A data transmission system based on screen display and image shooting recognition

Abstract

The application provides a data transmission system based on screen display and image photographing recognition. The system comprises a data processing unit, an image processing unit, a display unit, a camera and an inner and outer machine box; the data processing unit encodes data to be sent, generates an image, and displays the image in the display unit; the camera photographs the image displayed in the display unit in the same inner machine box; the image processing unit identifies the picture photographed by the camera, and stores the identified data. The inner and outer machine box is composed of an outer machine box and two inner machine boxes; the inner machine box comprises a supporting part for fixing the display unit and the camera; the inner machine box is fixed in the outer machine box; and the outer machine box comprises a machine box shell and an exhaust fan. The system realizes physical isolation of the inner network and the outer network in a complete sense by using the image recognition mode, and realizes safe, efficient and accurate transmission of data between the inner network and the outer network.

Description

Technical Field

[0001] This invention relates to the field of information security technology, and in particular to a data transmission system based on screen display and image capture and recognition. Background Technology

[0002] In the information age, to protect internal data security, such as the core internal data of enterprises and the important data stored by government agencies and institutions, it is usually necessary to establish an internal local area network (LAN). Meanwhile, the external Internet (EXB) that we use daily provides extensive and convenient network connectivity. The data exchange and transmission between the internal and external networks raises concerns about the security of core internal data and sensitive information. How to balance the ease of use of both networks with the data security of the internal network has become a challenge for enterprises, government agencies, and other organizations.

[0003] Currently, existing technologies for maintaining secure data transmission between internal and external networks include: installing both internal and external network cards on a single computer to create a dual-network system, protected by a professional firewall; and setting up a secure isolation gateway between two independent host systems, separating the internal and external network processing units through an isolation hardware switching unit. The connection between the dedicated isolation hardware switching unit and the two processing units is controlled by hardware circuitry, ensuring that only the internal or external network processing unit is connected at any given time, using an information "ferry" method to transmit data.

[0004] The disadvantages of the existing methods for maintaining the security of data transmission between internal and external networks include: once a virus invades the computer through the external network, there is still a significant security risk. Setting up a security isolation gateway is a complex process that requires reliance on professional hardware vendors. While it achieves physically isolated file transfer, security isolation gateways are expensive. Summary of the Invention

[0005] Embodiments of the present invention provide a data transmission system based on screen display and image capture recognition to achieve secure, efficient and accurate transmission of data between internal and external networks.

[0006] To achieve the above objectives, the present invention adopts the following technical solution.

[0007] A data transmission system based on screen display and image capture recognition, the entire system consists of a data processing unit, an image processing unit, a display unit, a camera, and internal and external chassis;

[0008] The data processing unit reads and encodes the data to be sent, generates an image, and displays the image on the display unit. The camera takes a picture of the image displayed on the display unit, and the image processing unit recognizes the picture taken by the camera and stores the recognized data.

[0009] Preferably, the data processing unit is used for data encoding and display calculations. When transmitting a message containing mCount characters, the transmission time is required to not exceed t0. The time consumed by calling the image recognition algorithm to recognize an image is t1. The time for processing the recognized data and storing it in the database is t2. The minimum number of times the same image is captured is set to k. The display time for each page of data on the screen is t3. The number of pages of images that need to be displayed on the screen to transmit this data is n. The number of characters displayed on each page of the screen is nCount. The above data must meet the following conditions:

[0010] t3=(t1+t2)×k

[0011] n×t3<t0

[0012] nCount ≥ mCount ÷ n.

[0013] Preferably, the data processing unit is further configured to fix the number of characters displayed in each line of data according to the number of characters nCount displayed per page of the screen and the font size f, to ensure that the number of characters displayed in each line is the same, and to set a limit on the number of lines of displayed data to prevent the displayed content from exceeding the screen display range.

