Ocr-based endoscope serial number recognition and decontamination process association system

Abstract

This invention relates to the field of medical device technology and discloses an OCR-based system for identifying endoscope numbers and linking disinfection procedures. The system includes a system configuration module, an image acquisition module, an OCR recognition module, a disinfection record query module, and a process control module. The system configuration module sets a preset display area for endoscope number information; the image acquisition module acquires device information images from the endoscope host screen based on the defined area; the OCR recognition module locates and identifies the endoscope number after image correction and noise reduction preprocessing; the disinfection record query module queries disinfection records in an external database based on the identified endoscope number and determines whether it meets standards; and the process control module controls the progress and interruption of the examination process based on the determination results. This invention improves the accuracy and success rate of endoscope number character positioning and recognition, ensuring the safety of medical operations.

Description

Technical Field

[0001] This invention relates to the field of medical device technology, specifically to an OCR-based system for identifying endoscope numbers and associating them with the disinfection process. Background Technology

[0002] Medical devices refer to instruments, equipment, appliances, in vitro diagnostic reagents and calibrators, materials and other similar and related items that act directly or indirectly on the human body, including the necessary computer software.

[0003] Currently, due to model differences and dynamic display of the endoscope host screen, when performing automatic endoscope number recognition, the image acquisition module relies on a preset fixed area to acquire images. It cannot adaptively adjust to capture dynamically changing and non-fixed endoscope number information areas on the screen. When the acquisition area is deviated, the endoscope number cannot be recognized subsequently, causing the process to be interrupted.

[0004] Therefore, we propose an OCR-based system for identifying endoscope numbers and linking the disinfection process to solve the above problems. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an OCR-based system for identifying endoscope numbers and associating them with the disinfection process. This system solves the problems mentioned in the background section regarding the differences in display models and dynamic display of the endoscope host screen, as well as the inability to adaptively adjust to capture dynamically changing and non-fixed positional endoscope number information areas on the screen.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an OCR-based endoscope number recognition and disinfection process association system, comprising: The system configuration module initializes the system in the endoscopy workstation system to set the preset display area position for endoscope number information; The image acquisition module is connected to the system configuration module. It acquires the device information image displayed on the endoscope host screen according to the preset display area position and transmits the image data to the OCR recognition module. The OCR recognition module is connected to the image acquisition module, receives the device information image, performs preprocessing on the image including correction and noise reduction, locates and identifies the endoscope number in the image, and transmits the identified endoscope number text data to the disinfection record query module. The disinfection record query module connects the OCR recognition module with the external disinfection record database. Based on the recognized endoscope number, it queries the disinfection process record corresponding to the endoscope device in the external disinfection record database, determines whether it meets the preset disinfection process standard, and transmits the determination result to the process control module. The process control module is connected to the OCR recognition module and the disinfection record query module respectively. It receives the endoscope number text data and the disinfection process judgment result, and controls the progress and interruption of the endoscopy examination process according to the judgment result.

[0007] Preferably, the system configuration module performs system initialization in the endoscopy workstation system. The initialization includes setting a preset display area position for the endoscope number information in the endoscopy video stream. The preset display area position serves as the priority area for the image acquisition module to acquire images and for the OCR recognition module to locate text regions.

[0008] Preferably, the system configuration module includes a parameter adjustment unit and a learning update unit; The parameter adjustment unit adjusts the parameters of the image preprocessing operation in the OCR recognition module and the acquisition range of the image acquisition module for the preset display area position according to different endoscope device models and host screen display interfaces. The learning and updating unit is connected to the OCR recognition module, collects the result data of mirror number recognition, and optimizes and updates the recognition model parameters of the OCR recognition module.

[0009] Preferably, the image acquisition module includes an image signal access unit and an image buffer unit; The image signal access unit receives video and image signals output from the endoscope host screen; The image caching unit is connected to the image signal access unit, temporarily stores key frame images containing device information in the video and image signals, and outputs them as device information images to the OCR recognition module.

