Ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission method and device, computer equipment and storage medium

Abstract

The application discloses an ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission method and device, computer equipment and a storage medium, and relates to the technical field of ophthalmic data processing. The ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission method and device, computer equipment and storage medium can collect data of different structures from various ophthalmic data sources, build a storage model and an interface, and realize the fusion storage of ophthalmic multi-source heterogeneous data. The seamless data transmission between different operating systems and various devices can be guaranteed. The ophthalmic data can be subjected to encryption processing, access permission setting and backup and recovery operation. The data flow and quality can be monitored in real time, and the system log can be effectively managed. The application can be integrated with a related big data analysis platform and an enterprise information system, solves the problems in the fusion storage and cross-platform transmission of ophthalmic multi-source heterogeneous data, and effectively improves the comprehensive utilization value of ophthalmic data.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of ophthalmic data processing, in particular to an ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission method and device, computer equipment and storage medium. BACKGROUND

[0002] In the field of ophthalmic medical treatment and research, data sources are increasingly diverse and complex in structure. They include structured patient basic information database data, structured detection data generated by ophthalmic examination equipment, such as intraocular pressure measurement data and vision examination data, unstructured ophthalmic medical record document data, such as doctors' diagnosis records and operation records, and multimedia data, such as fundus images, OCT optical coherence tomography images, FFA fundus fluorescence angiography image sequences, UBM ultrasonic biomicroscopy examination images and videos, etc. Due to different sources, the structures and formats of these data are significantly different, which brings great challenges to data management, storage and sharing.

[0003] Traditional data storage and transmission methods are mostly designed for a single data type or a specific platform, which exposes many problems in ophthalmic data processing. For example, structured databases are difficult to effectively store and manage unstructured medical record documents and multimedia ophthalmic image and video data. Due to incompatible data formats and inconsistent transmission protocols between different operating systems (such as Windows server systems in hospitals, iOS or Android mobile terminal systems used by doctors, etc.) and device types (such as desktop examination equipment, portable examination instruments, servers, personal computers, mobile terminals, etc.), it is extremely difficult to share and interact ophthalmic data between different platforms, which seriously limits the comprehensive utilization value of ophthalmic data and cannot meet the needs of precise diagnosis, remote medical collaboration and scientific research analysis for data processing.

[0004] Under this background, the present application provides an ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission method, device, computer equipment and storage medium. SUMMARY

[0005] (I) Technical problems solved

[0006] The purpose of the present application is to provide an ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission method, device, computer equipment and storage medium to solve the problems of difficult ophthalmic multi-source heterogeneous data fusion storage and inconvenient cross-platform transmission in the prior art, and to improve the comprehensive utilization value of ophthalmic data.

[0007] (II) Technical solutions

[0008] The purpose of the present application is to provide an ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission method, device, computer equipment and storage medium, comprising:

[0009] The data acquisition module is the key entrance of the ophthalmic data flow, and has data source identification and data extraction capabilities. The data source identification submodule uses intelligent algorithms and feature matching technology to quickly and accurately identify the data source. Whether it is the mainstream database system MySQL, Oracle used to store patient basic information and part of the examination equipment data, covering file formats under Windows, Linux operating systems, including various ophthalmic medical record document file systems, including web data, cloud storage data, involving data sharing in telemedicine collaboration or scientific research data acquisition of various network data sources, or specific application program interfaces such as hospital-internal customized ophthalmic examination software interfaces and medical record management system interfaces, they can all be accurately identified. For example, when identifying database sources, by deeply analyzing the port number, connection string and database-specific system table information features, the database type and version information can be determined within a very short millisecond time. For file systems, not only can common file formats be determined according to file header information, but also complex file storage systems can be identified through file directory structure and metadata information, such as accurately identifying the classification storage structure and related metadata information of ophthalmic medical record documents.

[0010] The data extraction submodule then applies extraction strategies according to the identification results. For structured database data, through carefully constructed SQL query statements or efficient database connection tools: JDBC, ODBC, the target data can be accurately extracted. When dealing with large-scale structured data, database connection pool technology and optimized query execution plan are used, which improves the extraction efficiency compared with traditional extraction methods. For example, when extracting a large number of patient's intraocular pressure measurement history data, the required information can be quickly and accurately obtained. For unstructured document data: Word, PDF, TXT format ophthalmic medical record documents, regular expression and DOM parsing based on regular expression and DOM parsing are used to deeply mine the text content. Taking the processing of large PDF medical record documents containing rich examination results and diagnosis analysis as an example, through multi-thread parallel analysis technology, the analysis time can be shortened, and the key information can be efficiently extracted. For multimedia data: fundus images, OCT images, FFA videos, UBM images and videos, professional multimedia data reading library: OpenCV for image processing, FFmpeg for audio and video processing, complete data information extraction. In image data extraction, the efficient pixel processing algorithm of OpenCV can obtain detailed pixel information, color features and size of the image within a very short time, and the processing speed is improved compared with ordinary image reading library, for example, the pixel distribution and morphological features of blood vessels in fundus images can be quickly obtained; for video data, the frame rate, resolution, encoding format and other key data of the video can be accurately extracted, as well as the key frame information in the image sequence, providing comprehensive data support for subsequent disease analysis and diagnosis.

[0011] The metadata description layer of the data fusion storage module uses a relational database table structure to store the metadata information of the ophthalmic multi-source heterogeneous data. These metadata, like the "identity card" of the data, record in detail the key elements such as the data source (e.g. a specific examination device from the ophthalmic clinic of a certain hospital, a specific medical record of a certain doctor, etc.), type (structured, unstructured, multimedia), format (e.g. database table structure, Word document format, JPEG image format, etc.), creation time, and correlation (e.g. the fundus image of a patient is closely related to the corresponding medical record document, and the OCT image is logically related to the intraocular pressure examination data).

[0012] The data storage layer stores the data according to its characteristics. The structured data is stored by means of a mature relational database storage engine (e.g. InnoDB, MyISAM, etc.). In the data insertion and query operations, the index structure is optimized by using B+ tree index and combining index coverage technology. The unstructured document data is stored in the distributed file system Ceph, GlusterFS, which utilizes the distributed storage and high fault tolerance features of the distributed file system to ensure the reliability and scalability of the data. Even if some storage nodes fail, the integrity and availability of the medical record document data can be guaranteed. The multimedia data is stored in the MinIO multimedia storage based on object storage, and a tight mapping relationship with the metadata description layer is established through the designed index mechanism to ensure the efficiency of data query and positioning.

[0013] In addition, the data index construction submodule adopts a combination of inverted index and B-tree index to construct a multi-level index system based on the keywords, metadata information, or data content features of the data, greatly improving the data retrieval speed. For the keyword index construction of structured data, a dynamic index update strategy is adopted. For the multimedia data content feature index, deep learning algorithms are used to extract the lesion feature vector of the fundus image and the retinal layer feature vector of the OCT image to construct a feature index, which improves the accuracy in similar lesion image and related examination image retrieval, and provides a powerful technical means for doctors to quickly compare and analyze the changes in the patient's condition and similar cases.