[0014] Preferably, the image processing unit is used to establish a deep learning text recognition algorithm framework and train algorithm parameters. It then uses the text recognition algorithm to perform text recognition on each image in the storage folder, compares the format of the data in the recognition result with the set data format, and deletes the image information if the comparison result indicates a format mismatch. If the comparison result indicates a format match, it processes the recognized data, deletes the corresponding placeholders, separators, and line breaks, extracts the page header name, and temporarily stores the recognized data. If the same recognition result occurs more than twice, the recognition is considered successful, and the data is written to the corresponding storage unit.

[0015] Preferably, the image processing unit is further configured to set the maximum and minimum focal lengths of the camera to fmax and fmin, respectively. max and f min Set the distance between the camera and the screen to S0. Place a chessboard image within the area where the characters are displayed on the screen. The chessboard image is composed of square black and white squares of the same size. Set each row of the chessboard to have m squares and each column to have n squares. Use the camera at a distance of s0 to vertically align with the center of the image and take a picture of the chessboard image. Calculate and process the picture of the chessboard image.

[0016] Let P be the black-and-white dividing point. i，j The index value corresponds to the ordinal number of the square it occupies. In the chessboard image, the bottom left corner P... 0，0 The pixel of the point is (X 0，0Y 0，0 (), upper right corner point P m，n The pixel of the point is (X m，n Y m，n If P, then m，n The point coordinates are (X m，n -X 0，0 Y m，n -Y 0，0 For each black-and-white separator P i，j The coordinate values ​​are calculated by comparing the x and y coordinates of each black and white dividing point with the coordinates of the upper right corner point P. m，n Divide the x and y coordinates to obtain the relative positions of the black and white dividing points in the actual captured image within the entire chessboard grid.

[0017] k s The parameter representing the degree of distortion in the captured image is calculated as follows:

[0018]

[0019] Where X i，j and Y i，j For each black and white dividing point P ij The x and y coordinates;

[0020] When k s When the value is less than the set threshold, it is determined that the accuracy of the image captured by the camera meets the requirements. If it meets the requirements, the current S0 is determined as the distance between the camera and the screen.

[0021] Preferably, the inner and outer chassis consist of one outer chassis and two inner chassis. The inner chassis includes a support portion for fixing the display unit and the camera. The inner chassis is fixed in the outer chassis. The outer chassis includes a chassis shell and an exhaust fan.

[0022] One side of the inner chassis shell provides a circular hole for mounting the camera, with the hole located at the center of this side and the diameter of the hole matching the diameter of the camera lens; the other side of the inner chassis is open, with the display unit mounted close to this side, and the chassis shell on this side can surround the display unit; the depth of the inner chassis is determined by the distance between the camera and the screen, the size of the side of the inner chassis is consistent with the size of the display screen, and the width of the inner chassis is smaller than the width of the outer chassis.

[0023] Preferably, the data processing unit, image processing unit, and display unit are implemented through an all-in-one machine. The data processing unit in the all-in-one machine connected to the external network reads the data to be transmitted from the external network server, processes the data, and displays it on the screen of the external network all-in-one machine. The external network camera takes a picture and transmits it to the all-in-one machine on the internal network. The image processing unit in the internal network all-in-one machine performs text recognition on the image, stores the recognition result, and sends it to the internal network server. The process of internal network data transmission to the external network is similar to the process of external network data transmission to the internal network.

[0024] As can be seen from the technical solutions provided by the embodiments of the present invention above, the present invention provides a data security transmission system based on screen display and image capture recognition, which enables secure, efficient and accurate transmission of data between internal and external networks while ensuring complete physical isolation between internal and external networks.

[0025] Further features and advantages of the invention will be set forth in the description which follows, some of which will be obvious from the description or may be learned from practice of the invention. Attached Figure Description

[0026] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0027] Figure 1 A structural diagram of a data security transmission system based on screen display and image capture recognition provided in an embodiment of the present invention;

[0028] Figure 2 This is a schematic diagram of a chessboard grid composed of alternating black and white squares of the same size, provided as an embodiment of the present invention.

[0029] Figure 3 This is a flowchart of an image text recognition algorithm provided in an embodiment of the present invention. Detailed Implementation

[0030] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0031] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this specification means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “coupled” to another element, it can be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein can include wireless connections or couplings. The term “and / or” as used herein includes any and all combinations of one or more of the associated listed items.