[0010] Preferably, the OCR recognition module includes an image preprocessing submodule, a text localization submodule, and a character recognition submodule; The image preprocessing submodule sequentially performs image correction and image denoising operations on the input device information image, and outputs an optimized image to be recognized; The text localization submodule is connected to the image preprocessing submodule and determines the text region containing the endoscope number in the image to be recognized. The character recognition submodule is connected to the text positioning submodule, which recognizes the characters within the text area and outputs the endoscope number text data.

[0011] Preferably, the image preprocessing submodule includes an image correction unit and an image denoising unit; The image correction unit detects the tilt angle of the text area in the device information image and corrects the image orientation through one or more of the following methods: rotation, affine transformation, and perspective transformation. The image correction unit performs orientation correction on the image, and its correction effect is evaluated by the text orientation consistency score. The quantitative assessment is conducted, and the score is calculated as follows: ; in, The number of character blocks detected in the text region. For the first The angular deviation between the main axis direction of each character block and the target horizontal direction The closer the value is to 1, the more consistent the direction of the corrected text. The image denoising unit is connected to the image correction unit, and performs noise reduction processing on the corrected image using one or more of the following methods: frequency domain filtering, wavelet transform, median filtering, and Gaussian filtering.

[0012] Preferably, the OCR recognition module further includes an image binarization unit; The image binarization unit is located between the image preprocessing submodule and the text localization submodule, and between the text localization submodule and the character recognition submodule, to convert the preprocessed image and the located text region image into a black and white binary image.

[0013] Preferably, in the preprocessing operation of the OCR recognition module on the device information image, the image correction specifically includes one or more of rotation, affine transformation and perspective transformation based on angular deviation, and the image denoising specifically includes one or more of frequency domain filtering based on Fourier transform and edge detection and removal based on gradient.

[0014] Preferably, the disinfection record query module includes a standard verification unit and a result generation unit; The standard verification unit compares the endoscopic equipment's cleaning and disinfection process records with preset standardized cleaning and disinfection process parameters. The result generation unit is connected to the standard verification unit and generates a judgment result signal representing whether the washing and disinfection process is qualified or unqualified based on the comparison result. The standard verification unit compares the retrieved disinfection process records with preset standardized disinfection process parameters to determine the degree of conformity. Calculated using the following formula: ; in, This represents the total number of key parameters in the standardized disinfection process that need to be compared. For the first The preset weights of each key parameter, and satisfying , For the first The conformity indicator value of each parameter, when the actual record meets the standard of that parameter. ,otherwise The result generation unit determines the result based on the compliance degree. The relationship with the preset threshold generates a judgment result signal.

[0015] Preferably, the process control module includes a process execution unit and a process blocking unit; When the process execution unit receives a judgment result indicating that the disinfection process is qualified, it generates a control command that allows the endoscopy examination process to continue and transmits the identified endoscope number to the endoscopy information management system through the endoscopy system interface. When the process blocking unit receives a judgment result indicating that the disinfection process is unqualified, it generates a control command to prohibit the use of the endoscope device and triggers an alarm.

[0016] Compared with existing technologies, this invention provides an OCR-based system for identifying endoscope numbers and associating them with the disinfection process, which has the following advantages: 1. In this invention, the system configuration module sets the preset display area position of the endoscope number information, guiding the image acquisition module to acquire images according to the preset display area position. Combined with the OCR recognition module, the acquired device information image is preprocessed, including correction and noise reduction. This solves the image quality problems caused by screen reflection, stains, and image distortion, thereby improving the accuracy and success rate of endoscope number character positioning and recognition, reducing the risk of recognition failure due to poor image quality and improper acquisition area, and providing reliable data input for subsequent processes.

[0017] 2. In this invention, the process control module controls the progress and interruption of the endoscopic examination process based on the judgment result output by the disinfection record query module, and triggers an alarm. This achieves a hard block on endoscopic equipment that does not meet the preset disinfection process standards. This design avoids the risk that alarms will be ignored by operators and continue to be used in the traditional mode, thus ensuring the safety of medical operations.

