Pathological HE full-automatic dyeing and sealing equipment and control system

Abstract

The invention discloses pathological HE full-automatic staining and sealing equipment and a control system, and relates to the technical field of automatic control, and the pathological HE full-automatic staining and sealing equipment comprises a slide identification module used for identifying slide identity information; the slide pre-processing module is used for micro-jet drip dyeing, dyeing environment temperature and humidity control, and operations of dropwise adding a slide sealing liquid and covering a cover glass on the dyed slide; the intelligent analysis module is used for generating exclusive optimized dyeing parameters for a subsequent to-be-processed pathological slide by using a graph neural network model based on the dyeing parameters; the central control module is electrically connected with and uniformly dispatches other modules; according to the invention, full-automatic and programmed control is realized through artificial intelligence real-time quality control and the central control module, artificial variables are eliminated, and artificial intelligence and a series of precise electromechanical control technologies are introduced, so that the quality, the efficiency and the standardization level of pathological slide preparation are fundamentally improved, the characteristics of different tissues are automatically adapted, and the working efficiency is improved. And continuous, stable and high-quality dyed sections are produced.

Description

Technical Field

[0001] This invention relates to the field of automation control, and in particular to a fully automated pathological hematologic staining and sealing device and control system. Background Technology

[0002] Hematoxylin and eosin staining (HE) is a commonly used staining technique in pathology, widely applied in histology and cytology for the observation and diagnosis of pathological sections. This staining method uses specific dyes to make the cell nuclei (which bind to Hematoxylin in HE staining) and cytoplasm (which bind to eosin) appear in different colors, helping pathologists identify cell and tissue structures and make disease diagnoses. The application of AI-automated equipment in pathology aims to improve work efficiency, ensure the stability and consistency of staining quality, and reduce human error, especially when dealing with large sample volumes or requiring high-quality consistency. Fully automated HE staining equipment not only needs high staining accuracy but also strong stability and compatibility, adapting to different types and sizes of specimens, and providing convenient data recording and result tracking functions.

[0003] Currently, Chinese invention application CN201810338846.4 discloses a pathology staining machine with learning and constant staining functions, including a central processing unit (CPU) and a display module electrically connected to the CPU; a learning platform, which includes a teaching module, an examination module, and a trigger module, all electrically connected to the CPU; and a constant staining reagent management system, which includes a QR code scanning module and a reagent parameter preset module, both electrically connected to the CPU. This invention facilitates user learning, eliminates the need for paper documents, and saves costs. It can recognize reagent QR codes, realizing an integrated closed-loop marketing model for reagents and equipment, ensuring staining quality and improving staining success rates. In the prior art, traditional automatic staining machines typically use fixed, preset programs, applying the same staining parameters to all types of tissues. This results in significant variations in staining effects due to tissue type, thickness, and fixation time, heavily relying on technician experience for manual adjustments. Staining, mounting, and digital scanning are usually separate steps performed by different devices, requiring manual transfer, which is inefficient and prone to errors. Manual operation or semi-automatic equipment is affected by the technician's experience, fatigue, and operating habits, resulting in large fluctuations in staining effects and mounting quality. Traditional methods are batch-processed, requiring the entire batch to be completed before proceeding to the next step, which cannot flexibly handle emergency samples, such as intraoperative frozen samples. Repeated use of staining tanks inevitably leads to cross-contamination of reagents, affecting staining results. Reagent consumption is high, replacement is frequent, and management is inefficient. Open or semi-open operations expose technicians to toxic reagent vapors such as formaldehyde and xylene, and the disposal of large amounts of waste liquid also puts pressure on the environment. Summary of the Invention

[0004] The technical problem solved by this invention is that existing automated staining machines typically use fixed, preset programs and apply the same staining parameters to all types of tissues. This results in significant variations in staining effects depending on tissue type, thickness, and fixation time, heavily relying on the technician's experience for manual adjustments. Staining, mounting, and digital scanning are usually separate steps performed by different devices, requiring manual transfer, which is inefficient and prone to errors. Manual operation or semi-automatic equipment is affected by the technician's experience, fatigue, and operating habits, leading to large fluctuations in staining and mounting quality. Traditional batch processing requires waiting for the entire batch to complete before proceeding to the next step, making it inflexible for handling emergency samples, such as intraoperative frozen samples. Repeated use of staining tanks inevitably leads to cross-contamination of reagents, affecting staining results. Reagent consumption is high, requiring frequent replacements and inefficient management. Open or semi-open operations expose technicians to toxic reagent vapors such as formaldehyde and xylene, and the disposal of large amounts of waste liquid also puts pressure on the environment.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a fully automated pathological hematoxylin and eosin (HE) staining and mounting control system, comprising a slide recognition module, a slide pre-processing module, an intelligent analysis module, and a central control module: The slide identification module is used to transfer the pathological slides to be processed from the loading station to the micro-spray staining module, and to identify the slide identity information at the initial transfer position and at the end of the sealing process. The slide pretreatment module is used for micro-spraying dyeing, controlling the temperature and humidity of the dyeing environment, and performing operations such as adding mounting solution and covering the stained slides with coverslips. The intelligent analysis module is used to receive the digital scan results generated by the scanning unit, and based on the digital scan results and the corresponding staining parameters, to construct and continuously optimize a graph neural network model for predicting staining quality, and to use the graph neural network model to generate exclusive optimized staining parameters for subsequent pathological slides. Based on the output parameters of the intelligent analysis module, the central control module electrically connects to and uniformly schedules the slide recognition module, the slide preprocessing module, and the intelligent analysis module.

[0006] Preferably, the slide recognition module includes an automatic feeding unit and a scanning unit: The automatic feeding unit is used to pick up the glass slides to be processed one by one from the standard glass slide holder and place the glass slides to be processed onto the linear transport track; The scanning unit includes a slide identification subunit and a digital scanning subunit. The slide identification subunit is used to read the slide identity information on the slide label in the initial segment of the linear transmission track of the fully automated pathological HE staining and sealing equipment. The slide identity information includes the slide serial number and the pathological tissue sample name. The digital scanning subunit includes digitally scanning the sealed slides to obtain slide pathology information, which includes digital scan images and pathology recognition information.