[0032] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless defined as herein.

[0033] To facilitate understanding of the embodiments of the present invention, the following will provide further explanation and description with reference to the accompanying drawings and several specific embodiments. These embodiments do not constitute a limitation on the embodiments of the present invention.

[0034] Example 1

[0035] This invention provides a data transmission system based on screen display and image capture recognition, enabling secure, efficient, and accurate transmission of data between internal and external networks while ensuring complete physical isolation between them.

[0036] A structural diagram of a data transmission system based on screen display and image capture recognition provided in this embodiment of the invention is shown below. Figure 1As shown, the system includes inner and outer chassis, a data processing unit, a display unit, a camera, and an image processing unit. The inner and outer chassis consist of one outer chassis and two inner chassis. The inner chassis include a support structure for fixing the display unit and camera; the inner chassis are fixed inside the outer chassis, which includes a casing and a mounting structure for a ventilation fan. The data processing unit encodes the data to be sent, generates an image, and displays the image on the display unit, which is typically a monitor connected to the host computer, but can also be an all-in-one device. The camera captures images displayed on the display unit within the same inner chassis. The image processing unit recognizes the images captured by the camera and stores the recognized data in a database. The recognition tool used in the image processing unit can be a self-trained recognition model or a third-party SDK tool.

[0037] The working process of the above-mentioned data transmission system based on screen display and image recognition includes the following steps:

[0038] Step S1: Design the specifications of the equipment installation chassis.

[0039] (1) Calculate the distance between the camera and the screen.

[0040] Set the camera's maximum and minimum focal lengths to f. max and f min Set the distance between the camera and the screen to s0. To ensure the clarity of the captured images, the range of s0 should be within the focal length range of the camera.

[0041] When the distance between the camera and the screen is small, the image clarity is higher, but the screen area captured by the camera is smaller, resulting in less data that can be recognized from a single image, and more significant image distortion. Conversely, when the distance between the camera and the screen is large, the image clarity decreases, affecting the recognition rate, and the design cost of the chassis also increases. Therefore, it is crucial to provide a user-friendly evaluation standard for the camera-screen distance.

[0042] Generally, in image recognition using a camera, camera calibration is the first step. However, considering the specific circumstances of this system, the target object is a character display interface on a screen, and the display area of ​​this character information on the screen is fixed. Cropping the image ensures the proportion of effective content in the captured image. Furthermore, due to advancements in modern cameras, image distortion has decreased. For 2D calibration scenarios with fixed camera position and display content area, distortion correction does not yield significant optimization results. Therefore, to reduce user complexity and the time spent on image distortion correction, the camera calibration process is omitted in this system design. Instead, referencing camera calibration methods, a method for testing image distortion is designed to evaluate the rationality of user-set camera-screen distance and camera parameters.

[0043] By placing a checkerboard image within the area of ​​the screen where characters are expected to be displayed, the checkerboard image is composed of alternating black and white squares of the same size, such as... Figure 2 As shown in the diagram. The chessboard is set to have m squares in each row and n squares in each column.

[0044] The process involves displaying a checkerboard image on a screen, taking a picture of the checkerboard image at a vertical angle with the lens positioned at a distance s0 from the center of the image, and then performing calculations and processing on the captured checkerboard image.

[0045] Measurements show that the origin P in the lower left corner of the chessboard image is... 0,0 The pixel of the point is (X 0,0 ,Y 0,0 (), upper right corner point P m,n The pixel of the point is (X m,n ,Y m,n ), then P m,n The point coordinates are (X m,n -X 0,0 ,Y m,n -Y 0,0 Similarly, for each black-and-white dividing point P... i,j The coordinates of point P are calculated to separate the points. i,j The subscript value corresponds to the block number of its position. This is achieved by comparing the calculated x and y coordinates of each black-and-white dividing point with the top-right corner point P. m,n Dividing the x and y coordinates gives the relative position of the black and white dividing points in the actual photograph within the entire chessboard grid.