[0018] 3. In this invention, the disinfection record query module automatically queries and compares the identified endoscope number with the disinfection process records in the external disinfection record database, and generates a judgment result based on the preset disinfection process standards. This achieves mandatory association and traceability between endoscope usage information and disinfection quality information. This process replaces the error-prone manual entry and verification, ensuring that each examination record can be accurately and automatically associated with the corresponding verified disinfection process through the process control module, thus meeting the traceability requirements of medical quality management and infection control. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the architecture of the OCR-based endoscope number recognition and disinfection process association system of the present invention; Figure 2 This is a schematic diagram of the disassembled structure of the OCR recognition module of the present invention; Figure 3 This is a schematic diagram of the system configuration module breakdown structure of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Please see Figures 1-3 The specific implementation of the OCR-based endoscope number recognition and disinfection process association system is as follows, including: The system configuration module initializes the system in the endoscopy workstation system to set the preset display area position for endoscope number information; The image acquisition module is connected to the system configuration module. It acquires the device information image displayed on the endoscope host screen according to the preset display area position and transmits the image data to the OCR recognition module. The OCR recognition module is connected to the image acquisition module, receives the device information image, performs preprocessing on the image including correction and noise reduction, locates and identifies the endoscope number in the image, and transmits the identified endoscope number text data to the disinfection record query module. The disinfection record query module connects the OCR recognition module with the external disinfection record database. Based on the recognized endoscope number, it queries the corresponding disinfection process record of the endoscope in the external disinfection record database, determines whether it meets the preset disinfection process standards, and transmits the determination result to the process control module. The process control module is connected to the OCR recognition module and the disinfection record query module respectively. It receives the endoscope number text data and the disinfection process judgment result, and controls the progress and interruption of the endoscopy examination process according to the judgment result.

[0022] The system configuration module performs system initialization in the endoscopy workstation system. Initialization includes setting the preset display area position of the endoscope number information in the endoscopy video stream. The preset display area position serves as the priority area for the image acquisition module to acquire images and the OCR recognition module to locate text areas. Among them, the overlap between the preset display area and the actual endoscope number display area. Evaluate using the following formula: ; in, The area of ​​the preset display area. This refers to the average area of ​​the actual mirror number display area, calculated from historical successful identification data. The closer the value is to 1, the more accurate the preset area setting is; Among them, the average area of ​​the actual mirror number display area The acquisition process includes: during the system initialization and learning phase, for device information images after successful recognition of the mirror number in multiple consecutive frames, the text location submodule of the OCR recognition module outputs the coordinates of the text region, calculates the area of ​​the region, and calculates the arithmetic mean of multiple area values ​​acquired over a period of time. .

[0023] The system configuration module includes a parameter adjustment unit and a learning and update unit; The parameter adjustment unit adjusts the parameters of the image preprocessing operation in the OCR recognition module and the acquisition range of the image acquisition module for the preset display area position according to different endoscope equipment models and host screen display interfaces. The parameter adjustment unit adjusts the acquisition range based on historical recognition success rates, and its adjustment coefficient... The calculation method is as follows: ; in, This is the average mirror number recognition success rate for a specific device model and interface within the previous statistical period. To improve the success rate of target identification, The learning rate coefficient is used as the preset value. The adjusted acquisition range boundary will be based on the original preset area according to the coefficient. Scaling; Average mirror number recognition success rate The calculation method is as follows: ,in This represents the number of times the endoscope number was successfully identified in the previous statistical period. The learning rate coefficient represents the total number of recognition attempts. The value range is set to (0, 0.1]; The learning and updating unit connects to the OCR recognition module, collects the results data of mirror number recognition, and optimizes and updates the recognition model parameters of the OCR recognition module. The optimization and updating adopts an incremental learning method, specifically: every preset period, all successfully recognized samples with a recognition confidence of higher than 0.95 in this period are added to the training set, and the last fully connected layer of the recognition model is fine-tuned using the stochastic gradient descent method with a learning rate of 0.0001, aiming to improve the model's adaptability to the font style of the endoscope host screen in a specific hospital. The learning update unit connects to the OCR recognition module, collects the result data of mirror number recognition, and optimizes and updates the recognition model parameters of the OCR recognition module.