[0007] Preferably, the digital scanning subunit specifically includes a high-resolution image sensing subunit and a precision motion control platform: The high-resolution image sensing subunit includes a high-pixel area array camera, which is used to acquire color images of pathological tissue samples at several magnifications, automatically identify multiple focal layers of the tissue sample before scanning, perform layered scanning during scanning, and superimpose and synthesize the layered scanning results into a full-depth clear image according to the three-dimensional solid dimension. The precision motion control platform is used to support and move the glass slide via a cross roller slide and a high-precision ball screw. It provides submicron-level positioning feedback through a submicron grating encoder and stitches together the full-depth clear images to obtain digital scan images of pathological sections on the glass slide surface and pathological identification information, including the glass slide number and the name of the pathological tissue sample.

[0008] Preferably, the slide pretreatment module includes a micro-spraying staining unit, a staining temperature control unit, and a sealing unit; The micro-spray staining unit is used to perform a non-contact and procedural multi-step staining and cleaning process on tissue samples on slides based on the staining scheme matched with the slide's identity information. The staining temperature control unit is used for automatic temperature adjustment of the independent staining platform during the internal slide staining process of the fully automated pathology HE staining and mounting control system. The mounting unit is used to add mounting solution and cover the stained slides with coverslips. The specific operations include: A high-definition camera captures an overall image of the slide. Based on the overall image, the contour and centroid position of the tissue sample on the slide are identified and calculated. An adaptive coordinate axis is established based on the contour of the tissue sample to obtain the two-dimensional coordinates of the tissue sample. A preset volume of environmentally friendly mounting liquid is added to the slide according to the centroid position using a micro-dispensing pump. A six-axis flexible robotic arm picks up a coverslip and places it on the slide that has been covered with the mounting liquid. The end effector of the six-axis flexible robotic arm integrates a vacuum suction cup, a pressure sensor, and a flexible joint. The vacuum suction cup is used to stably adsorb and release the cover glass slide in a negative pressure manner; The pressure sensor is used to monitor and provide feedback in real time on the force generated when the coverslip comes into contact with the sealing liquid and the slide during placement. The flexible joint is used to buffer the compliant displacement generated by the end effector when placing the cover glass.

[0009] Preferably, the micro-jet printing unit includes a multi-channel piezoelectric inkjet printhead array, a high-frequency piezoelectric oscillation subunit, a targeted spray cleaning subunit, and a waste liquid negative pressure recovery subunit. The multi-channel piezoelectric inkjet printhead array stores different reagents through several independent reagent channels. The reagents include hematoxylin, eosin, differentiation solution, blueing solution, gradient alcohol and environmentally friendly transparent agent. The multi-channel piezoelectric inkjet printhead array is used to spray reagent microdroplets onto the tissue sample area on the glass slide under the command of the central control module. The high-frequency piezoelectric oscillator subunit is used to apply micro-amplitude vibrations at the kilohertz level to the glass slide after the reagent is sprayed onto the glass slide; The targeted spray cleaning subunit includes programmable micro-orifice nozzles for pulsed, high-pressure liquid spray cleaning of the glass slide surface; The waste liquid negative pressure recovery subunit works in conjunction with the targeted spray cleaning subunit to immediately remove the waste liquid generated during the cleaning process.

[0010] Preferably, the intelligent analysis module includes: An initial pathological HE heterogeneous knowledge graph is established. The nodes of the pathological HE heterogeneous knowledge graph include slide feature nodes, staining parameter nodes, and staining quality nodes. The slide features include the pathological tissue sample name, slide number, and slide model. The staining parameters include the staining time and temperature of hematoxylin and eosin. The staining quality includes the cell nuclear clarity resolution and the percentage of cytoplasmic color separation effect. The edges of the pathological HE heterogeneous knowledge graph include the inherent correlation between the information represented by each node. The correlation is obtained from historical experimental data collected from the big data network. The pathological HE heterogeneous knowledge graph was used as a training set to train a graph neural network, which learned and quantified the nonlinear dependence between slide features, staining parameters and staining quality. When a new digital scan result is received, the new digital scan result is updated as new data to the pathological HE heterogeneous knowledge graph, and the graph neural network is iteratively trained and optimized. For the pathological slides to be processed, the graph neural network is used to infer and predict the optimal staining parameters based on the slide's feature node information.

[0011] Preferably, the specific scheduling process of the central control module includes: Step S1: The central control module instructs the automatic feeding unit of the slide recognition module to pick up a pathological slide to be processed from the slide holder and place it on the transport track. When the slide passes the slide recognition subunit, the central control module receives the slide identity information read by the slide recognition subunit. Step S2: The central control module sends the slide identification information to the intelligent analysis module, generates the optimal staining parameters for the slide, and sends the staining parameters back to the central control module. Step S3: The central control module schedules each module to execute the staining microcirculation according to the staining parameters; Step S4: After staining is completed, the central control module instructs the mounting unit to perform the mounting operation; Step S5: After the slide is sealed, the central control module instructs the slide to be transmitted to the digital scanning subunit to obtain digital scanning images and pathological identification information of the pathological sections on the slide surface. The digital scanning images, pathological identification information and staining parameters are then transmitted to the intelligent analysis module. Step S6: After receiving new digital scan images, pathological identification information and staining parameters, the intelligent analysis module updates the pathological HE heterogeneous knowledge graph and updates the graph neural network according to the updated pathological HE heterogeneous knowledge graph. Step S7: The central control module generates a complete digital archive of the entire process data of the slide and stores it in the pathological staining database. The entire process data includes the staining parameters used, the reaction time of each step, the reagent batch number, the mounting pressure curve and the operation timestamp. Then, process the next slide and repeat the above steps S1 to S7. Step S8: For the slide to be processed, the slide identification module obtains the slide identity information, retrieves the corresponding personalized staining parameters bound to the slide from the currently updated pathological staining database, and sequentially and cyclically calls the micro-spray staining module to perform micro-circulation steps. The micro-circulation steps also include hematoxylin nucleus staining, differentiation, blueing, eosin staining, gradient dehydration, and clearing. The staining temperature control unit is called for temperature control, and the mounting unit is called to complete the operation of identifying the contour and centroid position of the tissue sample, applying environmentally friendly glue, and covering the cover slide with zero bubbles. Digital files are generated for each processed slide, and the digital files are associated with the slide identity information, stored in and updated in the pathological staining database.