[0046] k s The parameter representing the degree of distortion in the captured image is calculated as follows:

[0047]

[0048] Where X i，j and Y i，j For each black and white dividing point P ij The x and y coordinates. After testing, when k... s Images captured with a value less than 0.05 achieve high recognition accuracy; therefore, when designing the distance between the camera and the screen, the corresponding deformation parameter k can be calculated. s To evaluate the rationality of the design interval and camera parameter settings.

[0049] (2) Design the internal chassis size:

[0050] The main purpose of the inner chassis is to provide a stable enclosure to ensure that the camera and monitor remain in their original positions within the chassis, while also providing a dark environment for the camera to shoot in, avoiding interference from external light. One side of the inner chassis should have a circular hole for mounting the camera, centered on that side, with the hole's diameter related to the size of the camera lens. The other side of the inner chassis should be open, with the monitor mounted flush against it, and the chassis shell on that side should be able to completely surround the display screen. Therefore, the depth of the inner chassis is determined by the distance between the camera and the screen, the side dimensions of the inner chassis should match the size of the monitor screen, and the width of the inner chassis should be slightly smaller than the width of the outer chassis to ensure smooth connections for internal power cables, network cables, etc.

[0051] (3) Design the size of the external chassis:

[0052] The external chassis is fixed inside the server rack in the data center, providing a fixed position for the internal chassis, facilitating connection with servers or hosts in the data center, and providing mounting slots for exhaust fans to dissipate heat from internal cameras and monitors; the dimensions of this part are mainly designed with reference to the dimensions of the server rack in the data center.

[0053] Step S2: Design the data processing and display unit:

[0054] (1) Automatic data reading method:

[0055] Considering the more common database storage methods currently available, the following discussion will primarily focus on this scenario; other methods may be used in practical applications.

[0056] In the main thread, the database count is read at regular intervals and compared with the count read last time to determine if there is an update. When there is an update, the database is read from the last record count to the current read count, thus obtaining the updated data.

[0057] (2) Design method for display content:

[0058] Preferably, to ensure that the text in the captured image is clearer and more prominent, the overall style of the display interface on the screen adopts a black background and the displayed text is displayed in white font;

[0059] Preferably, to avoid excessive distortion in the images captured by the camera, the text display area should be within the checkerboard area in step S1;

[0060] Preferably, to avoid blurry images captured by the camera during data updates, each page of data images needs to be captured and recognized multiple times. Therefore, the display time for each page of data images must ensure that the camera can capture the image at least three times. When there is no data update, it will continue to be displayed with a black background.

[0061] Preferably, the font is boldface; wherein, the font size f is related to the number of characters nCount to be displayed per page on the display screen, and the following is the calculation method for the font size requirement:

[0062] Figure 3 This is a flowchart illustrating the processing of an image text recognition algorithm provided in an embodiment of the present invention. During data transmission, different application scenarios have different real-time requirements for data transmission. Assuming a scenario requires the transmission of information containing mCount characters, with a transmission time not exceeding t0, the time consumed by the image recognition algorithm to recognize an image is t1, and the time for processing the recognized data and storing it in the database is t2. Based on the requirement for multiple shots of a single image, the minimum number of times the same image is captured is set to k, the display time for each page of data on the screen is t3, the number of pages of images that need to be displayed on the screen to transmit this data is n, and the number of characters displayed on each page of the screen is nCount. The above data must meet the following conditions:

[0063] t3=(t1+t2)×k

[0064] n×t3<t0

[0065] nCount ≥ mCount ÷ n

[0066] The number of characters displayed in each line of data is fixed based on the number of characters displayed per page (nCount) and the font size (f) to ensure that the number of characters displayed in each line is the same. A limit is set on the number of lines of data to be displayed to prevent the displayed content from exceeding the screen display range.

[0067] Preferably, to avoid blurring around the camera-captured images, the data on both sides will be replaced with special character formats as placeholders without specific meaning.