[0014] The data transmission module, to achieve seamless transmission across platforms, exerts force from multiple dimensions. The platform detection submodule adopts a combination of system detection technology and network scanning technology to comprehensively detect the operating system types of the source platform and target platform for data transmission, such as Windows system of hospital server, iOS or Android system of doctor's mobile terminal, detection of device types such as large ophthalmic examination equipment, desktop computer, laptop computer, tablet computer, smart phone, and detection of network environment information including bandwidth, delay, and packet loss rate.

[0015] The transmission protocol adaptation submodule selects the adapted protocol from the TCP, UDP, HTTP, and FTP transmission protocol libraries according to the detection results. In a low-bandwidth and high-real-time requirement scenario, such as when a doctor uses a mobile terminal to view the fundus image or video of a patient during remote consultation, the lightweight UDP protocol is preferred, and the application layer custom protocol is combined to optimize the transmission efficiency; in a high-bandwidth and strict data integrity environment, such as the transmission of a large amount of ophthalmic examination data and medical records between hospital internal systems, the reliable TCP protocol is selected, and the SSL / TLS encryption transmission is used to ensure data security and prevent data from being stolen or tampered with during transmission.

[0016] The data format conversion submodule constructs a dynamic format recognition and conversion mechanism for different ophthalmic data types, including structured, unstructured, image, and video. When the system identifies ophthalmic image data, it performs content detection on the image through a deep learning-based image recognition algorithm to automatically determine the best image compression and format conversion strategy. The intelligent format conversion based on deep learning and image content can compress the lesion area in the ophthalmic image, avoiding the loss of important diagnostic information during compression. This enhances the intelligence of the image processing process and improves the quality and efficiency of image transmission.

[0017] The process of constructing a dynamic format recognition and conversion mechanism includes: through the data acquisition module, the input ophthalmic data is source-identified and classified to ensure that data from different sources can be correctly classified and processed. The recognized data types include but are not limited to structured data, unstructured document data, image data, and video data. After the data is classified, the system needs to further identify the characteristics of each data format. For image, video, and document data types, the most suitable conversion method and format are selected. According to the identified image, video, or document type and the requirements of the target device, the system will perform corresponding compression and conversion to achieve the best transmission efficiency and user experience. To ensure the real-time performance of format conversion and the stability of transmission, the system should design a real-time monitoring and feedback mechanism to timely handle network fluctuations and data format incompatibility issues. The transmission request of images, videos, and documents is input, and the network bandwidth, delay, and data traffic are monitored in real time. When the system detects insufficient network bandwidth or excessively high delay, it automatically adjusts the transmission strategy, outputs the optimized image, video, and document data transmission stream, and transmits the converted and optimized data to the target device or platform.

[0018] The data security management module comprehensively safeguards data security. The data encryption submodule combines the AES algorithm in symmetric encryption and the RSA algorithm in asymmetric encryption, and performs layered encryption on sensitive ophthalmic data. First, the symmetric encryption algorithm is used to quickly encrypt a large amount of ophthalmic examination data and medical records, and then the asymmetric encryption algorithm is used to encrypt the symmetric encryption key for transmission, ensuring data confidentiality and key security.

[0019] The AES algorithm uses a 256-bit encryption key length, and the encryption strength is 2^128 times higher than that of a 128-bit key. In data encryption performance testing, the encryption speed can reach more than 100 MB per second, which can ensure data security and meet the demand for fast processing of ophthalmic data.

[0020] The access permission control submodule is based on the role-based access control (RBAC) model. According to user roles (such as ophthalmic specialists, general ophthalmologists, nurses, researchers, and patients) and data sensitivity (such as patient privacy information and research data on rare eye diseases), fine access permissions are set at the data storage and access level, covering read, write, modify, and delete permissions, and only users or systems with corresponding permissions are allowed to access specific ophthalmic data. For example, ophthalmic specialists can view and modify detailed medical records and examination data of all patients to make comprehensive diagnoses and develop treatment plans; general ophthalmologists can only view and modify the data of patients under their responsibility, ensuring the privacy and security of patient data; nurses can only view basic patient information and nursing-related data for daily nursing work; researchers can access anonymized ophthalmic data for research after authorization, promoting ophthalmic medical research; and patients can only view their own basic examination results and diagnosis information to understand their own medical conditions. Through the audit function of the permission management system, detailed records of permission operations are kept to trace and supervise data access behavior, preventing data leakage and illegal access.

[0021] The data backup and recovery submodule adopts a combination of full backup and incremental backup strategies. Regular full backups of stored ophthalmic data are performed, and incremental backups are performed when data is updated. The backup period is flexibly set according to the importance and update frequency of the data, with daily incremental backups of newly generated examination data and weekly full backups of all data. In the event of data loss or damage, the recovery algorithm based on backup data and data log-based recovery technology can accurately restore data to a specified time point, ensuring data availability and integrity. In data recovery testing, the average recovery time for ophthalmic data recovery requests at different time points can be controlled within 5 minutes, and the integrity and accuracy of the recovered data pass the verification with a 100% pass rate, ensuring that ophthalmic medical work and research analysis are not affected by data loss. For example, after system failure or accidental data deletion, data can be quickly recovered to ensure the normal operation of hospital ophthalmic services.

[0022] The data monitoring and management module strictly monitors and finely manages the entire process of ophthalmic data processing. The data flow monitoring submodule uses network flow monitoring technology and data flow calculation algorithm to monitor the flow of data in the storage and transmission process in real time, accurately counts the inflow and outflow of data and the data transmission rate, and according to the threshold curve based on historical flow data and ophthalmic data processing system performance indicators, i.e. the flow threshold model, when the data flow exceeds the threshold, an alarm is sent in time to quickly locate and solve the flow anomaly problem. During the flow monitoring process, the flow data real-time analysis algorithm can send an alarm within 1 second after the flow anomaly occurs, with a false alarm rate of less than 2%, ensuring the stability and reliability of ophthalmic data transmission, avoiding data transmission delay or interruption due to flow problems, and affecting the timeliness of ophthalmic diagnosis and treatment.

[0023] The data quality monitoring submodule uses a data verification rule library and a data cleaning and repair algorithm to deeply monitor the quality of stored ophthalmic data, and comprehensively checks the integrity (such as whether the required fields in the patient information table are complete, whether the ophthalmic examination data record is missing, etc.), accuracy (such as whether the intraocular pressure measurement data is within a reasonable range, whether the diagnosis information in the medical record is accurate and standardized, etc.) and consistency (such as whether the visual acuity data of the same patient generated by different examination equipment is consistent, etc.) of the data. For data that does not meet the quality standards, the corresponding processing flow is automatically triggered according to the data problem type, including data cleaning (such as removing duplicate ophthalmic examination records, correcting incorrect intraocular pressure data, etc.), data repair (such as repairing missing medical record information according to data association relationships, etc.) or data deletion (such as severely incorrect or outdated examination data), to ensure that the data quality meets the needs of ophthalmic medical and scientific research. In the data quality monitoring test, for large-scale ophthalmic data samples, the accuracy rate of data integrity verification can reach more than 99%, the accuracy rate of accuracy verification can reach more than 98%, and the accuracy rate of consistency verification can reach more than 97%, providing a reliable data foundation for ophthalmic precise diagnosis and scientific research, such as ensuring the accuracy and reliability of data in scientific research analysis. The data verification rule library includes integrity verification, accuracy verification, and consistency verification rules. Based on the data cleaning and repair algorithm, duplicate ophthalmic data is cleaned using a hash algorithm-based duplicate detection technology, and ophthalmic data entry errors are repaired using an automatic repair algorithm based on data type and business rules.