[0012] Preferably, the process of the central control module instructing the sealing unit to perform the sealing operation includes: Based on the two-dimensional coordinates, an instruction is given to the micro-dispensing pump to dispense sealing liquid at the centroid coordinate position. The six-axis flexible robotic arm is controlled to grasp the coverslip and tilt it at a preset angle. Based on the real-time feedback of the pressure sensor, the coverslip is smoothly turned from the tilted state to the horizontal state while maintaining the contact force in a closed-loop control manner. The air is completely expelled from one side by using the surface tension of the liquid. After the coverslip is completely horizontal and covers the sample, the negative pressure of the vacuum suction cup is released and the six-axis flexible robotic arm is controlled to reset.

[0013] Preferably, the steps of the central control module scheduling each module to execute the staining microcycle according to the staining parameters include: Based on the personalized staining parameters corresponding to the glass slide, the spray volume, spray pattern, reaction time, vibration frequency and duration of the high-frequency piezoelectric oscillation platform, and spray pressure and number of sprays of the targeted spray cleaning subunit in each channel of the multi-channel piezoelectric inkjet printhead array are controlled. Associating the digital file with the slide identity information includes: The digital archive records data associated with the slide's identity information, including: the name of the staining protocol used, the batch number and spray volume of the reagents used, the reaction time and temperature of each step, the pressure curve recorded by the pressure sensor during mounting, the operator ID, and the processing completion timestamp.

[0014] A fully automated pathological hematologic staining and sealing device includes a memory, a processor, and a computer program stored in the memory. The processor executes the computer program to realize a fully automated pathological hematologic staining and sealing control system.

[0015] The beneficial effects of this invention are: This invention can improve staining efficiency and stabilize staining quality through artificial intelligence machine staining, thereby reducing misdiagnosis caused by staining quality and lowering the risk of patient treatment. This invention's fully automated, programmed control eliminates human error, ensuring high consistency and repeatability in the processing of each slide, significantly improving the stability of slide quality. The single-slide flow design allows for sample processing on demand, continuous loading without waiting, and the ability to insert high-priority emergency samples and invoke rapid protocols at any time, significantly shortening report issuance time. Each slide uses fresh reagents sprayed from a nozzle, and the targeted cleaning and immediate waste liquid recovery system completely eliminates cross-contamination. Micro-spraying technology reduces reagent usage, and combined with batch traceability, it achieves refined and economical reagent management. The fully enclosed automated process completely isolates operators from chemical reagents. Furthermore, the invention mentions the use of environmentally friendly clearing agents and sealing solutions, combined with centralized negative pressure waste liquid recovery, significantly improving biosafety levels and reducing environmental impact. This invention not only simply automates the staining and mounting steps but also fundamentally improves the quality, efficiency, and standardization of pathological slide preparation by introducing artificial intelligence and a series of sophisticated electromechanical control technologies. A heterogeneous pathological hematologic staining (HE) knowledge graph was established, linking slide characteristics (such as tissue type), staining parameters (such as time and temperature), and staining quality (derived from digital scan image analysis). Graph neural networks were used to learn these complex relationships, enabling the system to predict the optimal personalized staining scheme for each new slide. After processing each slide, the system immediately performs a digital scan, analyzes the staining results, and feeds the "practice data" (slide characteristics + parameters used + final quality) back to the GNN model for retraining. This means the system becomes "smarter with use," continuously optimizing itself and improving staining quality. Automatic loading, staining, mounting, and digital scanning are integrated into a linear transport path. From slide entry to completion of digital scanning and archiving, the entire process is automated, requiring no manual intervention. The built-in high-resolution digital scanning unit (supporting Z-stack layered scanning to synthesize panoramic depth images) is not only for archiving but, more importantly, provides real-time, high-quality image data input for the closed-loop optimization of the AI ​​model. The system generates a complete digital profile for each slide, recording all information from mounting to output: slide ID, personalized staining parameters used, reaction times for each step, reagent batch number, mounting pressure curve, operator ID, and timestamp, achieving irrefutable end-to-end quality traceability. Through AI-driven personalized parameter recommendations and closed-loop optimization, the system can automatically adapt to the characteristics of different tissues, producing consistently high-quality stained slides, reducing reliance on experienced pathologists and enabling even novices to produce expert-level slides. Attached Figure Description

[0016] Figure 1 This is a basic flowchart of a fully automated pathological hematologic staining and sealing control system according to an embodiment of the present invention. Figure 2 This is a schematic diagram illustrating the basic scheduling process of the central control module of a fully automated pathological HE staining and sealing control system according to an embodiment of the present invention. Detailed Implementation

[0017] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0018] Reference Figure 1 As one embodiment of the present invention, a fully automated pathological hematoxylin and eosin (HE) staining and mounting control system is provided, including a slide recognition module, a slide pre-processing module, an intelligent analysis module, and a central control module: The slide identification module is used to transfer the pathological slides to be processed from the loading station to the micro-spray staining module, and to identify the slide identity information at the initial transfer position and at the end of the sealing process. The slide pretreatment module is used for micro-spraying dyeing, controlling the temperature and humidity of the dyeing environment, and performing operations such as adding mounting solution and covering the stained slides with coverslips. The intelligent analysis module receives the digital scan results generated by the scanning unit, and based on the digital scan results and the corresponding staining parameters, constructs and continuously optimizes a graph neural network model for predicting staining quality. It also uses the graph neural network model to generate exclusive optimized staining parameters for subsequent pathological slides. Based on the output parameters of the intelligent analysis module, the central control module electrically connects to and uniformly schedules the slide recognition module, the slide preprocessing module, and the intelligent analysis module.

