the feedback module through a wireless network.

Images