[0024] The system log management submodule uses a log recording framework: Log4j, Slf4j, configures the log output format and storage path through Log4j, records the running log of the software system in detail, covers every detail information of the ophthalmic data collection, storage, transmission, security management and monitoring operation, in order to carry out system troubleshooting and statistical analysis, such as the source address of data collection (accurate to the IP address of the ophthalmic examination equipment, the directory of the server where the medical record document is located), the collection time (accurate to milliseconds), the collection data volume (in bytes); the storage location of ophthalmic data storage (such as which table space of the relational database the data is stored in, which path of the distributed file system the file is stored in), the storage time, the storage data format; the source platform of data transmission (such as the source server host name or the examination equipment number), the target platform (such as the target mobile terminal equipment number or the receiving server name), the transmission protocol (detailed to the UDP protocol and the application layer custom protocol details), the transmission data volume; the encryption operation of ophthalmic data security management (involving ophthalmic data, encryption algorithm and key information), access permission setting (for user role and specific permission), backup and recovery record (when full backup is performed, the content of incremental backup and the time and result of recovery operation), so as to accurately locate the problem root cause when troubleshooting and comprehensively trace the ophthalmic data operation history when statistical analysis, provide basis for data management and optimization. Through the efficient retrieval function of the log management system, in the mass log data, the average response time of querying the log related to a specific event can be controlled within 2 seconds, which improves the utilization efficiency of log data, for example, quickly finding the detailed information of a data transmission exception.

[0025] An ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission device, comprising:

[0026] An ophthalmic data collection device for performing the function of the ophthalmic data collection module, collecting data of different structures from multiple different ophthalmic data sources;

[0027] An ophthalmic data storage device for performing the function of the ophthalmic data fusion storage module, constructing an ophthalmic data storage model and performing data fusion storage, and providing a storage interface;

[0028] An ophthalmic data transmission device for performing the function of the ophthalmic data transmission module, realizing seamless ophthalmic data transmission across platforms;

[0029] An ophthalmic data security device for performing the function of the ophthalmic data security management module, ensuring the security of ophthalmic data during storage and transmission;

[0030] An ophthalmic data monitoring device is used to perform the functions of the ophthalmic data monitoring and management module, and to monitor and manage the ophthalmic data processing process.

[0031] A computer device is composed of a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement each step of the above-mentioned method to achieve the ophthalmic data processing function and task execution requirement.

[0032] A computer readable storage medium has the ability to store a computer program. When the computer program stored in the medium is called and executed by the processor, each step of the above-mentioned ophthalmic data processing method will be executed in sequence to complete the operation process and achieve the expected effect.

[0033] (III) Beneficial effects

[0034] The application provides an ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission method, device, computer equipment and storage medium. Has the following beneficial effects:

[0035] The ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission method, device, computer equipment and storage medium can effectively solve the problems of ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission through close cooperation and fine operation between modules, and can play an important role in ophthalmic clinical diagnosis, remote medical cooperation, medical scientific research innovation, ophthalmic medical teaching and training, hospital information management and other key fields; The application solves the problems of ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission in the prior art, and improves the comprehensive utilization value of ophthalmic data. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 It is an ophthalmic multi-source heterogeneous data acquisition process schematic diagram in an embodiment;

[0037] Figure 2 It is an ophthalmic multi-source heterogeneous data fusion storage process schematic diagram in an embodiment;

[0038] Figure 3 It is an ophthalmic multi-source heterogeneous data transmission process schematic diagram in an embodiment;

[0039] Figure 4 It is an ophthalmic multi-source heterogeneous data encryption method process schematic diagram in an embodiment;

[0040] Figure 5 It is an ophthalmic multi-source heterogeneous data access permission control method process schematic diagram in an embodiment;

[0041] Figure 6 Figure 1 is a flowchart of an embodiment of an ophthalmic multi-source heterogeneous data backup and recovery method;

[0042] Figure 7 Figure 2 is a flowchart of an embodiment of an ophthalmic multi-source heterogeneous data flow monitoring method;

[0043] Figure 8 Figure 3 is a flowchart of an embodiment of an ophthalmic multi-source heterogeneous data quality monitoring method;

[0044] Figure 9 Figure 4 is a flowchart of an embodiment of an ophthalmic multi-source heterogeneous data system log management method;

[0045] Figure 10 Figure 5 is a structural block diagram of an embodiment of an ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission device;

[0046] Figure 11 Figure 6 is an internal structural diagram of an embodiment of a computer device. DETAILED DESCRIPTION

[0047] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of the present application.

[0048] In an embodiment of ophthalmic data collection, as shown in Figure 1, an ophthalmic multi-source heterogeneous data collection method is provided, specifically comprising: Figure 1

[0049] After the data collection module S11 is started, the data source identification submodule scans all potential data sources comprehensively, including database systems (checking port numbers, connection string characteristics, and other multi-dimensional information, such as the port and connection characteristics of an ophthalmic patient information database), file systems (analyzing file path formats, file header information, such as the relevant information of an ophthalmic medical record document file system), network data sources (according to network request URL formats, data return formats, and protocol characteristics, such as the network characteristics when obtaining data from a remote ophthalmic medical data platform), and specific application program interfaces (analyzing interface specifications and identification information, such as the specific identification of an ophthalmic examination equipment software interface). The data source type is determined according to the identification result. If it is a database system, go to step S12; if it is a file system, go to step S13; if it is a network data source, go to step S14; if it is a specific application program interface, go to step S15;

[0050] ​The S12 structured database data extraction submodule connects the database using a JDBC or ODBC database connection tool. SQL query statements are constructed and executed to extract structured data, including a patient basic information table, an intraocular pressure measurement data table, and a visual examination result table. After extraction is complete, the process proceeds to step S16;

[0051] Specifically, in a hospital network environment, it detects a data source from a MySQL database, accurately determines that it is a MySQL database type by deeply analyzing its port number 3306 and database connection string characteristics. Subsequently, the data extraction submodule connects the database using JDBC, carefully constructs and executes corresponding SQL query statements, and successfully extracts the patient's name, age, medical history information, and historical intraocular pressure measurement data, visual examination data, and other structured database data.