[0019] This invention's fully automated and programmed control eliminates human variables, ensuring high consistency and repeatability in the processing of each slide, greatly improving the stability of slide quality. The single-slide flow design allows for sample processing as it arrives, continuous sample loading without waiting, and the ability to insert high-priority emergency samples and call up rapid protocols at any time. This invention does not simply automate the staining and mounting steps, but fundamentally improves the quality, efficiency, and standardization of pathological slide preparation by introducing artificial intelligence and a series of sophisticated electromechanical control technologies.

[0020] The slide recognition module includes an automatic feeding unit and a scanning unit: The automatic feeding unit is used to pick up the glass slides to be processed one by one from the standard slide holder and place the glass slides to be processed onto the linear transport track; The scanning unit includes a slide identification subunit and a digital scanning subunit. The slide identification subunit is used to read the slide identification information on the slide label in the initial segment of the linear transmission track of the fully automated pathology HE staining and sealing equipment. The slide identification information includes the slide serial number and the name of the pathological tissue sample. The digital scanning subunit includes digital scanning of the sealed slides to obtain slide pathology information, which includes digital scan images and pathology recognition information.

[0021] Based on the unique slide identification information, the central control module retrieves the personalized staining and mounting process parameters bound to the slide from the preset scheme database. The scheme database includes exclusive staining schemes for pathological HE collected through big data. The exclusive staining schemes for pathological HE include the paraffin section-standard scheme, the frozen section-rapid scheme, and the special tissue-enhanced contrast scheme. The personalized staining and mounting process parameters include the full process parameters for multi-step staining, washing, dehydration, clearing, and mounting.

[0022] The automated feeding unit automatically picks up slides from the slide holder and places them onto the track, replacing the repetitive manual labor of placing slides. This significantly improves processing efficiency and throughput, laying the foundation for 24 / 7 unattended operation. By scanning slide identification information (such as tissue sample name), the system can retrieve specific process parameters optimized for that particular tissue type from the database. This overcomes the shortcomings of traditional equipment that offers only a limited number of fixed programs or a one-size-fits-all approach, ensuring optimal staining results for tissues of different thicknesses and origins, greatly improving staining quality and diagnostic accuracy. The database of collected data means that the system not only executes preset programs but is also an intelligent platform capable of learning and optimization. By continuously collecting and analyzing successful staining cases, staining protocols can be continuously improved and expanded, making the equipment smarter with use and adapting to the processing needs of more diverse samples in the future. By binding the parameters of all steps, including staining, cleaning, dehydration, clearing, and mounting, to the slide's identity, the system achieves "full lifecycle" digital management of the single slide processing process, providing a solid data foundation for subsequent quality traceability.

[0023] The digital scanning subunit specifically includes a high-resolution image sensing subunit and a precision motion control platform: The high-resolution image sensing subunit includes a high-pixel area array camera, which is used to acquire color images of pathological tissue samples at several magnifications. Before scanning, it automatically identifies multiple focal layers of the tissue sample, performs layered scanning during scanning, and superimposes the layered scanning results into a clear full-depth image according to the three-dimensional dimension. It can achieve clear image acquisition, support multiple magnifications, reduce color difference distortion, and automatically identify focal layers in digital scanning and perform layered scanning to ensure that every part of the image is clear. The precision motion control platform uses a cross roller slide and a high-precision ball screw to carry and move the slide, ensuring smooth and precise movement of the slide during scanning. It also provides submicron-level positioning feedback through a submicron grating encoder and stitches together clear panoramic images to obtain digital scan images of pathological sections on the slide surface and pathological identification information, including the slide number and the name of the pathological tissue sample.

[0024] The performance of AI models is highly dependent on the quality of the input data. This constraint ensures that the digital images provided to the intelligent analysis module are globally clear, distortion-free, and high-resolution. This is crucial for AI to accurately assess staining quality (such as cell nuclear clarity). It solves the problem of local blurring caused by uneven tissue thickness in traditional scans; high-quality panoramic digital slides are themselves valuable pathological diagnostic and archival data.

[0025] The slide pretreatment module includes a micro-spray staining unit, a staining temperature control unit, and a mounting unit; The micro-spray staining unit is used to perform a non-contact and procedural multi-step staining and cleaning process on tissue samples on slides based on the staining scheme matched with the slide's identity information. The staining temperature control unit is used for automatic temperature adjustment of the independent staining platform during the internal slide staining process of the fully automated pathology HE staining and mounting control system. The mounting unit is used to add mounting solution and cover the stained slides with coverslips. The specific operations include: A high-definition camera captures an overall image of the slide. Based on the overall image, the contour and centroid position of the tissue sample on the slide are identified and calculated. An adaptive coordinate axis is established based on the contour of the tissue sample to obtain the two-dimensional coordinates of the tissue sample. A micro-dispensing pump dispenses a preset volume of environmentally friendly mounting liquid onto the slide according to the centroid position. A six-axis flexible robotic arm picks up a coverslip and places it over the slide that has been dispensed with the mounting liquid. The core advantage of this module lies in its ability to identify the actual location and shape of tissue samples using a vision system and precisely apply mounting medium to the center of mass. This solves the problem of mounting failures caused by the fixed position of the mounting medium in traditional mounting machines (such as insufficient medium coverage or excessive spillage). This ensures that each dispensing is optimized, completely covering the sample while minimizing the use of mounting medium. Compared to traditional simple XYZ three-axis robotic arms, the six-axis flexible robotic arm offers greater degrees of freedom. It can more flexibly grasp and place coverslips, adapting to minor positional deviations of the slides on the track, significantly improving the system's robustness and operational success rate.

[0026] The end effector of the six-axis flexible robotic arm integrates a vacuum suction cup, pressure sensor, and flexible joint: Vacuum suction cups are used to stably adsorb and release cover glass slides using negative pressure. The pressure sensor is used to monitor and provide feedback in real time on the force generated when the coverslip comes into contact with the sealing liquid and the slide during placement; Flexible joints are used to buffer compliant displacements of the end effector during cover glass placement.