[0068] (3) Data processing methods:

[0069] Based on the limitation on the number of data characters displayed per page in the above steps, the obtained updated data to be displayed is cropped to ensure that the data content displayed on the screen does not exceed the data display range on the screen;

[0070] Preferably, certain data content is replaced, such as replacing punctuation marks in the data, to improve recognition accuracy;

[0071] By introducing placeholders and adding specific separators to the data when each transmitted data contains data that needs to be separated, a method can be provided in the recognition module to detect the accuracy of the recognition results. In order to further improve the recognition accuracy, in addition to displaying the data in forward order, a reverse order display method can be added. At the same time, a unique name is added to each page of data displayed on the screen and placed at the beginning to avoid duplicate data entry. Furthermore, the recognition results of images with the same name can be compared to improve the recognition accuracy.

[0072] Step S3: Design the software running on the receiving image processing unit. The main thread of the software is used to run the image acquisition module, while the recognition and data storage modules run in sub-threads. This mainly consists of the following two steps:

[0073] (1) Image acquisition algorithm design:

[0074] The camera is connected to the receiving host, enabling the camera to take a picture at regular intervals (related to the camera's frame rate setting, to avoid excessive blurring in the pictures), and the acquired pictures are named and stored in a specific folder on the host in ascending numerical order.

[0075] (2) Image text recognition algorithm design:

[0076] The image text recognition algorithm employs a Convolutional Recurrent Neural Network (CRNN) as its overall framework. This neural network mainly consists of three layers: convolutional layers, recurrent layers, and transcription layers. The convolutional layers in the CRNN text recognition network are transformed into multiple residual blocks, each containing multiple convolutional operations. The total number of convolutional layers in the neural network is 50. The CNN layers in the CRNN framework consist of Layer 0 to Layer 5, and each layer contains multiple residual blocks. Introducing the residual block structure increases the network depth, thereby improving recognition accuracy. The residual blocks extract as many text image feature sequences as possible, and then a Long Short-Term Memory (LSTM) model is used to predict the label distribution of each frame, enhancing the network's ability to process sequence data. The sequence labels output by the screen text recognition network are processed through the transcription layer and integrated using a Connectionist Temporal Classification (CTC) algorithm to finally obtain the text content corresponding to the test image. The prediction results of each frame of the text image are integrated into the final label sequence for output.

[0077] (3) During training, the character table is set to ten numbers and uppercase and lowercase English letters, the maximum length of the recognized character is 25, the batch size is set to 64, and the epochs are set to 600 rounds.

[0078] The training process for the image text recognition algorithm is as follows:

[0079] 1. Generate images using historical data and display them on the monitor. Use a camera to capture a large number of images as a dataset, and use the text corresponding to the images as text labels for the images. Use a portion of the images as the training set and a portion of the images as the test set.

[0080] 2. Train the established CRNN network model using the training set images. The training character set is set to ten digits and uppercase and lowercase English letters, with a maximum character length of 25. The batch size is set to 64, and the training iterations are set to 600 rounds. The CRNN network weight parameter file is obtained after training.

[0081] 3. Initialize the network using the weight parameter file obtained from training, test it using the test set, and put it into use once the recognition accuracy meets the requirements.

[0082] The recognition module runs in a sub-thread, repeatedly reading the first image from the folder where the captured images are stored.

[0083] Preferably, the first image in the storage folder is used to perform text recognition using a recognition algorithm;

[0084] First, the format of the data in the recognition result is compared with the placeholders, separators, etc. designed in step S2. If a mismatch is found, the image information is directly deleted. If the format is correct, the recognized data is processed, and the corresponding placeholders, separators, line breaks, etc., are deleted. The page header name is extracted. When a new page name appears for the first time, this measure is used to avoid blurry images due to screen refresh rate. If the page data name does not appear for the first time, the recognized data is stored in the default cache list. When the same recognition result appears more than twice, the recognition is considered successful, and the data is written to the corresponding storage location on the internal network, such as a database.

[0085] When a stored image is recognized again, the image will be deleted directly to avoid wasting recognition time.