[0052] S13 non-structured document data extraction. If it is a Word document, text reading technology is used to read the text content, and a multi-thread parallel reading strategy is used to improve reading speed. If it is a PDF document, the iText parsing library is called to parse the document, extract text, pictures, and tables, and use the Document Object Model (DOM) parsing technology to deeply analyze the document structure. After extraction is complete, the process proceeds to step S16;

[0053] Specifically, in a hospital file server, the data source recognition submodule acutely discovers some ophthalmic medical records documents. By carefully analyzing the file header information, it accurately determines that it is non-structured document data. For Word format medical summary documents, the data extraction submodule uses text reading technology and multi-thread parallel processing to orderly extract text content such as doctor's diagnosis analysis and treatment recommendations. For PDF format surgical report documents, the iText library is used in combination with DOM parsing technology for in-depth analysis to extract text descriptions, surgical pictures, postoperative recovery situation tables, and other content information.

[0054] S14 multimedia data extraction. For image data, the OpenCV library is called to read pixel information, color features, size, and other information, and efficient image pixel processing algorithms are used to quickly obtain detailed information. For audio and video data, the FFmpeg library is used to extract channel information, sampling rate, frame rate, resolution, encoding format, and other information, and hardware acceleration technology is used to improve extraction efficiency. After extraction is complete, the process proceeds to step S16;

[0055] Specifically, from the ophthalmic examination equipment (such as fundus camera, OCT equipment, etc.) or hospital image storage system, the data source identification submodule quickly determines the existence of some ophthalmic image and video data, such as fundus image, OCT image, FFA video, UBM video and other multimedia data sources, according to the file format characteristics and device identification information. For image data, the data extraction submodule calls the OpenCV library to accurately read the pixel information, color characteristics, size and other detailed information of the image, such as the distribution and morphological characteristics of blood vessels in the fundus image, the layered structure information of the retina in the OCT image; for audio and video data, the FFmpeg library is enabled to efficiently extract key data such as frame rate, resolution, encoding format, and key frame information in the image sequence, such as the key frame images of the contrast agent flowing in the fundus blood vessels in the FFA video, providing comprehensive data support for the diagnosis of ophthalmic diseases.

[0056] S15 Application program interface data extraction, according to the interface specification and protocol, using the corresponding interface call method to obtain data, using cache mechanism and asynchronous request technology to improve data acquisition efficiency. After extraction, go to step S16;

[0057] Specifically, the data source identification submodule determines that it is an application program interface data source of the ophthalmic examination equipment or medical record management system according to the interface URL path “ / api / ophthalmologyData” and the request header identifier “Ophthalmology-API-Token”. The data extraction submodule constructs a request parameter containing the patient number or examination item number, sends a request to the interface, stores commonly used data through a cache mechanism, and uses asynchronous request technology to concurrently process multiple requests. After receiving the returned data (JSON format), the examination details such as the parameter settings of the specific examination equipment and the detailed numerical values of the examination results are extracted for subsequent processing.

[0058] S16 Data transmission to the fusion storage module, the extracted data is transmitted to the data fusion storage module, and a high-speed data transmission channel and data verification mechanism are used to ensure the accuracy and integrity of data transmission, and the process ends.

[0059] In one data fusion storage embodiment, as shown in Figure 2 An ophthalmic multi-source heterogeneous data fusion storage method is provided, which specifically includes:

[0060] The data collected by S21 is transmitted to the data fusion storage module in a timely manner. The metadata description unit quickly starts work to collect metadata information of the data, collects metadata information of various types of data through the metadata description unit, and stores and manages according to the "Metadata" table structure definition. Whether the data is a structured patient information table, an unstructured medical record document, or multimedia data, there is corresponding metadata to describe its source (such as from a certain ophthalmic clinic examination equipment, a certain doctor's medical record, etc.), type (structured, unstructured, multimedia), format (such as database table structure, Word document format, JPEG image format, etc.), creation time and association relationship (such as the association of the patient's fundus image with the corresponding medical record document, the association of the OCT image with the intraocular pressure examination data, etc.) and other key attributes. If it is structured data, go to S22; if it is unstructured data, go to S23; if it is multimedia data, go to S24;

[0061] Specifically, taking the patient information table as an example, its data source is marked as a specific patient information table of the MySQL database, the data type is clearly structured, the format is recorded in detail as a database table structure, the creation time is accurately recorded as the creation time of the database record, and the association relationship is analyzed by an intelligent association analysis algorithm. The association between the possible related tables of other ophthalmic examinations, such as the association between the intraocular pressure examination data table and the patient number, etc. These metadata information is strictly inserted into the metadata description layer according to the "Metadata" table structure definition, and the metadata storage uses a compression algorithm to save storage space.

[0062] The data storage unit classifies and stores the data according to the data type.

[0063] S22, structured data, such as patient information table, intraocular pressure measurement data table, and visual acuity examination result table, is stored in a relational database storage area using InnoDB engine, strictly following the database design specification, and ensuring data integrity and consistency through B+ tree index combined with index coverage. Step ends and goes to S25;

[0064] S23, unstructured document data, such as Word format medical summary documents and PDF format surgical report documents, are stored in a specific directory under the Ceph distributed file system, according to the pre-designed directory structure, and are subdivided and stored according to factors such as department and medical record type. At the same time, the storage path information of the document is recorded in detail in the metadata description layer, and the high availability and load balancing characteristics of the distributed file system are used to ensure reliable data storage and fast access. Step ends and goes to S25;

[0065] S24 multimedia data, AMD-related fundus images, DR diabetic retinopathy fluorescence angiography videos, FFA fundus fluorescence angiography image sequences, FP fundus photographs, HR hypertensive retinopathy image data, MRI magnetic resonance imaging data, OCT optical coherence tomography images, RD retinal detachment image records, RVD retinal vein occlusion images, UBM ultrasonic biomicroscopy images and videos, FFA videos, UBM images and videos, based on MinIO object storage, stored in the form of objects in the multimedia storage area, each object is assigned a unique object key, and its mapping relationship with the object key and the resolution, duration, encoding format of the multimedia data are recorded in the metadata description layer. Metadata information is queried through multimedia data index construction technology based on the combination of hash index and inverted index. Step ends to S25;

[0066] S25 In the storage process, the index construction unit actively constructs data indexes, such as constructing indexes of patient information tables based on patient names, medical record numbers, and other keywords, using a dynamic index update strategy; unstructured document indexes are constructed based on document keywords, and natural language processing techniques are used to extract keywords such as disease names and treatment methods from medical records; multimedia data indexes are constructed based on image color features and OCT image retinal layering features, and feature vectors are extracted using deep learning algorithms to construct feature indexes, improving data query efficiency and helping doctors quickly retrieve and analyze patient ophthalmic data.