[0024] The image acquisition module includes an image signal access unit and an image buffer unit; The image signal input unit receives video and image signals output from the endoscope host screen; The image buffer unit is connected to the image signal access unit, temporarily stores key frame images containing device information in video and image signals, and outputs them as device information images to the OCR recognition module. The image caching unit performs quality assessment on the acquired device information images, and its quality score Q is calculated as follows: ; in, For the global contrast of the image, This represents the percentage of the area in the image that is determined to be blurred. and These are the weighting coefficients for contrast and sharpness, respectively. ,when When the image falls below a preset threshold, an image re-acquisition command is triggered. Among them, global contrast The calculation method is as follows: ,in and For the height and width of the image, For pixels grayscale value, The average gray value of all pixels in the image, and the percentage of the blurred area. The following steps are used to obtain the following: perform edge detection on the image, calculate the gradient magnitude of each pixel, and determine the pixel regions with gradient magnitudes below a set threshold as blurred regions. The ratio of the total number of such blurred regions to the total number of pixels in the image is the blurred region. ; Edge detection uses the Sobel operator, and the threshold is determined by analyzing the statistical distribution of gradient magnitudes of historical clear image samples.

[0025] The OCR recognition module includes an image preprocessing submodule, a text localization submodule, and a character recognition submodule; The image preprocessing submodule sequentially performs image correction and image denoising operations on the input device information image and outputs the optimized image to be recognized. The text localization submodule is connected to the image preprocessing submodule to identify the text region containing the endoscope number in the image to be recognized. The character recognition submodule connects to the text localization submodule, recognizes characters within the text area, and outputs the endoscope number text data; The character recognition submodule adopts an OCR recognition model based on a convolutional recurrent neural network. This model consists of a convolutional neural network for feature extraction, a bidirectional long short-term memory network for sequence modeling, and a connected temporal classification decoding layer. The model is pre-trained using a publicly available synthetic text dataset and fine-tuned in a supervised manner using a local dataset containing thousands of labeled endoscope screen images with mirror numbers, with the cross-entropy loss function and the Adam optimizer. The character recognition submodule outputs a confidence score for each recognized character. The overall recognition confidence of the entire mirror image text The calculation method is as follows: ; in, The total number of characters in the identified mirror mark text. For the first The confidence level of character recognition, when When the value falls below the set threshold, a review and re-identification process will be triggered. Among them, the confidence level of character recognition The OCR recognition model used by the character recognition submodule provides the character classification results simultaneously, with a value range of [0, 1]. This value represents the degree of certainty of the model's character recognition result and the overall recognition confidence level. The threshold value is configured during system initialization and ranges from [0.85, 0.99]. Weighting coefficient and The typical value is =0.7, =0.3. This set of values ​​was tested on a validation set simulating different lighting conditions using a grid search method, and it achieves the best balance between contrast and sharpness.

[0026] The image preprocessing submodule includes an image correction unit and an image denoising unit; The image correction unit detects the tilt angle of the text area in the device information image and corrects the image orientation by means of one or more of rotation, affine transformation and perspective transformation, so that the text area tends to be horizontal and vertical. The image correction unit performs orientation correction on the image, and its correction effect is expressed as a text orientation consistency score. The quantitative assessment is conducted, and the score is calculated as follows: ; in, This represents the number of character blocks detected in the text region. For the first The angular deviation between the main axis direction of each character block and the target horizontal direction The closer the value is to 1, the more consistent the direction of the corrected text. The image denoising unit is connected to the image correction unit. It performs noise reduction processing on the corrected image using one or more of the following methods: frequency domain filtering, wavelet transform, median filtering, and Gaussian filtering, in order to remove noise and interference in the image and improve image clarity and signal-to-noise ratio.