[0086] Preferred Example: This example addresses the data transmission needs of an internal and external network for a third-party developed freight train auxiliary driving system. The development and design of the auxiliary driving system require obtaining train timetable information from the dispatcher within the internal network system for algorithm optimization. Simultaneously, positioning devices installed on the freight trains provide the internal network dispatcher with real-time dynamic updates on the train's location. This example utilizes a data security transmission system based on screen display and image recognition, as provided by this invention, to address these two data transmission needs. The following example illustrates the design steps and operational effects of this system in practical application, using the acquisition of internal network phase plan information from the external network as an example. The acquisition of real-time train location updates from the external network from the internal network follows the same principle.

[0087] Example 2

[0088] Step S1: Design the size of the equipment rack for installation.

[0089] This example uses an Ethernet industrial area scan camera, model MVL-MF0828M-8MP. This camera was chosen primarily because of its high real-time frame rate of up to 9.4fps and the ability to adjust camera parameters via the camera's official software. The size of the external chassis is selected based on the actual application environment of the system. In this example, the system is planned to be installed in a server rack in a data center. The rack dimensions are 42U, 60cm wide, and 100cm deep. The external chassis should fit perfectly within the rack. The monitor used is 40cm wide and 25cm high, with a resolution of 1920x1080. Testing with checkerboard images showed that the optimal distance between the camera and monitor is 34cm. Therefore, the internal chassis dimensions are 36cm wide, 34cm deep, and 23cm high, while the external chassis dimensions are 56cm wide and 94cm deep.

[0090] Step S2: Design the data processing and display unit

[0091] (1) Automatic data reading method:

[0092] Consistent with the above method, automatic data reading and data processing / display are both designed as a single program. Every 1 minute, it reads the number of stage plan information tables in the database and compares it with the previously recorded number of stage plans. If the number read this time is greater than the recorded number, it is considered that there is a data update, and the characters in the data are processed and displayed. The page transition time interval is 3.75 seconds (i.e., the time t3 for each page of data to be displayed on the screen).

[0093] If no data is being transmitted, the default value for data transmission is "0000", which means there is no data to be transmitted at the moment, and it will be set to "0000" after each display ends.

[0094] (2) Design method for display content:

[0095] For a specific railway line, each stage of the planned data contains approximately 50 data entries, each with about 80 characters; the total number of characters, mCount, is 4000. Due to the relatively low real-time requirements of the external optimization equipment, the acceptable maximum time is approximately 120 seconds. In this example, the time t1 for recognizing a single image using the image recognition algorithm is approximately 2 seconds; the time t2 for processing the recognized data and storing it in the database is approximately 0.5 seconds. In this recognition algorithm, to avoid blurry images due to image data updates, the first image of each page is deleted. Furthermore, to verify the correctness of the recognition, the same image needs to be captured at least 3 times, k. Therefore, the time t3 for displaying each page of data on the screen can be calculated to be 7.5 seconds. The minimum number of characters displayed per page is 308, and the maximum number of pages n is 13. Testing shows that within the specified range for screen display, the font that meets the character requirements is boldface, size 36.

[0096] (3) Data processing and display:

[0097] Preferably, a program is written to refresh the timetable data content from the intranet host database and temporarily store the updated data in a cache variable; and since this example only requires timetable data within a specific line, data from other lines can be filtered out; in this example, an external monitor is used to acquire and display intranet data.

[0098] Preferably, a program is written to refresh the train dynamic information data from the external host database, and the updated data is temporarily stored in a cache variable; in this example, the external network acquires data on a server and displays the data through an external monitor.

[0099] The placeholders and separators used in this example are “K” and “t”.

[0100] In this example, the transmitted data is processed into a display data format. The stage plan mainly consists of timestamps, station names, and train numbers. The timestamp information is mainly in Arabic numerals, but it still contains special symbols ":" and "-", which can easily lead to recognition errors. Therefore, the two symbols are directly deleted for display. After recognition, the string is converted to datetime format when it is stored in the database. It should be noted that in actual applications, the datetime format may be missing the complete date form when converted to a string. For example, March 1, 2021, 12:10:10, converted to the string "2021031121010", is missing a "0" before the "1" (day), so it is necessary to judge and supplement the character.