[0067] In one data transmission embodiment, as shown in Figure 3 An ophthalmic multi-source heterogeneous data transmission method is provided, specifically comprising:

[0068] S31 Data transmission preparation stage, when there is a data transmission requirement, the platform detection sub-module of the data transmission module uses a combination of system detection technology and network scanning technology to comprehensively detect the detection source platform (including operating system, device type, such as Windows system of hospital server, iOS or Android system of doctor mobile terminal, etc.) and target platform (including operating system, device type, network environment information, such as operating system, device model of doctor's tablet computer or smartphone, and current network bandwidth, delay, etc.) of data transmission. Subsequently, the transmission protocol adaptation sub-module selects the appropriate transmission protocol according to the detection results and performs application layer custom protocol optimization;

[0069] Specifically, when data needs to be transmitted to a doctor's mobile terminal, for example, from a hospital server to a doctor's tablet computer. The platform detection submodule of the data transmission module quickly detects that the source platform is a hospital server (the operating system is Windows, and the device type is a server), and the target platform is a doctor's tablet computer (the operating system is iOS, the device type is a tablet computer, the network bandwidth is 3 Mbps, and the delay is 30 ms). The transmission protocol adaptation submodule decisively selects the lightweight UDP protocol according to this information, and carefully optimizes the protocol at the application layer to balance transmission efficiency and data reliability, to meet the doctor's needs for quickly viewing ophthalmic data of patients in a mobile scenario.

[0070] In the data format conversion stage S32, the data format conversion submodule constructs a dynamic format recognition and conversion mechanism for different ophthalmic data types. When the system recognizes ophthalmic image data, it performs content detection on the image through a deep learning-based image recognition algorithm to determine the image compression and format conversion strategy. If the transmission is interrupted, it enters the breakpoint resume process S33.

[0071] Specifically, in the data collection stage, first, intelligent data source recognition technology is needed to judge the source type of each type of data and classify it. The source type includes structured, unstructured, image, and video. This step includes:

[0072] Step 1: Obtain data from multiple ophthalmic data sources, including databases (MySQL, Oracle), file systems (PDF, Word documents), image data sources (fundus images, OCT images), and video data sources (FFA videos, OCT videos). The process includes: judging the data source through port scanning, file header analysis, and interface protocol analysis method. For database data, determine whether it is structured data (MySQL, Oracle) based on the port number and connection string characteristics. For file system data, determine whether it is unstructured document data through file header information and file type recognition. For image and video data, use image recognition algorithms (deep learning CNN) to analyze file types and output data types, including structured, unstructured, image, and video. Classify the recognized data according to type and enter the subsequent processing flow, including image processing, video compression, and document parsing.

[0073] Step 2: Format conversion and adaptation for different types of data according to the requirements of the target platform. This step processes the data based on its specific content and the requirements of the target device (display device, transmission environment). For image data processing, high-resolution ophthalmic images (fundus images, OCT images) are input. The process includes analyzing the image content using a convolutional neural network (CNN) to identify the lesion area (retina, blood vessel distribution) in the image. Based on the image content, it is determined whether compression or resolution adjustment is needed. For background images or images without lesions, a high compression ratio (WebP format) is used to save storage and transmission bandwidth. For images containing critical lesion areas, more details are preserved, and a lower compression ratio is used. JPEG format is used. The output is an image format adapted to the target device (WebP, JPEG).

[0074] For video data processing: Input ophthalmic videos (FFA video, OCT video). Analyze the video content: extract the video frame rate, resolution, and encoding format through the FFmpeg library. Adjust the video resolution and bitrate based on network bandwidth and device processing capacity. For low-bandwidth or small-screen devices, select low-resolution, high-compression video formats (H.264 to H.265). For environments with sufficient bandwidth and high quality requirements, preserve high-resolution video content and select higher-quality encoding (MP4). Output video formats suitable for device and network conditions (low-bitrate H.264 video, low-resolution MP4 video).

[0075] For document data processing: Input ophthalmic medical record documents (PDF, Word documents). Use PDF parsing libraries (iText) or document parsing techniques (regular expressions, DOM parsing) to extract key information from the documents. According to the display requirements of the target platform (mobile device or PC), select the appropriate file format conversion to convert the document from PDF or Word format to a structured format (JSON or XML) for machine reading and rapid processing. Output standardized structured document formats (JSON, XML).

[0076] Step 3: After format recognition and conversion, the system dynamically adjusts the compression method and format of the data according to the target device, network conditions, and data characteristics to ensure data transmission efficiency and user experience.

[0077] For image compression and adaptation: input: image data (fundus image, OCT image), if the target device supports high-resolution images, the system will store and transmit the image at a low compression rate (JPEG or PNG format), if the network bandwidth is low or the target device is a mobile device, the system will compress the image and select an appropriate compression format (WebP format) to reduce the amount of transmission data, output: compressed image (WebP format) or original image (JPEG format);

[0078] For video compression and adaptation: input: video data (OCT video, FFA video), according to the screen resolution and processing capacity of the target device, select the appropriate encoding format (H.265, H.264), adaptively adjust the video, compress the video (reduce code rate, resolution adjustment) during transmission, output: optimized low code rate video format (H.264 or H.265 video);

[0079] For document compression and adaptation: input: unstructured documents (PDF, Word documents), structured conversion of the document, extraction and conversion of the document content into appropriate file formats (JSON, XML), according to the complexity of the document content, choose whether to compress, retain key data, reduce irrelevant content, output: structured document data (JSON, XML format);

[0080] Step 4: In order to ensure the real-time and stability of the format conversion process, the system introduces real-time monitoring and feedback mechanism to deal with network fluctuations, bandwidth limitations, data flow and network monitoring, input image, video, document transmission request, real-time monitoring of network bandwidth, delay and data flow, when the system detects insufficient network bandwidth or high delay, automatically adjust the transmission strategy: higher compression of images, use more suitable format for low bandwidth environment (WebP), reduce resolution or code rate of video, use low resolution video streaming, output optimized image, video, document data transmission stream;

[0081] Step 5: Format adaptation and transmission, transfer the converted and optimized data to the target device or platform, input the converted and optimized data (image, video, document), according to the requirements of the target platform (mobile device, PC, hospital server), automatically select the appropriate transmission protocol (HTTP, FTP, UDP), according to the display characteristics of the device and network conditions, ensure that the data can be seamlessly adapted, realize data breakpoint resume transmission, when the transmission process is interrupted, the transmission can be resumed from the breakpoint, output the data successfully transmitted to the target device or platform;

[0082] The implementation process of dynamic format recognition and conversion includes automatic recognition of data source, intelligent conversion of format, data compression and adaptation, real-time monitoring and feedback mechanism. Through these steps, the ophthalmic data can be seamlessly, efficiently and accurately transmitted and stored in different devices and network environments, so as to optimize the data transmission efficiency and improve the user experience.

[0083] In the breakpoint resume stage S33, if the transmission is interrupted, the breakpoint resume unit records the breakpoint position, and continues to transmit data from the breakpoint position according to the record after the network is restored, so as to ensure the complete transmission of data to the target platform.

[0084] Specifically, if the doctor's tablet computer suddenly loses network connection during transmission, the breakpoint resume unit quickly records the breakpoint position of the transmission, such as accurately recording the transmitted data block number, file pointer position and other information. When the tablet computer is reconnected to the network, the breakpoint resume unit continues to transmit data from the breakpoint position according to the recorded breakpoint position, so as to ensure that the data is completely transmitted to the doctor's tablet computer, and the doctor can smoothly obtain the complete ophthalmic data of the patient, without affecting the continuity of the diagnosis work.