[0027] The end effector integrates a vacuum suction cup, pressure sensor, and flexible joint. The vacuum suction cup is a mature and reliable non-contact gripping method that can stably adsorb fragile coverslips, avoiding damage or contamination that may be caused by physical grippers. The pressure sensor provides tactile feedback for high-quality slide sealing operations. It monitors in real time the contact between the coverslip and the sealing fluid, as well as the force applied when finally placed onto the slide, allowing the robotic arm to operate with extremely gentle and controlled pressure. This effectively prevents crushing the coverslip or damaging the underlying tissue sample due to excessive force. The flexible joint brings compliance to the rigid robotic arm. In the delicate operation of placing coverslips, it buffers the impact of minor positioning errors or vibrations, making the placement process smoother and more stable, further reducing the risk of coverslip breakage and improving the success rate of slide sealing.

[0028] The micro-jet printing unit includes a multi-channel piezoelectric inkjet printhead array, a high-frequency piezoelectric oscillation subunit, a targeted spray cleaning subunit, and a waste liquid negative pressure recovery subunit. The multi-channel piezoelectric inkjet printhead array stores different reagents through several independent reagent channels. The reagents include hematoxylin, eosin, differentiation solution, blueing solution, gradient alcohol and environmentally friendly clearing agent. The multi-channel piezoelectric inkjet printhead array is used to spray reagent microdroplets onto the tissue sample area on the glass slide under the command of the central control module. The high-frequency piezoelectric oscillator subunit is used to apply kilohertz-level micro-amplitude vibrations to the glass slide after the reagent is sprayed onto the slide, so as to accelerate the uniform spreading and tissue penetration of the reagent. The targeted spray cleaning subunit includes programmable micro-orifice nozzles for pulsed, high-pressure liquid spray cleaning of the glass slide surface; The waste liquid negative pressure recovery subunit works in conjunction with the targeted spray cleaning subunit to immediately remove the waste liquid generated during the cleaning process, preventing liquid residue and cross-contamination.

[0029] The text details how the micro-jet staining module efficiently and accurately executes the treatment plan through sophisticated hardware. A multi-channel piezoelectric inkjet printhead precisely sprays reagents into the tissue area in micro-droplet form, rather than soaking the entire slide. This significantly reduces the consumption of expensive reagents (such as hematoxylin, eosin, and antibodies), lowering the staining cost per slide. Simultaneously, the non-contact spraying and independent channels avoid the cross-contamination and rapid deterioration problems associated with traditional staining baths. A high-frequency piezoelectric oscillation subunit provides micro-vibration to the slide after reagent addition, acting as a miniature oscillator or mixer. This accelerates reagent spreading and penetration on the tissue surface, breaking surface tension and ensuring more uniform staining and a more complete reaction, while also shortening the necessary reaction time. The targeted spray cleaning subunit's pulsed high-pressure spray, compared to traditional immersion rinsing, more powerfully and quickly removes residual reagents, resulting in better cleaning and less water consumption. Programmable micro-orifice nozzles can precisely target the tissue area, avoiding ineffective cleaning of the entire slide. The waste liquid negative pressure recovery system works in conjunction with the cleaning unit to instantly remove waste liquid. This not only prevents the accumulation and splashing of waste liquid inside the equipment and avoids the mixing and contamination of reagents from different steps, but also reduces the volatilization of harmful chemical gases (such as alcohol and xylene), creating a safer and more environmentally friendly laboratory environment.

[0030] The intelligent analysis module includes: Initialize the pathological HE heterogeneous knowledge graph. The nodes of the pathological HE heterogeneous knowledge graph include slide feature nodes, staining parameter nodes, and staining quality nodes. Slide features include the pathological tissue sample name, slide number, and slide type. Staining parameters include the staining time and temperature of hematoxylin and eosin. Staining quality includes the cell nuclear clarity resolution and the percentage of cytoplasmic color separation effect. The edges of the pathological HE heterogeneous knowledge graph include the inherent relationships between the information represented by each node. The relationships are obtained from historical experimental data collected from the big data network. The pathological HE heterogeneous knowledge graph was used as a training set to train a graph neural network, which learned and quantified the nonlinear dependence between slide features, staining parameters and staining quality. When new digital scan results are received, the new digital scan results are updated as new data to the pathological HE heterogeneous knowledge graph, and the graph neural network is iteratively trained and optimized. For the pathological slides to be processed, based on the feature node information of the slides, a graph neural network is used to infer and predict the staining parameters that produce the best staining quality.

[0031] This concept of optimizing staining parameters in the upper layer is put into practice in terms of the physical operations that the equipment can perform, such as reaction time and heating temperature.

[0032] It clarifies how the system learns and thinks, describing a closed-loop optimization process that enables the system to become smarter with use, continuously accumulating experience, automatically adapting to new tissue types and reagent batches, and continuously improving staining quality.

[0033] The specific scheduling process of the central control module includes: Step S1: The central control module instructs the automatic feeding unit of the slide recognition module to pick up a pathological slide to be processed from the slide holder and place it on the transport track. When the slide passes through the slide recognition subunit, the central control module receives the slide identification information read by the slide recognition subunit. Step S2: The central control module sends the slide identification information to the intelligent analysis module, generates the optimal staining parameters for the slide's staining quality, and sends the staining parameters back to the central control module. The specific quantization instruction set for staining parameters includes: Precise reaction time of hematoxylin and eosin (in seconds); During the staining process, the staining temperature control unit sets the target heating temperature (in degrees Celsius) and duration (in seconds). The heating target is the independent platform supporting the slide below the slide being stained, thereby achieving the purpose of controlling the temperature of the staining solution. The precise spray volume (in microliters) and spray pattern (such as scanning and dot matrix) of each reagent in the micro-spraying unit.