[0101] The train number consists of uppercase English letters and Arabic numerals; to avoid conflicts with the placeholder "K" in the train number, "K" in the data is replaced with "m"; and some characters have a high recognition error rate, so they are replaced accordingly, "B" is replaced with "b", "C" is replaced with "n" and "Z" is replaced with "r".

[0102] The station name consists of fixed Chinese characters, which will be replaced by corresponding Arabic numeral codes;

[0103] Preferably, to prevent the data from exceeding the screen area, the maximum number of lines displayed on the external network is 9, with a maximum of 35 characters per line; the maximum number of lines displayed on the internal network is 11, with a maximum of 35 characters per line. Therefore, each page can display a maximum of approximately 4 pieces of phase plan data.

[0104] Organizing the displayed data: First, the acquired data is stored in a table format, and then divided into multiple tables with 4 data entries per table. Each table is named sequentially. The data content of each table is converted into a string format and stored as a corresponding variable, and then connected using "t". The converted and connected strings of each stage plan are connected using "K", and the name string of the table is added before the entire string.

[0105] The display form has a black background, and the font color in the controls is white with the font size "Hei" 36.

[0106] Preferably, to allow the identification module sufficient time to identify and store data, the timetable display on the intranet will last for approximately 7.5 seconds. Since the external network updates the dynamic information of trains every 5 seconds, it has higher requirements for real-time performance, and the dynamic information display time on the external network will last for approximately 5 seconds.

[0107] Step S3: Image Processing Unit Design:

[0108] The main thread of the image processing unit is connected to the camera, which captures and stores the images taken by the camera and names them in ascending numerical order. The child thread reads the first image (i.e. the earliest captured image) in the storage folder corresponding to the captured images in a loop and calls the recognition model to perform text recognition on the image.

[0109] The identified data results are processed. Extensive testing and statistical analysis of the recognition results revealed that similar characters may be misidentified during the recognition process. Since these characters have already been replaced by other characters during the display step, they will not be generated in correctly identified cases. Therefore, these incorrectly identified characters are replaced: "O" is replaced with "0", "o" is replaced with "0", "k" is replaced with "K", "B" is replaced with "8", "Z" is replaced with "2", "z" is replaced with "2", and "C" is replaced with "0".

[0110] The header name is obtained by extracting a substring and recorded. This header name is then used to check if the content of the image is the same as the previous one. If both the header name and the recognition result are the same, the image is considered successfully recognized. The obtained string is then processed through substring extraction and replacement, and the data is written to the corresponding location in the database. Subsequent images with the same header name are directly deleted to improve processing speed. When images have the same header name but different recognition results, the recognition result of the latter image is temporarily stored in a temporary variable. Subsequent recognition results with the same header name and result are then processed and written to the database. If the header name of a subsequent image changes, the data stored in the temporary variable is written to the database. After each image is recognized, it is deleted.

[0111] System operation results:

[0112] The data security transmission method based on screen display and image recognition described in this invention is used to meet the data transmission requirements of internal and external networks. The data transmission accuracy can reach 99.6%, and it takes about 100 seconds to transmit a set of phase plan information. The external network can transmit dynamic information data once within 10 seconds, which can meet the data transmission requirements of internal and external networks in this example.

[0113] In summary, the data transmission method in the data security transmission system based on screen display and image recognition disclosed in this invention does not rely on an internet connection or a third-party medium. It achieves complete physical isolation through image recognition, ensuring high recognition accuracy. This invention is applicable to most internal and external network transmission systems, and the installation of the device does not affect the normal operation of the internal or external network, demonstrating strong compatibility.

[0114] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of one embodiment, and the modules or processes shown in the drawings are not necessarily essential for implementing the present invention.

[0115] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of the present invention.