[0027] The OCR recognition module also includes an image binarization unit; An image binarization unit is set between the image preprocessing submodule and the text localization submodule, and between the text localization submodule and the character recognition submodule. It converts the preprocessed image and the localized text region image into a black and white binary image to achieve separation of text and background. The image binarization unit adopts an adaptive threshold segmentation method, specifically Gaussian weighted adaptive binarization, to adapt to the uneven illumination phenomenon on the endoscope host screen. The separation effect of image binarization units is calculated by assessing the contrast enhancement effect between text regions and background regions in the binarized image. To evaluate: ; in, and These are the average gray values ​​of the foreground and background pixels in the binarized image, respectively. The standard deviation of the background pixel grayscale values. The larger the value, the better the separation effect between text and background after binarization; The distinction between foreground and background pixels is achieved through a threshold segmentation method used in the image binarization unit: pixels with gray values ​​greater than a threshold are classified as foreground, and the rest as background; the average gray value... , and standard deviation All values ​​were obtained based on statistical calculations of pixel grayscale values ​​in the binarized image.

[0028] In the preprocessing operation of the device information image performed by the OCR recognition module, image correction specifically includes one or more of rotation, affine transformation and perspective transformation based on angular deviation, and image denoising specifically includes one or more of frequency domain filtering based on Fourier transform and edge detection and removal based on gradient.

[0029] The disinfection record query module includes a standard verification unit and a result generation unit; The standard verification unit compares the endoscopic equipment cleaning and disinfection process records found with the preset standardized cleaning and disinfection process parameters. The result generation unit is connected to the standard verification unit and generates a judgment result signal representing whether the washing and disinfection process is qualified or unqualified based on the comparison result. The standard verification unit compares the retrieved disinfection process records with the preset standardized disinfection process parameters to determine their compliance. Calculated using the following formula: ; in, This represents the total number of key parameters in the standardized disinfection process that need to be compared. For the first The preset weights of each key parameter, and satisfying , For the first The conformity indicator value of each parameter, when the actual record meets the standard of that parameter. ,otherwise The result generation unit is based on the compliance degree. The relationship with the preset threshold generates a judgment result signal.

[0030] The process control module includes a process execution unit and a process blocking unit; When the process execution unit receives a judgment result indicating that the disinfection process is qualified, it generates a control command that allows the endoscopy examination process to continue and transmits the identified endoscope number to the endoscopy information management system through the endoscopy system interface so that it can be automatically bound to the current patient information, examination images and report data. When the process blocking unit receives a judgment result indicating that the cleaning and disinfection process is unqualified, it generates a control command to prohibit the use of the endoscope equipment and triggers an alarm prompt to arrange for the equipment to be cleaned and disinfected again.

[0031] The operation steps of the OCR-based endoscope number recognition and disinfection process association system are as follows: Step 1: System Configuration and Initialization First, the system configuration module initializes the endoscopy workstation system. Its core operation is to set the preset display area position for the endoscope number information. This position will serve as the priority area for subsequent image acquisition and text localization. In addition, the system enters the learning phase, and by analyzing historical successful recognition cases, it calculates the average area of ​​the actual endoscope number display area to evaluate and optimize the accuracy of the preset area. The parameter adjustment unit in the system configuration module will pre-adjust the processing parameters and acquisition range according to different device models and screen interfaces. The learning and update unit prepares for the continuous optimization of the subsequent model.

[0032] Step 2: Image Acquisition and Quality Pre-screening Based on the preset display area, the image acquisition module actively captures the device information image displayed on the endoscope host screen. This module receives the video signal from the image signal access unit and extracts and temporarily stores keyframes containing device information from the image buffer unit. Before transmission, the system performs a quality pre-screening on the acquired images, generating a comprehensive quality score by calculating its global contrast and the proportion of blurred areas. When the score is lower than the preset threshold, a re-acquisition command is triggered to ensure the reliable image quality of the input OCR recognition module.