[0034] Step S3: The central control module schedules each module to execute the staining microcirculation according to the staining parameters; Step S4: After staining is completed, the central control module instructs the mounting unit to perform the mounting operation; Step S5: After the slide is mounted, the central control module instructs the slide to be transmitted to the digital scanning subunit to obtain digital scanning images and pathological identification information of the pathological sections on the slide surface. The digital scanning images, pathological identification information and staining parameters are then transmitted to the intelligent analysis module. Step S6: After receiving new digital scan images, pathological identification information and staining parameters, the intelligent analysis module updates the pathological HE heterogeneous knowledge graph and updates the graph neural network according to the updated pathological HE heterogeneous knowledge graph. Step S7: The central control module generates a complete digital archive of the entire process data of the slide and stores it in the pathological staining database. The entire process data includes the staining parameters used, the reaction time of each step, the reagent batch number, the mounting pressure curve and the operation timestamp, so as to facilitate quality traceability. The next slide is then processed, and the above steps S1 to S7 are repeated to form a continuous and intelligent closed-loop processing flow. Step S8: For the slides to be processed, the slide identification module obtains the slide identity information, retrieves the corresponding personalized staining parameters bound to the slide from the currently updated pathological staining database, and sequentially and cyclically calls the micro-spray staining module to perform micro-circulation steps. The micro-circulation steps also include hematoxylin nucleus staining, differentiation, blueing, eosin staining, gradient dehydration, and clearing. The staining temperature control unit is called to control the temperature, and the mounting unit is called to complete the operation of identifying the contour and centroid position of the tissue sample, applying environmentally friendly glue, and covering the slide with zero bubbles. Digital files are generated for each processed slide, and the digital files are associated with the slide identity information, stored in and updated in the pathological staining database.

[0035] The central control module generates optimized parameters through a graph neural network model, and then controls the fully automated pathological hematologic staining and sealing control system in real time and with precision based on these optimized parameters. The collaborative relationship between the two and the working logic of the entire system are clearly defined.

[0036] It details the complete steps from film loading, recognition, parameter acquisition, execution, film sealing, scanning, feedback optimization, to archiving. The "optimization parameters" output by AI are concretized into physical quantities that the device can execute (seconds, degrees Celsius, microliters, injection modes, etc.). It clearly defines how information flows between various modules, forming a continuous intelligent processing loop. It organically connects the various high-tech "components" defined in the preceding claims, forming a collaborative whole. It describes how the brain (intelligent analysis module) directs the limbs (each processing unit) to perform precise operations and receives feedback. This makes the entire patent's technical solution logically clear, the process complete, and highly feasible.

[0037] The process of the central control module instructing the sealing unit to perform the sealing operation includes: Based on the two-dimensional coordinates, a command is applied to the micro-dispensing pump to dispense sealing liquid at the centroid coordinate position. The six-axis flexible robotic arm is controlled to grasp the coverslip and tilt it at a preset angle. Based on the real-time feedback of the pressure sensor, the coverslip is smoothly turned from the tilted state to the horizontal state while maintaining the contact force in a closed-loop control manner. The surface tension of the liquid is used to completely expel the air from one side. After the coverslip is completely horizontal and covers the sample, the negative pressure of the vacuum suction cup is released and the six-axis flexible robotic arm is controlled to reset.

[0038] Preventing air bubble formation is the most critical and challenging problem in automated slide sealing. By "tilting" and then "gradually leveling," the system mimics the manual operation of even the most experienced technicians, allowing the coverslip to contact the sealing liquid from one side and then slowly cover it, using the surface tension of the liquid to smoothly expel air from the other side. This method, based on physical principles, minimizes the formation of air bubbles, ensuring sealing quality. Real-time feedback and closed-loop control based on pressure sensors are the hallmarks of this intelligent process. Instead of executing a fixed, blind sequence of actions, the system dynamically adjusts the descent speed and posture of the robotic arm based on the actual contact force sensed by the sensors. This means that regardless of subtle changes in the viscosity of the sealing liquid or unevenness of the slide surface, the system can adaptively achieve smooth, air bubble-free covering, resulting in extremely high operational consistency and success rate.

[0039] The steps of the central control module scheduling each module to execute the staining microcirculation according to the staining parameters include: Based on the personalized staining parameters corresponding to the glass slide, the spray volume, spray pattern, reaction time, vibration frequency and duration of the high-frequency piezoelectric oscillation platform, and spray pressure and number of sprays of the targeted spray cleaning subunit in each channel of the multi-channel piezoelectric inkjet printhead array are controlled. Linking digital archives with slide identity information includes: The digital archive records data associated with the slide's identity information, including: the name of the staining protocol used, the batch number and spray volume of the reagents used, the reaction time and temperature of each step, the pressure curve recorded by the pressure sensor during mounting, the operator ID, and the processing completion timestamp, forming a complete and traceable quality control record.

[0040] Establishing irrefutable end-to-end traceability and recording detailed data creates a complete "digital ID card" and "historical record" for each slide. When any staining issues arise (such as pale staining, precipitation, or air bubbles during mounting), the record can be accessed immediately, accurately tracing back to the reagent batch number used, specific process parameters (accurate to the μL volume and seconds), and even pressure changes during mounting. This is crucial for troubleshooting, quality control, and meeting the stringent certification requirements of medical laboratories (such as CAP and ISO 15189).

[0041] These recorded data, especially process data such as "pressure curves," are valuable information assets. By statistically analyzing a large amount of archival data from slides, potential correlations between process parameters and final staining quality can be discovered, which in turn guides further optimization and improvement of staining protocols, forming a data-driven, closed-loop quality improvement cycle.

[0042] A fully automated pathological hematologic staining and sealing device includes a memory, a processor, and a computer program stored in the memory. The processor executes the computer program to realize the fully automated pathological hematologic staining and sealing control system.