[0085] In one data security management embodiment, as shown in Figure 4 , a data encryption method for ophthalmic multi-source heterogeneous data security management is provided, which specifically includes:

[0086] The data encryption submodule S41 identifies sensitive data and quickly locates sensitive data by using data classification and sensitivity marking technology, such as patient privacy information (including ID card number, home address, contact information, etc.), rare ophthalmic disease research data and high-precision ophthalmic examination image data, etc.

[0087] For sensitive data, the AES symmetric encryption algorithm is used to quickly encrypt the data content S42.

[0088] The RSA asymmetric encryption algorithm is used to encrypt and transmit the AES encrypted key, so as to protect the data security, and the optimized key management strategy is used to ensure the secure distribution and storage of the key.

[0089] Specifically, the data encryption submodule encrypts the sensitive ophthalmic data to ensure the security of the data. For example, for the high-definition fundus image data of the patient, the AES symmetric encryption algorithm is first used to quickly encrypt the image data content, and then the RSA asymmetric encryption algorithm is used to encrypt and transmit the AES encrypted key, so as to ensure the security of the data in the storage and transmission process. In the encryption process, the key management server is used to centrally generate and distribute the key, and the key is periodically updated and backed up, stored and distributed.

[0090] In one data security management embodiment, as shown in Figure 5, provides an access permission control method for ophthalmic multi-source heterogeneous data security management, specifically comprising:

[0091] The S51 access permission control submodule determines different roles according to the role-based access control (RBAC) model. If it is an ophthalmic specialist role, it goes to S52; if it is an ordinary ophthalmic physician, it goes to S53; if it is a nurse role, it goes to S54; if it is a research personnel role, it goes to S55; and if it is a patient role, it goes to S56.

[0092] S52 gives the ophthalmic specialist role full access permission (read, write, modify, and delete) to all patient data, and records the permission operations in detail through the audit function of the permission management system, so as to comprehensively control the diagnosis and treatment plan for difficult cases;

[0093] S53 grants the ordinary ophthalmic physician role read and modify permissions for partial data (such as basic information, routine examination data, preliminary diagnosis results, etc.) of patients under his / her responsibility, and uses a permission refinement strategy to assign accurate access scope according to the range of patients under the responsibility of the physician, thereby ensuring the privacy and security of patient data;

[0094] S54 sets the nurse role to have read permission for only viewing patient basic information (such as name, gender, age, etc.) and nursing-related data (such as nursing records, medication, etc.), so as to perform daily nursing work;

[0095] S55 sets the research personnel role to have read and analyze permission for anonymized ophthalmic research data, but without modify and delete permissions, thereby preventing data leakage and illegal access, providing a safe data access environment for research personnel through data desensitization technology, and promoting the progress of ophthalmic medical research;

[0096] S56 gives the patient role read permission for only viewing his / her own basic examination results (such as visual acuity, intraocular pressure, etc.) and diagnosis information, so that the patient can understand his / her own condition.

[0097] In one data security management embodiment, as Figure 6 , provides a data backup and recovery method for ophthalmic multi-source heterogeneous data security management, specifically comprising:

[0098] S61, the data backup and recovery submodule sets a full backup cycle according to data importance and update frequency, such as full backup of all new ophthalmic examination data generated in the day at dawn, and uses an efficient backup task scheduling algorithm to ensure that the backup task is executed on time;

[0099] S62, when the data is updated, such as changes in the patient's intraocular pressure examination data, timely record the changed data block for incremental backup, and use data change monitoring technology to capture data updates in real time;

[0100] S63 According to the data importance, determine the storage location, important data (such as the original examination image of the patient, long-term medical record data, etc.) is backed up to the off-site disaster recovery center, adopts dedicated line transmission and encryption storage technology to ensure the safety and integrity of the off-site backup data, and the ordinary business data (such as temporary examination record summary, etc.) is locally backed up;

[0101] S64 When the data encounters loss or damage, the data is restored to the specified time point by means of backup data and recovery technology based on data log, so as to guarantee the data availability and integrity.

[0102] In one embodiment of data monitoring and management, as shown in Figure 7 , a data flow monitoring method for ophthalmic multi-source heterogeneous data monitoring and management is provided, which specifically comprises:

[0103] S71 The data flow monitoring submodule is started, a flow monitoring tool based on the SNMP protocol is used, a sliding window algorithm is used to count flow data, high-performance flow monitoring equipment and real-time data processing algorithms are used to ensure accurate collection and rapid calculation of flow data;

[0104] S72 The flow of real-time monitoring data in the storage and transmission process is monitored, the data inflow, outflow and data transmission rate are counted;

[0105] S73 The real-time flow data is compared with the preset flow threshold model (constructed based on historical flow data and ophthalmic data processing system performance indicators), and the flow threshold model uses an intelligent dynamic adjustment strategy to optimize the threshold in real time according to the system running state;

[0106] S74 If the data flow exceeds the threshold, an alarm is sent in time, and an investigation process is started to locate and solve the flow abnormality problem, such as checking the network bandwidth occupation, judging whether there is malicious data transmission or network failure, etc., in the flow monitoring process, a flow data real-time analysis algorithm is used, which can send an alarm within 1 second after the flow abnormality occurs, and the false alarm rate is controlled below 2%.

[0107] Specifically, as shown in Figure 8 , an ophthalmic multi-source heterogeneous data quality monitoring method is provided, which specifically comprises:

[0108] S81 The data quality monitoring submodule loads a perfect data verification rule library and an intelligent data cleaning and repair algorithm, the data verification rule library uses modular design and can flexibly extend the verification rules;

[0109] S82 The completeness of the stored data is comprehensively checked, including checking whether the required fields in various data tables (such as patient information table) are complete, and whether the ophthalmic examination data record is missing;

[0110] S83, check data accuracy, check the numerical correctness and data format specification of key business data (such as whether the intraocular pressure measurement data is within a reasonable range, whether the diagnosis information in the medical record is accurate and standardized, etc.);

[0111] S84, perform data consistency comparison, check whether the same data of the same patient generated by different examination equipment, such as visual acuity data and intraocular pressure data, are consistent;

[0112] S85, when the data quality is found to be substandard, automatically trigger the corresponding processing flow according to the data problem type, such as starting the data cleaning process for duplicate records, removing duplicate data using an efficient duplicate detection algorithm; repairing the input error data, using data association relationship, ophthalmic professional knowledge and business rules for intelligent repair; deleting serious errors or expired data to ensure that the data quality meets the needs of ophthalmic medical and scientific research. In the data quality monitoring test, for large-scale ophthalmic data samples, the data integrity verification accuracy can reach more than 99%, the accuracy verification accuracy can reach more than 98%, and the consistency verification accuracy can reach more than 97%.