[0116] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for apparatus or system embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. The apparatus and system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0117] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A data transmission system based on screen display and image capture recognition, characterized in that, The entire system consists of a data processing unit, an image processing unit, a display unit, a camera, and internal and external chassis. The data processing unit reads and encodes the data to be sent, generates an image, and displays the image on the display unit. The camera takes a picture of the image displayed on the display unit, and the image processing unit recognizes the picture taken by the camera and stores the recognized data. The image processing unit is further configured to set the maximum and minimum focal lengths of the camera to f. max and f min Set the distance between the camera and the screen to S0. Place a chessboard image within the area where the characters are displayed on the screen. The chessboard image is composed of square black and white squares of the same size. Set each row of the chessboard to have m squares and each column to have n squares. Use the camera at a distance of S0 to vertically align with the center of the image and take a picture of the chessboard image. Calculate and process the picture of the chessboard image. Let P be the black-and-white dividing point. i,j The index value corresponds to the ordinal number of the square it occupies. In the chessboard image, the bottom left corner P... 0,0 The pixel of the point is (X 0,0 ,Y 0,0 (), upper right corner point P m,n The pixel of the point is (X m,n ,Y m,n ), then P m,n The point coordinates are (X m,n - X 0,0 ,Y m,n - Y 0,0 For each black-and-white separator P i,j The coordinate values ​​are calculated by comparing the x and y coordinates of each black and white dividing point with the coordinates of the upper right corner point P. m,n Divide the x and y coordinates to obtain the relative positions of the black and white dividing points in the actual captured image within the entire chessboard grid. k s The parameter representing the degree of distortion in the captured image is calculated as follows: in and For each black and white dividing point P ij The x and y coordinates; When k s If the value is less than the set threshold, it is determined that the accuracy of the image captured by the camera meets the requirements. If it meets the requirements, the current S0 is determined as the distance between the camera and the screen.

2. The system according to claim 1, characterized in that, The data processing unit is used for data encoding and display calculations. When transmitting a message containing mCount characters, the transmission time must not exceed t0. The time consumed by calling the image recognition algorithm to recognize an image is t1. The time for processing the recognized data and storing it in the database is t2. The minimum number of times the same image is captured is k. The display time for each page of data on the screen is t3. The number of pages of images that need to be displayed on the screen to transmit this data is n. The number of characters displayed on each page of the screen is nCount. The above data must meet the following conditions: 。 3. The system according to claim 2, characterized in that, The data processing unit is also used to fix the number of characters displayed in each line of data according to the number of characters nCount displayed per page on the screen and the font size f, to ensure that the number of characters displayed in each line is the same, and to set a limit on the number of lines of displayed data to prevent the displayed content from exceeding the screen display range.

4. The system according to claim 1, characterized in that, The image processing unit is used to establish a deep learning text recognition algorithm framework and train algorithm parameters. It uses the text recognition algorithm to perform text recognition on each image in the storage folder, compares the format of the data in the recognition result with the set data format, and deletes the image information when the comparison result is a format mismatch. When the comparison result is a format match, it processes the recognized data, deletes the corresponding placeholders, separators and newline characters, extracts the page header name, and temporarily stores the recognized data. When the same recognition result appears twice or more, the recognition is considered successful, and the data is written to the corresponding storage unit.

5. The system according to claim 1, characterized in that, The inner and outer chassis consist of one outer chassis and two inner chassis. The inner chassis includes a support part for fixing the display unit and the camera. The inner chassis is fixed in the outer chassis. The outer chassis includes a chassis shell and an exhaust fan. One side of the inner chassis shell provides a circular hole for mounting the camera, with the hole located at the center of this side and the diameter of the hole matching the diameter of the camera lens; the other side of the inner chassis is open, with the display unit mounted close to this side, and the chassis shell on this side can surround the display unit; the depth of the inner chassis is determined by the distance between the camera and the screen, the size of the side of the inner chassis is consistent with the size of the display screen, and the width of the inner chassis is smaller than the width of the outer chassis.

6. The system according to claim 1, characterized in that, The data processing unit, image processing unit, and display unit are implemented through an all-in-one machine. The data processing unit in the all-in-one machine connected to the external network reads the data to be transmitted from the external network server, processes the data, and displays it on the screen of the external network all-in-one machine. The external network camera takes a picture and transmits it to the all-in-one machine on the internal network. The image processing unit in the internal network all-in-one machine performs text recognition on the image, stores the recognition result, and sends it to the internal network server.

Images