[0043] This invention can improve staining efficiency and stabilize staining quality through artificial intelligence machine staining, thereby reducing misdiagnosis caused by staining quality and lowering the risk of patient treatment. This invention's fully automated, programmed control eliminates human error, ensuring high consistency and repeatability in the processing of each slide, significantly improving the stability of slide quality. The single-slide flow design allows for sample processing on demand, continuous loading without waiting, and the ability to insert high-priority emergency samples and invoke rapid protocols at any time, significantly shortening report issuance time. Each slide uses fresh reagents sprayed from a nozzle, and the targeted cleaning and immediate waste liquid recovery system completely eliminates cross-contamination. Micro-spraying technology reduces reagent usage, and combined with batch traceability, it achieves refined and economical reagent management. The fully enclosed automated process completely isolates operators from chemical reagents. Furthermore, the invention mentions the use of environmentally friendly clearing agents and sealing solutions, combined with centralized negative pressure waste liquid recovery, significantly improving biosafety levels and reducing environmental impact. This invention not only simply automates the staining and mounting steps but also fundamentally improves the quality, efficiency, and standardization of pathological slide preparation by introducing artificial intelligence and a series of sophisticated electromechanical control technologies. A heterogeneous pathological hematologic staining (HE) knowledge graph was established, linking slide characteristics (such as tissue type), staining parameters (such as time and temperature), and staining quality (derived from digital scan image analysis). Graph neural networks were used to learn these complex relationships, enabling the system to predict the optimal personalized staining scheme for each new slide. After processing each slide, the system immediately performs a digital scan, analyzes the staining results, and feeds the "practice data" (slide characteristics + parameters used + final quality) back to the GNN model for retraining. This means the system becomes "smarter with use," continuously optimizing itself and improving staining quality. Automatic loading, staining, mounting, and digital scanning are integrated into a linear transport path. From slide entry to completion of digital scanning and archiving, the entire process is automated, requiring no manual intervention. The built-in high-resolution digital scanning unit (supporting Z-stack layered scanning to synthesize panoramic depth images) is not only for archiving but, more importantly, provides real-time, high-quality image data input for the closed-loop optimization of the AI ​​model. The system generates a complete digital profile for each slide, recording all information from mounting to output: slide ID, personalized staining parameters used, reaction times for each step, reagent batch number, mounting pressure curve, operator ID, and timestamp, achieving irrefutable end-to-end quality traceability. Through AI-driven personalized parameter recommendations and closed-loop optimization, the system can automatically adapt to the characteristics of different tissues, producing consistently high-quality stained slides, reducing reliance on experienced pathologists and enabling even novices to produce expert-level slides.

[0044] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0045] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A pathological HE full-automatic dye sealing control system, characterized in that, The pathological slide automatic dyeing and sealing device comprises a slide recognition module, a slide pre-treatment module, an intelligent analysis module and a central control module. The slide recognition module is used for transferring the pathological slide to be treated from a loading station to a micro-spraying and dripping dyeing module and identifying the slide identity information at the initial transfer position and the end of sealing. The slide pre-treatment module is used for micro-spraying and dripping dyeing, dyeing environment temperature and humidity control and performing the operations of dripping sealing liquid and covering cover slides on the slides after dyeing. The intelligent analysis module is used for receiving the digital scanning result generated by a scanning unit, constructing and continuously optimizing a graph neural network model for predicting dyeing quality based on the digital scanning result and corresponding dyeing parameters, and generating exclusive optimized dyeing parameters for subsequent pathological slides to be treated by using the graph neural network model. The central control module adjusts and schedules the slide recognition module, the slide pre-treatment module and the intelligent analysis module according to the output parameter result of the intelligent analysis module.

2. The pathological HE full-automatic dye sealing control system according to claim 1, wherein The slide recognition module comprises an automatic loading unit and a scanning unit. The automatic loading unit is used for grabbing the slides to be treated from a standard slide rack one by one and placing the slides to be treated on a linear transmission track. The scanning unit comprises a slide recognition subunit and a digital scanning subunit. The digital scanning subunit comprises a high-resolution image sensing subunit and a precision motion control platform.

3. The pathological HE automatic dyeing and sealing control system according to claim 2, wherein The high-resolution image sensing subunit comprises a high-pixel area array camera, which is used for collecting color images of pathological tissue samples at several magnifications, automatically identifying multiple focal planes of the tissue samples before scanning, performing layered scanning in the scanning process, and superimposing and combining the layered scanning results into a panoramic clear image according to three-dimensional dimensions. The precision motion control platform is used for carrying and moving a cross-roller sliding table and a high-precision ball screw for the slide and providing sub-micron positioning feedback through a sub-micron grating encoder and splicing the panoramic clear image to obtain a digital scanning image and pathological recognition information of a pathological section on the surface of the slide. The slide pre-treatment module comprises a micro-spraying and dripping dyeing unit, a dyeing temperature control unit and a sealing unit.

4. The pathological HE automatic dyeing and sealing control system of claim 3, wherein, The micro-spraying and dripping dyeing unit is used for performing non-contact and programmed multi-step dyeing and cleaning processes on the tissue samples on the slides according to the dyeing scheme matched with the slide identity information. The dyeing temperature control unit is used for automatically adjusting the temperature of the independent dyeing workbench in the pathological HE full-automatic dyeing and sealing control system by a temperature sensor. The sealing unit is used for performing the operations of dripping sealing liquid and covering cover slides on the slides after dyeing. ​ An overall image of the slide is captured by a high-definition camera, the outline and centroid position of the tissue sample on the slide are identified and calculated based on the overall image, an adaptive coordinate axis is established based on the outline of the tissue sample, and the two-dimensional coordinates of the tissue sample are obtained, a preset volume of environmentally friendly mounting medium is added on the slide by a micro-droplet pump according to the centroid position, and a cover glass is picked up by a six-axis flexible mechanical arm and covered on the slide to which the mounting medium has been added; The end effector of the six-axis flexible mechanical arm is integrated with a vacuum chuck, a pressure sensor and a flexible joint: The vacuum chuck is used to stably adsorb and release the cover glass in a negative pressure mode; The pressure sensor is used to monitor and feedback the force generated when the cover glass is in contact with the mounting medium and the slide during placement in real time; The flexible joint is used to buffer the compliance displacement of the end effector when placing the cover glass.