[0033] Step 3: OCR recognition and processing of mirror number: After receiving the device information image, the OCR recognition module performs a series of automated processes. First, the image preprocessing submodule corrects and denoises the image to fix tilt and reduce interference. The image binarization unit uses an adaptive threshold segmentation method to convert the image into a black and white binary image, enhancing the contrast between text and background. Next, the text localization submodule identifies the text region containing the endoscope number in the preprocessed image. Finally, the character recognition submodule uses an OCR recognition model based on a convolutional recurrent neural network to recognize the characters in this region and outputs the endoscope number text data. Simultaneously, the module calculates the overall recognition confidence level for this recognition. When the confidence level is too low, a review process will be triggered.

[0034] Step 4: Decontamination record inquiry and compliance determination: After obtaining the identified endoscope number, the disinfection record query module immediately accesses the external disinfection record database to query the historical disinfection process records corresponding to that endoscope number. The standard verification unit within the module compares the queried records one by one with the preset standardized disinfection process parameters. The comparison process calculates the compliance degree. This is used to quantify and objectively determine whether it meets the preset disinfection process standards. The result generation unit then calculates the result based on the degree of compliance. Based on the relationship with the preset threshold, a clear judgment signal is generated representing whether the disinfection process is qualified or unqualified.

[0035] Step 5: Process Control and Execution The process control module receives the endoscope number text data from the OCR module and the cleaning and disinfection process judgment result from the query module, and makes a final decision based on this to control the progress and interruption of the endoscopy examination process. When the judgment result is qualified, the process execution unit will generate an instruction to allow continuation and automatically transmit the endoscope number to the endoscopy information management system to complete the binding with the patient information. When the judgment result is unqualified, the process blocking unit will immediately generate a control instruction to prohibit the use of the endoscope equipment, trigger an alarm prompt, and force the equipment to be cleaned and disinfected again, thereby achieving a hard block of medical safety rules.

[0036] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0037] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An OCR-based endoscope scope number recognition and decontamination process linkage system, characterized in that, include: The system configuration module initializes the system in the endoscopy workstation system to set the preset display area position for endoscope number information; The image acquisition module is connected to the system configuration module. It acquires the device information image displayed on the endoscope host screen according to the preset display area position and transmits the image data to the OCR recognition module. The OCR recognition module is connected to the image acquisition module, receives the device information image, performs preprocessing on the image including correction and noise reduction, locates and identifies the endoscope number in the image, and transmits the identified endoscope number text data to the disinfection record query module. The disinfection record query module connects the OCR recognition module with the external disinfection record database. Based on the recognized endoscope number, it queries the disinfection process record corresponding to the endoscope device in the external disinfection record database, determines whether it meets the preset disinfection process standard, and transmits the determination result to the process control module. The process control module is connected to the OCR recognition module and the disinfection record query module, respectively. It receives the endoscope number text data and the disinfection process judgment result, and controls the progress and interruption of the endoscopy examination process according to the judgment result.

2. The OCR-based endoscope scope number recognition and decontamination process linkage system of claim 1, wherein, The system configuration module performs system initialization in the endoscopy workstation system. The initialization includes setting the preset display area position of the endoscope number information in the endoscopy video stream. The preset display area position serves as the priority area for the image acquisition module to acquire images and for the OCR recognition module to locate text regions.

3. The OCR-based endoscope scope number recognition and decontamination process linkage system of claim 2, wherein, The system configuration module includes a parameter adjustment unit and a learning and update unit; The parameter adjustment unit adjusts the parameters of the image preprocessing operation in the OCR recognition module and the acquisition range of the image acquisition module for the preset display area position according to different endoscope device models and host screen display interfaces. The learning and updating unit is connected to the OCR recognition module, collects the result data of mirror number recognition, and optimizes and updates the recognition model parameters of the OCR recognition module.