[0113] Specifically, as Figure 9 , an ophthalmic multi-source heterogeneous data system log management method is provided, which specifically comprises:

[0114] S91, the system log management submodule initializes the efficient log recording framework Log4j, and adopts a log grading and classification storage strategy to optimize log management;

[0115] S92, detailed running log information is recorded at each key link of data processing, including the source address (accurate to the IP address of the ophthalmic examination equipment, the directory of the server where the medical record document is located), the collection time (accurate to milliseconds), and the collection data volume (in bytes) at the data collection link;

[0116] S93, the storage location (table space of a relational database, path of a distributed file system), storage time, and storage data format at the data storage link are recorded;

[0117] S94, the source platform (source server host name or examination equipment number), target platform (device number of the target mobile device or receiving server name), transmission protocol (details of the UDP protocol and application layer custom protocol), and transmission data volume at the data transmission link are recorded;

[0118] S95, the encryption operation (ophthalmic data involved, encryption algorithm and key information), access permission setting (user role and specific permission), and backup recovery record (time, content and recovery operation details of full backup and incremental backup) at the data security management link are recorded;

[0119] S96 records flow abnormal information (time of flow exceeding threshold value, flow peak value, etc.), quality problem data (ophthalmic data details with problems, problem type and processing result) of data monitoring link, so as to comprehensively trace data operation history for subsequent system troubleshooting and audit analysis. Through the efficient retrieval function of the log management system, the average response time of querying specific event related logs in massive log data can be controlled within 2 seconds.

[0120] In one embodiment, as shown in Figure 10 An ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission device is provided, comprising a data acquisition module, a data fusion storage module, a data transmission module, a data security management module, and a data monitoring and management module.

[0121] The data acquisition module is configured to acquire data of different structures from multiple different ophthalmic data sources, including acquiring data from an ophthalmic patient information database, an ophthalmic medical record document system, various ophthalmic examination devices (fundus camera, OCT device, FFA device, UBM device), and related network data sources and specific application program interfaces.

[0122] The data fusion storage module is configured to fuse and store the acquired ophthalmic multi-source heterogeneous data according to the data storage model, and provide a data storage interface to enable external systems to access the stored ophthalmic data, thereby facilitating data management and calling.

[0123] The data transmission module is configured to realize seamless data transmission between different operating systems (Windows system of hospital server, iOS or Android system of doctor mobile terminal) and different device types (large ophthalmic examination device, desktop computer, notebook computer, tablet computer, smart phone), thereby meeting the data interaction needs between diversified devices and systems in ophthalmic medical scenarios.

[0124] The data security management module is configured to encrypt ophthalmic data during storage and transmission, control access permissions, and perform data backup and recovery, thereby ensuring data security and integrity.

[0125] The data monitoring and management module is configured to monitor the situation of data during storage and transmission in real time, monitor the quality of stored ophthalmic data, and record the running log of the software system, thereby ensuring the stability and reliability of the data processing flow.

[0126] In the aspect of the related setting of the ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission device, the content defined in the foregoing method of the ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission can be referred to, and will not be described in detail here. Each module in the above multi-source heterogeneous fusion storage and transmission device can be realized by a software approach, a hardware method, or a combination of the two. These modules can be directly embedded into the processor of the computer device in the form of hardware, or stored in the memory of the computer device in the form of software, so that the processor can call and execute the operation process corresponding to each module when needed.

[0127] In an embodiment, a computer device is provided, which can be a terminal type. The internal structure diagram of the computer device can refer to the following figure. Figure 11 The computer device is connected with the processor, the memory, the network interface, the display screen, and the input device by means of the system bus. The processor undertakes the task of providing computing capacity and control function. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores the operating system and the related computer program, and the internal memory creates suitable environmental conditions for the normal operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device can enable the device to be connected with and communicate with other ophthalmic related terminals outside through the network. When the computer program is executed by the processor, the specific method of the ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission can be achieved. The display screen of the computer device is a liquid crystal display screen or an electronic ink display screen, and the input device can be a touch layer overlaid on the display screen, or a key, a trackball, or a touchpad arranged on the shell of the computer device, or even an external keyboard, a touchpad, or a mouse.

[0128] Those skilled in the art should know that, Figure 11 The structure shown in the figure is only part of the structure framework related to the scheme of the present application, and it cannot absolutely limit the computer device to which the scheme of the present application is applied. In fact, the specific computer device can cover more or fewer components than those shown in the figure, or some components can be combined for use, or have a different component layout form.

[0129] In an embodiment, a computer device is provided, which includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, each step in the above method embodiment can be implemented one by one, thereby realizing the efficient processing and management of the ophthalmic multi-source heterogeneous data.

[0130] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, once the computer program is executed by a processor, each step flow of the above-mentioned method embodiments can be executed in sequence, and a solid software running foundation is provided for ophthalmic data processing.

[0131] It can be understood by those skilled in the art that, in order to realize all or part of the operation flows in the above-mentioned method embodiments, a computer program can be used to issue instructions to the related hardware to complete the task. The computer program can be stored in a non-volatile computer readable storage medium. During the execution of the computer program, the flow contents contained in each of the above-mentioned method embodiments can be involved. It should be noted that any reference to memory, storage, database or other medium in each of the embodiments provided in the present application can cover non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. For example, RAM exists in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM) and memory bus dynamic RAM (RDRAM), etc.

[0132] It should be noted that, in this paper, the relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment. Without more limitations, the element defined by the statement "including a" does not exclude the presence of other identical elements in the process, method, article or equipment including the element.

[0133] While embodiments of the application have been shown and described, it is to be understood that the embodiments described are merely exemplary of the principles and application of the present application. Numerous modifications and adaptions can be effected without departing from the spirit and scope of the present application, which is not limited to the exact construction and arrangement described. It is intended, therefore, to cover all modifications and adaptions that fall within the scope of the claims and their equivalents.