5. The pathological HE full-automatic dye sealing control system according to claim 4, characterized in that, The micro-spraying and dyeing unit includes a multi-channel piezoelectric inkjet print head array, a high-frequency piezoelectric oscillation subunit, a targeted spraying and cleaning subunit and a waste liquid negative pressure recovery subunit: The multi-channel piezoelectric inkjet print head array stores different reagents, including hematoxylin, eosin, differentiation solution, blueing solution, gradient alcohol and environmentally friendly transparent agent, through a plurality of independent reagent channels, and sprays the reagent droplets to the tissue sample area on the slide under the instruction of the central control module; The high-frequency piezoelectric oscillation subunit is used to apply micro-amplitude vibration of kilohertz to the slide after the reagent sprays the slide; The targeted spraying and cleaning subunit includes a programmable microporous nozzle, which is used to pulse and high-pressure liquid spraying and cleaning of the slide surface; The waste liquid negative pressure recovery subunit works with the targeted spraying and cleaning subunit to immediately suck away the waste liquid generated during the cleaning process.

6. The pathological HE full-automatic dye sealing control system according to claim 5, wherein the control system is characterized by, The intelligent analysis module includes: Initialize the pathological HE heterogeneous knowledge graph, the nodes of the pathological HE heterogeneous knowledge graph include slide feature nodes, dyeing parameter nodes and dyeing quality nodes, the slide features include pathological tissue sample name, slide serial number and slide model number, the dyeing parameters include dyeing time and temperature of hematoxylin and eosin, and the dyeing quality includes nucleus resolution and cytoplasm color separation effect percentage, the edges of the pathological HE heterogeneous knowledge graph include the internal association between the information represented by each node, and the association is obtained from historical experimental data of a big data collection network; The pathological HE heterogeneous knowledge graph is used as a training set to train a graph neural network, to learn and quantify the nonlinear dependence relationship between the slide features, dyeing parameters and dyeing quality; When a new digital scan result is received, the new digital scan result is updated as new data to the pathological HE heterogeneous knowledge graph, and the graph neural network is iteratively trained and optimized; For the pathological slide to be processed, the graph neural network is used to infer and predict the dyeing parameters with optimal dyeing quality according to the slide feature node information.

7. The pathological HE full-automatic dye sealing control system according to claim 6, characterized in that, The specific scheduling process of the central control module includes: Step S1: The central control module instructs the automatic loading unit of the slide recognition module to grab a pathological slide to be processed from the slide rack and place it on the transmission track. When the slide passes through the slide recognition subunit, the central control module receives the slide identity information read by the slide recognition subunit; Step S2: The central control module sends the slide identity information to the intelligent analysis module, generates the optimal staining parameters for the staining quality of the slide, and returns the staining parameters to the central control module; Step S3: The central control module schedules each module to perform a staining microcycle according to the staining parameters; Step S4: After staining is completed, the central control module instructs the slide sealing unit to perform the sealing operation; Step S5: After sealing is completed, the central control module instructs the slide to be transferred to the digital scanning subunit to obtain a digital scanning image of the pathological section on the slide surface and pathological recognition information, and transmits the digital scanning image, pathological recognition information, and staining parameters to the intelligent analysis module; Step S6: After receiving the new digital scanning image, pathological recognition information, and staining parameters, the intelligent analysis module updates the pathological HE heterogeneous knowledge graph, and updates the graph neural network according to the updated pathological HE heterogeneous knowledge graph; Step S7: The central control module generates complete digital archives of the full-process data of the slide into the pathological staining database, and the full-process data includes the used staining parameters, reaction time of each step, reagent batch number, sealing pressure curve, and operation time stamp, processes the next slide, and repeats the above steps S1 to S7; Step S8: For the slide to be processed, the slide recognition module obtains the slide identity information, calls the corresponding exclusive individualized staining parameters bound to the slide from the currently updated pathological staining database, sequentially and cyclically calls the micro-droplet dyeing module to perform a microcycle step, which further includes hematoxylin nuclear staining, differentiation, blueing, eosin staining, gradient dehydration, and transparency, calls the staining temperature control unit for temperature control, calls the sealing unit to complete the operations of identifying the outline and centroid position of the tissue sample, environmentally friendly glue dropping, and zero-bubble cover glass, respectively generates digital archives of each completed slide, and stores and updates the digital archives in the pathological staining database after associating them with the slide identity information.

8. The pathological HE full-automatic dye sealing control system according to claim 7, characterized in that, The process of the central control module instructing the sealing unit to perform the sealing operation includes: According to the two-dimensional coordinates, a command is given to the micro-droplet glue pump to drop sealing liquid at the centroid coordinate position, the six-axis flexible robot arm is controlled to grab the cover glass to tilt at a preset angle, based on the real-time feedback of the pressure sensor, the cover glass is gently transferred from the tilted state to the horizontal state while maintaining the contact force in a closed-loop control manner, the air is completely discharged from one side by using the liquid surface tension, and the negative pressure of the vacuum chuck is released and the six-axis flexible robot arm is controlled to reset after the cover glass completely adheres to and covers the sample.

9. The pathological HE automatic dyeing and sealing control system according to claim 8, characterized in that, The steps of the central control module scheduling each module to perform a staining microcycle according to the staining parameters include: Based on the personalized staining parameters corresponding to the slide, the ejection volume, ejection pattern, reaction time, vibration frequency and duration of the high-frequency piezoelectric oscillation platform, and the spraying pressure and number of times of the target spraying cleaning subunit of each channel reagent in the multi-channel piezoelectric inkjet print head array are controlled; The association of the digital archive with the slide identity information comprises: Data associated with the slide identity information is recorded in the digital archive, and the associated data includes: the name of the staining scheme used, the batch number and ejection volume of the reagents used, the reaction time and temperature of each step, the pressure curve recorded by the pressure sensor during mounting, the operator ID and the processing completion timestamp.

10. A pathological HE automatic dye sealing equipment, comprising a memory, a processor and a computer program stored on the memory, characterized in that, The processor executes the computer program to realize the pathological HE full-automatic staining and mounting control system of any one of claims 1-9.

Images