4. The OCR-based endoscope scope number recognition and decontamination process linkage system of claim 1, wherein, The image acquisition module includes an image signal access unit and an image buffer unit; The image signal access unit receives video and image signals output from the endoscope host screen; The image caching unit is connected to the image signal access unit, temporarily stores key frame images containing device information in the video and image signals, and outputs them as device information images to the OCR recognition module.

5. The OCR-based endoscope number identification and disinfection process association system according to claim 1, characterized in that, The OCR recognition module includes an image preprocessing submodule, a text localization submodule, and a character recognition submodule; The image preprocessing submodule sequentially performs image correction and image denoising operations on the input device information image, and outputs an optimized image to be recognized; The text localization submodule is connected to the image preprocessing submodule and determines the text region containing the endoscope number in the image to be recognized. The character recognition submodule is connected to the text positioning submodule, which recognizes the characters within the text area and outputs the endoscope number text data.

6. The OCR-based endoscope number recognition and disinfection process association system according to claim 5, characterized in that, The image preprocessing submodule includes an image correction unit and an image denoising unit; The image correction unit detects the tilt angle of the text area in the device information image and corrects the image orientation through one or more of the following methods: rotation, affine transformation, and perspective transformation. The image correction unit performs orientation correction on the image, and its correction effect is evaluated by the text orientation consistency score. The quantitative assessment is conducted, and the score is calculated as follows: ； in, The number of character blocks detected in the text region. For the first The angular deviation between the main axis direction of each character block and the target horizontal direction The closer the value is to 1, the more consistent the direction of the corrected text; The image denoising unit is connected to the image correction unit, and performs noise reduction processing on the corrected image using one or more of the following methods: frequency domain filtering, wavelet transform, median filtering, and Gaussian filtering.

7. The OCR-based endoscope number identification and disinfection process association system according to claim 5, characterized in that, The OCR recognition module also includes an image binarization unit; The image binarization unit is located between the image preprocessing submodule and the text localization submodule, and between the text localization submodule and the character recognition submodule, to convert the preprocessed image and the located text region image into a black and white binary image.

8. The OCR-based endoscope number identification and disinfection process association system according to claim 5, characterized in that, In the preprocessing operation of the device information image by the OCR recognition module, the image correction specifically includes one or more of rotation, affine transformation and perspective transformation based on angular deviation, and the image denoising specifically includes one or more of frequency domain filtering based on Fourier transform and edge detection and removal based on gradient.

9. The OCR-based endoscope number identification and disinfection process association system according to claim 1, characterized in that, The disinfection record query module includes a standard verification unit and a result generation unit; The standard verification unit compares the endoscopic equipment's cleaning and disinfection process records with preset standardized cleaning and disinfection process parameters. The result generation unit is connected to the standard verification unit and generates a judgment result signal representing whether the washing and disinfection process is qualified or unqualified based on the comparison result. The standard verification unit compares the retrieved disinfection process records with preset standardized disinfection process parameters to determine the degree of conformity. Calculated using the following formula: ； in, This represents the total number of key parameters in the standardized disinfection process that need to be compared. For the first The preset weights of each key parameter, and satisfying , For the first The conformity indicator value of each parameter, when the actual record meets the standard of that parameter. ,otherwise The result generation unit determines the result based on the compliance degree. The relationship with the preset threshold generates a judgment result signal.

10. The OCR-based endoscope number identification and disinfection process association system according to claim 1, characterized in that, The process control module includes a process execution unit and a process blocking unit; When the process execution unit receives a judgment result indicating that the disinfection process is qualified, it generates a control command that allows the endoscopy examination process to continue and transmits the identified endoscope number to the endoscopy information management system through the endoscopy system interface. When the process blocking unit receives a judgment result indicating that the disinfection process is unqualified, it generates a control command to prohibit the use of the endoscope device and triggers an alarm.

Images