Claims

1. An ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission device, characterized by: include: A data acquisition module is used to collect ophthalmic multi-source heterogeneous data from multiple different sources, wherein the ophthalmic multi-source heterogeneous data includes ophthalmic structured database data, ophthalmic unstructured document data, and ophthalmic multimedia data; The data fusion storage module builds a data storage model for ophthalmic data, fuses and stores the collected multi-source heterogeneous ophthalmic data according to the data storage model, and provides a data storage interface; The data transmission module is used to achieve seamless data transmission across different operating systems and different device types. During the data transmission process, ophthalmic data format conversion and adaptation are automatically performed according to the characteristics of the target platform; The breakpoint resume unit records the breakpoint location and continues data transmission from the breakpoint according to the record after the network is restored, ensuring that the data is fully transmitted to the target platform; The data storage model in the data fusion storage module includes: The metadata description layer is used to store metadata information of multi-source heterogeneous ophthalmology data. The metadata information includes the source of ophthalmology data, data type, data format, data creation time, and data association relationship. The metadata information is stored in the "OphthalmologyMetadata" table structure in the form of key-value pairs; The data storage layer classifies ophthalmic data and stores them in different data storage areas according to their data types. A relational database is used to store ophthalmic structured data, which is stored in the relational database storage area using the InnoDB engine and strictly follows the ophthalmic database design specifications. A file storage system is used to store unstructured ophthalmic document data, which is stored in a specific ophthalmic directory of the Ceph distributed file system in an orderly manner and is subdivided and stored according to the preset ophthalmic directory structure rules. An ophthalmic multimedia repository is used to store ophthalmic multimedia data, which is stored in the ophthalmic multimedia storage area in the form of objects based on MinIO object storage. Each object is assigned a unique object key, and a mapping relationship is established between the data storage layer and the metadata description layer. Query is performed through multimedia data index construction technology based on a combination of hash index and inverted index, and metadata and stored data are mapped through data structures based on hash tables or index trees. The data transmission module includes: The platform detection submodule is used to detect the operating system type, device type, and network environment information of the source and target platforms for data transmission; The transmission protocol adapter submodule selects the appropriate transmission protocol for ophthalmic data transmission based on the information detected by the platform detection submodule. It uses a lightweight transmission protocol for low-bandwidth network environments and a reliable transmission protocol for high-bandwidth environments with high data integrity requirements. The data format conversion submodule builds a dynamic format recognition and conversion mechanism for different ophthalmic data types, and proposes intelligent format conversion based on deep learning and image content. It performs targeted compression based on the lesion area in the ophthalmic image to avoid the loss of important diagnostic information during the compression process. For images containing critical lesion areas, details are retained and a low compression ratio is used. When the system recognizes ophthalmic image data, it uses a deep learning-based image recognition algorithm to detect the image content and determine the image compression and format conversion strategy; The process of building a dynamic format recognition and conversion mechanism includes: source identification and classification of input ophthalmic data through the data acquisition module; after the data is classified, the characteristics of each data format are identified, and the conversion method and format are selected for the type of data. Compression and conversion are performed according to the type and requirements of the target device. A real-time monitoring and feedback mechanism is designed, and the transmission requests of images, videos, and documents are used as input. The network bandwidth, delay and data traffic are monitored in real time. When the system detects insufficient network bandwidth or excessive delay, it automatically adjusts the transmission strategy, outputs optimized image, video, and document data transmission streams, and transmits the converted and optimized data to the target device or platform.

2. The device for fusion storage and cross-platform transmission of multi-source heterogeneous ophthalmic data according to claim 1, characterized in that: The data acquisition module includes: A data source identification submodule is used to automatically identify the source type of ophthalmic data, including ophthalmic database systems, ophthalmic file systems, ophthalmic network data sources, and specific ophthalmic application program interfaces; The data extraction submodule uses the corresponding extraction technology to extract ophthalmic data from the data source according to the source type determined by the data source identification submodule. For ophthalmic structured database data, SQL query statements or database connection tools are used for extraction. Specifically, JDBC connection is used for specific ophthalmic databases, and precise SQL query statements are constructed based on the database table structure to extract data. For ophthalmic unstructured document data, document parsing technology is used to extract text content. For ophthalmic PDF documents, the iText library is used to parse and extract text, pictures, and tables. For ophthalmic multimedia data, a special multimedia data reading library is used for extraction. For FFA fluorescence angiography videos and UBM ultrasound biomicroscopy videos, the FFmpeg library is enabled to extract channel information, sampling rate, video frame rate, resolution, and encoding format.

3. The ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission device according to claim 2 further comprises a data security management module, characterized in that: The data encryption submodule is used to encrypt ophthalmic data during storage and transmission. It uses a combination of symmetric and asymmetric encryption algorithms to encrypt sensitive ophthalmic data. During the encryption process, a key management server is used to centrally generate and distribute keys, and the keys are regularly updated and backed up, and storage and distribution strategies are implemented. The access rights control submodule sets different access rights based on user roles and the sensitivity of ophthalmic data. Only users or systems with corresponding permissions can access specific ophthalmic data. Access rights include read, write, modify, and delete permissions. The role-based access control (RBAC) model uses a specific permission allocation table structure for permission setting and management. The data backup and recovery submodule regularly backs up the stored ophthalmic data and restores the data based on the backup data when the data is lost or damaged. The data backup strategy is flexibly set according to the importance and update frequency of the ophthalmic data. The database backup tool mysqldump is used every morning to perform a full backup of the core ophthalmic business data. When the data is updated, the changes in the ophthalmic data are recorded through the data log, and incremental backups are performed based on the log. The storage location is determined according to the importance of the ophthalmic data. Important ophthalmic data is backed up to an off-site disaster recovery center, while ordinary ophthalmic business data is backed up locally.

4. The ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission device according to claim 3 further comprises a data monitoring and management module, characterized in that: The data flow monitoring submodule monitors the flow of ophthalmic data during storage and transmission in real time, including data inflow, outflow, and data transmission rate. It issues an alarm when data flow exceeds a preset threshold. This module uses network flow monitoring technology and a data flow statistics algorithm. The preset flow threshold model is based on a threshold curve constructed from historical ophthalmic flow data and system performance indicators. The data quality monitoring submodule monitors the quality of stored ophthalmic data, checks the integrity, accuracy, and consistency of the data, and marks and processes ophthalmic data that does not meet the quality requirements. The processing methods include ophthalmic data cleaning, data repair, or data deletion. Among them, a data verification rule base obtained by storing verification rules in the form of database tables and a data cleaning and repair algorithm are used; The system log management submodule records the operation log of the software system, including detailed information on ophthalmic data collection, storage, transmission, security management and monitoring operations. It uses a log recording framework to record log information, including the source address, collection time, and amount of collected data for ophthalmic data collection; the storage location, storage time, and storage data format of ophthalmic data storage; the source platform, target platform, transmission protocol, and amount of transmitted data for ophthalmic data storage; encryption operations, access permission settings, and backup and recovery records for ophthalmic data security management; and traffic anomaly information and quality problem data for ophthalmic data monitoring.

5. The device for fusion storage and cross-platform transmission of multi-source heterogeneous ophthalmic data according to claim 4, characterized in that: include: The data acquisition module is used to collect data of different structures from multiple different ophthalmic data sources, including ophthalmic patient information databases, ophthalmic medical record document systems, various ophthalmic examination equipment (fundus cameras, OCT equipment, FFA equipment, UBM equipment), as well as related network data sources and specific application interfaces; A data fusion storage module is used to fuse and store the collected ophthalmic multi-source heterogeneous data according to the data storage model, and provide a data storage interface for data management and call; The data transmission module is used to achieve seamless data transmission between different operating systems and different types of devices, and is used to meet the data interaction needs between diverse devices and systems in ophthalmic medical scenarios; Data security management module, used to encrypt ophthalmic data during storage and transmission, control access permissions, and perform data backup and recovery to ensure data security and integrity; The data monitoring and management module is used to monitor the data in real time during storage and transmission, monitor the quality of stored ophthalmic data, and record the operation log of the software system to ensure the stability and reliability of the data processing process.

6. A computer device, characterized in that: Its structure covers a memory, a processor, and a computer program stored in the memory and capable of running on the processor. When executing the computer program, the processor can perform the functions of the ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission device as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that This storage medium stores a computer program. When the computer program is executed by a processor, the processor performs the functions of the ophthalmic multi-source heterogeneous data fusion storage and cross-platform transmission device as described in any one of claims 1 to 4.

Images